First off, lets see the cool stuff. Here is the results. I finished the meat of the blog post over a week ago, so this video is the result of the work with the turn table and a quick hack-edit on some chromakey work.
I am preparing for a big project coming up, and one of the things I wanted to do is build a turntable for this project. I already have linear, pan, tilt, and focus/zoom, so a rotary table seemed like a welcome addition.
Rotary tables are fantastic for this sort of work, it can add everything from a spinning to arcing or swooping feel to some movement. By adding linear and rotation you can get some excellent effects. The problem? I dont have a lot of time to R&D a new project these days, so it is pretty difficult to get the time put aside to develop something new.
I started looking at turn key options, the very first that comes to mine was the Tt from eMotimo
I am a fan of Brian and eMotimo. They make some excellent products. Sometimes I wish there was a little bit more flexibility to customize the things way work, but I have to admit, they do work extremely well within the boundaries you are given. I do have a eMotimo TB3 and it is a fantastic product. It is also one of the reasons we have not really put a lot of effort into a pan tilt over at The Chronos Project. We never set out to be a motion control company, TCP was pretty much birthed due to the inadequacy of current offerings, but things have changed quite a bit in the past few years. I am in no hurry to re-invent things that already exist.
As for the Tt. I would have controlled it with one of my own controllers, but to be 100% honest for me it is a bit pricey and I have some stability concerns. I do not like having the weight sitting on the motor like that. My impression is this is really designed for lighter work, it will undoubtedly have a little bit of backlash (i have used dozens of motors just like the one there and they all have a tiny bit of backlash). I think the Tt would be excellent for product shots and such, but not for moving around a large 50lb set. It also only holds a 27:1 gearing ratio, and is only repeatable to 1/100th of a degree. I am not sure how they come up with that number to be honest. It has a 27:1 ratio with a 200 step motor is 5,400 steps per full 360. 16x microstepping gives it a resolution of 86,400 steps per full 360, or 240 steps per degree. That SHOULD be 1/240th of a degree, I don’t quite get how the mathematics allow for 1/100th?Unless anything has changed, they use the same a4988 stepper drivers we use. The website says it is good for 75lbs. It may be, but that would be one hell of a balancing act for the type of work I plan to be doing, and off-setting the weight is gonna play hell on those gears. I am sure it can handle 75lbs, it just seems like it would be difficult trying to get the set pieces I plan to build to balance on it. 75 lbs directly online the center of axis is one thing. Having a 75lb payload spread out over 5ft? Small changes in weight can have large impacts on balance, im sure it would not enjoy having 15lbs off centered 3 feet to the left.
It may sound like I am ripping on the Tt, that is far from the case, the fact is this project I am working on is going to be very demanding, I cannot have backlash, I need at least 3-4 times the movement resolution, and I need something that will handle moving larger set pieces. I am sure for most other work the Tt is more than adequate.
I looked for some other options for Photography Turntables, and there are quite a few selections, but anything made for video work generally will not have the resolution demands for timelapse work. Video work requires fluid motion, not precise starting/stopping ability. The specs for many of these tables are usually not readily available, they probably use DC motors, and the prices are flat out ridiculous. 3 grand for a turntable? are you high?
I wrote down some requirements
1) at least a 400 step per degree resolution
2) heavy payload capacity, minimum 100lbs.
3) Stepper driven
So using Designspark Mechanical I started coming up with a few ideas, and this is the one that I settled on. The idea is to use some legs attached to a base, on top of that base is some lazy susan hardware to hold the weight of the payload, then a Large/small/large wheel configuration to create a large pulley. The stepper motor would be held to the side and use a belt around the turntable. This keeps the weight off the motor, offers a larger surface area to make it easier to put larger heavy set pieces on it, and by using a geared stepper with a small pulley I can easily get the levels of resolution I want.
I needed some help designing the bracket to hold the motor which would be 3d printed, so I sent the design to Kyle to see if he could help me out. Of course, Kyle is better at design than I am. He quickly returned the file seen above with the green motor mount, but quickly followed it with a new version he punched out which not only used less pieces, but was geared with a god damn harmonic drive.
Harmonic Drives are brilliant. Whoever invented these gears was absolutely genius. It uses a flexable gear withing a rigid gear and an oval in order to move one tooth with each rotation. This is what you would see if looking down into the gear.
Click on the image to see the animation.
Harmonic drives are super awesome because there is ZERO backlash, and they are dead quiet. They are also very expensive, however Kyle had one laying around that he had been experimenting around with, and decided to donate it to this cause.
Of course the harmonic drive is just the gearing portion, so he also printed out a bracket to attach a small nema 11 stepper motor (its soooo tiny) into the back end of it, and then proceeded to print out a hub for the shaft.
The files came back, and the turntable was taking shape.
You can see the new design has the motor and 100:1 harmonic drive right up the middle. There is a 9 inch Laay Susan bearing between the two plats, and all the weight is supposed by the legs and bottom plate. The motor/harmonic drive sits beneath it and spins the top plate without taking any of the weight load.
Now the design is complete (thank you Kyle!) it was time to fire up the CNC router. I decided to just make it out of MDF. As long as it stays dry it should be fine. MDF is a fine material for this sort of thing. It machines extremely easy and fast, and it is very inexpencive. I paid $9 for a 2ft x 4ft section which was more than enough to do this.
I am a bit rusty on my CNC skills, so after about an hour of failure I finally managed to get everything up and running.
If anyone is interested, this CNC is a Probotix V90 Fireball. Fully loaded with all upgrades, it came in around 2 grand. That includes the CNC platform, motors, drivers, spindle, and linux box, and software. All I had to do was assemble it and add a mouse, keyboard, and an old monitor. This is one hell of a toy, it is extremely precise and pretty user friendly (well…. as user friendly as it gets for using a CNC machine, there is always a learning curve)
Seen avove, the top plate with the Lazy Susan bearing on top of it. I needed everything 100% perfectly centered, which can be quite a challange. While you could probably do it with a carpenter square and a pen, I find it is far easier to draw the design up in DesignSpark Mechanical (or your favorite CAD program) and just have the CNC align the holes perfectly.
Once the parts were cut out, the harmonic drive installed, lazy susance hardware installed, and everything set together I glued on the legs and let it sit overnight.
Here is another shot of the stepper motor and harmonic drive along with the printed bracket. See the pretty green cable?
Kyle had this quite controller cut out quite some time ago and it had been sitting on a shelf. I dont like using parts we have set aside for customers for my own projects so I will normally grab a couple bad controllers, or a shield that did not pass testing and start de-soldering and re-soldering them into a single working controller. Our controllers are perfect for this sort of work, I put a LOT of time into making these as friendly for DIY projects as possible. Once I had the controller setup it only took about half an hour of tweaking values in order to get things dialed in.
And here is the finished result.
It has 890 discreet positions per degree for a final maximum resolution of .00112 degrees per movement.
Once i got the values dialed in, I tested a variety of programs and it resulted in perfect results each time. Then I tossed a 15lb dumbell on it both centered, and off center, and the results were still identical. The motor is more than strong enough to move the top plate, all the weight is held by the frame. The entire thing was build with spare parts laying around, if building from scratch it could be done for under $100 if you use a geared stepper motor rather than a harmonic drive. If you wish to use an HD it would easily be 300-400.
Testing is complete. It went very well, I couldn’t be happier. I shot these dandelions for over a week, one camera was focused on the turntable and the other one was just collecting data for a project I am working on. Rather than to re-hash everything I mentioned in the video at the beginning, I think I will go ahead and call this blog post finished. I have another one I need to start working on for a splitter I designed.