Coyote - The Robot Wars Chronicles

Hey everyone,

It's been a while since I've touched my regular RogueTwo Robots build diary as I haven't done much stuff outside of RW since the start of 2016, and it's always been my plan to document the build of Coyote on here so I figure a new thread is the best option.

Coyote was my first scratch-built heavyweight. I've been doing this since 2004 (2001/2 if you count numerous failed builds) and in that time I've never managed to build my own heavyweight. I joined Gary and PP3D for Series 1/8 of Robot Wars and that was a tick off the bucket list, but building my own heavyweight for the show was still the dream. I'm going to detail the whole process as much as I can for those interested in it or looking to do the same. I'll warn you now, it'll be several posts and a lot of them will likely warrant TL;DRs but, well, tough :proud:

The Beginning

There were rumours at the end of Series 1 filming that Series 2 was going to be filmed in August/September time. Gary wasn't sure he'd be able to get PP3D ready for then but I was still desperate to be involved in the show so I decided to finally get cracking on my own heavyweight. I'd sold all my previous HW gear about two months before the reboot was announced (typical!) so starting from scratch was the only option.

Since the reboot would be on Series 2, and S2 of the classic era is one of my favourites, I immediately decided I wanted to have a chainsaw. Ineffective weapon? Yep. Massively outdated? Yep. Did I care? Nope! I reasoned that as it was designed to be used with wood, and that HDPE can be worked with wood tools, then it may actually be able to cause some cosmetic damage. So that was the first weapon. I say first because I knew a chainsaw wouldn't be good enough for a primary weapon.

I then bought a pair of actuators off Gary, ones that he had bought with plans to fit into Smidsy but no longer needed. They were compact but powerful and reasonably fast, so I knew I'd find a use for them. I think it was around this time that I also bought the Scot-Bots middleweight/Tanto Jr. off Tony. Making a functional drive system was always one of the biggest hurdles for me building a HW as, more often than not, it requires a lathe and I didn't have one, or easy access to one. So the MW gave my a fully built drive system that I could build a HW around.

Attachment 6856

I picked up a cheap chainsaw on eBay and plundered the bar, chain and sprocket setup from it. At this stage I wasn't sure whether to use the actuators for a lifter or some sort of grabber, but I knew the chainsaw would be a tail, actuated to move up and down like the original Matilda.

After some deliberation, I decided to use the actuators for a clamp and lift weapon with dual, independently operated jaws, akin to Diesector from Battlebots:

Attachment 6857

I reckoned this would be more interesting than a lifter and therefore more likely to be successful when applying for Robot Wars. The Tanto chassis was just a bit neat width wise to fit the actuators and expected gubbins in, so I shortened the bulkheads supporting the wheels and motors and rotated them 90° on the baseplate which gave about 100mm extra width. It may not sound like a lot but it was just what I needed for things to fit.

I also needed some extra height allowance to be able to run inverted as I didn't anticipate the jaws being capable of self-righting, and the original chassis wheels were a bit on the short side, but thankfully the hubs were a perfect fit for the foam wheels we used on PP3D in S1, spares of which I had lying around. So on they went, and I started to feel like this was bordering on HW territory :)

Attachment 6858

Next up was to make an MDF prototype for the jaw assembly to work out ideal locations for mounting holes and pivot points. I knew getting the right geometry would be tricky, so the MDF would allow for corrections!
As it turns out, I got it spot on first time. Drawing out the shapes of the jaws took a few goes but once I was happy with them, I cut them out and rigged them up and they seemed to work well.

Attachment 6859

And a video of them in operation:
https://www.youtube.com/watch?v=HIYB6-YxcUA

Overall I was really pleased with this. I could see that the jaws moved at different speeds depending on whereabouts in the actuator's travel they were at, but I was confident that it could be accounted for and wouldn't cause problems. This was is May last year, so only a couple of months since Series 1 filming and I was happy and surprised at the rate of progress. I knew there would still be a lot to do but that it was achievable. If I remember rightly, by this point talk of S2 filming had been changed to October-ish so it gave several more weeks build time.

That's enough for now. Follow up will be turning the MDF prototype into something battleworthy and more chassis developments.

Really cool stuff Jamie, good read :)

I like this design, reminds me a bit of Pitbull, a very successfull robot from the
early days of Robot Wars, it's only downfall was the lack of scrimex but it looks
as if your robot works both ways up so it shoulden't matter.
Only one question-what make and model are your actuators and where can I
buy them?.

Thanks for the kind words guys!

Colin, I'm afraid I have no info on the actuators beyond the brand being Servomech. I've tried for ages to find replacement ones online with no luck. In hindsight, they're not really tough enough for HW combat as the bodies and mounting points are cast ali, so unless you were to look at machining replacement bodies from proper stock, then going for a stronger alternative is a safer bet.

I remember Pitbull well. The first year Robots Live did their event at the Air Museum just outside Edinburgh, I went and chatted to a father and son post-show to see how they enjoyed it, and it turned out to be none other than Ivan from the Pitbull team, in the area for a holiday :)

Looking forward to the rest :D

Interesting to get a behind the scenes look at these sorts of builds. Hope to see more, not just from Coyotes build, but from other bots too!

Jamie this is awesome. You need to develop this into a more cowelled front end, sort of a wedgy skirty front plate that the jaws embed into for spinners.

From Prototypes to Pretend-O-Bots

It was a dark and stormy night!
Actually it was the complete opposite, a lovely warm sunny day in July. I was sitting watching the Wimbledon final (go on Andy!) and finally decided to sort out the CAD for the jaw assembly. I'd decided to opt for some sort of CNC cutting, either waterjet, plasma or laser, for getting the jaws cut out of Hardox. I could have tried something by hand but it wouldn't have been as fancy, as neat or as quick. After procrastinating and putting off the CAD work multiple times, I finally forced myself to get it done.

I drew top and bottom jaws similar to the MDF ones so that I could make sure that the mounting holes and pivots were all in the correct places. I realised with the prototype that the jaws were a bit on the small side for a heavyweight, so took this moment in CAD to enlarge them slightly while still retaining the hole geometry. I then decided that the top jaw needing a bit of characterising to make it more visually appealling. I already had the name Coyote by this point, so this was when a conscious decision was made to focus on giving it an animalistic look. Therefore the top jaw would have to resemble some sort of head. This was the first effort:

Attachment 6867

It wasn't too bad and remained like this for a while, but there were a couple of bits I wasn't happy with and it just didn't feel right looks-wise. I then sat down again and refined the design, added more serrated edges, added more obvious 'ears' as well as bracing pieces, and once that was done it felt much better:

Attachment 6868

The one part I hadn't factored into the CAD was outrigger arms to support the robot when the bottom jaws lift up. This was due to the way the base chassis attached to the jaw assembly, I wasn't able to work out where they needed to be without actually having the robot in front of me, so they would be made at a later stage.

When it came to cutting, I opted to go down the plasma cutting route. From the above CAD, the various pieces were laid out in a 2D dxf file that the cutting company could feed to their machines. Several monies, a few days and 32kg later, this lovely jigsaw puzzle arrived:

Attachment 6869

This was the first time I'd gone down this route on such a scale, and it was a weird but fun feeling seeing bits in the flesh that I'd only previously dealt with in computerised form. The size of the parts surprised me too, as the prototype arms were that bit smaller. Lifting each new part out of the box was like being a kid at Christmas again, as illustrated by my expression:

Attachment 6870

It was at this stage Rachel got involved. We both did a volunteer project in Peru several years ago, building a 2kW wind turbine from scratch for a rural community. She picked up welding really quickly while working on the project, so I let her loose on the jaw assembly:

Attachment 6871 Attachment 6873 Attachment 6874

I think it took about three days to get the whole lot welded up. I had a week off at this point (lead up to Insomnia) and I remember thinking I wanted to get the robot about 90% by the end of it. Turns out that was way wide of the mark as a possible achievement :lol:
But getting the jaw assembly done was a big hurdle out of the way, and allowed the first Pretend-O-Bot to happen which, at this stage, I thought looked cool as hell!

Attachment 6875

The size of the thing hit home at this point, taking up a lot of my workbench and measuring approx 1.4m tip to tail. Love it!

The next step was to make some small brackets to weld onto the jaw assembly bulkhead so that the actuators could be mounted. Once that was done, it threw up a slight error. Even when you use CAD, you can still make mistakes if you don't check things or use accurate models. The barrel of the bottom actuator was interfering with the base brace of the bottom jaw, preventing it from fully retracting. So after all the time spent fabricating the jaws nicely, I had to go ghetto with an angle grinder to rectify it:

Attachment 6876

Live and learn, and it all worked out fine in the end :proud:

So with all the Hardox stuck together, it seemed an opportune moment to stick everything we had on the scales:

Attachment 6877

Not too shabby at this stage. It still needed armour, the majority of its internals and the chainsaw mechanism, but with 45kg to play with, we were confident about getting it in weight. The main factor now was time, as that last photo was taken at the start of October, and it was looking extremely likely that filming would now be the start of December, leaving just seven weeks of spare time and weekends to get everything done. Full blown nerves weren't setting in yet, but it wouldn't be long....

Seems a good place to stop for now. Next time, relocations, applications and loads-of-other-stuff-ations.

Relocation, application and determination

We pick up this tale as we venture into the dark depths of October in northern Scotland. Cold all round, therefore perfectly acclimatising us for the Robot Wars warehouse!

I say northern Scotland (for reference, my workshop is in south-west Scotland) because I relocated the robot, several boxes worth of tools and accessories, and all the spare parts and materials to Gary's workshop. Time was marching on and with me moving onto a winter shift timetable at work in Aberdeen, there were barely any weekends I could head down the road to work on the robot. Gary graciously offered me the use of his workshop in the run up to filming, making it a much shorter distance to travel to work on the robot on odd days off here and there. We categorically wouldn't have been able to complete Coyote on time had it not been for this, so I can't thank Gary and Sarah enough for their hospitality during this time!

So, by this point in proceedings, Series 2 had been confirmed and applications were open! This is the photo of Coyote I included in the application:

Attachment 6900

It's obviously far from completion, so in order to give the production team a rough idea of how it would/should look when finished, I attempted a CAD/MS Paint/Photoshop crossover. Emphasis on the 'attempted' :proud:

Attachment 6901

As I already had a Matilda-esque chainsaw, I decided to try and incorporate some other small house bot elements, so the back was to feature spikes like Dead Metal (representing the hackles of the coyote as it's preparing for a fight) and a small copper exhaust stack reminiscent of Shunt. The copper was to try and help style the machine in a steampunk theme, plus I work with copper in my day job as a gas engineer and wanted to have something that I could tie into the robot from work. Anything that could make our application more appealing.

On that note, for anyone looking to apply for future series, along with the above pictures I also included a video cut together featuring clips of the robot driving and the jaws working (separate clips as they both weren't working together at this point) as well as explaining in real time what was planned for the robot in terms of armour and looks. We also tried to make the application fun; bombarding them with generic stats in a serious tone will get repetitive for the production team when they've got hundreds of applications to sift through, so keeping it light but informative can help your chances of being noticed.

With the application in, we needed to get cracking. We wouldn't hear for a couple of weeks whether we'd been accepted or not, but we couldn't hang around twiddling our thumbs until then. The first port of call was to deal with the internal components. Coyote's main chassis is incredibly small, 410mm x 265mm (some featherweights are larger), and we had to fit in two drive motors, a battery, a linear actuator for the chainsaw, various wiring gubbins and no less than six speed controllers! Five brushed (two weapon actuators, two drive motors, one chainsaw actuator) and a brushless (chainsaw spin). Thankfully the majority of the controllers were compact and/or low profile, which reduced the headache slightly.

The biggest single component after the drive motors was the battery, so installing that first would determine how much and what space we would have for the other components. We opted for a single Optipower UAV lipo pack to power the whole machine; a 22.2V/6S, 8000mAh pack. I made a simple HDPE retainer and the battery fitted perfectly with padding:

Attachment 6902

The motors pictured are just cans from Bosch 400s. The innards were sent off to be rewound for 24V, however it took ages for the job to get done and we ended up having to buy some second hand Bosch 750s further down the line to substitute for them in order to let us test and tweak the drive as necessary. Speaking of the drive, the motors were to be controlled by a pair of Victor 883 controllers. I've long since dispensed with the fans that come with this range of speed controllers as the fan blades always break in battle and become useless. Removing the fans leaves a low profile speed controller, which meant that they fitted perfectly on a plate which doubled up as the battery cover:

Attachment 6903

The space between the motors was earmarked for the chainsaw actuator, so that left the gap on the left hand side free for the brushless ESC and three TZ-85s for the actuators, as well as the radio receivers, any power distribution bars, wires and a possible link location. Thankfully the TZ-85 controllers are easily mounted in any orientation, so I stacked them all side by side next to the brushless ESC and clamped them all down. As a speed controller unit it was solid, but I wanted a bit of give when they were installed in the robot to prevent shock damage, so the plate was located on some M6 threaded rod with springs to allow for bounce:

Attachment 6904 Attachment 6905

Below is the chassis with all the main parts bar the chainsaw actuator in. Very compact as mentioned, but workable. The chassis also needed some beefing up for heavyweight combat, so you can see a couple of pieces of steel angle welded to the base and 8mm aluminium plates cut for the back. This would not be the outer back armour so we could get away with ali. The front of the chassis would also need a steel plate welded on so that the jaw assembly could bolt to the chassis:

Attachment 6906

By this point we'd received confirmation that we had been accepted onto the show and production had been in touch to arrange our VT.

For those interested in what's involved, for ours we had Joe (camera) and Sam (sound) arrive at about 9.30am. They set about making the workshop suitable for filming, which involved blacking out the windows and door, bringing in various artificial lighting, setting up the camera and deciding what colour filters to use to match the robot (orangy brown for us). They also used a smoke machine for added effect. The day began with them interviewing me, asking about general details, day job, the why and how of the robot and getting some soundbites. They then swapped us around and did some more interviewing, this time with Rachel. Once they were happy with that, they filmed us working on various bits of the robot, shots of welding, grinding, soldering etc, then did our hero shots where we pose with power tools while trying to look intimidating. The last part involved getting clips of the robot working as much as possible. We still had no drive motor internals so we rigged up the jaws to operate and ran them up and down several times while they filmed from different angles. This led to a great shot in the edited VT from the episode where the jaws opened and they overlayed the sound of a dinosaur roar, which looked really cool. We also 'crushed' a spare go-kart wheel that was lying about which again made it into the edit. Sam then needed to get back into Aberdeen to get a train home so everything was wrapped up nicely by about 3.15pm. Several hours spent for a 30 second snippet in the episode but it's good fun to do!

Attachment 6907 Attachment 6908 Attachment 6909

So concludes today's chapter. Tune in next time as things start to get frantic with only three weeks to go!

You're welcome Jamie!

Fun fact, our VT was filmed at Laurence's workshop so that it wouldn't look like PP3D and coyote were built in the same workshop when they were :)

Frantic Finale!

Apologies for the lack of updates, life got in the way a bit. But time for one final update as we reach the finish line ready for Robot Wars! I may have to split this over two posts if I reach the photo limit.

So last time, we'd relocated to Aberdeen, made some progress and shot our VT. At this point we had three-and-a-half weeks to get the robot finished and in a working condition.

Expansive to-do lists were a common feature at this point, and one of the main outstanding items remaining was to make outriggers. These would be needed to stop the robot tipping forward when trying to lift an opponent, helping to maintain traction between the wheels and the floor. Due to a whole heap of variables that hadn't been determined at the time of the CAD design of the front jaw assembly, the outriggers weren't incorporated into the plasma cutting process so had to be designed, fabricated and fitted by hand. The process started off with getting a rough shape from holding a piece of cardboard up to the side of the robot to work out angles and shapes. Once we were happy with the look, they were transferred to metal:

Attachment 7054

We then clamped the pieces to the robot to check the fitting before continuing with the fabrication. The outriggers would have to double up as mounting points for the armour panels as, again, these weren't designed in CAD. By this point, any new work on the robot was being done ad hoc. We also realised that the two outrigger plates that would be the outermost panels on the left and right would have to have an extra crescent shape cut out of them to let us slide the wheels on and off without having to unbolt the jaw assembly every time. In addition to this, they needed some bracing plates across their width to try and stand up to horizontal spinner (cough, Carbide) attacks. Finally, to reduce the friction between the outriggers and the floor when lifting a robot thereby not impairing our ability to drive, we decided to fit Nylon blocks to the base of the riggers. When you put all of that into a picture, it looks like this:

Attachment 7055

Who needs CNC? All cut and shaped by hand with a grinder, all holes marked by hand and drilled on a pillar drill. Without blowing my own trumpet, I surprised myself a few times during the build with what could be achieved by hand with some care, precision and effort.

The outriggers then had to be welded on to the front jaw assembly, and once they were, we could test they worked properly:

Attachment 7056 Attachment 7057

Attachment 7058

What a relief that was! I was worried about the ability of the actuators to lift heavy weights and while I'm only 76kg, it didn't appear to struggle too much so I was confident in its fighting capabilities.
The eagle-eyed among you may also notice bike tyre screwed to the wheels. Having had problems with the same tyres on PP3D in Series 1/8, which were solved with bike tyre, we opted to go straight to that solution before even being near the arena. It also helped to give the robot a slightly more industrial look.

With robot builds, getting the big things out the way is important, but often the little things take up just as much time or are easy to forget. Another thing to tick off the list was power LEDs. We opted to go for 12V orange LED strips (not as easy to find compared to the standard colours!) which were mounted in the underside of the top jaw. We put a thin layer of polypropylene over the eye holes to diffuse the light and give a menacing glow. Looked great in a dark room but we found during filming that it was easily washed out by the studio arena lights, so wasn't that effective in practice:

Attachment 7059

Another LED feature that didn't really get covered on the show was a couple of RGB strip LEDs running down the back panel of the robot. The original plan for these was both to try and convey the 'mood' of Coyote and to double up as a rev indicator for the chainsaw (dim being low revs, bright being high). Rachel created the circuitry for them using an Arduino and bluetooth module which allowed the LEDs to be controlled from a phone app, giving a full range of colours and intensity. Again in practice and the heat of battle, it proved difficult to match the LEDs to the situation, so they became primarily aesthetic:

Attachment 7060 Attachment 7061

Before starting work on cutting the armour panels, which were to be a 5mm steel sub-structure with 10-20mm HDPE on top, we stuck everything we had on the scales to see where we were sitting:

Attachment 7062

Talk about close! Thankfully the wooden board we put on the scales to increase the surface area weighed 6kg so that, along with weighing full sheets of material but knowing we wouldn't use it all, meant we should be comfortably in weight. Fingers crossed!

We cracked on with the armour. The steel plates took the longest to cut and shape as they were pretty large and a bit awkward, though the brackets to mount the armour to the chassis at the rear were equally as time consuming despite being a fraction of the size. Nevertheless we managed to get them done, as well as knock together a top panel of aluminium and angled HDPE. Nothing great, certainly not Thor or Terrorhurtz proof, but it would give us protection at least. Pretendo-bot time:

Attachment 7063

Adding these panels gave a sense of scale to the robot previously missing. Coyote's on the small side, especially when it's sharing a garage with PP3D, but finally it looked like a heavyweight robot size wise! I couldn't help but grin when looking at the robot like this. Still a long way to go though.

I've reached the picture limit so will continue this in another post below.

Frantic Finale! (continued)

With the armour finally out of the way, the focus turned to the one last big challenge, one that I had been putting off again and again; making the chainsaw gearbox. I had a plan in mind but the trickiest part would be working out how to connect the chainsaw driver tooth to the gearbox and incorporating the chain bar support, all while making sure everything lined up with regards to bearings, pulleys, belts and shafts. The motor powering it was the same one that was in Drumroll for the drum, so 1100kV and ~2kW. At 6S, the motor would be kicking out 24,420rpm and the target for the chainsaw was 3000rpm (this was taken from the Series 1 house robot intros where Matilda's chainsaw is quoted as 3000rpm. Finding actual realistic chainsaw rpms was difficult as most gave figures in m/s and a long oval bar isn't as easy as a circle for converting those figures). We opted for a 9:1 ratio, putting it just under 3000rpm, so two stages of 3:1 was looking like the most compact way to achieve this. First up was playing around with the pulleys to sort the layout:

Attachment 7064

The whole mechanism would be sandwiched between three layers of 10mm aluminium plate. First thing was to get the pulley centres spot on. Devised a simple setup for this:

Attachment 7065

I didn't take too many photos of the process since time was tight, so a few drilled holes and some pockets bored out on the mill later, and it was starting to look semi-decent:

Attachment 7066

A few more holes, grub screws and fixings later, and voila:

Attachment 7067

Another thing I was proud of making from scratch. It wasn't the smoothest mechanism around, but it worked and didn't fall apart, so I'm counting that as a success. With the mechanism complete, attention then turned to mounting it on the robot. This part was no-frills, just a couple of pieces of steel angle bolted to the back with a pivot hole drilled in. A small 50mm stroke actuator conveniently attached to the bottom threaded rod of the gearbox and allowed the chainsaw to move through a decent distance:

Attachment 7068

This picture was taken about 10pm two days before we had to pack and travel down to RW. You can see the 'hackle' spikes on the top panel and the angled brackets on the rear corners for attaching the armour to (all of which had rubber in between to try and reduce shocks). The spikes were meant to be more like Dead Metal's but they were a last minute job so had to improvise and simplify. Things still to do at this stage included completely wiring up the insides, devising and fitting some sort of link mount for top and bottom, assembling the whole robot including Loctiting all connections, painting the thing and making locking bars. May not sound like a lot but all very time consuming. There were various other small things still on the to-do list that would have helped in general but that weren't too important and had to be sacrificed as we just couldn't do it all in the time we had. A 10-hour shift at the day job the following day was followed by working from about 8pm til midnight in the workshop, most of which was spent trying to wire it. I took this picture at the start of wiring then became too frantic to document the rest until we got to the warehouse:

Attachment 7069

The Thursday involved an early start to go and collect the hire van (a horrible-to-drive Fiat Doblo), getting everything packed up at my flat then heading over to the workshop for some finishing touches. Despite planning to leave by 5pm-ish to head down the road, the finishing touches took from midday til 9pm non-stop (which I didn't realise until I did stop!) followed by a 4-hour drive back home to my own workshop. The only finishing touch I did manage to photograph was a bit of painting on the front jaws:

Attachment 7070

We had to be at the warehouse for 8am on Friday, so after just a few hours sleep, and with several things still to do on the robot to pass tech check, I threw a few last things in the van from my workshop and headed to Glasgow. We then spent 15 hours in the warehouse, from half 8 in the morning til half 11 at night, getting Coyote finished. This included completing the wiring, cutting the rest of the HDPE to layer over the metal armour, working out and making locking bars for the front jaws, chainsaw blade and actuator and little cosmetic bits and pieces. Finally, after all that, we had a robot that had passed tech check and was ready to fight......Coyote.

Attachment 7071

Thanks for reading if you made it this far. Apologies if the content seems rushed towards the end but it perfectly sums up what it was actually like at the time! I won't go into details of how we got on in the competition as I'll mostly just be repeating what was shown on TV, but upgrades are on the cards after Coyote got well and truly mangled. I'll share the changes in update posts sometime soon. Til then, happy roboteering :)

Great write up, thanks Jamie

thank you for the write up, made for some excellent lunchtime reading.