Team Antiquarian build diary

All things antweight

Moderators: administrator, BeligerAnt, petec

MySolderIsOlder
Posts: 82
Joined: Mon Jul 30, 2018 1:22 pm
Location: Kent
Contact:

Re: Team Antiquarian build diary

Post by MySolderIsOlder » Mon Aug 13, 2018 8:51 am

BTW, In the spirit of 'keep it simple, keep it safe', here's the hi-tech flipper-return mechanism - just a length of 'powergum' from my sea-fishing goodies box:
FlippinUgly-4.JPG
FlippinUgly-4.JPG (95.28 KiB) Viewed 1952 times
FlippinUgly-5.JPG
FlippinUgly-5.JPG (82.61 KiB) Viewed 1952 times
Having messed about with Flippin' Ugly, several lessons were learnt:
1) The wheels need to be further back. Ideally right at the back and protruding out of the rear of the case. Otherwise it keeps falling over backwards and the flipper won't go back far enough to self-right.
2) The 600 RPM motors I used for this experiment were way too fast, bot was far too skittish during the brief periods when it wasn't lying flat on its back.
3) My wedge shape was way too tall - I could have made it half the height and still fit everything in. Next one will have a much sharper triangular section, which should more stability as well as making it easier to get underneath an opponent.
4) I need to be a lot more precise with measuring, marking and cutting -and especially with the bending. This first test was largely done by eye, working from a rough cardboard template. For the next one I want to work everything out with a CAD tool first. So... time to learn Autodesk Fusion 360 (which will also come in handy when I eventually treat myself to a 3d printer.
5) This is FUN! Haven't enjoyed problem-solving so much in years.
Stuart (Anthony's dad)

MySolderIsOlder
Posts: 82
Joined: Mon Jul 30, 2018 1:22 pm
Location: Kent
Contact:

Re: Team Antiquarian build diary

Post by MySolderIsOlder » Mon Aug 13, 2018 9:02 am

BTW, this was my very first (and very heavy) attempt at at robot - the Proof of Concept, or 'PoC-Bot'. Smiley face was added just so driver knows which end is forwards. Eventually he may gain a lifter/flipper arm one end, a bucket the other and a Pi or Arduino board in the middle, before being reborn as an autonomous 'house robot' - but that's a longer term project. I'd also have to find a reliable way to stop those pimoroni moon-buggy wheels falling off so readily...
PoC-Bot.JPG
PoC-Bot.JPG (142.59 KiB) Viewed 1951 times
Stuart (Anthony's dad)

User avatar
EddieJ
Posts: 79
Joined: Sun Apr 08, 2018 2:11 pm
Location: Dorset

Re: Team Antiquarian build diary

Post by EddieJ » Mon Aug 13, 2018 10:54 am

Looking great! Im envious of your speedy progress :D Wish my projects moved this hastily forward
I always love the look of see through bots like this. Definitely agree you can lower the wedge angle. Good luck with the next steps!
Bots:
DisinfectANT (Antweight), Cilit BANG (WIP Antweight)

Dazmund
Posts: 23
Joined: Wed Dec 27, 2017 10:38 pm
Location: Deptford, London
Contact:

Re: Team Antiquarian build diary

Post by Dazmund » Mon Aug 13, 2018 10:11 pm

Nice work so far! I just wanted to ask if you're sticking to the 4 inch cube rule on the size of your robot? I'm asking because you actually don't need to have the robot sit square in the cube. My own robot is 4 inches wide, 2 inches high, but 5 inches long, but it'll still fit in the cube if I laid it on its side within it.

MySolderIsOlder
Posts: 82
Joined: Mon Jul 30, 2018 1:22 pm
Location: Kent
Contact:

Re: Team Antiquarian build diary

Post by MySolderIsOlder » Tue Aug 28, 2018 1:23 pm

Having spent a few weeks mucking about with ideas and test-beds, I decided it was time I got at least one of the half-dozen antweights I hope to build before Christmas finished. So... without further ado, I give you...

"Shove-All"


(this is my first attempt at embedding imgur image-links, so fingers-crossed...)
Image

I had a Pololu Zumo chassis lying around from first attempts at making an R/C bot. An unfortunate incident with some bare leads shorting on something metal while it was sitting in a pile of junk on the bench resulted in four very angry AA Duracells and a plume of smoke, which rather messed up the Zumo's battery holder (as well as stinking out the workshop). So in search of a quick win, I dug the remains out of the scrap pile, cut away the damaged or superfluous bits and bent what was left back into shape with the help of a hot-air gun. Then I roughly cut and drilled a piece of 1.5mm Lexan to replace the singed top. I've used the original Zumo 1mm steel shovel (hence shove-all) blade as a pusher. With some careful adjustment, the top sits fractionally below the upper track surface - so it still gives good front protection but the bot can drive happily upside down if it needs to (which compensates to some extent for the lack of any self-righting mechanism).

Image

Innards comprise 2x100rpm generic N20 motors, DasMicro 2S6 ESC, Redcon 4-Ch DSM2 receiver and a Vapex 120mah 6V Ni-MH battery. Bit of a squeeze but it all went in eventually and for now it's still possible to use the original clip-in hatch on the bottom to remove the battery for charging. Obviously that power switch was a mistake - in a vulnerable position and it's too big, which made fitting everything else in a lot harder than it needed to be - so ordering up some smaller replacements. Might also fit a charging port so I can then glue the hatch closed.
Total weight currently is 153g. For domestic use I'm not fussed about the excess baggage, though it would be easy to trim a few grams off and get it ant-legal if we ever needed to. Fits inside the 4" cube with loads of room to spare.

Image

Thanks to the rubber tracks and low-rpm motors, traction is v good. No trouble shoving wooden blocks 3-4 times its own weight around the workbench. Should be even better on a clean surface once I've washed all the fine sawdust off the tracks.

So, that's the easy one done. Only five more to go. Next bot is going to be a lifter...
Stuart (Anthony's dad)

MySolderIsOlder
Posts: 82
Joined: Mon Jul 30, 2018 1:22 pm
Location: Kent
Contact:

Re: Team Antiquarian build diary

Post by MySolderIsOlder » Mon Sep 03, 2018 3:00 pm

On reflection, decided that my "Flippin' Ugly" experiment (see above) might be salvageable, if I could lower the aspect ratio and move the wheels back a bit. So... out with the tin-snips and on with the face lift. Basically chopped off the top third of the wedge;
Image

Ten minutes hacking with dremel cutting disc and file to extend the wheel slots up the back of the case meant I could shift the wheels further back from the centre of gravity. Now it doesn't flip on it's back every time I drive forwards!
Image

Flipper hinge is just a couple of holes with an M3 bolt, nylon washers and spacers, held on by a lock nut. Simple but effective.
Image

While the flipper linkage is still just a simple servo-arm extension, using a length of 'Power-gum' for the return mechanism (that stuff, normally used for tying stop knots on sea-fishing rigs, has a breaking strain of around 10kg, so should be ideal for this role).
Image

Now I have less height to play with, fitting it all in is a bit more of a struggle (if only one could buy a Nano Two!) so I think I might try cutting/desoldering the pins on the receiver and wiring it direct to the ESC and Servo - with some kind of jumper for binding..


After an afternoon of plastic surgery, I wouldn't say that my Flippin' Ugly duckling has matured into a beautiful swan - but it does at least function as a fighting robot. It drives pretty well, can self right effectively and shift small obstacles. Best of all, it only weighs 134g - so scope to replace the generic 9g weapon servo with something that has a bit more kick. Then just need to find a more effective way of securing the innards so the stay in place and we're ready to rumble.
Last edited by MySolderIsOlder on Mon Sep 03, 2018 4:08 pm, edited 1 time in total.
Stuart (Anthony's dad)

MySolderIsOlder
Posts: 82
Joined: Mon Jul 30, 2018 1:22 pm
Location: Kent
Contact:

Re: Team Antiquarian build diary

Post by MySolderIsOlder » Mon Sep 03, 2018 3:55 pm

Actually managed to get quite a lot of time in the workshop this weekend (and with TMS on the radio too - sheer bliss) so I've been experimenting with alternative ways to build using polycarbonate sheets. I have some large sheets of the stuff and various off-cuts lying around, left over from DIY secondary glazing panels (used to have a very drafty house), but they're 3-4mm thick and don't bend easily. I really wanted something that would give me nice clean right angle edges. Experiments using small brackets of bent polycarb were fruitless and I don;t have a 3d printer (yet). Then I noticed the Nutstrip on Fingertech's website https://www.fingertechrobotics.com/prod ... d=nutstrip - loved the idea but didn't fancy the cost, the shipping time, nor the imperial threads - so had a crack at making my own. Ordered up a cheap 1-metre length of 1/4" square bar in 6082T6 Aluminium, drilled a series of 2.5mm holes at right angles, tapped with an m3 thread and cut it into short chunks. Not without its challenges (if I carry on with this I need to improvise some better jigs to ensure holes and cut-offs are are more reliably equidistant - and Aluminium really gums up taps) but I think the idea has potential. Very light but strong way of joining two sheets at right angles. If I can figure out a suitable drilling jig, should also be possible to have a third hole (x-y-z axes) for joining three sheets at a corner.

This is the prototype for ServAnt - a basic (cheap!) all-servo flipper that I'm using to test the edge joints. Currently looks like a bowl of psychedelic spaghetti but I treated myself to a simple crimp tool and some Futaba connectors recently, so those servo leads are facing savage cut-backs.
Image

Detail of an edge joint, holding 3mm Lexan sheets at right angles with 6mm m3 screws (should have used 9mm but don't have any at the mo). Obviously cut it too close the to hole on the left side, though it still feels pretty strong :
Image

Also using my threaded joints for the flipper hinge - obviously needs spacers and washers and a lock-nut on the end of that bolt:
Image

ServAnt's flipper blade is a 10cm square of 0.5mm titanium foil. Really just need to sort out the flipper linkage and it will be usable. The elephants in the room though are those blasted cart-wheels - at 60mm diameter they're far too big, so yet again my wedge is too tall. So far I've not managed to find any smaller ones that fit a continuous servo's shaft. Really need something around 30-40mm. Worst case I might try cutting discs from the end of an aluminium (or nylon?) rod and drilling the mounting holes - then improvising tyres. For the latter I did think about sticking on carborundum paper for grip but that might not go down too well with arena owners! More practically, I found some large diameter shrink tubing with an almost tacky finish that might do the job and could have great traction - but that's an experiment for another day.
Image
Stuart (Anthony's dad)

MySolderIsOlder
Posts: 82
Joined: Mon Jul 30, 2018 1:22 pm
Location: Kent
Contact:

Re: Team Antiquarian build diary

Post by MySolderIsOlder » Fri Sep 07, 2018 3:25 pm

Current work in progress is a flipper called (for reasons that will become clear in due course) "Project Box Bot". Space for innards is even more constrained than usual so decided to see what I could do to slim down the Redcon 4CH DSM2 receiver + DasMicro 2S6A ESC combination. With some trepidation, snipped off then desoldered the header pins from the rx and wired direct to the ESC on very short leads. With all other leads in place, wrapped the Rx in heat-shrink first, then folder it back over the ESC and sealed the package with another layer of heat shrink. End results are satisfyingly compact - positively elephantine compared with a Nano but a great improvement on my previous set-ups.

Blue & yellow wires are for the two motors (controlled by AILE & ELEV channels on the Rx), red JST connector on the left is to the battery input, black socket on left is female servo connector (on THROTTLE channel - will shorten that lead later), while the two green wires are soldered to the normal bind terminals - plan is to connect those to a combined charge/bind port and switch on outside of the case so I don't have to remove the battery for charging.
Image

Close up of the ESC side of the package:
Image

...and the Receiver side (with the LED still visible from outside)
Image

Tested both motor channels and the servo - everything works fine though will be interesting to see how well it holds up in battle.

May order up some OrangeRx R614XN nano receivers next and try those, simply because it was such a chore desoldering pins from the Redcon. Bit of a fiddle to assemble but the saving in space and weight definitely made it all worthwhile.
Stuart (Anthony's dad)

MySolderIsOlder
Posts: 82
Joined: Mon Jul 30, 2018 1:22 pm
Location: Kent
Contact:

Re: Team Antiquarian build diary - resumed!

Post by MySolderIsOlder » Fri Dec 07, 2018 10:46 am

Well, it's been a while... I was getting frustrated by my fundamental inability to bend and join materials neatly so decided to take some time out and experiment with different construction methods. Toyed with the idea of buying a 3-d printer but given how much I need to get done by Christmas, can't afford the learning curve (not to mention insufficient funds or suitable space for one.) So decided to stick with polycarb for now but find alternative ways to use it.
As mentioned in previous post, I'd been trying to make brackets out of aluminium bar, drilled and tapped for M3 machine screws, so I could mount sheets at right angles. That idea seemed to have potential - but in the end I got fed up with the tap threads constantly clogging up with the soft metal - and it all felt a bit too bulky. So instead I decided to try using thicker polycarb for chassis parts, which I could drill and join on to with M2 self-tapping screws. Prototype build using this method is a Tracked Lifter, which ended up massively overweight (175g without a battery), so you'll never be seeing this beast at AWS. Having learnt a few lessons from this build though, I'm fairly confident I could make a Mark II that would scrape into antweight range. Something for the new year perhaps.

Meanwhile, here's TL Mark 1...

Image
I've used the Pololu 22-T track set, DasMicro ESC, RedCon 4-ch receiver and a Vapex 230mah Nimh battery (sadly it seems they've stopper making the little 120mah ones).


Image

This was also my first attempt at a 4-bar linkage. In case it wasn't obvious, the rear-hinge design was inspired by the Nuts n Bolts flipper. Main difference being that whereas Alex's flipper mechanism is beautifully optimized and neatly printed, mine is heavily over-engineered and rough-hewn out of 6mm Lexan. Does work though - which was enormously gratifying, even if most of the Turnigy's power is going into shifting all that weighty plastic.

Image

Nylon stand-offs on top are just to help self-righting. Without them it's a beached whale.

Image

The top-plate is probably thicker than it needs to be but since it's the main structural member, I wanted something that would not flex. If I ever redo this design, I'll carve out unwanted areas of that chassis with a Dremel cutting tool.
I spent ages trying to figure out how best to mount a servo until I twigged it just needed a little scrap of thick polycarb either side and some self-tapping screws. Sometimes even the smallest off-cuts can be useful!
The underside is protected (and all the innards held in) by a thinner sheet of polycarb:
Image



Because this current batch of bots are meant to be played with at home, I wanted them to have easily accessible power switches - but I also wanted to be able to recharge the batteries AND re-bind to one or other transmitter (so we could switch bots between the A and B teams) without dismantling anything. After a LOT of head-scratching and worrying about wiring, I eventually came up with a combined switch/charging/binding port design which I'm now trying to incorporate in all my builds. I'll include more detail of the wiring in my next post - but here's a preview.

Image

Basically, it's an SPDT switch super-glued to a Futaba J connector (keyed so you can't get things the wrong way round). When the switch is on, it connects battery +ve to the ESC & Servo. When it's off, it connects battery +ve to connector Pin 1 (right of picture). Pin 2 goes to the ground lead - so with a simple adapter I can use pins 1 & 2 as a charging port. Pin 3 goes to the signal pin on the receiver's Bind channel - so shorting pins 2 & 3 puts it into Bind mode.

Image
I've crimped up a couple of these simple adapters with a Futaba J on one end and JST/BEC on the other. The yellow lead (looped around the charge lead so I don't lose it) is another Futaba J plug for Binding.
Image

Image


Main lesson from this build was that although using M2 self tapping screws into drilled polycarb is a huge improvement on my aluminium nut-strip, it does have one huge problem - unless you're super careful (and sometimes even then), the heads of the self tapping screws shear off really easily, leaving you with a real problem. Forget about trying to drill them out - in most cases I just had to grind the stump off and drill another hole. Although I tried three different types, from three different sources, self-tappers just don't seem to be as strong as equivalent sized machine screws. Luckily I had a 'plan C' which worked much better - but I'll save that for another post.

Cheers all,
Stuart
Stuart (Anthony's dad)

MySolderIsOlder
Posts: 82
Joined: Mon Jul 30, 2018 1:22 pm
Location: Kent
Contact:

Re: Team Antiquarian build diary (Ffythiana)

Post by MySolderIsOlder » Fri Dec 07, 2018 1:55 pm

Having given up on self-tapping screws due to the problem with heads shearing off, for my next build I decided to try cutting threads directly into the polycarb. Plan this time was to have thick side panels into which I could screw thinner top/bottom plates. Like this:
Image
(that's 4mm Lexan for the side and the top bar, with 1.5mm for the skin)

Spent a few quid on an M2 tap and did some experiments with different pilot-hole sizes before settling on 1.6mm. Delighted to discover that not only was tapping threads in polycarb incredibly quick and simple compared to aluminium (no need for cutting fluid and the tap threads never seem to get clogged up) but the resulting fixings are really strong. Moreover, even with cheapest M2 machine screws, I've not had a single one break on me yet. So, with my M2 tap securely mounted in an old pin vice and a 1.6mm bit in the Dremel, it was on to the next build - this time, for the sake of variety, a front flipper.
Don't have time to think of a clever name at the moment, so here is 'Front flipper yet to have its amusing name assigned', or 'Ffythiana' (although I quite like the name 'Ffythiana'. Reminds me of Wales...)

I wanted to have a low wedge shape, with the wheels as far back as possible. That meant the usual motor brackets were not an option (besides which, I'd run out). Then while fiddling idly with an N20 motor one day I noticed that there was a bit of the gearbox where a line drawn on the tangents across the two brass pillars wouldn't actually impinge on any of the gearwheels. That got me thinking about cable ties - and in no time at all I had a new favourite motor mounting bracket. Cheap (few pence each), weigh almost nothing, don't take up much space and they're very 'strong & stable'. Just make sure they cut across the gearbox in the right place or you'll be going nowhere fast.
Image

Having received four duds out of four in my last batch of DasMikro's, and not having the time or patience to get replacements (and since I hadn't kept all the packaging, BangGood didn't want to know), I had to revise my choice of ESCs. The one you see above is the "2/3S 5A Dual-Way Brushed ESC" that I'd bought from The Atom Club - and although it's twice the size of a DasMikro, so far it's proved a lot more reliable so for now I'm going to standardise on these. For my next build I'll try snipping off the servo pins from the receiver and solder the wires on directly. Hopefully the space/weight saving will compensate for the larger ESC.

While on the subject of electronics, here's a more helpful pic of the combined power switch and charge/bind port I mentioned in my previous post:
Image
The white lead is the signal wire for the bind channel. Red & black for +ve and ground - the heat shrink tubing sections in the middle are where the power supply leads are spliced in with the servo and esc leads.

And here it is with the spaghetti folded away and everything secured with another magic cable tie...
Image


A big problem with my previous builds has been the rather irregular/uneven shapes resulting from my inability to make straight or parallel bends in polycarb. I figured this time it might help me keep things neat if the side were actually symmetrical, so I roughed out the design in Fusion-360 (which I've started learning) then printed just the side panel life size and roughly cut it out with some space all around. I then cut two blanks of 4mm thick Lexan (with the protective films still in place on both sides), fixed them together with double-sided tape and stuck the cut-out design on top in the same way. Then just clamped the lot in a vice and cut around the pattern carefully by hand with a fret saw, finishing off with a file and sandpaper. Finally drilled a hole straight through for the front hinge, before peeling the sandwich apart (removing the protective film). Hey presto - two identical side plates. Or at least they were until I had to botch out wheel arches because I'd miscalculated the overall width... But never mind - the basic method works a treat.
Image


4mm sides/top bar and 1.5mm back/base.
Image


Image
Top panel/flipper was originally made from a 10cm square sheet of 1mm titanium, off ebay. Was going to have teeth along the bottom lip to help grip the arena floor when flipping. I figured that if I attached the top of the acetate half way up, it could still get under an opponent but then flex during the flip. Unfortunately while trying to bend lugs to make the hinge I managed to snap them off instead - so for now I'm using 2mm polycarb. Might revisit the titanium idea though if I ever get the chance to bring this to an AWS event. If nothing else it would make pretty sparks while the spinners demolish it :-)

The nut & round-headed bolt you can see in the middle of the floor panel is a simple adjustable caster to help it slide. Without that one of the wheels was half a mil off the floor due to the flipper lip being very slightly out of true with the wheel axis. Adding that bolt makes it a tripod so both wheels are firmly grounded and although it raises the lip slightly, the acetate should compensate for that.

Note how the servo arm has a little spur off it that pushes the flipper directly for the first 15-20 degrees of its arc - after which the normal 4-bar linkage takes over. Means I can have a too-acute-to-push angle on the linkage in the resting position but more welly when it's extended. I allowed an extra 1mm of play in the elbow joint to accommodate that. Took a bit of fettling to get the right shape (and a long time before I could convince myself it was going to work) but now seem to have got it in the sweet spot.
Image

Image
Detail of the font hinge (just an M4 screw threaded into a hole in the side panel). Although I'm just using bog standard M2 screws for most of the bodywork, I also got some sweet little s/s countersunk machine screws from Modelfixings for where I need a low profile. They're not cheap - but exquisitely made and perfect for joints like this where I don't want anything sticking proud of the surface.


So... that's ffythiana. Not tried it in combat yet but drives well and self rights with great gusto. Will see if I can figure out how to load a video...
With the NiMH battery it's overweight but I also tested with a nano 2S lipo and it came in at 147g - so I finally have a bot that could eventually be entered in some future AWS (assuming one ever comes within range of Kent)

Thanks for reading,
Stuart
Stuart (Anthony's dad)

Post Reply