Register To Comment
Page 1 of 8 123 ... LastLast
Results 1 to 10 of 72

Thread: Ensign Wedgeley Crusher (FW Sportsman)

  1. #1
    McMullet
    Guest
    Hi folks, and thanks for looking in!

    This thread is a retrospective build diary for the robot we took to the Robodojo Sportsman competition on the 27th. This was our first time building a robot and was quite a learning experience. I intended to start a build diary here pretty much every day for the last month, but every day I was too busy building or designing the robot, trying to find/order the right bits on RS, or sulking because something had gone wrong. Hence this will (mostly) be a look back at what I did at the time, probably more for our benefit than for anyone else's interest.

    First, a brief introduction… we are team Designed for Demise, consisting of me, Becky (RainbowCello) and Simon (stl100). All long-time Robot Wars fans, first time builders. If anyone’s interested in the team name by the way, it’s derived from a design approach to ensure that broken satellites don’t hit people or other satellites (see here). When I first heard the term I thought it sounded like a great name for a metal band, hence the gothy font on our team T-shirts.

    I always fancied having a go at combat robotics (like, since the first episode of RW in 199, and when I left my job a year or so back, my colleagues (knowing this) very generously gave me a DX6e as a leaving present. Since then I have bought a few bits and wired up some gubbins in a Lego chassis, but found it hard to start building a proper ‘bot. Here's a Lego sort-of-beetleweight (no ESCs or battery at the moment):
    DSC_0570.jpg

    The Robodojo Sportsman competition offered a slightly lower entry bar, and so when the second FW event was posted up at the end of last year we rather impulsively signed up, on the basis of a discussion lasting about 30 seconds:

    • Me: Can we build some kind of crappy robot in a month? For that Sportsman thingummy.
    • Her: Errrrr... yeah probably
    • Me: OK I’ll sign us up then. Shall I say definite? It says priority for definite attendees. I’d better say definite.


    Ten of your Earth pounds later and we were IRREVOCABLY COMMITTED to definitely build said crappy robot, with absolutely no way of backing out. Of course, we didn't have a month to build the robot, because we had visitors the first week of January and then I was away for work until the 9th. Still, that left nearly 20 days. Of which I would be away for 9 days. So actually, 11 days.

    So, on to the robot itself… To give us something to start with, we set ourselves a few design constraints (in addition to the FRA/Robodojo rules):

    • Use 2x Argos drill motors for drive with a FeatherTwo ESC
    • Weapon should be a simple "grabber" arm run off a gearmotor, as that seemed like the easiest thing to bodge together quickly
    • Basic design has to be able to run as just a wedge in case we don't have time to finish the weapon (likelihood: high)
    • Make from mostly HDPE as it's easy to work with (admittedly this was a theoretical assumption, I'd never tried before)


    PREP - up to 9th Jan
    I ordered a cheap gearmotor from eBay in the first week of Jan, as that was the only "long lead" item we had to worry about (though in the end it arrived from China in a few days). I selected 30 RPM by sort of moving my arm around in a 180° arc in what felt like about a second and trying to imagine that my arm was the weapon on a hypothetical robot. On reflection, I do not recommend this technique for component selection.

    We also spent a bit of time discombobulating Argos drills; the first one went a bit wrong, as I didn’t really read the motor removal instructions and just attacked it, eventually ending up with this:
    DSC_0569.jpg
    I may have lost my temper with this one. If anyone has any idea how to get that last bit of chuck off, I’m all ears.

    Finally, I ordered a relay switch and some HDPE from Direct Plastics.

    10-11th JANUARY (Design)
    Having absolutely zero points of reference for how to start designing a robot, I had to spend some time vaguely figuring out the layout for the thing in my head. I've been trying to figure out Fusion 360 for the last year so I fired it up and started sketching. My current level of expertise with F360 consists of frowning at the computer and wondering why it has to be different from SolidWorks.
    The two drive motors and the weapon gear motor were the only known quantities, so most of the chassis was designed around fitting them in. I then added a box on the back that could accommodate something around the size of a 4s 3000 mAh LiPo plus some extra space for other electr(on)ic gubbins.
    The fruits of my initial CAD work, after a couple of days' work:
    CAD1.pngCAD2.png

    Mostly made of HDPE, wheels at this point completely unknown provenance (other than "about 120 mm diameter"). The front prong/wedge and sides of the overhead grabby arm were meant to be made from some steel bar bought from eBay, with teeth made from M6 stud. The mount for the gearmotor is designed to be cut from a piece of scrap ally that I saved from the bin at work.

    Up next: actual building, which I suppose is what this forum is meant to be about

  2. #2
    McMullet
    Guest
    12-13th January (Build days 1-2)
    Over the weekend, we started cutting out some of chassis parts and fitting them together. I started badly by cutting the side plates 20 mm too tall. Lesson learned, measure twice cut once and all that. However, after a couple of days we had something not entirely dissimilar to part of the CAD model:
    DSC_0575.jpg
    Most of this time was just figuring out how to cut and fit the parts together accurately, or at least accurately enough for the thing to function. Despite checking, quite a lot of things ended up being slightly the wrong size, or attached in slightly the wrong place. "Slightly" here may be a bit of a charitable word to use. Although we had some bits after this weekend, we spent most of it experimenting with assembly techniques.

    We did all the the fabrication and assembly at Nottingham Hackspace, by the way - there’s a great selection of tools there so we had access to a pillar drill, bandsaw, disc sander, etc… If you're in the area and haven't been, check it out!

    I was then off away with work for 3 days. I ordered a few bits from HK and RS but couldn’t do any real work.


  3. #3
    McMullet
    Guest
    17-20th January (days 3-6)
    Around this point I realised that we really rather urgently needed some wheels that would fit on the drill shafts. I had originally planned on using the 5” Colson wheels, but I couldn’t find anywhere that could actually deliver them in time. Turns out there are a lot more 4" wheels out there, but we had already built a robot that was about 80 mm high, so that would be cutting it fine with the Robodojo area including a permanent 20mm pit.

    Other options were therefore required. I chose scooter wheels in the end because, 1. They’re designed to take impacts, and 2. They’re widely available. The problem is of course, scooter wheels normally just sit on a dead shaft with bearings.

    To fit the wheels onto the ⅜” UNF drill shaft I chose these little things. Very handy. That sentence took about 15 seconds to write, but it took me about 4 hours of searching to find them.
    Over this weekend we managed to get the following done:

    • Mounting system to hold the motors in place (Becky and Simon sorted this). I’ll show it in some more detail when I put up the post-event teardown.
    • Wheels and hubs all fitted together and mounted to the motors. This was actually pretty straightforward; just knocked out the bearings from the wheels, then removed the flange in the middle with a 22mm spade drill. The hubs were then just an interference fit (which is one of my favourite engineering euphemisms; whoever made "we wodged it into the hole with a vice" sound so professional deserves a medal)
    • Connectors soldered to the motor and ESC as required. I used Anderson PP30s for the most part because I like the modularity (also because I’m familiar with them from work and I needed something comforting in this sea of unfamiliar things), and some XT60s.

    Of course, that meant that we could actually plug the drive system in for a quick test:


    What you see there is pretty much all we had finished at this point. Admittedly it doesn't look very different to the previous weekend. I did try to cut out a weapon mount but drilled ALL the holes in the wrong place which made me VERY SAD.

  4. #4
    McMullet
    Guest
    24th-25th January (days 7-8 )
    The two sources of panic for this week (alongside the general “there’s still essentially a whole robot to build aaaaagh” panic) were:

    1. The OrangeRx receiver seemingly throwing some kind of wobbly. I initially thought the FeatherTwo had died, which would’ve been game-over at this point. I posted about this on the forum and reached a state of “this will probably be OK” thanks to everyone else’s reassurances.
    2. Fuses. I ordered a bolt-down fuse from RS, and after a long boring series of mixups by RS I eventually determined that I would not receive it in time. A lack of fuse would’ve meant a failed tech check and long, sad journey home. Fortunately, VW used to use bolt down strip fuses, and my local autoparts shop had a few 80A ones in stock. I don’t think the guy behind the counter quite knew what to make of my joyous celebration at the sight of 3 strip fuses….


    With that final component in place, I was finally able to sort out the wiring. At this point I should say, I made absolutely zero consideration of the harness layout inside the robot when sizing the box at the back. It was just “approx size of a battery + a bit”. I’m used to working with 20 or maybe 18 awg wire which takes up no space. 12 and 14 awg does not bend nearly as much. It didn’t help that I forgot quite how big SB50 connectors are (that's what I used for the link).
    In the end it all JUST fit in, but there wasn’t much space to spare and some of the wiring was curved rather more than I’d have liked. I used ring terminals and M5 bolts glued into the base to keep everything secure.

    It was a late addition to strap the battery on to the lid of the electrics compartment; that was basically the only way to fit it in, but in retrospect I like the outcome because it adds some “poke-yoke” safety to the setup - basically to remove the lid, you have to disconnect the battery, so you can’t forget to unplug it when opening up and leave something live by accident.
    Note the extremely shoddy addition of safety lights from the drills; these are only legal because the lid doesn’t fit properly, so they shine out of the gap at the back:



    Next job was to plug some actual motors into everything and see if they turned the right way.
    SUCCESS! The drive worked.
    DISASTER! The worm motor for the weapon didn’t do anything.
    A quick spot of fault-finding revealed that the relay was switching, the power supply to the switch was fine, and the motor still worked when plugged into something else. That narrowed it down to the connections in the relay. I took a wild guess at how I thought the switch was meant to work, soldered a bit of wire in place, and struck gold! We now had a fully functioning electrical system.

    26th January (day 9)
    This was crunch time - the day before Robodojo. We were staying nearby overnight so had to travel up this evening. At this point we had no components for the weapon system other than the motor, so there was still plenty to do.

    First job for me was to cut and drill out the mounts for the motor. This went slightly better than last time, as I only drilled *most* of the holes in the wrong place. I need to work on my measuring and marking skills. Untidy as it was, with some extra clearance it all fitted together, and gratifyingly both were wrong in the same way so at least the two piece matched up.

    Meanwhile, Becky and Simon were cutting and fixing the remaining HDPE sections, and making the front prong that forms the lower jaw of the weapon. This was originally meant to be made of steel but we decided given the time available that HDPE was all we could manage in the time available.

    The final fiddly job was to make the two bosses that fit the weapon gearmotor shafts. These came from more bits that I rescued from the scrap bin at work - I just had to drill out an 8mm hole down the centre, then drill and tap three M5 holes in each (one as a set screw to fit the D shaft of the motor, two to attach the upper jaw.

    Having gotten all that sorted, all that remained was the aforementioned upper jaw. At this point it was probably getting on for 7 PM, so again fabricating it out of steel was not on the cards, as was cutting up stud to make metal teeth. The original plan of an extra wide jaw went out of the window, and I just cut two HDPE jaws on the bandsaw, with a largely aesthetic pattern of teeth.

    Still, shortly after 7:30 PM on Saturday, nearly 13 hours full before tech checks started, we had something resembling a completed robot:
    DSC_0579.jpg

    Quite a lot of wonky bits and some bits that straight up didn’t fit properly, but it was finished and it operated! Note I didn't say "worked", as that would imply it was effective in some way. I think the cradle was the only thing I was truly happy with...


    All that remained was to go home, eat dinner, pack and still get to the pub before last orders, which we just managed (it was a close thing thanks to an accident on the M1 and some bad advice from google maps).
    Last edited by McMullet; 31st January 2019 at 18:18.

  5. #5
    Ocracoke's Avatar
    Team Kaizen

    It is heartening to see a few newcomers come into this sport, go out to achieve something and then coming out with something that actually looks like the design they envisaged, so well done. It can be quite a learning curve to get to grips with what works and what doesn't. I guess it being in the Sportsman event would have come back (hopefully) in one piece?
    Team Kaizen - Build Diary for all the robots

    AW: Amai, Ikari, Lafiel, Osu, Ramu
    BW: Shu!, The Honey Badger
    FW: Azriel
    MW: Jibril, Kaizen

  6. #6
    Well that was a really enjoyable read. I'm not on here often much anymore but glad I had a look-in today.

    Great job on getting your first bot together so quickly and as Ocracoke said, looking very close to your original design. The funny thing about wonky holes and such is that you know exactly where they are, but to most observers, they couldn't spot them. Looking at your final pic before leaving for the event, everything looked like it was where it was supposed to be. Almost every robot has a wonky something (with exception of the Nippers and the Dutch guys probably ) so don't worry too much about it!

    Looking forward to seeing how it evolves

  7. #7
    McMullet
    Guest
    Quote Originally Posted by Ocracoke View Post
    It is heartening to see a few newcomers come into this sport, go out to achieve something and then coming out with something that actually looks like the design they envisaged, so well done. It can be quite a learning curve to get to grips with what works and what doesn't. I guess it being in the Sportsman event would have come back (hopefully) in one piece?
    Thanks!

    Yep, we didn't have too much damage; just enough dents and scratches to feel like we'd "been there", and a slightly busted up motor after one fight.

    Quote Originally Posted by RogueTwoRobots View Post
    Well that was a really enjoyable read. I'm not on here often much anymore but glad I had a look-in today.

    Great job on getting your first bot together so quickly and as Ocracoke said, looking very close to your original design. The funny thing about wonky holes and such is that you know exactly where they are, but to most observers, they couldn't spot them. Looking at your final pic before leaving for the event, everything looked like it was where it was supposed to be. Almost every robot has a wonky something (with exception of the Nippers and the Dutch guys probably ) so don't worry too much about it!

    Looking forward to seeing how it evolves
    Cheers, glad you liked it! And thanks for the inspiration, I think there's more than a little hint of Coyote in EWC.

    I think the only really wonky bit was the front of the wedge, because the front plate is misaligned from the base by about 10 mm (curiously that's also the thickness of the HDPE, it's almost like *someone* measured it wrong then didn't have time to fix it). Other than that you're probably right, it's only noticeable to use because we've spent so long looking up close...


    Next, quick lowdown on the Robodojo event...

    We arrived just after 830, still stressed from the day before and mostly excited about the prospect of an egg butty from the van outside (they offered the addition of a hash brown to the egg butty for a small fee, which was a cause of great excitement).

    Still, we were technically there for ROBOTIN', not eating fried food, so we staked out a quarter of a table and set out the Ensign for tech checking. This went fine in the event; turns out the AR610 does shut down the drive on loss of signal, or maybe the FeatherTwo does or something, but in any event the wheels stopped turning when the Tx went off so GOOD TIMES. We just needed a locking bar for the weapon, which was swiftly bodged from a piece of scrap HDPE (we took all the offcuts with us, I'm gonna claim for things like that, it was really because I couldn't be arsed to unload the car the night before). After that, we put the spare battery (bought from HK 3 days earlier) to charge, and obtained an egg butty with a hash brown in it which lived up to ALL my expectations. It was then a matter of waiting for the first fight, and chatting to our fellow competitors (who were all excellent folks). We drew straws to see who would drive; Simon lost, so he had to do it.
    EDIT: in the finest Scrapheap Challenge tradition, we also spent this time decorating the robot (using sharpies).

    FIGHT 1: Tantrum
    Tantrum was a really quite intimidating thwackbot, with a nasty chunk of broken saw blade as a weapon. If one were to stretch to the assumption that Wedgeley's weapon and drive would function exactly as intended, there wasn't much for us to grab onto and Tantrum's tyres were huge and grippy, so grabbing was unlikely to give us much control.
    Anyway, here's the fight:

    Fair play to Simon here, no one had ever driven this machine before and his credentials for the job consisted of "had driven his nephew's RC car at Christmas". Managing to exact some semblance of control was quite an achievement, but the conclusion was probably inevitable given how twitchy the robot was to steer.
    Nevertheless, in this fight the robot demonstrated effective basic functionality - drive wheels made robot move, weapon control (sometimes) made weapon move. (Sometimes the weapon just didn't seem to respond, which is annoying; sometimes the relay switch just doesn't seem to react. Still investigating that...)

    After the fight, Eoin (Barróg) made a helpful suggesting, to add some wheely bars to the back of the robot to get a bit more control. We dug into the offcuts bag again and, with some manual reshaping came up with these:
    DSC_0591.JPG
    Note also the gouges on the bottom corner, courtesy of Tantrum...

    FIGHT 2: Midas
    This was another robot that looked pretty scary, a 1500N lin-ac driven crusher that looked about 4 times the size of our machine (more on this later). Becky lost the coin-toss and was selected to drive this time. I think the rear skids helped a bit but Ensign Crusher was still pretty undriveable:

    In the pub the night before, Simon set us the objective for the weekend of at least starting the second fight. So at this point we had achieved our goal for the weekend, and anything else from hereon was a bonus...

    We'd also had a sweepstake between the three of us on the drive up as to how heavy the robot was. We'd very roughly estimated around 8 kg, based on the time-honoured engineering technique of rounding everything up. Sometime around here, Dave the tech checker asked how much the robot actually weighed. I said I didn't really know, but it was probably a bit underweight, so we went to put it on the scales. At this point I became irrationally paranoid that it was going to be massively overweight; turned out I needn't have concerned myself, as Ensign Crusher tipped the scales at a puny 3.2 kg - we could literally have built 3 duplicates and entered them as a FW cluster. That explained the controls being so twitchy at least.

    Fight 3: Mattock

    At this point I was starting to wonder if I'd offended someone, as we kept getting drawn against all the really mean robots. Mattock is pretty beefy looking and has a really nasty pickaxe. By elimination it was now my turn to drive, so I took the controls with some trepidation. At least I had the second-hand experience of watching Simon and Becky try to drive it.

    My take-aways from this fight were, a) we lasted more than a minute and b) we got pushed out of the arena rather than driving out under our power. That's PROGRESS!
    However, the robot got pretty banged up in this fight; one of the motors was damaged and had to be swapped out, and one of the wheels got pushed off its hub, requiring some gentle re-positioning using a large G-clamp (I think Chris from Luna-Tic lent me the clamp, so ta!). In addition, though we didn't realise it, the weapon gearmotor was toasted in this fight due to a Good Hit from Mattock. More on this later!

    Fight 4: Chimera's Revenge
    Sadly Chimera's revenge had to retire from the contest, so we won this one by default. It would've been preferable to have the fight, even though we'd have struggled against another big chunky thwackbot, but I suppose the silver lining is that we had a win of some sort from the weekend...

    Fight 5: 8-way rumble
    There was some talk of us getting a final whiteboard fight to make up for missing our 4th fight, but in the end we participated in a final melee of all the Sportsmen (Sportspersons? Sportsrobots?), which sounded like more fun anyway. I lost the dice roll so drove again. There were 8 robots in the arena for this one:

    Unfortunately we got flipped after a couple of minutes and couldn't self-right. I think the gearmotor could've flipped us back over if it'd been in good condition, but (as I'll show in the next post), it was not in a condition I would describe as good. The last driving gear in the box had sheared in two, hence the sporadic waggling seen at 1:49 - every half turn of the driving wheel, it would catch and turn the shaft, then slip back again.


    Overall, we had a fantastic time at Robotdojo. I was really chuffed that we managed to get there with something that worked, regardless of how well we did. I think the Sportsman competition was a great idea; I can't see that we'd ever have gotten around to building a robot otherwise. It's not just that we didn't have to worry about taking the robot home in a bin bag, I also like the fact that the pressure to perform is not too high - it's specifically meant for beginners, so you don't feel like there are too many expectations. That's a great relief when you're not even sure you'll have a robot, never mind a good robot or the ability to drive it competently. So thanks to all the people involved in running the event!
    Last edited by McMullet; 4th February 2019 at 11:05.

  8. #8
    McMullet
    Guest
    Today I got around to stripping the robot down to look at the damage done.


    First up, the previously-seen weapon arm flapping issue:


    Pulling this apart, I found a smashed up gear (already spoiled above). The broken half had embedded itself in a lump of grease at the top of the 'box, which I suppose was lucky:

    Not much I can do about this right now, but it's actually quite a nice little piece of kit so might try and replace that bit some day, if I can.
    P2023779.jpgP2023781.jpg

    I'm not sure if this was caused by Mattock or my cack-handed overtightening of the bolts securing the weapon to the shaft:
    P2023789.jpg

    Also within the weapon subsystem, for curiosity here is the BEAUTIFUL rework I did on the relay switch:
    P2023777.jpg
    I know what you're thinking, what rework right? There's no visible bodging on there AT ALL.


    Next, drive. Both sides were pretty messed up. My bodged-together wheels just about survived the day, but I think some refinement is needed. First, there's not enough contact between the reverse-threaded screw that holds them in place and the inside of the hub, so one of them just came undone by the frankly rather unsubtle means of stripping the thread. I mean, rude or what.
    P2023767.jpgP2023765.jpg

    The other problem is, press-fit components can obviously be un-press-fit, with sufficient application of force:
    P2023783.jpg
    Sufficient forces were evidently applied. This wheel was also hard to get off, since obviously the motor can turn when you try and unscrew it. In the end I jammed something against the motor fan to hold it still and then it came off.

    Both motors were a bit wobbly. They sit in a hole drilled in the side of the chassis and a U shape in an internal bulkhead, then there's some complicated clamping gubbins to grab hold of the various protuberances on the gearbox and motor to hold them both steady.
    P2023771.jpg
    The big flangey screws (1) sit in the slots in the motor mounting face (A) to stop the whole thing slipping axially, while the two countersunk screws (2) sit in the slot in the top of the motor (B) to stop it rotating.

    Two blocks on the base stop the gearbox from rotating:
    P2023769.jpg

    Again, this worked pretty well but a few things need refining. First, Argos are not so kind as to make all their motors the same:
    P2023786.jpg
    Don't think a 4mm screw head will fit in that slot...

    Second, the axial restraint puts a lot of force on the flangey screws, which has damaged the slot in the motor, and on the bayonet-style fitting from between the motor and 'box, resulting in breakage, and consequent wobblage:
    P2023772.jpgP2023773.jpg

    All in all though, this would probably keep trucking for a few more fights (maybe with some extra duct tape to be sure).


    Finally, a couple of shots I never took while building, the wiring, including battery mounted for combat, and location of the F2 ESC:
    P2023775.jpgP2023776.jpg


    A few lessons learned:
    • Mount the wheels and drive motors better - those things take some whacks.
    • Cheap eBay gearmotors cannot take direct impacts (who'd have thunk it). If we want to go clampy, we may need to go big cogs or preferably linear actuator.
    • Do a proper mass budget. Although in some ways I was happier not knowing the first time around.
    • Sort out the mixing on the Tx so the robot doesn't rotate faster the Magnetar's drum when you try and go around a corner (thanks again to Eoin for the hint there - I'll have to look through the Tx manual)
    • Give the robot a shorter name - understandably, it rapidly became known as just "Wedgeley".



    We've definitely got the bug for this so expect to see us again at Robodojo (probably in May) with something bigger and better (or failing that, heavier and repaired).

    Cheers for reading!

  9. #9
    Al_'s Avatar
    Member

    This has been a thoroughly entertaining read, thanks for going to the effort of typing all this up! EWC both looked and performed great, especially for a first bot at your first event.

    If you want motor mounts, have a search for Roboteer Nat on here, he sells mounts for the Argos drill motors. (I don't think Shakey sells them anymore but please feel free to correct me if I'm wrong?) As you use HDPE for the bot, a lot of people make their own HDPE wheels of whatever size/thickness they need and use various different means of putting tread on them, again, lots of into on here for that. Weight budget is also good, 3.2kg is very light for a feather, are we sure those scales were reading correctly!?

    Anyway, looking forward to see how this progresses, and see you again at Robodojo!

  10. #10
    Ocracoke's Avatar
    Team Kaizen

    Nice write up. I suspect Eoin may have already suggested this but you may want to look at your TX to see if it can do exponential stuff with how much power your motors get in relation to where your controls are. I had to do this on Kaizen/Jibril to get the axe to fire at a good rate without much input from me. It might make EWC that little bit more controllable.

    Yes, definitely consider adding more weight to it, perhaps over the front for balance?
    Team Kaizen - Build Diary for all the robots

    AW: Amai, Ikari, Lafiel, Osu, Ramu
    BW: Shu!, The Honey Badger
    FW: Azriel
    MW: Jibril, Kaizen

Register To Comment

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •