[Nicholas]

Hey, everyone. I thought I'd share a project I've been working on for the past few months. For years (literally) I've wanted to build a 3-axis skull, and this last Halloween season I finally started to work on the project. Here you can see my third attempt in motion on my workbench:

As you can see briefly in the video (when it shows the plate without the skull), it's based on halloweenbob's design.

V1

My first attempt at the project was in the months leading up to Halloween. I dove head first into the project without a ton of planning. I knew there was a lot of knowledge I'd have to obtain by just doing. I also was under some tight time constraints if I wanted the prop working for the 2014 season.

I created the clear plexiglass plate by hand in my garage, using mostly a plunge router:

I made a few mistakes along the way though. The biggest mistake I made was using too narrow of a rod. The rod was too narrow to really handle the stress I was putting on it, and the animatronic ended up only being active for one night of the haunt. For the one night that it was active, I had a programming bug that prevented the character from being fully animated. And I was out of time -- when I had any chance of finally rectifying these issues the season was over.


V2
After Halloween, I went back to the drawing board. I realized that my attempts at cutting the plexiglass in my garage were not precise enough. The slop in the design meant that I couldn't get the servos exactly where I wanted, they wouldn't be held in place as tightly as I had hoped, and overall the entire system just lacked the precision I needed for the stress I was putting the animatronic under. It was time to pull out the big guns. I contracted a laser cutter:

The laser cutter milled my second design with sub-millimeter precision into 0.25" plexi. What a world of difference! The plate was gorgeous.

Designing purely in 2D and then imagining how it would assemble in 3D is a tricky thing to get right the first time though. I missed some measurements, misjudged how wide the laser kerf would be, and overlooked a few other issues in the design. The plate was good for taking measurements off of for refining the design, but it couldn't actually be used in an animatronic.

V3
And that brings us to the third design, the one in the video. After some revisions to the design I sent it off to the laser cutter again, and received this some time later:

The design still isn't perfect, but it's entirely usable (as seen in the video above.) The main rod comes in at a hefty 3/8" thick. The skull is a standard Lindberg skull. The four servos are being controlled by an Arduino with an Adafruit servo board. I do software development for a living, so running this project with an Arduino was just the natural choice.

The project is still far from complete. The software still needs a lot of refinement, and I've got some some other cool features in mind that I'll post as they progress.

 

[Abunai]

Nice.
I'll be following this thread.
I built four of the 3-axis skulls while the HalloweenBob version was being developed here.
They've held up pretty well over the years; only needing a few minor part replacements.
Eventually, I'll need to upgrade to some new ones.

 

[bobzilla]

Very cool!
Nice work so far

 

[Nicholas]

Thank you both.

Today I managed to get some time to work on this project a bit. I mentioned that I was worried about the stress being put on the animatronic, and so I was overbuilding a bit. This may help to clarify why that's such a big concern...

Yup... that's a wheelchair that the character is sitting in. My objective is to actually make the character mobile. This may be a bit of a crazy goal, but I figure that the worst case scenario is that I end up with a stationary but very well built animatronic.

I picked up the wheelchair at a local thrift store. As luck would have it, the chair actually has a rigid detachable back. This makes it a perfect place to attach the head. So today I focused on building an attachment point for the neck rod:

For the body I'm just using a Costco skeleton. I carved out a portion of its spine to fit the head attachment mechanism.

The strange things seen in a haunter's home:

 

[threader]

Very nice! I am impressed.

Did you decide not to use the Costco skull (that came attached to the body) because it didn't look as realistic as the Lindburg or where there complications with the Costco skull?

It looks like you have an Arduino Uno with an audio shield. Is the servo driver what is connected to your breadboard? How did you get the jaw to sync with the audio? Do you have a Scary Terry board inside the skull or were you able to do it with software (Arduino code)?

This is very cool and is something is something I am trying to pull off for this year, but haven't been able to put the time in yet.

 

[Nicholas]

Very nice! I am impressed.

Did you decide not to use the Costco skull (that came attached to the body) because it didn't look as realistic as the Lindburg or where there complications with the Costco skull?

It looks like you have an Arduino Uno with an audio shield. Is the servo driver what is connected to your breadboard? How did you get the jaw to sync with the audio? Do you have a Scary Terry board inside the skull or were you able to do it with software (Arduino code)?

This is very cool and is something is something I am trying to pull off for this year, but haven't been able to put the time in yet.

Thank you.

I actually had the Lindburg skull in-hand before I ever knew what I was going to be doing for a body. I figured I'd end up with a bucky body, and I knew I didn't want the weight of a full bucky skull. The Costco skeleton bodies ended up being good enough though (especially for the price) that I don't think I'll end up with a bucky.

That's correct. I found a little hack for the Adafruit Wave Shield where you can tap off a certain pin on the board to grab the audio level. You then feed that into one of the analog pins on the Arduino and can read off of it in the software. It takes a little bit of additional software smoothing to make it look decent, but I was glad to not have to involve additional hardware.

I have a bunch of parts arriving from Hong Kong over the next few weeks, so I'm hopeful I can post some updates soon.


Edit: In case someone needs it, this is the audio level hack that I used: http://forums.adafruit.com/viewtopic.php?f=31&t=33020

 

[weaz]

Do you have any close ups on the neck build and/or a list of parts used? This is where i seem to be stuck.

 

[Nicholas]

After waiting for several shipments of parts from all over the world, I finally had what I needed to progress on this project. Here's a quick video of the character running 100% on battery power:

And a still of the electronics case so far:

The amp still needs some work, so I'm running off the little portable iPod speaker, but hey, it's still completely battery powered.

That's a 12v, 20ah battery sitting to the left of the electronics case. It should be enough to run the servos and the electronics for an entire night, with plenty left over (I could actually probably get two nights out of it.)

Do you have any close ups on the neck build and/or a list of parts used? This is where i seem to be stuck.

Hmm, I can see if I can get a pic, but with the head fully assembled and on the wheelchair it's not real exposed. You can see it pretty well though in the video I included in the first post.

 

[pcman312]

Where did you get the ball joint that you are using? I've been trying to find something like that, but I can't seem to figure out the proper name for it as my google-fu is failing me.

 

[Nicholas]

Where did you get the ball joint that you are using? I've been trying to find something like that, but I can't seem to figure out the proper name for it as my google-fu is failing me.

Yeah, they're tricky to find unless you know the proper name. They're known as "steel ball joint rod ends", or just plain "rod ends". McMaster Carr sells them, but they charge a dishonest and shockingly high amount for shipping. The one I got from them ended up being really gummed up and required too much force to move. I think the one I ended up actually using was just off Amazon.

 

[TheFletch]

Hi,

you may want to share your link to the steel ball joint rod ends that you use. you have a following here and this part is to important for them to get wrong.

Great project.

Fletch

 

[HalloweenBob]

Nice work! Thanks for the shout-out on the design. Very professional design on the machining. Is that what you do for work? Are you not using VSA? You created a custom program for this? I have so many questions. Great work, I am impressed.

 

[BobbyA]

Yeah, they're tricky to find unless you know the proper name. They're known as "steel ball joint rod ends", or just plain "rod ends". McMaster Carr sells them, but they charge a dishonest and shockingly high amount for shipping. The one I got from them ended up being really gummed up and required too much force to move. I think the one I ended up actually using was just off Amazon.

They are $3 to $4 on feebay.

 

[Nicholas]

Hi,

you may want to share your link to the steel ball joint rod ends that you use. you have a following here and this part is to important for them to get wrong.

Great project.

Fletch

I believe this is the one I'm using right now:
http://www.amazon.com/gp/product/B00CF6AZ9I

I've brought multiples of most parts though, so it gets a bit tricky keeping track of what is from where. Also again, this is a 3/8" size, which is probably bigger than most people need.

Nice work! Thanks for the shout-out on the design. Very professional design on the machining. Is that what you do for work? Are you not using VSA? You created a custom program for this? I have so many questions. Great work, I am impressed.

Thank you, Bob. Your project was definitely the inspiration for a lot of this.

I do software development for a living. The red plexi plates were laser cut by a third-party. I designed them digitally, but didn't operate the machine myself.

Yeah, there's a handful of custom programs that I'm working on for this rather than using VSA. One for the arduino, one for the real-time control device, and possibly one more that I'm toying around with...

They are $3 to $4 on feebay.

Yeah, I put up with McMaster Carr because I needed a couple other parts from them, and because they carry a unique rod-end that includes the spacers in it. I was hopeful that the rod end would allow me to not have to fashion my own spacers, but it had too much friction to be usable at all.


I'll be going silent on here for the next few weeks while I'm out of country.

 

[The Halloween Lady]

Very very cool! I'm jealous of your skills.

 

[Nicholas]

Hey all. Time for an update on this project; I've been rather busy the past few months...

After extensive testing of the mechanism, I decided that I needed to go back to the drawing board (again). Let me preface by saying that Halloween Bob's design really is a great design that has a lot of advantages. It's compact, simple, and it works for a lot of situations. My situation is unfortunately not one of those situations, I'm afraid. For my situation, I wanted more control than the design allowed (since the axis tend to cross-talk a lot.) I was also having issue with the fact that the movement goes straight from the servos to the center post, which can put a lot of strain on the servos and basically limits the strength of the entire thing to the gearing of the servos. Lastly, I was having a difficult time figuring out a way to counter the high/forward center of gravity. So I went back and completely redid the design. What emerged is this:

The new design utilizes 1/2" tubing as axles, which takes a lot of the load off the servos. Each axis is independent of the other, eliminating cross-talk. Pitch and roll are geared at a 2.5 to 1 ratio, while yaw is geared at 2 to 1, which makes the entire thing a lot stronger with the same servos. The design actually accomplishes a lot of what I was wanting. But there is one issue...

... it's too big...

Due to the complex shape of the inside of the skull, I misjudged how much room I had and it just doesn't fit. D'oh!

So now I'm contemplating where to head from here. I've yet to find a skull that's slightly larger than a human adult skull (please let me know if you know of one!) I'm considering how I might go about developing such a skull. I'm also considering selling these two mechanisms so that I can reinvest it in further experimentation. The two mechanisms I've developed so far are both useful in many situations; they just don't meet my extremely specific requirements. If you're interested in purchasing them at cost, let me know. I'd love to see them go to a good home, as right now they're just going to get shelved.

 

[HalloweenBob]

My Rack and pinion design utilizes independent movement of each axis and does fit in a skull. I can post pictures and video if you are not aware of it.

Bob

 

[Nicholas]

My Rack and pinion design utilizes independent movement of each axis and does fit in a skull. I can post pictures and video if you are not aware of it.

Bob

If you have them handy, or a link to them, I'd be very interested in seeing.

At this point, my current thought is to just downsize this design, utilizing smaller servos (they're currently the limiting factor), but I'm open to all suggestions.

 

[HalloweenBob]

Here's the original video of my prototype from 2009: https://www.youtube.com/watch?v=TIjkcSjkDek

It was later revised slightly to save space so it would fit in the skull better. The mechanics are the same, just removed the spaces between the plates.

I'll post clearer pictures and another video if I can find it later. The recording in the video has auto gain on for the volume. It is not anywhere near as noisy in real life.

 

[Nicholas]

Here's the original video of my prototype from 2009: https://www.youtube.com/watch?v=TIjkcSjkDek

It was later revised slightly to save space so it would fit in the skull better. The mechanics are the same, just removed the spaces between the plates.

I'll post clearer pictures and another video if I can find it later. The recording in the video has auto gain on for the volume. It is not anywhere near as noisy in real life.

Ah, that's a really fascinating way of addressing the problem. Very nicely machined. Thanks for sharing.

As long as you're here: How did you go about addressing the center of gravity issues? It seems like most animatronics have a far-forward center of gravity in their head. I attempted a few different methods for mounting springs to counter the weight when using the previous design, but never found anything I was happy with.