...final part (hopefully!)

June 2019

With Glasgow Maker Faire also cancelled, I decided to keep the holidays I'd booked for the two events and use the time to get Shunt fully working.

I made a bracket to support the gas bottle from a portion of a large diameter steel tube that I cut up, welded that to some angle steel to bolt to the plywood and secured the bottle with two hose clamps. In OG Shunt, the bottle is on the opposite side but since I had the link on the left-hand side, I figured it would make sense to put the bottle there and have all the safety controls on the one side.

I was a bit unsure about the safety of the gas bottle opening. I cut a hole in the front of the chassis above the link for fitting the 17mm tool in to open the bottle, but this was very close to the scoop. Speaking to the UK EOs, it wasn't really ideal but there was no other way to locate the bottle that would be accessible. To reduce the risk of injury during gassing up, I fitted a quarter-turn isolation valve inline right after the regulator so that the gas bottle could be opened without allowing gas to the system, then once the gassing up tool was removed, the valve could be opened, making the system live. It wouldn't be permitted in a competitor bot, but as Shunt was more for show/demo then an exception could be made. Original Shunt didn't have this so I suspect they turned the gas on then had to fit the shell on before it could go into the arena!

shunt_build (184).jpg Annotation 2020-05-05 005801.jpg

In terms of solenoid valves, I had two 3/2 valves, a 5/2 valve and a 5/3 valve. I settled on a 3/2 valve for the bucket as it would retract under gravity. I didn't want a 5/2 valve for the axe as this would mean the axe would be pressurised in the retract position as soon as the gas was turned on. I opted for the 5/3 valve as this would use gas for fire and retract, but then dump the ram gas once the stroke was complete and leave the axe unpressurised. Or so I thought. It didn't really behave in testing properly and seemed to still be pressurised once retracted. Not sure if it was user error on my part or not, but I didn't want the hassle :P

Instead I rejigged things so that the two 3/2 valves were used for the axe (one for fire, one for retract) and then the 5/2 valve was used for the scoop but only for fire. I blocked off the retract port so basically turned it in to a 3/2 valve too. Now presumably original Shunt had its bulky SLAs in the bottom of the chassis but mine was running on a single 6S pack on the electronics board, so I had pretty much the entire bottom of the robot just for a couple of valves. It looks....sparse

shunt_build (186).jpg

This was before I switched to three valves so it's a little busier now, but not by much. To control the pneumatics, I wanted something like the old Electronize relay switcher that Technobots used to sell but they don't do them anymore. Instead I found a unit from Forge Electronics which was essentially the same thing but in a nicer and more compact unit. I got two, one for both axe movements and one for the scoop which only used half the unit:

shunt_build (178).jpg Annotation 2020-05-04 221017.jpg

They were good, but a couple of times in testing, they wouldn't engage and it would take a few flicks of the switch for something to fire. I decided to persevere though. Another issue was that the solenoids generated noise or back EMF which would sometimes cause the Vantec to cut out; a similar thing led to Coyote's loss to Expulsion in Series 10. The main suggestion was to put a diode across the solenoid terminals to prevent backflow of noise (I believe was the idea). After much debate about the flow of direct current, what side of the diode was the blocking side and what way round it should go, I eventually soldered them in. Next test of the pneumatics......nothing, except a slight crackling sound and smell of burning electronics - I'd blown the switchers. I thought it was my fault, that I'd put the diodes the wrong way round, so flipped them round, repaired the switchers (wasn't as terminal as I thought) and tried again. Once more, nothing but burning smells. The switchers had blown elsewhere this time. £60 down the drain less than a day after receiving them! In my annoyance, I ripped them out and rigged up some servos and micro-switches. The wiring and layout for the switchers was pretty, the wiring and layout for the servos was not:

shunt_build (194).jpg shunt_build (195).jpg

In hindsight I realised a diode across the solenoid terminals was basically creating a short. I'm not sure the diode was the right thing. It presumably would either be a capacitor that was required or I wired it wrong and it should have been in series. Anyway, expensive lesson learned. I got rid of the diodes when I fitted the servos and just decided to see how the Vantec would go. It wouldn't always cut out so might be okay for demos etc.

Eventually I got everything wired in electrically and plumbed in pneumatically, and the robot was ready to run:

shunt_build (188).jpg

I did a few tests at home to get the pneumatics dialled in, but they were working pretty sweet. All that was left to do was to take it down to the industrial estate, get it fully clad and bash away at a microwave in order to make a video for the masses:



The Vantec cut out a few times during the test (as well as a chain falling off the drive right near the start of the afternoon) but thanks to the magic of video editing, you'd never know :P
Unfortunately a bucket fitting also blew and couldn't be repaired on site. And my video editing isn't good enough to magic that in. I managed one decent lift which just made it into the video.

shunt_build (205).jpg shunt_build (202).jpg

So that was Shunt fully operational, much to my glee. A 2-3 year backburner project that was done in 13 months. Oops, but not complaining! I took Shunt to the July RoboDojo event for display and a demo; this turned out to be a bit lacklustre due to issues with the BEC and the Vantec cut-out issue but still managed to smash up my parents' old TV.

It also went to RoboNerd in August; by this point I'd switched out the Vantec for a RageBridge which eliminated the cut-out issues but early in its demonstration it threw the drive chain again (you can see it jump off in a spectator's video when I first fire the axe) so it spent most of the time trying to crab-walk its way to its prey - one of these days I'll do a demo with it that's flawless!
It was completely worth it though for getting a snap with the Robotazia Matilda:

shunt_build (211).jpg


So that covers Shunt up to the end of summer last year. As it was pretty much done, it spent the rest of the year in my storage unit while I focused on other things (namely a full size RC Dalek and my mechanoid RORI).

However, as most of you know well, a robot build is rarely completely finished, you always find things you want to change or improve. Those of you on Twitter and Instagram will know I've been using lockdown to tidy up a few bits and pieces of Shunt that were bugging me. I'll document those changes on here in the next few days once I get the photos off my phone. After that, I'll just need to remake the plough and scoop slightly (to a more accurate shape) and then maybe, just maybe, it will be finished