With Thanksgiving just the corner, I figure that it is an excellent time to describe how the tentacle mechanism I wrote about, back in July, was adapted for use as an animatronic turkey neck.
Link to AMODINO
A few weeks ago I received a call, from someone I frequently do work for, inquiring about an animatronic turkey puppet that needed a mechanized head and neck. The schedule was tight and a system was needed that would provide lots of expressive movement and it needed to be operable by a single puppeteer. “I have just the thing”, I told him, and in very short order I was able to send him images demonstrating how an already existing tentacle mechanism can be modified into a turkey neck. Here is an overview of the process I went through to modify the 2-Stage Tentacle Mechanism into a turkey neck.
The puppet in question needed to operated by a single puppeteer. One arm of the operator passes through the body of the puppet from the rear, supporting the base of the neck and providing gross movement to the entire body of the puppet. The other arm of the puppeteer would be used to operate a single cable controller connected to the mechanized head and neck.
So, to adapt the 2-Stage Tentacle Mech, the top stage needed to be removed and replaced with a head. Pretty straight forward.
I snapped a quick photo (see image 002) to demonstrate the basic concept. I also created a CAD model of the mechanics and laid it over a photo of a turkey head to further illustrate the idea (image 003).
TIL (today I learned) the dangly bit hanging off of the turkey’s face is called a “snood”.
A pre-painted taxidermy turkey head was obtained and a mold was made. It was decided that all the red areas of the turkey would be of .25” thick silicone. The head from the ears forward would be a fiberglass shell, except for the snood, which would also be cast silicone. A core for the skin molding process was made by carving down a hydrocal casting of the head. All the soft bits were carved away to a depth of .25”, representing the areas to be of silicone. The remaining parts of the hydrocal head were left unmodified.
After the modifications to the hyrocal skin core form were finished, I handed it off to the mold maker. I then turned my attention to designing the head mechanics.
I always start with a quick sketch to help establish things like what moves and how far. Then I move onto the computer to finalize the mech design in a CAD program. Fusion 360 from Autodesk is now my CAD program of choice.
I will spare you the gorey details about the design of the turkey head, but here are a few things I try to keep in mind whenever I design a mechanism:
Keep it simple. Too many complicated parts means it takes longer and is more expensive to build. I sometimes have to stop and count the individual parts I am coming up with just to keep myself from going off the complicated end.
Don’t over build. Make it only a little bit stronger than it needs to be. Lightness counts, especially when it is to be mounted on the tip of a cable controlled tentacle.
Keep things modular. If one part isn’t what it needs to be, be able to quickly remake or modify the part without having to fuck with the rest of the system.
Use the same sized fasteners everywhere. That way the chance of drilling a hole incorrectly is reduced and assembly is simplified. 4-40 screws were the screws of choice for this project.
Keep in mind that fiberglass head shells need to be mounted to the mechanism. Placement of screw holes and gluing surfaces need to be anticipated and provided.
With only some minor modifications, the tentacle mechanism has been converted into a very functional animatronic turkey neck.
Over the years I have developed a healthy paranoia about mechanisms not doing what they are meant to do, and I like to conduct little tests along the way. Frankly, cable actuated tentacles are not very good at supporting weight at their tip. It’s just the way the physics work. So as soon as I got the this thing cabled up and the head mech installed I had to make sure the silly thing was going to work. The test was successful and I moved onto installing the fiberglass head shells.
A quick rundown of the head mech:
A clevis is mounted at the end of the neck tentacle
A cable actuated bell crank pivots within the clevis
A pair of struts extend off of the sides of the clevis, to support the head plate
The head plate pivots between the two struts
A rigid link connects the bell crank to the head plate
I purposely included technical jargon like clevis and bell crank. It does help to use the precise technical terms for things when trying to describe a mechanical system. That being said, I didn’t know any of these terms when I started my career in animatronics and picked up the jargon along the way. I may well be using these terms incorrectly; my degree is in Studio Arts, not Mechanical Engineering.
The fiberglass head shells were then mounted to the mechanics. Mounting shells is an art. The shells are there to support the skin and act as an interface between the mechanics and the flexible skin material. The job of the animatronics mechanic is basically to move skins in a convincing way. The shells need to mounted securely and in such a way that the skins gets stretched and compressed in a naturalistic way without funky wrinkles. Mounting shells is an art.
Move test time, once again.
I do believe it is going to work
It is now ready to handed off to the art people to cover with silicone, foam, and feathers.
Here is the final video test of the Turkey Neck Mech, ready to install. Happy Tentacular Turkey Day!
It looks like I will be cutting back on the time I spend on writing blogs posts for a while, so I thought I would outline some of the ideas I have for future posts. Just so you know I am thinking of you.
This is a mash up of my 3d printed Crawling Terminator Hand, from six or eight posts ago, and the Endoarm posted to Grab Cad by Simo83. Simo83 saved me a huge amount of work by creating a beautiful model of the classic Terminator Arm and posting it online. Yay for Simo83.
https://grabcad.com/library/endoarm-stl-1
I modified the Endoarm design by making the fingers mechanically functional and then 3d printed the whole thing up in order to create a replica of the T-800 Terminator arm capable of dragging itself along by its fingertips.
I got myself a 3d scanner awhile back and put it through its paces to see what it could do. It could do quite a bit until it stopped working just as its warranty had expired. That’s what I get for being an early adopter but I learned a lot. I will demonstrate the process I went through to create a 3d model of my hand and how I was able to use that data to create a custom fit, 3d printed Telemetry Glove.
The use of practical creature effects for film and television has become so “old-school” that there is a whole new generation of filmmakers who don’t know how to best utilize it. I am going to create a primer on some of the basic techniques involved in the process of filming the performances animatronic puppets.
A few years back a videographer friend of mine and I made a short film featuring animatronic puppets called The Escape. A wide range of techniques came into play during filming and I am going to deconstruct the whole process and explain how all the various shots were made to work.
Stay tuned!
Now that the parts are electroplated, it is time to put everything together.
The neck is the heart of the Brave New Pink Flamingo. It is a 2-stage tentacle mechanism with a 1/8th inch diameter speedometer cable as its core.
When I started seriously considering the use of 3d printing for animatronics, my first thought was “how awesome would it be to just print up a mechanism and all I needed to do was drop the servos in?”. Well, it never really worked out that way, but I made a good attempt at that goal with this system. The body was 3d printed as three different layers, and when screwed together, accomodated three servos with pulleys and cable housing terminations, and integrated perfectly with the neck mechanism.
Ther servos I chose for this project were three Hitech 805bb servos. They are fairly strong and really cheap (~ $40 each). There are servos out there that are two thirds the size and are three times as strong, but they cost four times as much. So there you go.
Flamingos have long legs and the Brave New Pink Flamingo is no different. This part of the project actually ended up being a tremendous lesson for me in the structural limitations of 3d printed parts.
Aw Snap! I designed and printed the components meant to serve as the legs and hips of the BNPF. After assembling these parts I didn’t like what I saw. So I gave it the “I wonder how easily I can break it” test. Oops, too easily. Back to the drawing board.
One of the great things about 3d printing is the ability it grants to try out different iterations of and idea without a lot of man hours involved. In the picture above you can see where the first leg structure failed and how I beefed it up in the second. If I had spent the time machining these parts and had this kind of failure, there would have been much wailing and gnashing of teeth. As it was, it was more a matter of “huh, look at that…”. It was a good learning experience.
When one learns to design mechanical systems made out of metal, like I did, there are certain things one takes for granted, like strength and durability. So, moving forward with this project, I had to make it a point to stop and think about what I was asking the materials to do. It’s not a difficult thing to do, but it was an interesting process, at the time.
Here is an example of the shift that took place in my design approach. The project called for the whole body and neck assembly to elevate 120 degrees on a .25” diameter steel shaft suspended between the hips of the BNPF. The idea was to use a .5” bore metal shaft collar to clamp down on a printed plastic hub in order to hold that steel shaft. On the left, is the first hip hub part and the off-the-shelf shaft collar I intended to clamp down over the part. On the right is the redesigned hip hub with a much beefier seat for a much bigger shaft collar. The new shaft collar was something I had to custom make.
In my new found paranoia about breaking plastic components, I was very concerned about my plan to press fit bronze bushings into my 3d printed parts. I am pleased to announce that, yes, you can press fit bronze bushings into 3d printed parts.
Here is a good work around for the fact that 3d printed parts cannot be threaded: press fitted nuts in hexagonal recesses. Works great!
Raising and lowering the body of the BNPF was accomplished with a linear actuator from C.K. Design Technology Inc.
http://www.ckdesigntech.com/wseriesfb.html
It was during the construction of this part of the BNPF that my 3d printer started giving me problems. I had gotten most of the plastic parts I needed so I didn’t it slow me down and continued on using more traditional/old-school machining techniques. In terms of the levers, cranks, and clamps needed for this part of the project, there wasn’t much 3d printing was going to do for me, anyway.
The head is a relatively simple yet critical aspect of the BNPF. It is actually made from a model of the jaws of a dragon fish. How cool is that?
The neck had to be completely installed and cabled before the head could be dealt with. There is a lot that goes into a 2-stage neck mech, especially if there is going to be a cable actuated head mounted on the end of it.
To facilitate bringing all the various elements together, the BNPF needed to be mounted on its base. I wanted this critter to stand stand fairly high in relation to the eye line of its viewers, so I chose to mount it upon a metal pedestal that was once part of a decorative lamp post. It is to be displayed in a group art show and these things need to be considered.
In addition to supporting the BNPF and displaying it to its best advantage, the base houses all the various electronics needed to bring this project to life (power, microcontroller, sensors, motor control) . So, it not only needs to be sturdy and look good, it’s interior needs to be accessible while the electronics are installed. The enclosure for the electronics actually required a surprising amount of work.
The electronics for the BNPF required two different power supplies and they needed some sort of decorative enclosure. I modified an old radio I picked up at a flea market to serve that purpose.
As an art piece the BNPF looked really cool. However, it was meant to be an interactive, robotic art piece, and as such, it was less than ideal. The problem was that I was using a Basic Stamp 2 as a microcontroller and its capabilities were just too limited. At the time, it was the microcontroller I knew best. The inherent limitations of the Basic Stamp 2, as well as the fact that I had run up against the deadline for the art show, meant that its performance was less than satisfactory. Ah well.
The system I had put together to control the performance of the BNPF consisted of the Basic Stamp 2, a motion sensor, and an array of three sonar range finders. The idea was that the motion sensor would alert the Basic Stamp 2 of the presence of people, the array of sonar sensors would located the location of the nearest target within range, and then the BNPF would respond with some behavior appropriate to the direction and proximity of the nearest target. I have some success in the past with this exact system with animatronic tentacle creatures, but alas, the BNPF was just a bit too complex for it to work well. It needed to be able to respond to its environment with the same level of interactivity as a pet parrot on a perch, which it really didn’t.
I was discussing this situation with Jon McPhalen, who is a big proponent of the Parallax Propellor microcontroller and he has me convinced that the Propellor is the way to go with this kind of interactive, robotic sculpture. Microcontrollers like the Basic Stamp and the Arduino are capable of doing only one thing at a time: check the sensors, move a servo position, move another servo position, check the sensors again, ect… not really what was needed. The Propellor has parallel processing, which basically means it has 8 individual processors working simultaneously. That means sensors could continuously be scanning the surroundings of the BNPF, servos could be going through complex little behavioral subroutines, and multiple emotional states could be qued up and ready to go once the sensory inputs indicated it was appropriate. Sound great right?
I know people who seemingly eat new computer languages for breakfast. I am not one of them. I have sat down a number of times with a chunk of time set aside for learning to program the Parallax Propellor in the Spin language, and every time it was a miserable experience. I found the learning curve for Spin to be brutal. The Brave New Pink Flamingo stands lobotomized in a corner of my studio to this day. So sad.
After that loonnng series of posts I just finished (Wonderful World of Tentacle parts 1 through 5) I thought it would be nice to keep this one short and sweet. This project I am about to describe features a tentacle mechanism, but I cover new information such as the the use 3d printed mechanical parts and how to strengthen printed parts with electroplating.
The Brave New Pink Flamingo was originally created to be part of the Conjoined 5 group art show curated by Chet Zarr at the Copro Gallery in Santa Monica. It was inspired by the tacky pink plastic flamingos people sometimes use to decorate their yards. “Wouldn't it be so much cooler if they were robotic,” I thought to myself, and so was born the Brave New Pink Flamingo.
The original concept started off as a retro-futuristic-rocket-ship-looking robot flamingo featuring 3d printed pink parts. However, it evolved into something scarier, probably because I was binge listening to the H.P. Lovecraft Literary Podcast while I was building it. So it goes.
At the time, I was still getting to know my 3d printer and what it could do. I decided the Brave New Pink Flamingo (BNPF) was going to feature a servo operated tentacle mechanism. I like tentacle mechs because they are relatively simple yet can be very expressive. This was for an art piece so simple and expressive were desirable features. In addition to the tentacle mech neck, I wanted it to have jaws. A robotic sculpture with a long, sinuous neck and big, toothy jaws trying to bite people: how cool is that?
I had the opportunity to try out a new material: carbon fiber-filled PLA from Proto Pasta. It seemed like a good choice for mechanical components. I also experimented with electroplating as a way to strengthen and stiffen the 3d printed parts. Electroplated plastic works really well for mechanisms as well as art. Not only does it strengthen and stiffen the parts it gives it a beautiful and durable finish. Most of the animatronic art pieces I've done over the years have featured electroplating. I just can’t seem to help myself.
A microcontroller and an array of an electronic sensors were incorporated into the BNPF to give it some robotic interactivity. Animatronics for use in film usually involves an operator/puppeteer but as a stand-alone art piece I wanted this to be a robot, not a puppet. Alas, this was probably the least successful aspect of this project due to my limitations as a programmer. Improving my programming skills in order to bring the Brave New Pink Flamingo to life is still very much on my bucket list.
All of the printed parts for this project we're created on Woody, my Type A Series 1 3d printer. One look at the photo will tell you why I named it Woody. As a material for mechanical components, carbon fiber-filled PLA has some nice characteristics, primarily it's stiffness and it seems to warp less than regular PLA. However, the main drawback is the wear-and-tear the carbon fiber PLA inflicts upon all the metal parts of the printhead. I managed to get most of the way through two rolls of the filament before the little knurled wheel that feeds the filament through the hot end of the 3d printer was worn smooth. That's not a huge deal if one is prepared to replace printer parts on a regular basis for the sake of using carbon fiber-filled PLA, but I was still unfamiliar with the technology and I was up against a deadline. So, the experience of being in the final phase of the project and having all my prints unexpectedly turn into crap pretty much turned me off to carbon fiber-filled filaments. I ended up resorting to more traditional machining and model-making techniques to finish this project. Specifically, the feet, the head, and the body shell are fabricated by methods other than 3d printing. They turned out pretty cool, but the added aggravation was not appreciated.
A common misconception amongst people unfamiliar with 3d printing is that the parts come out of the printer in pristine condition. This is not the case. There is still a considerable need for what machinists call benchwork. In 3d printing there tends to be loose strands of filament, wonky edges where the parts were adhered to the print bed, and all the holes end up being a little bit undersized. These issues all require some trimming, standing, drilling, and filing. The great part about 3d printing mechanical components is how accurate the fabrication process is. The holes may be a little undersized but they all line up perfectly with each other. I love that.
Another underappreciated aspect of the 3d printing process is the characteristic texture that everything ends up with. There may be some really high-end machines out there that can make some beautifully smooth and flawless prints, but I don't own one of those. When I first started working in the film industry I was introduced to the concept of “if you can't hide it, feature it”. In the case of 3d printing, this means you should learn to love that funky texture because it is not worth the hassle of getting rid of it.
I was probably 90% of the way through the printing needed for this project when the prints began to fail. I didn’t realize it at the time, but the carbon-fiber-filled PLA really wears upon any metal components of the print head it comes in contact with. The benefits of using carbon fiber-filled filament are just outweighed by this fact, in my humble opinion. The filament strands are indeed strengthened by the addition of little chopped up pieces of carbon fiber but the inherent weakness between the layers of the printed part is not mitigated in any way. The junction between these layers are where the parts are weakest, so there is no real benefit gained by using the carbon fiber-filled PLA filament, though I have to admit, I do like the matte black color of the finished parts.
As useful as 3d printed parts are, there are certain applications that require metal. Specifically, anything in the body of the BNPF that is going to have threads cut into it will be made of aluminum. Printed plastic parts can be tapped but it won't take too much tightening for a screw to just rip those threads out. There are certain applications where one can get away with that, but for the most part, I try to avoid tapping plastic.
For the BNPF project I turned some simple standoffs on a lathe for securing the various components of the body together. I probably could have used off the shelf threaded standoffs but a little lathe work seemed like a nice change of pace. Additionally, I drilled and tapped a bunch of small aluminum gear blanks for mounting the tentacle segments. The gear blank modifications in particular were a little labor-intensive: lots of little holes to drill and tap. For future 3d printed tentacle projects I have figured out different mounting techniques involving the use of heat set threaded inserts that eliminate the need for the gear blanks, but that is for another project.
In addition to the aluminum standoffs and the modified gear blanks there are other mechanical components machined from metal. These were the parts involved with moving the whole body and needed to be particularly strong. We'll discuss these when we get to the body elevation aspect of the project.
A pile of parts is pretty uninspiring. So I find it helpful and informative to assemble the components as they are produced, even though I know full well I'm going to have to take everything apart again. The fact is, the more complicated the mechanism, the more assembly and disassembly will take place. One of the guys who taught me how to do animatronics, way-back-when, told me that a project isn't done until you’ve taken it apart and put it back together at least 8 times. And 20+ years later I pretty much have to agree. It's best to keep it to a minimum, for the sake of one's own sanity, not to mention that of your employer, but a certain amount of it is unavoidable. But it's pretty damn cool to see your stuff coming together during that initial pre-assembly.
Servos were used to drive the neck of the BNPF. The servos I chose were three HiTec hs-805BB servos because they are strong, durable, inexpensive, and have nylon gears instead of metal gears. Nylon gears are not as strong as metal gears but they have better wear characteristics and will last longer if not put under too much load. The design of the 3d printed body framework is laid out so as to provide a place to mount the neck, hold the servos in place in relation to the neck, and to provide a structural pivot point for the elevation of the body. This should all become more clear as all the components come together. The elements of the body structure are designed in layers to facilitate the 3d printing process. The the aluminum standoffs connect these different layers together.
Once I am satisfied with how the various body and neck elements go together I then have to take it all apart for the electroplating process. Yay.
I have seen examples online of people electroplating 3d printed plastic parts but never for the purpose of making them more structural: only to make the parts prettier. I have utilized electroplating in the past with cast resin parts to make them both stronger and yes prettier. There is a huge opportunity or making lightweight 3d printed parts more suitable for use in mechanisms, and now I am going to let you in on the secret.
Primarily, the individual components of the neck and the neck strut will be electroplated. These comprise the most visually dominant elements of the BNPF and would benefit the most from being strengthened and stiffened. Some of the other visual elements will also be plated but the neck and the strut are the most mechanically crucial parts to undergo the process.
Once the components are cleaned up, the first step of the electroplating process is to give them a coat of primer. I use an automotive primer as it is of a better quality than the more generic types of primer available from your local hardware store. After the primer is applied, a fairly heavy gauge (18 to 16 gauge) of copper wire needs to be attached in order to suspend the parts in the electroplating solution. The plastic parts will want to float in the plating solution so the copper wire needs to be strong enough to keep the parts submerged during the plating process. Then, a layer of electrically conductive paint is applied to the part. Electrically conductive paint is available from plating supply companies online.
I have found it helpful to apply the conductive paint with an airbrush for anything but the smallest objects. Airbrushing helps to apply the paint evenly but it can be tricky. The conductive paint needs to be thinned down enough to pass through an airbrush but not so thin that it becomes non-conductive. My approach to this is to thin the paint just enough to be able to spray it with an airbrush and no more than that. The texture of the paint can get a little stippled, adding some roughness to the final finish of the metal plate, but I have learned to live with it. The alternatives to applying the paint with an airbrush is to use a paint brush (very labor intensive and inconsistent) or to dip the parts in conductive paint (requiring a lot of paint).
My electroplating system does really well with copper. Copper is the base metal for any plating operation weather it is for gold jewelry or chromed hot rod parts. I allowed the copper layer to build up on the parts for about 12 hours.This creates a layer that is plenty strong for the purposes of the BNPF.
I also had a small amount of nickel plating solution that I thought I'd make use of, though the finish has never been quite as bright and shiny as it might have been. Don't know why, but it always comes out a little funky. We're making art here not a hot rod; I have embraced the funk. The nickel plate was allowed to go on for about 30 minutes. It came out of the plating tank looking brown and funky, but with some buffing with a soft wire wheel in a dremel tool it shined up fairly well. However, it still retained some of that funky stuff. In future projects, the funkiness increased to the point where the solution just became useless. Perhaps some chemical element was becoming depleted in the solution. Don’t know. Not worried about it. Moving on.
The tentacle mechanism functions on the principle that one cable on one side gets longer while the cable on the opposite side gets shorter. The squash-plate controller works exactly the same: one side the cable gets longer, other side the cable gets shorter.
An affordable universal joint suitable for use in this project can be found online at McMaster Carr:
www.mcmaster.com/#k1/=1dte6o7
The u-joints actually shown in the photos is this one, www.mcmaster.com/#k77/=1dte87j , and it is surprisingly pricey (about $50 each). They were in my stache of left over parts from jobs long over with. This u-joint had a predrilled set screw hole while the less expensive one above needs to be drilled and pinned to the shaft. Easy and expensive or cheap and requiring more work, such are the choices in life.
Two controllers are needed, each with a universal joint. Cut four 6” lengths of .25” steel rod and secure them into the two u-joints. If you have the expensive u-joints, set screw them. If you have the cheap ones, drill and pin them; or solder them, or weld them (but don't go crazy with the heat).
The cast urethane plates for the squash-plate controllers will probably be fairly flat after the casting process, and the .25” diameter center hole needs to be drilled through fairly straight, which could be accomplished with a cordless drill and a steady hand. I, however, chucked up the plates on my lathe, faced off the flat surfaces, and precisely drilled the .25” center hole, dead nuts straight and centered, because it just takes less effort on a lathe.
The cast urethane plates need to be finished by drilling holes in them. Let’s take a look at the finished product so that we know what holes get drilled where and why.
Here is a run down of the important features of the squash plates:
Top Squash Plate:
The top plate of the squash-plate controller has four clearance holes for .5” long ¼-20 hex head bolts with two washers and a nut on the upper side. This hardware is for capturing the ends of the cables coming from the tentacle. There are four additional 1/16th inch holes, equally spaced between the .25” clearance holes. These holes are pass through holes for the cables. It is important that these holes be equally spaced from the center of the squash plate for the controller to work properly.
Bottom Squash Plate:
In the bottom squash plate there needs to be a way to terminate the cable housings from the tentacle mech. To accomplish this I used ferrules made from ¼-20 bolts. To secure the bolts in the plate I drilled and tapped ¼-20 holes. What’s a ferrule? Basically, I drilled a hole in a metal bolt to capture the end of the cable housing, rather than drill a capture hole in the cast part. The modified bolt/ferrule is then held in the cast part in the threaded hole. It’s a very strong connection for the end of the cable housing.
Had I been thinking, I could have eliminated the need to drill more holes by using 5/16-24 hardware and just tapped the .25 in holes that already existed in the cast parts. Sometimes I work faster than I think.
Base Plate:
This plate needs is a set screw to hold a .25” steel rod in the quarter inch center hole. I used a 10-24 machine screw because the threads are large enough to be used in cast urethane parts. Finer threads have more of a chance of stripping out.
The upper squash plate needs four holes drilled for the ends of the cables to pass through. The lower squash plate needs four holes drilled that will be threaded to accept ¼-20 screws. We are going to do this by using the .25” diameter holes already in the plates as drilling templates.
The squash plates need to aligned so that the .25” holes are offset from each other by 45 degrees. The existing holes will serve as a guide so that the new holes can be drilled accurately using a cordless hand drill.
Use a piece of .25” diameter rod to pin the upper and lower plates together so that the holes are 45 degrees out of alignment from each other.
Center drills are typically used in machining processes to accurately place a pilot hole to be drilled through with a regular drill bit. A pilot hole is simply a shallow hole placed where you want to make a bigger hole. Center drills are useful because they are very stiff and the tip won’t wander around before actually biting into the material to be drilled. We are using the center drill to accurately place a series of pilot holes by fitting it into the existing .25” holes in the squash plates. The .25” center drill will drill right down the center of the .25” holes. Drill four shallow pilot holes in both the upper and lower squash plates right between the existing .25” holes. That’s a total of eight pilot holes.
Use the pilot holes to drill four .062” holes through the upper squash plate. Now that we have the hole placements established with pilot holes, this can be done with a hand drill or a drill press. These holes are pass through holes for the ends of the cables coming from the tentacle mech.
Drill the lower squash plate just like the upper plate, but make the holes .188”. These four holes are to be tapped with ¼-20 threads.
I used my drill motor to tap these holes with ¼-20 threads. Use a tap handle if that makes you more comfortable.
Both of the squash plates need to be secured to a .25 steel shaft. A set screw is what is called for to hold things together. To drill the hole I clamped the squash plates in the vice of my milling machine. In this case, the mill serves as a glorified drill press. The same can be done with a regular drill press and a machinist’s vice, or you can just go at it with a hand drill, though I would still recommend holding the plate in a vice. In the photos you can see my set up on the mill. It was a little bit overkill, but I you want super accuracy, this is a way to do it.
I ended up drilling and tapping for a 10-24 set screw, but a ¼-20 screw would work just as well (maybe better). I wouldn’t go too much smaller as we are cutting threads in cast urethane resin and they won’t be very strong.
Slide the upper and lower squash plates onto the steel rods until the butt up against the u-joint. Secure them with the set screws.
Looks good but I almost forgot the ferrules. What are ferrules you ask?
If you do a search online for ferrules you will discover that a ferrule is a metal doo-dad that slides over another less strong doo-dad to make it stronger. Got it? Me neither. O.K. our controller squash plates are basically .25” thick cast urethane parts and we need to press-fit the ends of our cable housings into them and everything needs to secure and strong. As it is, it aint gonna be secure or strong. That’s why we are going to use ferrules.
I am going to demonstrate how to make ferrules out of ¼-20 hex bolts, and I am going to do it on a lathe. Sorry, but I just don’t know a better way to do it. I’ve looked for off-the shelf alternatives and the best I can find is called a Derby ferrule #620- from www.flandersco.com . It looks like it MIGHT work as a capture for the spring housing IF it is modified by drilling it out to the correct size (.128”), AND you use it in conjunction with the hex head adjuster #620-. I don’t know. It might work or it not. That’s why I have a lathe. Flanders is great, by the way: brake cables and housing for motorcycles and stuff like that. Too heavy for use in tentacles but good for bigger cable controlled projects.
Grab a ¼-20 hex nut in the jaws of the lathe chuck. Then thread a .5” long ¼-20 hex bolt into the nut. A washer between the nut and bolt will help to keep everything held straight and true.
We are using .5” long hex bolts, so we need a .128” diameter hole drilled to a depth of .375”. Drill in from the head, obviously. Then drill the rest of the way through the hex bolt with a .062” drill. Drill eight ferrule bolts. Then chamfer the the holes real good so there are no sharp edges.
Thread the finished ferrules into the ¼-20 threaded holes, as shown.
There needs to be a way to secure the ends of the cables on the upper squash plate. In each of the four .25” holes in the upper plate, put a .5”long ¼-20 hex bolt, two washers and a nut to secure them. The cables will be captured between the washers.
The controllers need to secured to a base plate along with the tentacle. In the photos you can see that I bolted everything to an aluminum plate. A piece of plywood, or even an old wooden crate, would have worked just as well. Use what you got and what is appropriate. I had an aluminum plate. Securing the tentacle to the base plate is a block of delrin with some holes drilled in it. A chunk of wooden 2x4 could have served as well. Do what you gotta do. Just make sure the tentacle is mounted close enough to the controllers that the housings can be inserted into the ferrules.
If your setup looks anything like this you are good to go.
The last thing to do is to connect the tentacle cables to the cable controllers. Are you as excited as I am? Let’s do this!
I am going grab all this stuff in my little vice so I can flip it up at a better angle for the sake of photography and the crick in my neck.
The trick to cabling tentacles is to get the right cable in the right hole. If you pull back on a controller you want the tentacle to also pull back. Otherwise, controlling the tentacle will just confuse your brain. Other things to keep in mind: lower segment on one controller, upper segment on the other controller, and keep the front/back, left/right cables together. As this goes together it will all make sense.
Start with the cable that connects to the front of the upper segment of the tentacle. It connects to the rear part of the controller (on the left, in this case). This way, when the controller pulls back, the cable to the front of upper tentacle segment will lengthen and the upper part of the tentacle will pull back. Awesome huh? If it doesn’t make sense when you read this, it will once you start putting it together.
Anyway, feed the cable into the correct ferrule and press in the housing till it bottoms out in the ferrule. Then do the same for the opposite cable. That is, if you do the front cable first, do the rear cable second. Then continue with the side-to-side cables. Don’t worry about connecting the cables yet, just get the housings press fitted into the appropriate ferrules. Do this for both the upper and lower segments. Be prepared to have to stop and redo your work occasionally. I almost always brain-fart my way through this process.
Thread the cables though the little holes drilled in the upper squash plate. Then, while holding the controller handle upright, pull a cable between the washers on the nearest bolt. Pull the cable only enough to move the tentacle into a “neutral” position (or straight up). Don’t worry about getting it perfect, there will be some fine tuning as you go. Just try to get close. Once the cable is positioned correctly (or close to it), snug down the bolt so the cable can’t slip.
Do this for all the cables on both controllers.
When it is finished, both of the controllers and the tentacle should all be pretty close to standing straight up. It doesn’t have to perfectly straight. We are talking tentacles here. Who ever heard of straight tentacle?
Now for a test run!
O.K. not bad, but do you see how the tentacle segments bend more at their lower ends than they do at their higher ends? That's because there is more mechanical leverage exerted by the cables at the bottom than at the top. That's just my best guess, anyway. Maybe some engineer or physicist out there could say for sure, but whatever. It kind of bugs me me, however. The way to fix that is to add "stiffenators".
Certain areas of the tentacle are more bendable than others. The way to adjust this is by stiffening the overly bendy areas of the tentacle with strategically placed pieces of music wire. There are holes in the tentacle discs that allow for the insertion of lengths of spring steel wire (a.k.a. music wire) that can serve to adjust the flexiness of areas of the tentacle. That is what we are going to do.
These are the area that need some stiffening, to my eye anyway. There is an art to tuning a tentacle mechanism. Lets see what we can do.
Here I have inserted a piece of .025” music wire through the lowest six discs in the upper segment plus the termination disc, to get a feel for how the tentacle will react to stiffenation. It seems about right.
It looks like 4” lengths of music wire will work for the upper segment, and 4.5” lengths for the lower segment.
With the cable cutters, cut four 4” pieces of .025” music wire, and four lengths at 4.5” long.
Give the music wire pieces a 90 degree bend about .25” from the end. These bends will help capture the wires within the tentacle.
Insert the 4.5” long stiffeners into the lower segment of the tentacle. There are a total of ten discs in the lower segment, so start from between disc 6 and 7 (counting from the bottom). Feed them down through the same holes that the cable housing pass through. That way, the ends of the music wire stiffeners can be inserted into the holes in the tentacle base. This will increase the stiffness of the lower 60% of the lower segment.
The 4” long stiffeners get inserted into the upper segment. Start feeding the stifferer wires up through the aluminum termination disc. They should come up though the tentacle to the point between discs 7 and 8. Poke the ends of the wires out enough to be able to give them another bend at their tips. The bends at the ends of the wire should keep the wires from falling out during operation of the tentacle, but there should be enough play that the tentacle can freely move and the wires slide around. It is somewhat of a fine balance, but tentacles are all about the fine balance.
All this monkeying around with music wire stiffenators will allow you to control the bendiness of the tentacle. The results were pretty good, but like so much in animatronics for film and television, it is possible to futz around for ever in the quest for perfection. The tentacle moves pretty well. Could it be made to work better? Undoubtedly, but I am calling it done. Yay! High fives!
Here are the various components needed that can be obtained from www.McMaster.com
(1) .125” bore shaft collar #t3
(1 pkg. of 100) .125” bore grommets #k11
(1 pkg. of 100) .25” bore grommets #k14
(1 pkg. of 50) 6-32 threaded inserts #a113 (only one of these is needed, but they come in packages of 50, oh well)
(8, 36” lengths) extension springs .125” o.d. #k12
These are the things needed from www.versales.com
Microcable #90.0 (about 40’ is needed, as the project is presented)
(1 pkg. of 100) Copper stop sleeves #871-32-b
Other components (with suggested suppliers)
(40’) PTFE (Teflon) Tubing ~1.5mm i.d., ~1.8mm o.d. (you will have to resort to Ebay on this one, this supplier seems like a good bet: www.ebay.com/str/ATOPELEC?_trksid=p.l
Speedometer cable (all that is needed is the steel flex shaft, not all the sleeving and connectors, any auto parts store will have it, for example: www.autozone.com/drivetrain/speedometer-cable/pioneer-speedometer-cable/__
(1’) Polyurethane tubing .25” i.d., .375” o.d. (available at any hardware store)
Here is a quick run down of the tools needed to make a 2-Stage Tentacle Mechanism.
From top left to bottom right:
Micro torch
Nut driver (7/16th inch)
Crescent wrench
Tap handle
Needle nose pliers
Felco cable cutters www.versales.com
Nicopress crimpers www.versales.com
Ball-end Allen wrenches
Calipers
Tweezers
Xacto knife
Tape measure
These are the various drill bits, counter sinks, and center drills needed.
The very first power tool I would recommend that anybody should buy would be a cordless drill motor. Don’t leave home without it.
I picked this little bench grinder up from Harbor Freight Tools a couple years back and it has proven itself well worth the 50 bucks I spent on it.
This drill press also came from Harbor Freight. I love Harbor Freight.
Sometimes machined, metal parts are the only way to go. It is usually best to try to minimize the need for machined, metal parts, but occasionally you have to go there. In the case of this tentacle mechanism, I fabricated two parts using a lathe and a milling machine with a small rotary table. It is possible for you get by without them, but in this case, we're going for the best possible results. I'm not above cutting some corners now and again however in my opinion, these two exceptions where unavoidable. One of the greatest advantages of aluminum parts over plastic is that aluminum can be threaded to take a screw: plastic, not so much.
In a tentacle mechanism, it is important that the discs do not rotate on the central flexible core. If the parts do rotate the tentacle will become less controllable. The best way to eliminate the rotation is to make all the discs out of aluminum so that each individual disc can be secured on to the central shaft with a set screw. That is a lot of machining and tapping and set screwing. Satisfactory results can be gotten by set screwing the disc at the very tip of the tentacle and the disc at the transition between the top segment and the bottom segment of the tentacle (known as the termination disc). The central shaft can also be secured in the base of the tentacle with a set screw, but this is optional. Securing these points of the tentacle will ensure its controllability.
The other reason for machining parts is to be able to press-fit the steel cable housing into the termination disc between the two segments. Plastic parts will either not be able to hold the press fit housing or will simply break.
For this project, a lathe is required for turning the disks to the proper thickness and diameter. A milling machine with a rotary table is necessary or drilling the holes in the disc evenly and accurately. It is possible that somebody who is a whiz with a hacksaw and a hand drill could get adequate results but using machine tools will always get the best results.
The lathe I used to make tentacle parts is a 70 year old Logan. They started making these right after World War 2 when steel was available for things other than the war effort. It was meant to be the type of lathe one would have around the house for those little metal fabrication chores that commonly pop up (?). Yeah, anyway… They were built to last and there are still plenty of them around, as well as parts, information, and very active website forums full of machinists who dig these old machines. It is a good machine for making animatronic puppets.
www.loganact.com
In my little shop I have what is effecionatelly known as a “Baby” Bridgeport. That is a J head Bridgeport mill with a 32” table, rather than a 40 “ standard table. When Bridgeport milling machines came out (early 60s?) everyone realized what turds mills had all been up to that point, and all manufacturers of milling machines basically started making Bridgeport clones. Awesome machines with only one down side: they weigh pounds. Try getting one of those into your apartment or dorm room.
www.bpt.com
However, there are options to the towering mountain of mechanical awesomeness that is the Bridgeport milling machine. From personal experience, I can recommend the Sherline benchtop milling machine. I had one for years and made a lot of cool stuff with it. I used my Sherline rotary table clamped in the vise of my Bridgeport mill to drill the termination disc for the 2-Stage Tentacle Mechanism we are discussing.
www.Sherline.com
So, in the spirit of sticking with the Instructables format… Step 1!
Yes, it is possible that these two parts could be produced via casting or 3d printing, but they won’t last long. These parts really should be made of aluminum and that means machine tools. I will not go into the step by step process of how to make these parts with a mill and lathe: presumably, if you are a machinist , you can figure it out.
Once all the various discs are finished it is time to string all the parts together on the central shaft. We do this initial pre-assembly so as to be able to determine the lengths of the cable and cable housings. Also, it’s kind of exciting to finally see things coming together.
The entire tentacle is going to be about 12 inches long. The speedo cable will run the entire length of the tentacle so cut it to be about 15 or 16 inches long. It will extend into the base and be secured there. The excess will be trimmed later. To cut speedo cable I use either cable cutters or a dremel tool with an abrasive cut off wheel.
Start assembling the upper segment by installing the machined aluminum tip on the speedometer cable with a set screw. Then, begin sliding the discs onto the speedo cable, from smallest to largest, like beads on a string. Each disc is separated from the others with a rubber grommet of appropriate size (.125” i.d.). For the upper segment of the tentacle I am using three of each of the five different sizes of discs to create a tapered form. This will give us a length of approximately 6 inches.
The lower segment tentacle discs are strung on a piece of .25 inch (outer diameter) polyurethane tubing. The steel speedometer cable of the upper segment will be inserted through the polyurethane tubing of the lower segment. The lower segment needs to be stiffer than the upper segment because it supports the weight of the entire tentacle mech and needs a little help. The urethane tube provides the needed, additional stiffness.
It is time to string the lower segment discs onto the urethane tube, each separated with the appropriately sized grommets (.25” i.d.). The lower segment is going to be about 6 inches long, so cut the polyurethane tubing about 7 inches long. The extra length will be needed for mounting the tentacle into its base. For the lower segment, we are using 10 cast urethane discs and one machined aluminum disc to get a 6 inch length. There are four different sizes of cast discs on the lower segment: three each of the smaller sizes and only one of the largest. Start by mounting the aluminum disc on the tubing and then stack the grommets and cast discs behind it, going from smallest to largest.
I have seen other tentacle mechanisms online where all the discs are of the same diameter. This is obviously done for the sake of simplicity. In nature, tentacles are tapered: fat at the base and pointy at the tip. There are two reasons for this arrangement: the fat base has more mechanical advantage and the skinny tip weighs less and is easier to move. So, while the manufacture of various different sizes of tentacle discs is a lot of work, it does make for a better tentacle.
Once the components of the lower segment are assembled, insert the end of the tube extending from the bottom of the tentacle into the tentacle base. Then, insert the excess length of speedo cable extending from the bottom of the upper segment into the polyurethane tube of the lower segment. The pre-assembled tentacle allows us to exactly determine the lengths of the cables and housing.
For the purpose of this demonstration I am keeping the lengths of the housing relatively short. There is going to be only about 18 inches of housing between the base of the tentacle and the cable controllers. The cable housings can be made considerably longer without negatively affecting the performance . 12 to 15 feet are the maximum lengths I would recommend. Cables will stretch a bit during use and the longer the length of cable the more the stretch. Stretched cables will require tightening and tuning to maintain controllablity.
This cable housing is made of spring steel which is tough stuff. To cut the spring housing I either use my cable cutters or I use a dremel tool with an abrasive cutting wheel. I then deburr the ends with my trusty little bench top grinder. It is important that there be no little sharp bits on the ends of the housing to impede the movement of the cables or interfere with inserting the housing ends into their capture holes.
There will be two sets of four housings, of different lengths. The longer set will be for the upper tentacle segment and the shorter lengths are for the lower segment. In this case, the housings for the lower segment are 18 inches long, while the housings for the upper segment are 25 inches long. These lengths do not have to be exact. One of the great things about cable actuation is that there is some fudge factor when it comes to lengths. As long as the lengths are relatively close to what is needed (give or take and inch), all is well.
The next step is to run lengths of Teflon tubing inside of the steel spring housing. When working with short lengths of housing (as in this case) I prefer to cut lengths of Teflon tubing off the roll first a few inches longer than the cable housing. For longer lengths, I will keep the liner on the roll, for the sake of tidiness. Insert the Teflon tubing into the housing, being careful not to create any kinks, and feed the liner all the way through till it protrudes out the other end. Once the liner has been strung completely through the housing, trim the Teflon tubing so that there is approximately 1” of tubing sticking out of either end of the housing. These 1” excess lengths of Teflon tubing will be split with an Xacto knife so that when the cable housing is inserted into a capture hole the split ends of the Teflon tubing will fold back over the exterior of the cable housing and create a nice snug fit. This really is a snazzy technique for terminating cables housing.
Both of the upper and lower segments of the tentacle have their own sets of four cables and housing. The cables that go up to the top segment must pass through the lower segment without adversely affecting the movement of the bottom half of the tentacle. To help facilitate the movement of the housings within the lower segment I used a Dremel tool to open up the lower disc clearance holes a bit. When the housings run up through the lower segment of the tentacle they do a little quarter turn helical spiral from one end to the other. By opening up the clearance holes for the housing to pass through, we will relieve some of the tension that would otherwise interfere with the movement of the tentacle. The discs nearest the aluminum termination disc between the two segments will require the most pronounced modification with the Dremel, while those farthest will require none. Dremeling out the top 3 or 4 discs of the lower segment will probably be sufficient.
The lower segment of the tentacle gets assembled on a piece of quarter inch polyurethane tube. The aluminum termination disc is to be set screwed on to this tubing. In order to give the set screw something to bite into, we are going to press fit a threaded insert into the end of the tubing. The threaded insert will cause the tubing to expand so we will need taper the end of the tubing with a piece of coarse sandpaper. Otherwise we will not be able to slide the discs over the tubing. Inner diameter of the threaded inserts will need to be opened up with a drill so that the eighth inch barometer cable of the upper segment they freely pass through. I drilled the insert out with a #29 drill bit (.136” diam.) to allow the .125” speedo cable clearance to pass through. I put an inset into both ends of the tube but only the end with the termination disc was necessary. After all these years, I am still making it up as I go.
We are now going to install the cable housings for the upper tentacle segment into the aluminum termination disc. With all the grommets and cast resin discs removed from the polyurethane tubing, secure the aluminum termination disc to the end of the tubing with a set screw. Use the threaded metal insert pressed into the tubing to give the set screw something to bite into.
There are two sets of cable housings, one set longer than the other. The longer set are the cable housings for the upper segment. There should be about 1" of Teflon liner poking out of either end of each housing. These ends need to be split before inserting them into the termination disc. Use an Xacto knife to carefully slice end of the Teflon liner as shown in the photograph. When when press-fit into the termination disc the to split halves of the Teflon liner will fold-back creating a nice snug fit within the capture hole.
The outer diameter of the cable housing we are using is .125". The folded tabs of Teflon liner will add a couple additional thousandths of an inch to that dimension. If the capture holes in the aluminum termination disc were drilled correctly (.128” to .130” diameter) the ends of the spring housing should press snugly into the holes without too much problem.
Once the cable housings have been press fitted into the termination disc it is time to install the rest of the discs on to the lower segment of the tentacle. Start with the smallest discs first and work your way down to the largest disc, keeping each disc separated with a rubber grommet. Feed the four cable housings from the termination disc through the access holes drilled in the cast discs. The housings should slide freely through these holes. Otherwise, they will impede the movement of the tentacle. Once all the discs and grommets have been slid onto the polyurethane tubing there should be a 1” length of tubing left uncovered. This end will be secured in the tentacle base.
Feed the spring housings, extending out of the lower tentacle segment, through the clearance holes previously drilled through the base piece. Then, seat the end of the polyurethane tube into the center hole of the base piece. Everything should seat down snugly with no gaps showing between discs and grommets. If the urethane tube is too long trim it with an exacto knife.
Press fit the shorter length of cable housings into the previously drilled capture holes of the cast urethane tentacle base. Use the liner splitting technique discussed before, to prep the ends of the Teflon liner. Once these last four housings are installed in the tentacle base, we should have a total of 8 housings, of approximately equal length, poking out the bottom of the tentacle. They don't have to be exactly equal length, just fairly close.
Now, feed the speedo cable extending out of the bottom of the upper segment into the polyurethane tubing of the lower segment. It may take some doing to work the speedo cable past the aluminum inserts pressed into the urethane tubing. The bottom end of the upper segment seats down upon the upper end of the lower segment, and a short length of the speedo cable should be protruding from the bottom of the tentacle base.
Find the length of speedometer cables protruding from the bottom of the tentacle base and slide in 1/8 inch inner diameter shaft collar over the speedo cable. Secure it in place with the set screw, and trim off the excess speedo cable with the cable cutters.
Get in there with the Felco cable cutter and snip off the excess speedometer cable.
This tentacle mechanism is nearly functional. We are now going to mount the tentacle on a flange which will allow us to secure the tentacle to the cable controllers (we have yet to construct). Basically, we want a length of quarter inch aluminum armature wire poking out of the base of our tentacle so that we can orient the tentacle in whatever direction is desired.
While photographing the assembly process, I was taking things apart and putting them back together fairly often. I found it helpful two number all of the holes and cable housings. Being organized does sometimes help.
Basically, we need clearance holes for the cable housings (8 total), a receptacle for the shaft collar securing the end of the speedometer cable, and clearance holes for the ¼-20 bolts we are going to use to fasten the mounting flange to the tentacle base. The eight holes for the housings to pass through should be at least .156” diameter, and the ¼-20 bolts need at least .25” diam clearance. For the shaft collar receptacle hole, I just made a .25” deep, .5” diameter hole with the countersink bit I have been using to chamfer the holes in all the cast parts.
Run the cable housings through the clearance holes in the mounting flange. All the various holes in the tentacle base and the flange should line up. If not, well, you are very special.
A couple of ¼-20 bolts and nuts should do it.
For the sake of our little demonstration, I jammed a length of aluminum armature wire into the center hole of the flange. I didn't worry about set screws or anything like that. Ultimately, how a tentacle mechanism gets mounted is entirely a matter of how it is going to be used.
At the heart of our tentacle mechanism are the cables. Cable actuated systems are very common in animatronics, because they are so flexible, both literally and figuratively. When creating mechanisms that mimic organic movement, flexibility and compliance become very important issues. So, for those who are interested in such things, this is some tasty information to have.
You may be familiar with cables from bicycles, but bike cables and housings tend to be too stiif for use in a tentacle mechanism. With the exception of housings, most of the cable and cable related things I need, I can usually find it at www.versales.com . They carry a lot of rigging supplies used onset for film and television. Specifically for this project, I used the following Ver Sales items:
Microcable #90.0 ( 7x7 , .027” diam.)
Nicopress Copper Stop Sleeves (Crimps) #871-32-B
Nicopress Hand Tool (Crimpers) #17-BA
Felco Cutters (Cable Cutters) #C7
The tentacle mechanism measures a little under 3 feet long from one end to the other, so let’s cut each of the eight cables to 48 inch lengths. The first thing to keep in mind about cutting cables is that the cables are of high carbon steel and do require special tools to cut effectively. I strongly recommend that you get the Felco Cable Cutters from Ver Sales mentioned above. The trick to cutting the cables without them fraying at the ends is to anneal the cables with a micro torch before you cut them. Only a small area of the cable needs to be heated for this to work.
With the Nicopress Hand Tool (a.k.a. crimpers) from Ver Sales, crimp a Nicopress Copper Stop Sleeve (a.k.a. crimp) onto the end of each of the eight cables. Insert the end of the cable into the copper crimp and squeeze it real good with the crimping tool. When you crimp down on the crimp, hold it tight for a second or so. That way, the copper sleeve bonds well with the cable so it won’t slip. Then, with the Felco Cable Cutters, cut that little crimp in half on the cable. We are dealing with tight space restrictions within this tentacle and half of that crimp is still plenty strong for our purposes.
O.K. the time has come to cable the tentacle mechanism. Start at the tip if the tentacle and run four lengths of the steel cable through all the discs and down through the spring housings. It is work that requires one to be methodical and have a sure hand. Periodically, the ends of the cables will fray when being strung strung through the little holes in the discs. When this happens, anneal the frayed end of the cable with a micro torch and trim off the offending bit of cable with the cable cutters, and continue on. Once the upper segment is cabled up, continue on with the lower segment. It’s pretty much the same deal.
Alright, at this point the tentacle is assembled and all the cables are installed. Now we need a way to control the tentacle. In Part 5 we will cover cable controllers.
Now that we have our cast parts demolded, they need to be cleaned up. The excess urethane needs to be removed from the castings and the rough edges need to be smoothed before the holes can be drilled out.
A piece of 60 grit sandpaper will work for our purposes. Place the sandpaper on a flat surface and sand each of the discs until they're all nice and pretty.
Most of the holes (if not all) will need to be drilled out because they will be partially filled after the casting process. The holes need to pierce all the way through the cast urethane discs. I used my cheapy little Harbor Freight drill press for this but a hand drill would probably work as well. I also used a countersink bit to chamfer the holes. You might be able to get away with not doing this step, but I recommend getting rid of all sharp edges on the discs to facilitate unimpeded movement.
Center Hole: .125” diam.
Outer Cable Clearance Holes: .062” diam.
Inner Clearance Holes: .156” diam
Before getting into the how of drilling out the holes of the tentacle base, I am going to discuss some of the reasons of why they need to be drilled as I describe.
The base serves the purpose of tying all the elements of the tentacle mech together. The tentacle mechanism is going to be actuated by using cables. This is the same technology used in bicycle brakes and derailleurs: one end of a cable is pulled through a housing, exerting force at the other end. To function properly each end of the cable housing needs to be securely terminated (that is, not allowed to move). We will accomplish this, at the tentacle end, with a press-fit into the tentacle base.
The trick to terminating the housing is to drill a hole of the appropriate depth and diameter to secure the end of the cable housing. The cable housing used in this project are lengths of steel extension spring. The use of steel spring housings is a standard technique in animatronics for film and television, but long lengths of spring housing aren't necessarily easy to come by. There are companies capable of doing custom orders of long lengths of spring housing (upto 25 feet long, typically), but for our purposes I recommend getting the stuff sold online at McMaster Carr. The cable housing commonly used in bicycles tends to be stiff and will inhibit the movement of the tentacle for this particular application. So spring housing is the way to go.
https://www.mcmaster.com/#k12/=1dbqbn1
Pulling steel cables through steel housing, during the operation of the tentacle, generates a lot of friction. To help mitigate that, PTFE (Teflon) tubing will be used to line the inside of the housing. Here is a link to some tubing I found on Amazon that looks like it will work.
https://www.amazon.com/gp/product/B076QBDQFN/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1
The dimensions of this PTFE liner is 1.5mm inner diameter and 1.7mm outer diameter (.059” i.d x .070” o.d.). The Amazon source looks a little dicey, as there isn’t much available at the time of this writing. There seems to be plenty of sellers of this tubing on Ebay, but they are all in China, so it will probably take a week or three to get the stuff. The important thing to keep in mind when sourcing this liner tubing is that the outer diameter must be small enough to fit into the spring housing (.081” i.d. in the case of the spring housing from McMaster Carr) and the wall thickness needs to be pretty thin so that it is flexible. Note: I ordered some teflon liner from China via Ebay and I received 3 weeks later. It looks like it will work just dandy. Why I have to order this stuff from the other side of the freaking planet, I just don’t know...
While I am at it, I will give you the info about the cables. The cable used for this project is from Ver Sales and it is the 7x7 non-jacketed, .027” diam variety with the part number #90.
http://www.versales.com/ns/wire_rope/minicable.html
Isn’t all this technical minutia fun? All these details need to be determined so that we know what size holes to drill in the base in order to capture the end of the spring housing securely.
Here is the spring housing with an outer diameter of .125 inches.
Here is a piece of the Teflon liner protruding from the spring housing. The end of the liner has been split with an X-acto knife so that when the housing is inserted into the hole we are going to drill in the tentacle base, the split ends of the liner will fold back over the spring housing creating a nice, snug fit.
The split ends of the liner folded back over the housing add about .005” to the overall width. So in this case I would use a #30 drill bit (.128” diam.) to create the capture hole in the tentacle base. If the housing has a different diameter than .125” then the hole size is adjusted accordingly. Now that I’ve gotten all that explaining out of the way, it is time for a little less talk and little more action.
If you are using parts cast from urethane resin it will be necessary to drill out the holes of the tentacle base. If you are using a tentacle base that was fabricated on a 3d printer, the holes will be in the correct places but will be a little under-sized, so the following information still applies. However, if you happen to using parts fabricated in a machine shop, your holes are already perfect (presumably) and you may proceed onto the assembly of the 2-stage tentacle. For our purposes, we will assume you are working with cast parts in need of some refinements.
Here is a diagram of the tentacle base showing the layout of the holes. Most of the holes simply pass through the base. It is important that the 4 housing capture holes (marked in red) do not go all the way through, because they need to provide a physical stop for the cable housings. The .062” holes go all the way through so as to allow the cables to pass through.
Once again, I used my little drill press in order to keep the holes straight.
This photo shows the tentacle base prior to any holes being drilled. The .25” inch holes around the perimeter managed to be reproduced during the molding and casting process, but all the other holes need to drilled out.
Here is the base after I have drilled out the center hole and the four clearance holes around it. The center hole will capture the central, flexy-shaft of the tentacle while the four holes will allow the cable housings for the upper segment of the tentacle to pass through.
Here is a photo of the bottom of the tentacle base showing the central hole and the clearance holes.
The next step is to drill the .062” holes all the way through from the top.
After the casting process, all that remains of the .062” diameter holes, through which pass the cables controlling the lower segment of the tentacle, are some little dimples. These dimples will serve as pilot holes. Use a .062” drill bit (or anything reasonably close to that size) and drill all the way through from the little dimples on the top side, through the base. Then flip the base over, so that the cable capture holes can be drilled from the other side.
Using the .130” drill bit, drill down each of the .062” holes to a depth of about .75”. This is the cable housing capture hole. Do not drill all the way through, or it is not a capture hole. It is an entirely different kind of hole.
At this point, I am REALLY bored with describing how to drill holes. And because everything is all about me we are moving on to the assembly phase of the project.
I am going to show to you how to make a 2-Stage Tentacle Mechanism. This design is largely based upon the tentacles I learned to make while working on the film Species 2. A large number of tentacle mechs where required on that job, so a standard, easily replicable approach was developed to facilitate the manufacturing of many multiples of parts. In this project I am going to outline for you, cast urethane components are used extensively. Some machining is required because sometimes metal parts are the only way to go, but we will keep it to a minimum. The process I am going to describe involves technology that is decades old. 3d printing definitely has a role to play in the creation of tentacle mechanisms, but for our purposes here, I am mostly going to show you how to do it old-school (mostly).
Initially, I intended to post this information on the website www.Instructables.com . I didn't really have a good grasp the scope of this particular undertaking, and more I got into the creation of a step by step set of instructions for the creation of a tentacle mechanism, the more I realized this went way beyond a simple how-to article. I am retaining the format of an Instructables article, but it is going to end up being more like a book. Happy reading!
When many multiples of the same part are needed, molding and casting are tried and true techniques for generating those parts. Tentacle mechs are mostly made up of a long series of discs, so molds will be used in this demonstration to produce the discs.
However, before a mold can be made, one must have an object to mold. These objects are called patterns. The original tentacle disc patterns molded for use in the Species 2 movie were made using a lathe and a milling machine with a rotary table. Currently, 3d printing presents an alternative to machining the disc patterns. If the technology of 3d printing had been available 20+ years ago that would have been the way to go for generating those initial disc patterns for molding. That being the case, couldn’t all the discs needed for the movie have been made using 3d printers, rather than going through the molding and casting process? The answer is yes, but hundreds (if not thousands) of parts were required for that project and it still would have made sense to mold and cast all those parts because casting is faster than 3d printing when you scale up to hundreds (or thousands) of parts. However, 3d printing would have been much faster than machining to fabricate that initial run of patterns.
So, if you need a single tentacle, 3d print those parts. If you need a dozen tentacles, molding and casting is the way to go, but 3d printing is the most efficient way of creating the patterns to be molded. In any case, some patterns will be need to be generated, whether they are printed or machined. To this end, I have uploaded a complete 3d model of the 2-Stage Tentacle Mechanism to Grabcad.
https://grabcad.com/library/animatronic-2-stage-tentacle-mech-1
For the purpose of this mold making demonstration, I already have parts ready to be molded. Years ago, after all the tentacles were made for Species 2, there were plenty of leftover parts. So I took a selection of discs home and made my own molds for future use. I will now show you how I made those molds.
Mold Silicone
First and foremost, to make a silicone mold, one requires silicone. Specifically, mold making silicone. Makes sense, right?
Yeah, well. Looking over my little stache of mold making supplies I realized that I didn’t have any mold making silicone, but I did have some silicone that I had bought a while back for the purpose of experimenting with skins. The shelf life of this particular silicone was about to expire, so what the hell, while I try to teach others to do something the right way I might learn something new (by possibly doing it wrong). Looking at the package of this sample-sized silicone kit, it sure seemed like it could work. So, like I said, what the hell.
It turns out, it does work to use skin silicone to make molds. Maybe not as good as mold making silicone... but beggars can’t be choosers. Nor can people who don’t want to go get the right kind of silicone (ahem).
This is what I used (this time): www.reynoldsam.com/product/ecoflex/
This is what I’ve always used before (and recommend): http://www.silpak.com/pdfs/ECONOSIL25PDS.pdf
Mold Release
Mold release makes the life of the mold maker easier. Ask any of them. When I bought that skin silicone sample kit I had also bought a can of spray mold release. The skin silicone I am using is platinum-based so it can be a little sensitive to the chemicals it comes in contact with, unlike the tin-based silicones typically used for making molds. A possible recipe for disaster? Possibly but screw it. Live dangerously. And I really didn’t feel like driving into Burbank to get the correct stuff. Does that make me lazy? Probably. www.reynoldsam.com/product/ease-release-200/
Casting Urethane
I did have the correct casting urethane. The bottles were really dusty and crusty, and this stuff must have a shelf life too, but it still works. This what I used: http://www.silpak.com/pdfs/QUICKCAST.pdf
Isopropyl Alcohol
Cleanliness counts, and for cleaning up goopy residues isopropyl alcohol works great. It also works amazingly well for removing hot glue.
So here is a run down of the various tools and implements I used to make my box mold:
Hot glue gun
Scale (digital or triple beam)
Measuring cup
Plastic container (walls for the box mold)
A pane of glass (a surface to make the box mold on)
Paper towels
Paper cups (waxed)
Mixing sticks
A blob of clay (water based clay prefered, softer than oil-based)
X-acto knife
Small flashlight
Small sturdy knife
Small sharp scissors (cuticle scissors)
Small pokey thing (like a piece of wire or a tiny screwdriver)
Razor blade
If you want to learn more, please visit our website Animatronics Dragon.
Fat felt-tipped pen pen
Needle nose pliers
Step 1: Select and Modify the Plastic Container
The first thing to do when making a box mold is to make the box. A plastic container of a suitable size is what I will use for this demonstration. The container needs to be big enough to allow all the tentacle discs to be arranged inside and still have at least a quarter of an inch of space between all the parts. The plastic container will serve as the walls of the box mold while the pane of glass will serve as the bottom of the mold. The bottom of the plastic container needs to be carefully cut off with an exacto knife.
Step 2: Secure the Discs to the Glass
Once the plastic container has been modified, arrange it upside down on the glass surface so that it’s previously removed bottom faces up. Then trace around the perimeter of the plastic container on the glass with a felt-tip marker. This is the space in which the discs will be molded. The discs should be arranged so that no two discs are closer than about a quarter of an inch. Then fire up the glue gun and secure each disc to the glass with a small blob of hot glue. It does not require more than a small amount of hot glue to secure the discs so don’t go crazy with it. It does need to come apart later.
Step 3: Secure the Plastic Container to the Glass
After all the discs have been glued to the glass it is time to glue down the edges of the plastic container. Place it within your traced boundary and run a bead of hot glue all the way around it so that there will be no leaks. It is important that there are no leaks because the silicone for the mold will be poured into this cavity. You really don't want the silicone leaking out; this qualifies as a minor disaster. Once the hot glue has set up give everything a quick spritz with the mold release. Then it is time to mix up the silicone.
Measure, Mix, and Pour the Silicone
Step 4: Pour Equal Amounts of Parts A and B Into Mixing Cups
The silicone comes in two parts: part A and part B. You will need to estimate how much silicone is needed to cover the discs with at least a quarter inch of silicone mold material. Estimating the amount of silicone needed is one of those things that is easier to do once you have done it a few times. So for now, overestimate the amount you will need. Too much silicone is better than not enough. Pour equal amounts of part A and part B in two separate mixing cups. The amount you pour should be enough so that when the two parts are mixed together and are poured into the mold they cover the discs to the desired depth.
Step 6: Mix Silicone Parts A and B Together
Once you're satisfied that you have equal amounts of part A and part B in the cups, pour them together into the same cup and mix them up thoroughly with a stirring stick. Make sure the silicone is thoroughly mixed, including the material in the bottom corners of the cup. We don't want any uncured silicone in our mold.
Step 7: Pour the Silicone
After the silicone is thoroughly mixed it is time to pour it into the mold. In case there is leakage around the bottom edge of the mold it is a good idea to have a blob of clay standing by that can be used to plug the offending hole. Water based clay works better than oil based clay (like plasticine) because it is softer. Silicone seeping through the cracks will get messy real fast so having a piece of clay to slap over the hole can save the day. Hopefully there are no holes. There really isn’t any other effective means of plugging a hole with silicone seeping through it. More hot glue won’t work and neither will tape because uncured silicone is a slimy mess once it escapes from where it is supposed be.
The only trick to pouring the silicone into the mold is to do it slowly so as to minimize any air bubbles captured in the nooks and crannies of the disc patterns being molded. Take it slow. There is no hurry. The silicone will remain runny for a fair amount of time. However, the holes in the disks are small enough that it's pretty much guaranteed there will be air bubbles trapped inside.
Step 8: Get the Bubbles Out (Some, Anyway)
The air bubble situation can be mitigated, somewhat, by using a narrow piece of wire to break the air bubbles loose. I used a little jeweler's screwdriver but even a paperclip would serve the purpose. Just poke the “poker” down into the little holes of the discs and stir the silicone up a bit. Some air bubbles are unavoidable, but the less bubbles we have at this point of the process the less cleanup the cast discs will need later on.
There are more effective ways of eliminating bubbles in the silicone involving subjecting the mold to a vacuum to suck the air out, but that is a little beyond what I am presenting here. Some bubbles ain’t gonna kill us. Infact, my little “stirring the bubbles out” technique may be utter hogwash, but that’s how I roll.
Step 9: Mold the Base Too
In addition to the discs, I am also going to make a mold of the base of the tentacle, if only to use up the rest of the silicone I have on hand. As I mentioned earlier the stuff has a limited shelf life and this particular batch has been sitting around for awhile. The process for molding the base is exactly the same as with the discs: secure it to the glass with a small dab of hot glue, create a wall around the base disc with some sort of container, hot glue that wall down to the glass, give it a spray with mold release, and pour in some silicone.
Step 10: Let the Silicone Cure
Let the silicone set up overnight. The instructions that came with this silicone kit stated that the mold will be set up in 4 hours but it will be even stronger if allowed to cure overnight.
Step 11: Remove the Plastic Container
Once the silicone is fully cured it is time to remove the disc patterns from the mold. The first thing to do is to remove the hot glue securing the edges of the plastic container to the glass. This is easily accomplished by soaking the area where the hot glue adheres to the glass with isopropyl alcohol. The alcohol will break the bond of the hot glue and make it easy to peel off. Then carefully remove the plastic container from the silicone mold leaving the silicone and the disc patterns still on the glass.
Step 12: Separate the Discs From the Glass
The discs will still be stuck to the glass with hot glue and the silicone mold will have thoroughly encapsulated the disks so removing the discs from the mold must be done with care. You don't want to damage the mold. So go ahead and use the alcohol trick to help break the bond of the hot glue between the discs and the glass. Carefully use the tip of a sturdy knife to pry the discs up off of the glass. The alcohol will help immensely.
Step 13: “Surgically” Remove the Disc Patterns From the Silicone Mold
Once the mold has been separated from the glass, the silicone will require some careful trimming so that the disc patterns can be demolded with as little damage to the mold as possible. I used a sharp X-acto knife and a razor blade to surgically remove the disks. I then trimmed the edges of the mold with a small pair of cuticle scissors. Some imperfections in the mold are very likely to be present but they're not a big deal. We'll just have to clean up the discs after they have been cast.
Once I had gone through the mold making process with this particular silicone I really started to wonder if I had made a mistake. Like I mentioned earlier, this silicone is not necessarily meant for making molds. It is specifically meant for casting skins for animatronic creatures or masks. Sure enough, the packaging of the silicone kit describes the mold making process but this really is not the best silicone for making molds. However, it looks like it will work. Comparing it to some older molds I made with actual mold making silicone it is apparent these molds are significantly softer but they will serve for demonstration purposes.
Step 14: Determine the Volume of the Molds
The molds are ready for use. We now need to determine the amount of urethane casting resin needed to fill the molds. My favorite technique for determining the volume of casting material needed is to fill the mold with water and then measure the volume of the water in a measuring cup.
Step 15: Figure How Much Urethane to Mix
In this case, the base mold had a fluid volume capacity of 2 .75 fluid ounces. The base will require considerably more urethane resin than the discs, so once the volume requirements of the base mold is known I am just going to estimate the additional amount of resin needed to fill the discs mold. Let’s call it 1 fluid ounce.1 fluid ounce plus 2.7 fluid ounces divided in half equals about 1.9 fluid ounces. So, to fill the base mold and the discs mold, we will need 1.9 fluid ounces of part A and 1.9 fluid ounces of part B of the casting urethane resin measured out in separate cups.
Step 16: Prep the Mixing Cup or How to Not Gunk Up Your Measuring Cup
Urethane resin is fairly yucky stuff, and if you pour it into a measuring cup, that measuring cup is pretty much unusable for anything else, ever again. So for the sake of the longevity of your measuring cup, we are only going to be pouring the urethane into disposable paper cups. What we want is 1.9 fluid ounces of part A of the urethane in a disposable mixing cup. So, measure out 1.9 ounces of water in your measuring cup and pour that into a paper mixing cup. Then, mark the level of the liquid on the side of the cup with a felt tipped pen. Use a small flash light to shine a beam of light through the paper cup so you can see the level of the water in the cup, and then make the mark with a felt tip pen.Then, pour the water out of the mixing cup. Be sure to get all of the water out of the cup before using it to mix urethane because moisture in the curing urethane causes lots of funky little bubbles and voids in the finished cast part. At this point, we have a mixing cup marked at the level it will be filled with part A of the casting urethane resin.
Step 17: Measure Out Equal Amounts of Parts A & B by Weight
Now, fill the mixing cup with part A to the level indicated by the mark. Then, weigh the cup of part A on the scale. Then, fill another cup with part B that weighs the same as part A. We now have two equal parts of the casting urethane resin ready to be mixed and poured into the mold. The part A and part B constituents of the casting resin need to be measured out in equal proportions according to their weight, not their volume. For this reason I don’t recommend just eyeballing the amounts, unless you have some experience using urethane. In the past, I have often just eyeballed the part A and part B amounts allowing for a greater amount of the lighter part B and not had any problems, but for the sake of this demonstration I am going to go ahead and measure the part A and part B with a scale. In the side-by-side comparison photo you can see that the clear part A is of a larger volume than the amber colored part B.
Step 18: Prep the Molds
Once the A and B parts of the urethane resin are measured out it is time to prep the molds for casting. Arrange them on the glass in case there is some spillage. The glass is easier to clean than a benchtop. Then, give the molds a spritz with the mold release.
Step 19: Mix, Pour, Repeat…
In a mixing cup, combine the part A and part B of the urethane resin and stir vigorously. Be sure to work somewhat quickly because the urethane will catalyze within just a few minutes. Once it is mixed, pour the urethane into the molds. You have only a minute or three before the urethane sets up so don’t lollygag.
Once the urethane begins to set up it will get hot. It will be completely cured once it has cooled down to room temperature. That should take no more than 10-15 minutes. Then you can remove the castings from the silicone.
Now repeat the process until there are enough parts to make a tentacle mechanism. We need three of each size of the discs, and we need a total of NINE of the base components, because, in addition to the tentacle we are going to make cable controllers.
Now, after all that, we get to cleanup and “bodyshop” all the cast parts. Yay! And yes, bodyshop is a verb.
I was introduced to tentacle mechs early on in my animatronics career. One of my first jobs was on the movie Species 2 (). The first Species () did well enough that, when the time came for a sequel, it was decided that Species 2 was going to be an all out special effects extravaganza. This was before computer graphics had really come into its own so a whole bunch of practical effects was thrown against the wall to see what would stick. I will refrain from commenting on the efficacy of that approach, but I will say that I sure did learn a lot on that job.
The creature effects team at XFX had already done a bunch of R & D on tentacles for the previous Species, as well as for the film Anaconda (), so by the time I got there, they had tentacles pretty much dialed in. I will share with you some of these techniques.
Before going any further, however, I want to discuss the use of tentacles in the movie Little Shop of Horrors (). I just watched it again for the first time in decades and the performance of the tentacles still impress. They worked so well for two reasons: the tentacle puppetry was really well rehearsed, something that just became more and more rare as time went on, and the performance of the animatronics was filmed at a slow speed so it could be sped up on playback. These two factors really made all the difference with the quality of the performance.
So, with all that in mind, let’s get into some of the nitty-gritty of tentacle mechanisms.
Over the years I have employed tentacle mechs in a wide variety of applications:
Tentacles (duh)
Cat tail
Horse neck
Snake bodies
Demon tongue
Robot monkey arms
See the theme?
There are a number of ways to make a mechanical tentacle. I will concentrate on the approach that is best suited for the creation of animatronic puppets.
Flexible central shaft
Series of discs
Cable actuation
Two stacked segments
At the core of the tentacle mech is a shaft of material that is flexible, doesn’t twist, and is tough. I’ve used nylon rod and urethane tubing but the best stuff to use are shielded hydraulic hose (for the bigger projects) and the steel flex shafts from speedometer cables (for smaller projects). The shaft needs to be durable enough to be clamped onto or have set screws bite into it without kinking or being damaged and the speedo cables and hydraulic hoses are the bomb.
The discs are made of rigid material and are perforated with holes for the central shaft and cables to pass through. In a tentacle, the discs are strung upon the central shaft like beads on a string. For the optimal functioning of the tentacle, it is important that the discs not spin upon the central shaft. Set screws and shaft collars are good methods of securing the disc to the shaft. Not all the discs need to be secured, but the longer the tentacle mech, the more important it is to secure all the discs. Another thing to keep in mind is that the cables literally saw back and forth in the holes of the discs. So the more durable the material the discs are made of the better. There are ways of mitigating the sawing cables issue, but I will discuss more on that in another post when I get into the specifics of tentacle construction.
There is a specialized disc called the termination disc. It is the disc where the cable housing terminates and is secured. These are generally made of aluminum, the reasons for which will become obvious when we examine tentacle construction in detail.
Cable actuation is the lifeblood of many animatronic creatures. This is especially true in tentacles. Again, I will go more into the details of using cables to move things when discussing specific applications.
A single tentacle segment is frequently all that is called for a particular application, but sometimes you need to go that extra added segment to really bring a project to life. More than two tentacle segments becomes a self-defeating proposition, as the segments are impossible to control and just work against each other. The next post will deal with all the gorey details of creating two-segment tentacle mechanisms.
I was always interested in 3d printing for the purpose making functional mechanisms. I still had all my mechanical drawings from that previously mentioned Terminator project and decided to 3d print an articulated robot hand. Not just any robot hand, but something capable of dragging itself along under its own power. I wanted to see what 3d printed mechanisms could do.
The Terminator hand design was based upon the T-800 Endoskeleton Arm that used to be available from Sideshow Collectibles. That model was apparently molded directly from a working terminator arm made at Stan Winston Studios for use in the movies. It had the screw heads and accommodations for finger linkages already laid out and the dimensions were exactly what they needed to be to recreate the functional mechanism.
The mechanisms for the T-800 hands were originally fabricated using traditional machining techniques ( of course). This meant there were plenty of places to grip the parts in a vise or a 4 jaw chuck for the machining process. These features translated well to the technology of FDM 3d printing, where it is desirable to have flat surfaces to attach to the print bed.
I was first exposed to stereolithography (SLA) back in . This was before the term 3d printing was in common use. It was on a job involving the recreation of a full-scale Terminator robot for the Sarah Connor Chronicles television series. Certainly, a lot of man hours were saved in sculpting but the parts were of a pretty low resolution and required extensive cleanup and body shopping (sanding and bondo). The resin was also very brittle. It was interesting but I didn't think too much of it at the time.
A couple of years later, I started hearing rumors about 3d printers. “Oh great,” I thought, “more brittle plastic parts eating countless hours of bondo and sanding time.” It was a different technology, however, similar to a computer controlled hot glue gun, called fused deposition modeling (FDM). Promising, but still not ready for primetime.
Then in , Make Magazine came out with a special edition featuring 3d printer reviews. In , based upon these reviews, I bought a Type A Series 1 3d printer. Then in , I actually pulled it out of its box, taught myself to use a CAD program and started printing. Can you say no longer under warranty?
Machining has been the mainstay of animatronics fabrication for a long time. Aluminum, brass, and steel were some of the materials of choice in the creation of animatronic mechanisms, and for many applications still are. However the machining process can be labor-intensive and milling machines and lathes are big, heavy, and pricey. The introduction of 3d printing technology has changed the situation.
Robot time versus man hours: Complexity not an issue. An intricate part and a simple part of the same size require the same amount of time to make.
Only requires occasional supervision.
Rapid prototyping/quick iterations: different ideas can be tried out in rapid succession.
The equipment is more affordable, accessible, and mobile than metal working equipment.
A wide variety of materials are available, each with their own potential specific applications.
Makes for a lighter finished product.
Plastic is weaker than metal.
Plastic is less stiff than metal.
The layers of weakness inherent in the printing process.
Parts need a flat side to attach to the print bed.
Plastics generally don't machine and thread well.
The equipment can be finicky and lacks the longevity and durability of milling machines and lathes.
Susceptibility to failures:
wrong settings
clogs resulting from contaminated filaments ( dust, moisture, etc.)
sensitivity to variations in temperature
and don't bump the machine when it's working, for God’s sake!
Does not scale: there no advantages to making a run of multiples. Its all about the prototypes.
Maintenance of equipment. You will become an expert in troubleshooting these things if you stick with it for any length of time. Hope you enjoy tinkering and talking to tech support. You will be doing a lot of it.
Here is a run down of the materials I have used in my 3d printing and what I can say about them.
Body shops great (bondo, sanding, filing)
Gluing and welding excellent
Prone to warping during printing
Cheap
Needs a heated bed to work best
Biodegradability not a selling point when going for durability and longevity
Texture is glossy and waxy. Not conducive to body shopping
Not as prone to warpage as ABS
Does not need a heated bed. Probably it's best feature
Save the wear and tear on your machine and don't use it
Has improved stiffness characteristics
I like the matte black carbon fiber color
Tough
Very flexible
Great for special applications:
small joints
user interface pads ( hand grips, wrist pads, headbands, etc)
Has all the desirable qualities of nylon but it is easier to work with.
Tough
Somewhat flexible
Holds a memory/ snaps back into shape when flexed
Slick surface = less friction (makes a good bearing surface)
Threads well
Good wear characteristics
Strong
Machines and threads well
Slightly gummy texture does not body shop as well as ABS
Needs high temperatures and a heated bed to print
Stinky and fumey during the printing process
Functional Design for 3d printing by Clifford Smyth
The Zombie Apocalypse Guide to 3d printing by Clifford Smyth
These books are full of great information and practical tips for anyone interested 3d printing functional components for animatronics. The the first book is concerned primarily with aspects of the design, as you may have figured out from the title. The Zombie Apocalypse book is more about getting optimal results from your 3d printer. There is some overlap between the two books but both are worth having. The author also maintains they very informative blog at Threedsy.com .
We live in interesting times.
3d printing is becoming a mature technology and can be very useful and making mechanisms. I once heard 3d printing described as a lifestyle enhancement. Speaking as someone who is tired of expending my valuable hours planted in front of a lathe or milling machine, I heartily agree.
Man hours are expensive. Robot hours not so much. The more work that can be delegated to some form of CNC (computer numerical control) machine, the better.
3d printing has its strengths and it has its weaknesses. The trick is to play up its strengths and avoid the weaknesses. Over the past several years I've been experimenting with 3d printing parts for animatronics mechanisms with mixed results. Sometimes plastic just won't do and metal fabrication comes back into the picture. However, there is a lot plastic can do.
Since I built the original thrashing torso, all those years ago, a few things have changed.
3d printing is one of these changes. Another change is the introduction of new off-the-shelf products meant for use in Halloween haunted attractions. I've recently been introduced to the Spider Joint from Spider Hill Prop Works. It is a versatile plastic joint used in conjunction with 1 inch pvc pipe to create body armatures. The new Thrashing Torso also uses a 12-volt motor from Frightprops, another haunted attraction oriented business. Additionally, a plastic halloween skeleton is used. Some modifications are required and this proved to be the most time-intensive part of the build.
A few other refinements have been incorporated into the design. Bungee cord is used instead of springs and clothesline from the hardware store is used instead of steel cable.
All of these changes make the finished mechanism much easier to make, the parts are easier to acquire, and everything is much lighter.
Man Hours Required: ~30 hrs.
I spent about a half a day designing the mechanism in Fusion 360 and I estimate another half a day setting up for the 3d printing and cleaning up the parts as they came out of the printer. Only another day was spent assembling the mechanism. Modifying and mounting the plastic skeleton took the better part of two more days, which includes futzing around with clearances and tensioning the bungee cords. Considering how much time the first Thrashing Torso took to build, this is great.
Cost: ~$158 (total)
Plastic Skeleton $40
Fright Props motor $25
Spider Joints (x8) $28
Misc. Components and Fasteners ~$40
ABS Printer Filament Roll $25
There has been quite a bit of hype about 3d printing over the past few years. 3d printing is incredibly useful but let me come right out and say that it can be a pain in the ass. The technology is getting better and better all the time, but like any fabrication technique, it takes time to master. This is especially true when 3d printing functional mechanisms. In the next post (or three) I am going to discuss 3d printing and its applications to making animatronics, as well as some of the pitfalls you may be able to avoid.
Time Required: ~80 hrs.
Cost: ~$200
These are estimates based upon my recollection of this project from 25 years ago, so take them with a grain of salt.
The design process was been broken down to the basic steps and then applied to this specific project.
Translating the geometry of organic movement into mechanical movements is a fundamental aspect of the design process so I tried to be as clear as possible about how this is done.
Always try to use off-the-shelf components whenever possible.
Aluminum angle and channel stock available from any hardware store was used throughout the Thrashing Torso project.
Repurposed electrical cable and PVC pipe was used for the rib cage.
That electroformed skull and jaw were definitely NOT off the shelf items, and a plastic Halloween prop skull would probably work just as well.
Screws versus rivets. Basic rule: rivets are good for things that will never need to come apart. Screws for everything else.
A very rudimentary application of cables to drive the mechanism was demonstrated. This was suitable for this application but not a typical example. More refined applications of cable control will be explored in future projects.
Gravity was used to move the Thrashing Torso into its slumped position with the assistance of springs. It is best not to rely too much on springs. They can be finicky and prone to fail. A trained gunsmith or an engineer can make use of springs to their greatest advantage but the rest of us are generally just guessing when we slap springs into a project.
The way I chose a motor for this project 25 years ago is pretty much the same way I do it now. I get as big a motor as I can to do the job, within reason. Motors that don’t have to work hard last a long time.
The use of cranks should be well understood by anyone who has peddled a bicycle.
Make the best use of leverage. Initially, I learned to use leverage by doing it badly. Best to understand leverage and how to use it to its best advantage as soon as possible. A lack of leverage and mechanical advantage will result in jerky, uncontrolled movements, if anything moves at all. The leverage in the mechanical Thrashing Torso is achieved by keeping the drive cable up off the segmented spine mechanism with metal angle brackets with holes drilled in them. It was crude and involves a lot of friction but it worked.
Controlling a complex puppet is often dependent upon limiting movements to what is absolutely necessary. Sometimes this can be determined only through trial and error. Start with your best guess as to what is necessary and adapt your approach as needed. Let the mechanism tell you what it needs. All the various joints of the Thrashing Torso move in the same plane, forward/back, and each joint is limited as to how far it can pivot forward and backward.
Fabrication is a huge topic and I really didn’t want to get into it here. There are a couple of books I will recommend for those who are interested in learning the basics of fabrication.
“Making Things Move” by Dustyn Roberts
“Robot Builder’s Bonanza” by Gordon McComb
For those with only limited experience these books are full of really valuable information.
I’ve created a CAD model of the Thrashing Torso and uploaded it to GrabCad. It includes some minor refinements in the design but it is fundamentally the same as what is presented here. Anyone interested in making their own Thrashing Torso can download the files and get all details and dimensions.
grabcad.com/library/animatronic-thrashing-torso-1
I am going to deconstruct the Mechanical Thrashing Torso in an attempt to provide some idea of how it was originally made. This is not going to be a step by step tutorial, as this project is several decades old, and frankly, I just don't remember what the steps were,
Someone once said a picture is worth a thousand words. It is time to put that to the test.
Stripped Down Spine Mechanism
This is an uncluttered view of the segmented vine. The full range of motion allowed by the mechanical stops can be demonstrated to full effect.
Joint Assembly
Here are a series of images demonstrating how the segmented joints come together. Pop rivets were used in assembling the joints because they're cheap and strong and will not rattle loose with use. A piece of steel tube, slightly longer than the width of the joint, provides the bearing surface upon which the joints pivot. 1/4-20 bolts and nuts secure the segments together without the need for lock nuts or Loc-Tite. Tightening the nuts and bolts down upon the ends of the steel tube insert was enough to keep everything from shaking apart during use. No lubrication was ever required either. Clanking and rattling was all part of the effect I was going for.
Rib Cage (front back and interior)
These images detail the rib cage. Extruded aluminum channel provides structure(sternum and backbone). Bolted to the aluminum are pieces of PVC pipe, and inserted into the pipe are segments of thick insulated electrical wire. This assembly of randomly found crap made a pretty satisfactory rib cage. Aside from the motor and batteries this is the heaviest part of the whole figure, which actually helped in getting the torso to convincingly flop around.
Motor, Mounting Plate, and Crank Arm
The drive motor is a 24 volt electric motor I bought surplus. I didn’t know what its specs were and I selected this motor on the basis that it was big and beefy. I was told that this particular motor was for an electric gate. Basically, a nice, hefty motor that is geared to operate slowly, say around 1 or 2 rotations a second, has a good amount of power. The motor is simply mounted to a piece of quarter inch thick aluminum plate which is attached to the base with a bracket.
The rotational movement of the motor is translated into the linear motion of the torso my means of a crank. As the crank rotates to its highest point the cables slacken, allowing the spinal column to slump forward, and when the crank rotates down to its lower position, the cables tighten up pulling the spine into an upright position.
Cable Guides
In order to get enough leverage to move the torso, the cables need to be held away from the segmented spine. The drive cable pulls through holes in steel "L" brackets. A crude but effective set up. Eventually the stainless steel cable will saw through the softer metal brackets, but so far so good.
Cable Guide Breakage
Shown in this image is a point of failure an aluminum bracket that became over stressed when I was running the thrashing torso as fast as it could go. This is why some sort of speed control on the motor is a very good idea. This thing could tear itself apart pretty quickly if it was allowed to. Now I get to fix this.
Lower Cable Termination
This was a quick and dirty way of securing the loose ends of the cables and creating a mechanical connection. This is a small rope clamp I picked up at the hardware store with a piece of brass attached to it. The hole in the brass allows it to be connected to the drive motor cam.
Upper Cable Termination
Two separate cables run up the back and attach to the uppermost spine segment and the base of the skull. At the lower end both cables terminate together but are routed separately up the spine so that the torso and the head move independently of each other.
This image shows the upper cable terminations for both cables as well as the steel brackets serving as cable guides. There are much better ways of doing this but it got the job done.
Neck Detail
Here is the torso with ribs removed so that the positioning of the cables is clearly seen. You can also see the cable that secures the jaw to the spine, causing the mouth to open when the spine goes erect. To operate correctly the cables required a bit of fine-tuning to get the lengths correct and the travel of the mechanism something close to what I wanted. Also shown is how the return springs were mounted, which cause the torso to tend to curl forward into a fetal position.
Head Details
I allowed myself to get a little artsy-fartsy with the head. The skull is an electroformed copper shell. The copper was deposited a upon a wax sculpture coated with conductive paint. Once the layer of copper plate was thick enough I melted out the wax. That technique is called the lost wax technique. When I first heard of the lost wax technique I wondered how do you use a technique that's been lost. Yeah, pretty stupid. I attempted to plate over the copper with nickel but it didn't really turn out. It gave it a funky, industrial appearance, so I was happy with it. The eyes are 12 volt incandescent indicator lamps.
The Mechanical Thrashing Torso was my first attempt at designing and fabricating a mechanical system for emulating organic movement. As such, it is a good starting point for a discussion about the creation of animatronic figures. The process I went through to create the Thrashing Torso is the same I've used ever since.
The Mechanical Thrashing Torso could accurately be described as a single axis, cable-actuated tentacle mechanism with spring-assisted gravity return and a high-torque electric motor with a crank. That description is accurate but unnecessary. At the time I made the Thrashing Torso I was not familiar with most of these terms. However, I was familiar with the things like levers, pulleys, and springs, thanks to a childhood spent disassembling my toys and bicycles. This experience, plus a rudimentary knowledge of tool use and fabrication techniques, was enough for me to figure things out.
The design process for an animatronics figure can be broken down into four basic steps:
-Establish Form
-Determine Movement
-Decide How
-Make A Plan
Form follows function is a basic rule of design. However in animatronics this axiom usually gets reversed. Typically, you are given a form and from it you figure out the functions. Whenever I've been called upon to create an animatronic figure the form has usually been decided upon and is presented to me as a sculpture or other type of concept art. This usually works out but it is important to keep in mind some fundamental rules of physics. For example, long spindly legs or giant wings may look great from an aesthetic standpoint but they are mechanically difficult to move. Leverage and mass dictate what moves and how. An octopus can’t gallop and giant flying dragons don't fill the sky. If there is not a good example of what you want to do in nature then it probably can't be done. Stick with what already works and function will follow form.
Once you have a thing you must decide what the thing is going to do. What do you want to move, how far, and how fast? The success of an animatronic project is largely a matter of movement quality. Quality movement is the movement that best pleases the eye and meets or exceeds the expectations of the viewer.
We are all creatures who have evolved surrounded by other creatures. We are all hardwired to respond to those other creatures in very fundamental ways. So everybody is an expert on how living things should move and behave. The animatronic creature needs to play upon that fundamental programming we all share. Is it familiar, is it new, threatening, or friendly? Reference photos and videos are invaluable in determining appropriate movement. Real life examples are even better.
Range of movement, speed of movement, and control of the movement are all determined by basic mechanical and biomechanical principles. You don't need to study engineering to learn how to use them but there are some basic laws of physics that can not be ignored.
In animatronics there are some tried-and-true techniques for achieving a desired performance. Cable control, servos, and direct physical manipulation (a.k.a. puppeteering) are all common means of moving animatronic figures. We will explore a wide range of techniques when discussing specific projects in future posts.
This may be as simple as where to start and how to end. Parts and materials need to be sourced and obtained. Then everyone involved in the project needs to be coordinated with in order to keep everyone on the same page. A simple project = a simple plan. A complex project with no plan = problems.
The mechanical thrashing torso was created to be part of a haunted house attraction for halloween. As such it needed to have three primary characteristics:
-Be relatively simple
-Be low maintenance
-Have big, scary movements
The initial idea was to have a limbless human torso wrapped in plastic trash bags lie motionless until an unsuspecting victim passed by, at which time it would begin thrashing around. The garbage bag idea was discarded once it became apparent just how cool the mechanics looked by themselves. Perhaps the amputee in a trash bag was the scarier concept but I decided to feature my handiwork instead.
I used myself as the model for the Thrashing Torso. Measurements of my own anatomy determined the size and proportions..
All that was required was to approximate the motions of a spasmodically thrashing human torso. One big cyclically repeating motion operated by a single electric motor. Simplicity is always best.
I am a visual thinker so I always start with a drawing. For the Mechanical Thrashing Torso I created a drawing of a human torso in two positions: fully erect and fully slumped. This served as a graphic representation of where the movement begins and ends. The spinal column was divided up into jointed segments that approximated the articulation of a human backbone.
In order to control the motion of the spinal column each joint needs a mechanical stop, limiting how far each joint can pivot. The physical stops in each joint define the configuration of the spinal column at the erect and slumped positions.
The movement of the spine is limited to a single plane (or axis) and each joint is limited in its range of movement. This allows the mechanical thrashing torso to move in a controlled way. Any more axes of movement and the thing will flop around like a rag doll.
Once the spinal column is assembled and the full range of movement is established, the length of the driver cable can be determined. Cable travel is the length of the pull required to move the spinal column through its full range of movement. Once the travel of the cable is known the length of the crank arm on the drive motor can be determined. The placement of the motor in relation to the torso should also be determined at this point.
When I build an animatronic figure, I find it is helpful to design only up to a certain point and then start building. If I try to design everything out completely, and then start building, all too often much of my design has to be reworked as the build proceeds. That is wasted effort. In the case of the Mechanical Thrashing Torso, I designed and built the spinal column, and then figured out how the cable and motor would work to the best effect.
Not much to it. I had very little in terms of a budget, so for materials I scrounged up what I could and bought what I had to. The first thing to build was the segmented spine, followed by the base, the cable/motor drive system, and then the head. My plans always tend to be somewhat vague and consist of broad conceptual strokes. This because unforeseen issues always arrse and sometimes one must zig when when the original idea was to zag.
Introducing the Mechanical Thrashing Torso! It was created back in my art student days as part of a haunted house attraction. It is a good example of what can be accomplished with limited experience, tools, and materials. The Thrashing Torso consists primarily of a segmented spinal column driven by a cable pull from a single electric motor. After 25 years it is still functional, though it has seen some maintenance and upgrades in that time.
Hacksaw
Electric hand drill
Drill press
Hand file
Bench vice
Pliers
Screwdrivers
Crescent wrench
Rivet gun
Tap handle
Taps
Drill bits
Extruded aluminum square tubing
Aluminum flat stock
Misc. fastener hardware (nuts, washers, screws, pop rivets)
Misc. springs
Misc. wire and connectors
.125” steel cable
PVC pipe
Salvaged chunks of electrical cord (rib cage)
DC motor
Power supply
Misc. brackets and “things”
This materials list is left pretty vague because a lot of scavenged materials went into this project and it was a long time ago.
A quick glance at my workbench gives me a good overview of the hand tools I use the most. Here is a rundown, in order of frequency used (sort of):
Cordless drill/screw gun
Allen wrenches (ball-end)
Calipers (digital or dial)
Pliers (small needle nose, big needle nose, channel locks, vice grips)
Screwdrivers (Phillips, flat, all sizes from tiny on up)
Exacto knife
Drills, center drills, counter sinks, and taps
Tap handle
Drill press vice
Adjustable crescent wrench (small)
Ball peen hammer (small)
Dremel tool w/ bits
Wire cutters and dykes (small on up)
Forecepts
Micro torch
Clamps (various sizes of c-clamps, Kant twist clamps, and spring clamps)
Files (from tiny needle files on up)
Cable cutters
Crimpers
These are generally what is in my tool bag when I show up on a job.
Here is a list of the power tools and equipment used in the making of animatronics:
Bench grinder (small)
Bench grinder
Drill press
Bandsaw
Belt sander (small)
Belt sander
Air compressor
Welder
Chop saw
4 inch angle grinder
3d printer
3d scanner
Lathe
Milling machine (bench top)
Milling machine
An individual with a small workshop isn’t necessarily going to have all of these tools, but some of the tools listed will prove invaluable from the start. I would include a small belt sander and drill press in that category. Others, while very useful, can be difficult to accommodate in a typical home workshop (like a full-sized milling machine).
Making things generally requires tools. “The right tool for the job” as the saying goes, but what is the right tool? The tool that gets the job done without being too expensive is the right tool. Here are some sources for tools:
If you are looking for more details, kindly visit Simulation Animals Supplier.