Swallow - BW Crusher

qbnut · 25 August 2023 17:25

Having built a number of different ‘meta’ beetles I decided to give something else a go for the BBB Subterranean Showdown 2023 event. This wasn’t ever meant to be competitive in the sense of a competition winner, and for me was just a bit of fun and a ‘could it work’ kind of deal.

I had the idea for a crusher floating around for ages, so with the spare parts from my ‘competitive’ (I say loosely) egg beater bot, Contradiction, I set about on a design. Once again this was pretty front loaded design in Fusion 360 so that the manufacture and assembly was just a jigsaw puzzle - I wouldn’t have to make much bar the front wedge by hand.

You’ll see some bits here that didn’t make the final cut - the wings that would prevent it from rolling over never got made as weight ran out. But I did have the idea, honest!

Before I set about making the thing, the beak mechanism went through several prototypes:

The first made out of PLA, because at the time my other printers were busy doing other things. Did it work? Yes, for about 30 seconds.

The motor also was a scavenged 2836 with dubious quality, and a bit bulky.

Several other iterations later (that I didn’t take photos of), and we ended up with this 2 stage design:

The bolts went ‘out’ rather than ‘in’ and I found out later this made it a pain to assemble! I changed this in the final final design so that the nuts were inside the mechanism

At this stage the beak design wasn’t locked in, but I wanted to try out the concept. So I make some bulkheads and put it together:

A test video got posted to the OOTA Discord of this test if you want to go hunting

This was around May, and it got sat on for a few months.

Around July I picked it up again, and started on the CNC parts:

I had the old screws from Contradiction, but these were countersinks, and given what happened with Contra at the BBB Summer event, I wasn’t going to make the same mistake. So button heads were put on order (the ones in Contra also got changed)

We were getting close to the POM deadline now, but we made it in time: https://www.youtube.com/watch?v=8u60uYA2mJU&ab_channel=TeamDarklightcombatrobots

I just had to finalise the beak, and it was at this point I realised I was out of white 5mm HDPE. The rest had been built from the stock I had leftover from the last Contra build so I hadn’t had to buy any. So black it was going to be!

The printed TPU sides were also now done - not the tidiest that they’ve ever come out, but time wasn’t on my side to print them again. And besides, why did it even matter, it was going on a Combat Robot after all!

Also appearing is the wedge I never used in the end - I had this for Horizontals but never met any.

Between this photo and the event was the mad scramble to the end. The gearbox was still PLA but the plan was it was Nylon for final. However at this point, I realised I hadn’t dried the Nylon filament, and had no time to leave it in a drier for an extended period (the print room is also my day job WFH office so it couldn’t just stay on). With mere days left, I picked the next best thing - I had some ASA for reasons, and this seemed to work OK in other applications, so ASA it was

All the bits that go to make the gearbox ready for assembly. I dod however keep the PLA version around - which came in very handy later on.

I wasn’t 100% happy with the gearbox, but you can’t be choosy this close to the event. It could have done with maybe a few more iterations, but… ehh

So here we are, at the event:

We now have forks and a different wedge - this actually worked out really well apart from when it didn’t

So I lost all my fights - which went as well as to be expected! The first fight melee I got turned upside down and hit in that place. You know how in a lot of video games, the boss always has a weak spot? Well for Swallow it’s the motor on the top of the gearbox. You’d think it was out the way enough, wouldn’t you? Well actually no - Delta V (who - spoilers - won the event in the end) hit it and the motor just exploded.

That was that for fight 1 - no way to get up from that.

Having cribbed the motor off the other PLA gearbox, and the PLA pinion with it (which surprisingly actually held up) we went into the next two fights - Grabbing defeat from the jaws of victory and being pushed around by a brick.

All in all the bot left the event, defeated but relatively unharmed.

I did get a nice award though!

It’ll be back, but entirely new beak gearbox with more (far, far more) torque for those thick tops, and let’s fix the self righting and getting upside down to start with, shall we?

hhills · 25 August 2023 20:18

Fantastic effort for a really well executed design. 100% deserving of the spirit award.

Are you planning to separate self righting from the weapon mechanism or is it still going to come off the claw motion?

qbnut · 26 August 2023 12:12

My current thinking, which might change, is that if it can use a cam action from the beak it would be fine, so long as the mechanism isn’t exposed like it is now. The motor should have been tucked underneath rather than on top in hindsight, when I get the gearbox revised it’ll be covered up

McMullet · 26 August 2023 14:51

Lovely! So good to see a crusher at any weight, and especially in beetles.

Congrats on the beermat!

qbnut · 16 December 2024 13:38

Long overdue update!

Swallow has been rebuilt not once, but twice - from scratch - since the BBB Sub Showdown event

Let’s start with mk2…

So post the first (and only) event that Swallow ever went to, I decided that it needed to be… better. So I started work on a new revision.

Pretty much everything was completely redone, from the beak to the drive. The front ended up with these articulated wedglets so that it could get under things much easier, to solve the issue that although mk1 had forks, it wasn’t as easy or effective as it could have been.

So little was shared with mk1 that for a short period of time, they co-existed:

The weapon was also changed so that the pivot was lowered and it pushed from the top, not the bottom. This was more as a weight saving measure, as the mounts could be much lower. Combined with this was a new gearbox based around steel rather than printed gearing, allowing for a much higher ratio.

I also wanted to experiment with a different motor solution, leading me down the route of trying a NEMA11 and a TMC2208, most commonly found in 3D Printers. The idea was that, whilst this was much heavier than other solutions, it would have a large hold force. And whilst in theory, this worked and would have been viable, I ended up hating it for a variety of reasons.

I created some custom controller PCBs, that preformed the duties of managing the driver and also electronic end stops, triggerred by two microswitches located at the end of the arm travel. This was to stop the thing bending itself at the extremes of the arms travel.

Around April 2024, the build was ‘completed’ - or at least enough that I could finish it for an upcoming event:

And then… it sat on a shelf…

I wasn’t happy with mk2 - the thing had been a complete pain in the ass to assemble and maintaining it at an event was going to be less than fun. So after 6 months had passed, I started again again, on mk3.

… to be continued!

qbnut · 16 January 2025 16:35

So onwards to mk3, and the big start again.

I didn’t gel with the design of mk2 at all, so the CAD was basically binned and started again, only taking across some common components.

Some of the parts can be reused, I didn’t want the whole thing to go to waste - so those that do have some value have been adapted to fit Contradiction, which will also get a new revision soon. Will diary that journey separately later on.

What I wanted was to radically change the size and geometry of the beak, such that it laid at a much shallower angle, and thus was easier for stuff to get under it, ala Razer (how could it be a crusher and not be a little inspired by it?). This was sort of a step back to how mk1 was, in that being a metal front wedge with attachments, rather than just a pretty steep bit of TPU.

Alongside this, the whole beak mechanism and the drive mechanism are two functional components, with the beak able to articulate to always match the floor.

In the excitement to get it manufactured, I made a number of CAD ‘oversights’ shall we say, that will come back and bite me later…

This design decision meant I had to do something either entirely stupid or clever, depending on which way you look at it - we now have 90 degree gear gearboxes which fit between the wheel base so that these free up space in the main body, and allow the decoupling:

Onto the jaw/beak - how could the lead screw design fit in? In the end, and it order to get a much higher reduction, I opted to change out the lead screw for a first reduction off of a 19:1 BE1806 Rotalink conversion gearbox, to a 20:1 worm gear, to a 3:1 final stage - a 1320:1(!) reduction! This all runs off a BE1806 motor.

As the NEMA is gone, and brushless is the way forward, the standard limit switch setup you’d use on brushed, of two switches and some diodes, isn’t going to cut it. So, more custom PCB shenanigans ensued. I hadn’t made it easy on myself.

This sits between the RX and the ESC, and will interrupt the PWM signal if the switches are triggered, bringing the ESC to failsafe position. I also added some RGB indicators because why not?

Trying to cram it all in, and ‘it worked in CAD’ but not in reality, we ended up with some cramped internals:

So cramped in fact that it wouldn’t all fit

The wedge thus needed a design tweak and reprint:

The armour sides this time around are TPU rather than HDPE. I could run a prototype off in 40 minutes in some end of spool green PETG on one of my faster printers, the TPU would need nearly 4 hours per side as I’d ran out of the high-speed Ninjatek Cheetah TPU so had to slow it right down to work with the Polymaker TPU I had on hand. So it was prudent to make sure it was right first.

Which brings us to today - wiring has started to get it in motion ready for the POM for BBB Beetle Bash '25