Sunday, 12 March 2017

Hobgoblin 2016/2017

Hey. This.

I made this.


Yes indeed, it is that program about Robots having some kind of War. Fantastic. I somehow managed to blag my way into an entry slot despite not being qualified to even consider thinking about designing and building such a thing. This is how I managed it. Kinda.

Heavyweights are stupid. I half thought this beforehand but the design and build of one confirmed it. The sheer size, weight, complexity and strength required is absolute nonsense. It's a super hardcore challenge to get everything to work and function. That was the lesser problem, if you invest yourself and enough time and effort you can work through anything, finance on the other hand was an issue.

Simplicity and reliability has a price tag, and it's a fairly large one. NPC motors, vyper ESC's and a few Optipower packs and you'll have a  fairly bulletproof platform which can easily power itself to victory. However this is several thousand pounds. From the time between the start of the design and the competition (2.5 months ?) I worked out how much of my paycheque I could divert and still make rent, coupled with savings I had. Worked out to about £1400 tops. Seems like a lot, but heavyweights are stupid.

Ebay became a really close friend and ally during this whole adventure. I managed to get some great deals on things like motors, wire, connectors and some material stock.

I managed to snag some motors which ticked a lot of the "heavyweight drive motor" boxes, 24v, 500w (very on the low end but enough to move a weapon around) and had what looked like a sensible pinion cut into the shaft. Oh. and they were £25. Take that £350 NPC!

It turned out the pinion was very sensible indeed, it was a nice metric pitch and a low tooth count so I could have a single stage of gearing. Simpler and cheaper! I worked out that a 5:1 ratio would give me about 11mph which, again is low but should be okay to trundle a weapon around. I then found I could fit a slightly larger gear in the height and space so the end ratio was closer to 5.5:1. Slower but I wanted to give the system as much advantage as I possibly could.

In other Bargain Brushed Bonuses I managed to acquire some very large, very powerful motors to use for the weapon. The Lynch 170 is an immense piece of engineering, and that's the smaller one! 24v and 4kw constant and they have absolute bags of torque.

They came with a V pulley and trantorque which I removed (after some effort)

I could then start pushing around the pile of parts and proceed with Pretend-O-Botting posthaste

Can see the heap of Hobbyking Lithium that I intended to power the beast with. Some people have their issues with the Nanotech range, but from personal experience they have been absolutely flawless, better than "proper" brands I've used.

Having a rough idea of the layout and enough parts to accurately model things, I began work on the CAD model of the robot. Slapdash is probably a bit kind. I smashed out a bare bones model of what I wanted, because the deadline for RW applications was drawing close - and applying with an artists impression is better than just a pile of bits!

The model ended up being largely broken as my PC conked out 80% of the way in. Luckily I was able to retrieve a copy of the main assembly and shove it onto my work laptop but although  some
detail was lost, I managed to hammer it back into
a working state.

For the size and complexity of the robot the CAD model is hugely basic. Avert ye eyes Ellis Ware

Because of the sheer scale and holyfuck-ness of this project there was so many things that could go wrong. Stuff could just not fit or not be assembled in the way I anticipated.

One of the hallmark catchphrases that has been hammered into my head over and over during my short time in engineering is "trust, but verify" I had faith that I hadn't designed a total broken mess of aluminium and steel that would only be able to put together on a screen... But I'd better double check... for sanity's sake.

Luckily in my current line of work a lot of dicking about prototyping is done with 3D printing and we had a pretty decent setup. I set about making an exact scale HobGob, one tiny solid model then a larger "kit" at quarter scale. This would let me have something physical and colourful to shove under the collective RW nose but, more importantly I could see IRL that shit was where it was supposed to be.
Don't talk to me or my son ever again.


Long story short, it went together fine!

Somehow with enough smoke, mirrors and #buzzwords Robot Wars decided to accept my application. It was an amazing opportunity that I'm still stupidly grateful for, especially considering there were so many fantastic robots that didn't make the cut or pulled out due to time constraints. It was an honour to be a part of it.

And so, the build began.

This is the state the shed started out in... it got much, much worse as the panic build set in.

For quick assembly, as this robot had to just exist within like 6 weeks (heavyweights are stupid!) I elected to go with the tried and tested waterjet jigsaw puzzle style of frame.

Its a mishmash of 30mm, 25mm and 20mm aluminum and 4-8mm Hardox

Much of the material was suplied from stock by the cutting place but they didn't carry any alu thick enough, or the right grade. Not for the last time, eBay to the rescue! I managed to snag a pretty intense deal on 30mm 6082 plate as it was an offcut that I could fit both bulkheads onto.

 It seems every build I trump my previous record for "most bullshit heavy parcel" but bugger me, was this a nice suprise to come home to.

A big o pile of 6082, just what the doctor ordered!

 A quick once over showed that all the holes were where they should be, and I was (still am) hugely impressed by the cut quality on the 30mm profiles. On the right hand one there is bugger all taper, and the left one has a bit of a rough patch but is miles better than I was expecting


Its nearly done now, right?

Here I am just filing the slots for clearance so they fit together and add a bit of mechanical strength so the bolts holding the frame together are not in shear. I'm also marking out the holes for drilling and end tapping. Still constantly mocking the bot up to see if there are any cock ups...none so far!

Not in chronological order but so much of this build happened ontop of itself that its just easier to do it by category.  I'll just go through and detail the secondary machining and finishing of the aluminium parts here

 To accommodate the faceplates of the motor, the gearbox blanks needed to be faced down by about 12mm. Fairly simple operation and a decent amount of material removal. After this the hole pattern was picked up and drilled out and the bore adjusted for the 'lip' on the front of the motor.
 One freshly faced housing vs the untouched watercut blank. If I was doing this now and had more time I would have just totally machined these from plate but I was under pressure so needs must!

I didn't trust myself not to cock this up royally so a helpful reminder that this was a MIRRORED PART.

Forget what HR departments tell you, belligerent notes scrawled on any available surface  is an excellent way of communication.

All the countersinking was done with Super Happy China Import mini drill. A gift from my dad a few years ago and its come in super handy for just simple hole enlarging and countersinking when I can't be arse to slog the thing up to work to use the big mill.

Yes, before you ask I am painfully aware of the fact the part weighs more than the rest of the drillpress and I find it hilarious.

Counterbores were milled into the bulkheads to give the dead shafts a little extra strength. Due to general incompetence (and laziness!) the largest slot drill I had was 22mm and I didn't fancy enlarging it with the boring head.

Solution? Just make a stepped shaft. 25 down to 22 isn't a huge deal in my brain but its still an unbrilliant solution.

Finally my love of boring on the mill could be restrained no longer and I whipped out my head and set about precision mangling the Bulkheads to accommodate their rather chunky bearings.

These parts were brought too and from work on public transport... so if you were travelling around the south coast of England up to London around October time and saw some poxy skinhead waving around some big lumps of aluminium... that was me!

Bad photo (in keeping with tradition) but the Lynch motor mount had been countersunk and cleaned up. I elected to mount it using the 8 (10?) outer M6 holes as it meant the whole assembly was marginally lower profile and I trusted it more than the 4 closer M8.

I mean, the engineer or clever person would use both sets but I'm not either so.... moving swiftly on!

This is supported in the frame by 4 M8 bolts end tapped into the front and a large 30mm dia aluminium standoff which ties it to the frame via M10 countersunk.

Better idea of how the motor mounts, if you were wondering. Plus you can see the crap slowly starting to fester!

Righto, onto the drive system. Simplest and arguably the best solution to wheels that have intense grip is a plastic hub with bike tire screwed to it.

Ebay was my source of bike tire

And it's green. A E S T H E T I C

I also picked up some 5" HDPE round from my favourite online plastics store   alongside some suitably green plastic and some polycarb, both of which ended up not being used

Commence wheelification!

Finished wheel with the two dead shafts that go up the front of the robot and totally act as a structural member   They are 25mm dia with a step down to fit into the counterbores in the bulkheads. They have been tapped to M10

Then came the cutting of the tire. I just used a box cutter to slit the tread (AND SPENT WAY TOO LONG TRYING TO KEEP THE PATTERN ALIGNED) and trim the fat off the side of the tire

They were then just screwed around the hubs. Nothing fancy or simple and it yielded quite a pleasing result. A knobblier tire would have been better... but they weren't as green!

Continuing the drive theme, here is one of the gearbox shafts.

Yo dawg, we heard you like walkin, so we put steps on yo step, so you can step while you step.

This was one of those stupid solutions to stupid problems caused by money/time being in rather sparse supply. There was supposed to be a step down to 12mm at the end of the 25mm shaft to go into a bearing to support it that end (all 25mm bore were too large and would have interfered with other bores) The reason for the 17mm step is because I couldn't afford a 25mm keyway bushing, and I had no time to lose. The largest I had was 17mm, so 17mm it was!


Onward to  more lathework. I bored out the MOD 1.5 gears to 17mm (after buying a replacement as I already bored one out to 25 before realising my chronic bushing deficiency)

Once the bore was done they were faced and the boss machined off.

They game from gearsandsprockets. Who, as the name suggests, sell novelty pens and pantyhose.

Broach o clock!

One of the shafts had a ratty keyslot milled in it, and keeping in traditon with my high quality standards I only have a picture of that one.

Ignore the coupling. That just ended up on my desk at the time as a result of a sporadic outburst of actual engineering.

An finished(ish) drive assemby!

It done all went together like wot how it woz supposed to guv, honest!

However, the time had come to do something I was dreading. Talking on camera. Yes, it was at this point as the filming date loomed closer that film crews came around to interview the teams, and probably to check up that there was a robot there that looked vaguely working.  I quickly assembled old Gobbo into a semi reasonable state and swept the worst of the roboshit off the bench.

Ultimately it wasn't as bad as it could have been, the two guys who came down to film me were excellent, I'm just not the best at speaking or general human interaction at the best of time, and on camera its even worse! I'm just glad it won't ever be seen due to the lacklustre performance of HobGob... #spoilers

Buoyed up by them seeming semi impressed by goblin, i.e. not just pointing and breaking into unstoppable laughter, I had a nice little energy injection into the build. Fuck knows it needed it!

Work had me holed up in Scotland for a few days which sucked a bit, but their pay cheques allow me to do all this robot lark so I can't bitch too much.

However when I returned I had a package from Ferrous Frankie's Steel 'n' Rust Bargin Bin

Yep, it was a load of watercut hardox which made up the outer shell of my goblin

Here is my wedge, another slab of hardox which has been bent from a single piece. No welds to break and no extra heat into the material.

Panels slowly being fitted,  a fair amount of drilling and tapping for bolts to hold it all together.

All the armour plates on goblin are held on with a mixture of M10 and M12 buttonhead bolts

The guards that protect the drive wheels and pulleys has been  ground and filed to slot in to the aluminium and be captured by the top and side armour.

The top is being lined up for its mounting holes.

G clamps make up for a lack of hands and team members. One man builds can sometimes be a juggling act!

Beefy! Starting to look pretty tank-like by this point!

I'm sorry? What was that about beef? THIS is the beef. 40mm of Hardox 500 expertly plasmacut.

It is starting to all get a bit serious. Seeing the frame in the flesh did give me a slight twinge in my water about what the hell was I doing but when this beater dropped on my doorstep it suddenly all seemed very, very real.

Beater on floor, you can get a better idea of scale with the broom and you can see it has a super crisp cut quality for being plasma cut.

Now I've machined beaters before, as this picture of the successful featherweight "pendulum"weapon I machined a few years back shows but something this heavy, brutal and just so damn intense is a whole other kettle of fish.

Long story short, I didn't die.

That was the biggest twat of a machining operation I have ever done though.

The impossibly immense twattery stemmed not from the actual machining, but from the lathe being *just* slightly too short to be happy. I solved this by removing the thread chasing dial and a couple covers, and jacking the cross slide way the fuck out.

Machining the rather hench HTD8  pulley out to fit on the beater. I left a portion of the beater un turned down, then machined a square bore into the pulley.

This is a really horrible way of doing it, I mean, if it was done properly it would be fine. I kind of had no option as I this pulley was about half the price of the cheapest billet of aluminium to make a toothless pulley with more support

Needs must, and it got done just fine in the end. Post machining the beater needed a touch up with a grinder to fit the bore nicely as it wasn't totally square (plasmacutting fun)

The completed assembly. Things were starting to reach fever pitch at this point and I left basically the defining feature of my robot until the last possible moments. BECAUSE I HATE MYSELF?

Oh. And this tank track belt is just mind hobboggling. I can fit in it, and I'm not in any way small!

Heavyweights are stupid.

With much of the fabrication and physical build out of the way I could unleash my borderline artistic tendencies. The "vibrant" colour scheme of the renders I had thrown on during my RW application are probably a good 50% of why I got on. Boring robot painted purple becomes interesting robot... who am I to judge?

At this stage it all basically devolved into  PAINT THE BACK WHILE I WORK ON THE FRONT.

Stuff was hanging from the ceiling to dry. Meaning Bumblebot-Prime here kept backing into them and ending up with purple streaks on my everything. Dangers of the job I spose.

Coming together. Notice the contrast between the watercut edge finish on the main bulkheads (pretty sexy) and the armour mounts (total arse)

The top panel lining up with the link mount nicely, getting ready to lay out the holes for drilling and tapping.

The semi-pointless shoulderpads were added with some nice #branding in the way of vinyl decals cut at work... thanks Jim!

You get the idea, I like purple shit.

Shout out to muh homebois montana gold. Take note kids, and you too can replicate the hobgoblin look!

Fast forward a little ways and suddenly Hobgoblin is ready to take its first steps!

Sub optimal performance (but hey, why break tradition) but it fuckin moved. Unimpressive to everyone but me but it actually moves. When I want it to. Like some kind of robot or something.

Seriously this was a genuine achievement at this point. In the back of my mind the whole way there was serious doubt, shit like "it won't even move it will just burst into flames as soon as you poke it" the fact it trundled about quite happily in the driveway was a really nice moment.

If only we drove it about for more than 3 minutes..... OH WELL.

No time to stop as there were about 3 days left before it was off to Scotland for the filming!

Just a few minor parts left to make, like these little bits of steel which kinda like, hold the damn front on. Yep. Totally minor, nonessential parts!

Getting the wedge mounted, and hey look, An Heavy Hardox Beater!

To counter this all too progressive step forward, I took two back. The eagle eyed among you may notice a drive motor is ever so slightly not in the robot at this point.  It was such a blur at this point I genuinely forget why. Noob shit probably.

Near enough to being done, for me to quit taking photos. Had bolts to put in and a wedge to sharpen but other than that it was as done as it was gonna get.

Upside down goblin! Like the robo-slut it is it seemed to love being on it's back so much it spent the majority of it's television career in the same position. DAMNIT GOB I BELIEVED IN YOU.

And thus, the build concludes, it is all a bit of a mess and as I mentioned at the start the timelines are bound to be at least a little cocked up basically due to so much stuff happening. And Hobgoblin is a relatively simple robot!

Other scratch builds like the new Pulsar are magnitudes more complex and cutting edge. So much effort  and thought goes into these things, I think the fans and audience sometimes fail to appreciate this, because before I did one I sure as hell didn't!

I won't go into the event details too much as quite frankly, due to a mix of fatigue and alcoholism my memory is patchy at best. Most of the photos I took there could potentially contain the cardinal sin of ~SPOILERS~

It involved being in a very cold warehouse, marvelling at the wonderful (and weird!) creations people had made, lots of waiting, photos, interviews then finally trundling back to the hotel and propping up the bar talking shop with various people till 3 or 4 in the morning.

I am so very,very happy to have been a part of it all.

I'll address the elephant in the arena now I guess. "why didn't goblin work good or at all" As painful as it is to admit, the old adage PEBCAK (or rather PEBRAT for robots) In short, I done fucked up.

In an attempt to turn up/off the current limiting on the ESC to give the drive more grunt and must have turned it the wrong damn way. Bruh. Bruhhhhhhh. This turned it all the way down so the robot really struggled to move forwards and just noped at any attempt at turning. I found this out in the arena.

Well I know not to do that ever again ever. I love it when your dumbshit moments are on national television!

As soon as we got back from losing our fight we had to pack up and give our pit table over to another team so we couldn't open or check the robot then and there. First thing after unpacking (after sleeping then going back to work the next day!) I whacked the current limiting off and it was.. well

Well it does the thing at least... I know what to do now. Double up on the drive escs, no limiting, and jack the drive voltage up to 8s

Even better. I hope to have time to work more on goblin soon, I think it has potential to be a half decent robot. Its never gonna win titles but its not designed to do that. I build what I wanted to build within the budget I had.

Thanks for slogging through this general butchery of the English language with blatant cynicism and forced dank memes - at least the pictures were pretty.

Any questions, comments or particularly inventive insults can go to

Don't panic, and try not to die



  1. I personally was a huge fan of Hobgoblin, was a real shame you never got to hit that beater with anything. Good luck on the upgrades and hope to see you on the show again!

    1. Whilst being upside down, we managed to shave a lot off the floor and walls, I'm still prying hardox shards off the beater, which itself is untouched. Thanks for the comment!

  2. Wait, this thing only cost you about 1400 pounds? That's insanely cheap for a heavyweight. The design is solid and you clearly have the skills to build it, I hope you find some more cash to make the next version awesome!

    1. £1400 is a lot when that's all you have! Being 19 and just over minimum wage work don't leave an awful lot spare for hobbies, especially the expensive ones like robot wars. Cheers, its cheaper to improve than rebuild, so I should have a better run if I get another chance.

  3. Any plans to take it to Extreme Robots?

  4. What a great experience to be involved in. I've been a huge fan of the show for years and marvelled at the ingenuity and dedication of the teams that make it to the finals. I think even for something as complicated as your project, the underlying principle should always aim for KISS - Keep It Simple, Stupid!

    Arthur Greene @ CentralMM