Embedded Files
Wecome To The New Home Of TheEPROM9
The past part of the mechanism based on the original Skutter mechanical design.
The claw design in Blender. Took an evening to put together.
Skutter claw in the slicer ready to join the physical realm.
The printer doing it's thing, replicating things from a computer world to our own.
The main endo printed & assembled.
The electronics for the claw added to the claw.
Claw ready for testing.
Partner worked on the ESP32 side & PlayStation controller control with I2C bus for communication between the Arduino Nano & the ESP32.
Head tilted left.
Head tilted right.
Head looking up.
Head looking down
Test print of eye fit & hinge mechanism.
Eye & top half of the head printed, eye mounted & open.
Underside of top half of the head with mounting brackets for the magnets & servo for the blinking eye mechanism.
Top half of head with eye closed.
The head after a week of printing the external case. Took a while to design as I am use to designing engineering models. The test fit highlighted two errors I made in the design. The first was the claw mounts fouled with the rotational mech. The second was the neck was two short. The first was solved by rotating the neck 180*, the second was solved by printing a longer neck. There were quite a few bits that had to be printed due to the design proccess. Have to test your work before commiting.
Front of the head fully asemabled.
Back of the head fully asembeled.
The 25x25 millimeter aluminum tube that will make up the main neck. I used a cupboard door hinge for the joint, the original creator also did a similar thing on the show Skutter.
The rubber boot test fit. Sadly does not fit over the servo motor so i really need to source a larger rubber boot.
The old neck servo linkage was to weak & would bend & break. So I took it apart & designed a new one.
Some springs & mounts from an old desk light were kept on standby for if the servo was to weak to lift the head.
One evening later the new joiner & linkage model was complete. Took a while to print but when done I was able to test it.
With the new servo mount & linkages the head mounted to the head mounted to the neck/arm, the solution looks quite nice & worked very well.
Powering the system up showed the servo was finally able to hold the heads full weight, meaning it was a final solution.
The base was made of ply-board & aluminum extrude that once was a very shit laser engraver that was just taking up space in my loft/attic. I test fitted everything together before cutting anything as you can only cut once. The base board is the same dimensions as K9 18" 45.7cm X 30" 76.2cm.
Once the base frame was put together I was able to mark out where the wheels should go so I could have them aligned right & not fuck up & have to buy another bit of ply-board.
Wheel holes cut & wheels being test fit though the holes, all was done right.
A center line was drawn on the frame & board so I could center it properly. I then bolted the frame to the board. The improved the rigidity of both even more.
The screws on the base board holding the two sections together.
Found some caster wheels at B&Q that were tiny. They tuned out to be the exact right size. Lucky guess.
With the rear caster wheels installed. The level was perfect.
I then went though the nightmare of mounting motor drivers on the base board. They hooked up nicely & I was able to take the Skutter base for a test drive. It was fun until the battery literally kicked the bucket.
The last part of the main frame for the bottom of the robot was to make the framing that would hold the rotational mechanism for the head & neck/arm. I went with using angle brackets just for east of construction, however the frame is bolted to the main frame with bolts that got though the bottom frame. Similar to how most 3D printers are constructed.
Remounting the motors as the original solution was not very strong leading to them becoming loose.
Some testing with the wheels running at full speed.
Wheels not running.
The battery died during testing. So I followed the warning label of "No User Serviceable Parts Inside" & user serviced the battery pack. After taking the pack apart & doing some testing. I could not find an issue with the BMS so i measured the voltage of each cell. All except one were measuring 4V. The odd one out was measuring 0V, indicating a dead cell.
I replaced the dead cell, resembled the battery pack, hooked it up to the charger to reactivate the BMS system & charged the pack. It has worked fine ever since.
My partner designed these rather stunning PCB's for the main control chip an ESP32 & the Arduino for the body motors.
The rear of the PCB, looking mighty fine.
After my partner had populated the board, looks very neat.
The rear of the board populated. So far all is testing well.
The head needed some work. The Arduino had found a way to fry its self & was no longer receiving I2C data, after measuring with the scope & testing with the wheels Arduino it was determined the fault was with the Nano in the head. Replaced it & the problem went away. I also added in a noise cap on the Arduino just to filter out any noise spikes on the line just in case that was the cause of the issue.
So for the eye's I was thinking of using an Arduino which would of had config options, but I decided that would just be feature creep. So I just went with a 4017 & an NE555 chip. This is the first application I have had to use the cap on pin 5 of the 555 chip. Don't know why out of all the things I have stuck it in this was the thing that needed it. So I was not going to argue & installed the cap.
The bottom of the board. You can see the diode steering I installed to get the LED light chasing effect. At this point, the other half of the LED's had not been wired up. No custom PCB was made as I could not be assed.
All the pieces drying after painting on my really tidy & pristine workbench. Totally not a hoard of shit sitting on it. It took a good while to fully dry. Me & my partner went shopping while it dried. Was done when we got back & had dinner so I was able to assemble it.
A few more hours later & I had it installed in the head. I used the old eye case I had printed to fit & design the hinge to fit the LED board. Once I had the design that worked I mounted it in the final eye case.
After that was all done I took it apart & painted it a dark navy blue. This is the blue they look when they are on the show, sadly there is not much info in general how the original Skutters were built, so a lot of guessing is required. Once the second coat was applied & dried I added some red light gel to the eye which compl8imented the LED's nicely I then reassembled the head. It now looks fantastic.
So I finally mounted the head to the arm, 8 screws in total as there are two each side. The USB cable & power/data cable also comes up though the grommet. A rubber steering boot bellows is used over the joint which it just fits over the arm mechanics.
Rubber boot in its final place where it hides all the crimes of the not so good engineering. Over all the neck piece is turning out to be really good. The head nod is operating at the servos limit so some springs might have to be used to help the motor a bit.
The head fully painted & mounted on the arm. Looking mighty fine if you ask me.
So the next part is put a spring on the neck to give the servo some much needed tension to help the struggling motor & figure out why the Arduino keeps crashing in the head. Also need to remove all delays from the code. This will mess with the servo smoothing I want to implement so the head is less jerky.
I ended up remounting the wheels to the base. As the original solution was just crap & vulnerable to being stood on which is a real risk at a convention.
Ones the wheels had been mounted I then remounted the top part of the frame.
So the next part to work on was the arm to body mount. The arm used panels cut out from a bus that once toured passengers round my local town. The bearings were from some old army night vision camera.
The side with the bearing mount in the middle.
Inside you can see the bearing mount that goes into the inner race of the bearing.
Bottom shot of the finished assembly.
Front view of the assembly.
How the sides are attached to the base.
Right side of the assembly.
The Back of the assembly. I used a small off-cut section of tubing so I could get the dimensions right.
After my first session using On-shape I had the core of the rotational mechanism printing. It took almost two days to print, but the thing needed to be solid enough to take the weight of the whole head & arm as well as the mechanics. It also had to be strong enough to take any rotational forces that the head would be subject to. I took the build it like a brick shit house method of construction.
I used a bearing from a photocopier that I got back when I was in the early years of secondary school, so about 15+ years ago. Always horde the good stuff, might take a few decades but it will come in handy.
The rotational part printed, also printed & mounted a bottom part & called the whole thing the missile silo as it reminded me of the Titan series of ICBM's used by the US back in the cold war.
Test fitting the arm connection piece to the silo
The linear actuator mounted in the silo just like a ICBM back in the day ready for deployment.
The next part was mounting the silo in the main body, this would allow the head to rotate. I just mounted it in the bottom bearing which was not enough to support the top of the unit.
I then modeled & printed some top bearing mounts, it was originally meant to be just raw bearings. At the same time I replaced the guide wheels in my 3D printer, so I just used the wheels from the 3D printer as the bearings were still good & plastic would make a good tier.
As the PS3 remote was giving us grief we have started designing our own controller. It will allow us to do some extra functionality & also implement some safety features so it does not loose control & take someone out.
Skutter's gona fuck ya ;-)
The arm mounted to the base mount.
The arm & base mount connected & hooked up to the linear actuator.
Some testing of the linear actuator. Worked a treat. Also more than strong enough to lift the head & it's associated hardware.
The head & upper arm mounted to the lower arm & the rest of the body, it's really starting to take shape now.
Full body shot of the Skutter with the head finally mounted to the body. This was a major step in the project.
Even though the servo can lift the head, it does struggle a bit. So I just added a ballast spring to head so the servo would not struggle. It worked a treat & now it has no issue lifting the head. The spring does poke out from the rubber boot, but oh well, such is life.
I was able to install the small 180N linear actuator in the arm, however it still needs tension from a cord or spring as it is not strong enough to lift the head. I have had to repair it once as the weight of the head broke it.
As the bungee cord did not provide enough tension, I added a spring to the cord to give some more tension. The motor now struggles but it can pull the head up to it's full height. A zip-tie is used to adjust the tension of the whole thing.
The single section of this project that has given me the most problems. High talk in a small space is not an easy engineering challenge.
Mounting some of the electronics for control of the linear actuators, the PWM signals are going to go back up to the head for control.
Page updated
Google Sites
Report abuse