Let's start with ACE (Ambulatory Canine Emulator), the quadruped. My recent objective was to get at least the hindquarters fully assembled and test hock joint operation with the new motors (17HS08-1004S from StepperOnline). I'm pleased to say it worked out. The old motors could barely flex the joints back and forth without stalling against the friction of the tendons rubbing against their channel. The new ones operate the joints more smoothly, and can extend them to lift the hindquarters (raising their own weight and the weight of the other pair of motors for the hip joints).
No demo video for this one, because I didn't have enough hands to aim a camera while manually switching power to the motors and keeping half a robot from falling over. It still can't get up from a full kneel (joint at maximum flexion, lower leg flat on floor), but I was expecting that; the amount of leverage available from the tendons in that position is just really poor. Even then the motors don't stall - my homemade plastic gearboxes start skipping teeth first.
I didn't want to make a wiring harness before I was even sure the stepper motor approach was viable, but wow, what a mess to run all of TWO motors. I still have lots of my lean custom stepper boards that only need three wires for power and two for control, but they can't handle the current these motors need, so I bought some from Adafruit. They need a bunch of extra pins on them grounded.
Oh yeah, and literally at the end of the final test I wanted to do, the tendon support spar on one of the hock joints snapped. Now I know what I need to shore up before I print the rest of them ...
Next we look to the hydraulic spider, which I've nicknamed "Rachner." I completed the 3D model, fully constructed one leg and one quadrant of the chassis, and made a new batch of bladders to fit the legs joints ... oops, I mean two new batches. I tried using the 2 mil transparent plastic at first, because I thought it would look cool, and it turned out to be too fragile.
I printed polypropylene valve stems and heat-sealed the flanges to the insides of the bladders, and also still attached the rings on the outside of the bladder with urethane sealer, for strain relief. Yet ALL FOUR bladders that I made leaked from around the valve stem. I fell back to the tougher 4 mil black plastic, and also began using washers like little irons to help distribute the heat from my soldering iron tip into a round seal. This worked better, and I obtained three bladders that seemed sound. I tried one in the leg and was able to operate both joints, inflating and draining the bladder with a syringe.
I know of some improvements I need to make already. The rotor joint at the base of the leg isn't achieving its full range of motion on the return stroke. This is partly because it needs a stronger spring (don't worry, I have lots of candidates), but the angle at which the tension element pulls on the vane is sub-optimal also; I need to modify the parts a bit, so that the guide holes line up better. This is why I started with only one leg!!
When I operate the knee joint, I'm letting gravity provide the force that returns the joint to its neutral position when I drain the bladder. In the complete robot, this will be reversed; gravity pulling down on the chassis will tend to make the legs splay. I'll attach springs to balance the gravitational force and hold the joints closed, and then inflation of the bladder will act against the springs to push open one at a time and swing its leg up. (This is similar to how real spiders work - by default the legs are contracted, and fluid pressure extends them. Their legs curl after they die because the pressure is gone.)
This feels like precious little progress to talk about, but life has been busy and it's what I've got.
Until the next cycle,
Jenny
