Birth Of The Impulse Portal
It finally has a name. The Impulse Portal. It took a great deal of time to come to, but it was worth it.
Recently, I have had a little down time from the project to clear my head. Instead I’ve been learning how to code in Unity3D, making small video games. It’s been fun, but I still want to make progress on the sim. So this is the perfect time to detail some history on the build so far.
The rotational mechanism is made up of three main sections, the inner gimbal (which rotates the roll axis), the mid gimbal (pitch axis) and the outer frame support.
Before I started anything, pretty much the whole build was modeled in OpenSCAD to make sure all the dimensions were OK, and that all the parts meshed and didn’t clash.
Shown here are the frame corners and midpoints for the mid gimbal. They have M8 nut inserts built in, and I used M8 threaded rod for the structural support. This is the first of the prints that began the whole assembly. There are four corner pieces, and four T pieces.
The T piece has a through hole down the length for the threaded rod, so that I could move it to the centre of gravity. Then it gets locked into place with the washers and nuts. I had to move them around a few times, because I kept adding components. I should have waited until the end.
Below is an end cap for the inner gimbal, one for each end. It has four M6 nut traps for threaded rod to join each end, a mounting position for a Nema 17 style stepper motor, and a housing for a 608ZZ bearing in the middle.
After the mid and inner structures were completed, I needed a way to translate power through to the inner motor (and anything else that might be going in there). I looked around for slip ring solutions to buy, but none of them with a reasonable price tag had a through bore to allow a support via bearings. So a made one. I know it’s not fantastic, but it is good enough for a prototype. At first I tried to make it out of copper pipe, cut down into rings with long tabs, but it proved to difficult to cut straight. Plus the ones that I did try needed the long tabs to be thin enough to slide inside the channels between the rod and the other rings. Bending this down and flat again was too much for the copper and they either snapped or were to wide.
I stripped down some cable, which has annealed copper, making it easier to bend and twist without breaking. The downside is that I no longer have a single flat surface to make contact with. Without an alternative it was worth a shot, and if it didn’t work I could always go back to the drawing board. I made channels through the edges of the slip ring assembly to allow each ring to pass through to the end and allow connection, without making contact with each another, or the centre supporting rod. By having openings at each end for the copper cores, each ring won’t have a catch point, but it did make the construction a lot trickier. Lastly I made a cap piece to enclose the exposed ends of each core.
At first I chose five rings to accommodate 12VDC, GND, VCC, Step signal, Direction signal, which would be enough to have the stepper driver inside the inner gimbal. I also left open the option to change the last two for a serial comms line like I2C, but that is only if I needed more devices inside. At first I was worried about signal transmission though my slip rings, and wanted the ability to drive the motor locally. In hindsight I should have made six rings to include a separate GND for the VCC, but this works for now. The first test I did with this was to have the stepper driver on the bench, and pass just the 4 pole signals through (I am using bipolar steppers). That worked fine, so I left it. All that worrying for nothing. Later I’ll show how the contact arms work.
Here I have the bracket for the support frame. It has nut traps for a tripod style stand, mounting holes for another Nema 17 stepper motor, and a cutout for another 608ZZ bearing.
Missing from these photos is the gear pairs printed for each motor. Each motor has a gear made for the motor shaft, and a gear fixed to the axis. They can be seen attached to the machine in the following photos.
All together now…
Seen here, the cables for the slip rings are just connected from one axis to the next. There are no motors on the mid gimbal.
The gears were printed using modified “Gregs Accessible Wade Herringbone Gears” as the inspiration (and his OpenSCAD code). The files that I used came from: “http://www.thingiverse.com/thing:18379”.
The inner motor looked like it was weighing down a bit on it’s back end, so I made a bracket to hold it. Feet were printed to fix the tripods to a base (at a later stage), and then the whole assembly is almost finished.
The first test now was to see if the exterior pitch motor did it’s job. I fire up the arduino, dial up the current on the drivers to max (as per the motor, not the driver) and give it a try. All I hear is whirring, but not much movement. The rig wobbles a bit but doesn’t go anywhere. I thought maybe the gear ratios were a bit too optimistic. I had made them 1:1 and was hoping it would be OK. So I make up a set of new 2:1 gears and give that a try. Still no luck. The little Nema 17 was just not going to cut it. Without many other choices, I order in a Nema 23. but one with a lot of torque (there are quite a few variations in the Nema range) and a driver to suit. Next I needed to make a modification to the motor support stand and gears to suit the new size motor.
Setting it all up again, I give it a test and everything works great. Onto the next part.
This is when I started working on the contact arms for the slip rings. I had the idea to just use M3 brass bolts with long unthreaded shafts. Unfortunately they don’t make those in brass. That stuck me for a while, as my whole design was built around this idea. Maybe some research before building was a good idea, some more hindsight. All the brass products I could find were in imperial measurements and were much larger than I had planned. I needed smaller than 4mm, as that was the size I made the slip rings to fit M3 bolts. Any larger and I would need to disassemble, redesign and rebuild.
After a lot of research and pacing, my brother suggested I buy some brass rod (smallest I could get was 3.125mm, perfect) and simply cut my own threads onto the end. I bought a 3mm die, and off I went. I had to grind down one end of each short rod to just smaller than 3mm for the die to work properly, but it worked better than I thought. I made dual arms for each ring to allow for continuous connection during rotation. Each pair of arms were offset from the next to prevent catching and short circuits. They also have small hooks to hold little rubber bands which pull the contact pairs towards the centre. The stand by which the arms pivot was made to fit into the same mounting holes as a Nema 17 motor, which meant I could reuse models for each side. The pivot was to allow each arm to rock independently to overcome the wobbles in the slip rings, but are still held tight by the rubber bands. The purpose of the thread on each rod was two fold. I needed to securely fix the rods in place, as well as bolt cable lugs to the ends for a good electrical connection. The end result looks quite the part and I’m really proud of it.
After all the contacts are finished, and the cabling completed, it needed to be mounted onto a plywood base for stability.
This is where I started testing software, and uploaded my first videos to youtube. My wife and kids thought this was all pretty cool, but it didn’t look right without someone riding inside.
They were totally right. Now it works.