T_Y's Build Log

[T_Y]

So, Robodojo's Christmas social is coming up next week so I'm busy getting my bots ready! For now, here is Have You Tried Turning It Off And On Again, version 2:





There's a load of changes from the first version. The wheels have been hollowed out to cut weight, and the motor mounts have been integrated with the brackets to hold the back panel and the top and bottom. This also brings the wheels out so they protrude out the back, allowing me to drive off other bots' wedges.

The front end has been made a lot wedgier, so it can get under bots more easily. There seems to be an issue with the construction so the front wedge isn't actually flush with the sides, but it works well enough. I've somehow managed to cut 30g off the bot so now its only around 136g.

https://youtu.be/uB-XOSm9D2w

I also managed to fix up Percussive Maintenance v2 so it actually works now. And, it can self-right... sort of. On my desk, it sometimes fails to self right, but on other surfaces such as at Sheffield it seemed to work just fine, so I'm not too sure whats going on there.

Next step is to actually finish the new front hinge flipper! It turns out there's already a US antweight called Blue Screen of Death, so I'm thinking of changing the name to something else.

[Matt]

Don't worry about my robots- we can't make it.

[T_Y]

New year, new plans!

I'm planning on having my bots ready for BotFest 2. The front hinged flipper is now called "It's Not a Bug, It's a Feature!". I've ordered a fast servo to go into that (unfortunately, it looks like pink servos are short in supply).

"Have You Tried Turning It Off And On Again" is basically already ready for that, I just need to print a slightly modified lifter arm to close the small gap at the front of the bot - a spinner can get underneath the lid and strike the lifter arm instead of deflecting off the polycarb. The front linkage will also be modified, so when the lifter extends, the front of the bot will lift off the ground slightly so it doesn't get stranded on one of the screws at the front.

"Percussive Maintenance 2" will replace the older PM, using HYTTIOAOA as a basis. I'm planning on running a servo-based axe this time, exploiting gear ratios. Sheffield Roboteers Society recently showed a servo-based hammer bot which shows its definitely doable.

[T_Y]

Right! So, I've been working on getting INABIAF to fighting condition. I've printed off a servo harness to mount the servo to the base as the wings on the servo are not oriented correctly to let me screw through the base of the bot and into the screw holes. The harness sits around the servo and serves as the screw point.

This afternoon I've been soldering the wires and components together. The good news is nothing exploded! The bad news... drive doesn't seem to be working. Debugging the Arduino, it seems to be working fine, so I suspect it's something to do with the motor driver. I checked the motors by plugging it into the motors for HYTTIOAOA but that didn't work, so I don't think the motors are the problem. I have a spare copy of the motor driver from Hello, World! and Percussive Maintenance so I'll use that I guess, as I know it works (it's the same chipset, but different board manufacturer). Should get that done tomorrow and hopefully have it working.

At least the servo works. Unfortunately, it can't self right - the flipper doesn't quite extend far enough to tip the bot back over. But this is where those 'cat ears' you can see in the CAD come in. I've designed it so that these are separate components and are screwed in, so if they get damaged, I can swap them out. Or in this case, I can redesign them so that they are taller, and print off new versions to mount. This should solve the self righting issues.

Oh, and HYTTIOAOA has new lifter arms which bring the front plate flush with the sides now.

[T_Y]

I've been preparing for Robodojo's Antweight Open which is tomorrow. So, I have It's Not a Bug, It's a Feature! and Have You Tried Turning It Off and On Again? ready for combat.



Here is INABIAF in action: https://youtu.be/F4DeGyrom84

There's a problem with INABIAF's wiring. If I reverse direction (e.g. full left to full right) quickly, the drive cuts out for a second or two (one of the arduino lights goes out, the one I control in code), but the flipper still operates. I suspect either I've got a faulty motor driver, or that my wiring is too thin... is 26 AWG for N20s too thin? Previously I was using 22AWG and that was fine...

There's some improvements I want to make already. I want to work out how to machine aluminium, that way, I can a) give HYTTIOAOA a aluminium front plate for spinners, if I've got the weight, and b) an aluminium wedge for INABIAF, designed to be interchangeable with, say, a set of forks.

[peterwaller]

If the flipper servo runs directly from the receiver it might be that the the arduino is being reset by a dip in the power rail due to the sudden change of motor direction drawing a large current surge.

[MarkR]

Maybe try to reroute the power to the servo, add some more decoupling capacitors on the mcu power supply, if you have space/weight.

I think you can't go much wrong with a 10uF cap on the Arduino, which is probably much bigger than the built-in one

[peterwaller]

All my controllers have 100 uf across the 5 v supply you can get some really small 6.3 v ones these days.
https://uk.rs-online.com/web/p/ceramic- ... %3Atrue%7D

[T_Y]

@peterwaller @MarkR Thanks for the suggestions. Would the capacitor go across the power and ground wires between the battery and the arduino then? And I guess across the power and ground wires between the arduino and the motor driver?

Small-ish update before BotFest 2 tomorrow. I haven't had a lot of time to sort out my bots, so I've only got HYTTIOAOA ready. INABIAF is still in need of repairs after I shorted it out at Robodojo, and I've been having problems with the design of PM2. What I have done is added some forks to the front of HYTTIOAOA instead of using acetate:



Hopefully they work well enough. For now, they're made of HDPE as I can cut points easily with a Stanley knife. Though the forks might compromise the defence against horizontals a bit...

[MarkR]

The capacitor would go between the power and ground, as close as possible to your microcontroller.

The idea is, that if there are spikes in the main power supply caused by (e.g.) motors turning on and off, they are smoothed out by the capacitor. As wires have resistance (and inductance), you want the cap to be close to the digital / sensitive load.

Usually these boards already have a decoupling capacitor on the microcontroller's power supply, but maybe it is insufficient.

What's happening - I guess - is that if you quickly reverse the drive on a motor, then it's still spinning in the opposite direction, which means it's generating *against* the power, so that it can draw quite a lot of current, causing a power spike.