|This is our first ant-weight robot:
"Robot Of No Name Yet"
17 February 2002
Went to Modelworld at the Brighton Centre and saw a variety of antweights on display and being demonstrated. We were impressed by what could be achieved and thought it looked great fun! We spent ages talking to Pete Collier about building Ants and modifying servos, his enthusiasm was infectious :o) We left with a copy of his plans for "Pants" and ideas of building our own Ant.
Two weeks pre-occupied with thoughts of robots and trying to find some reasonably-priced R/C gear. Also tried to find a cheaper alternative to modified servos for drive. Came to the conclusion this option is not simple as small, light, high-power, low-cost motors are not easy (possible?) to find, then you have the mechanical problem of building a gearbox, you need a speed controller, etc etc. Figured out it's better to use modified servos, at least to get started with.
2 March 2002
We took the plunge and made the biggest investment necessary to get started - bought a Hitec Ranger (3-channel 40MHz FM) and two HS-81 servos. James spent his pocket money on a pair of 1.75" aeroplane wheels which looked about the right size for a robot.
I went to Maplin in search of batteries but was told the 4.8V 150mAh PCB-mounting type had been discontinued - despite the Maplin website claiming thay had some in stock! Ended up with four 700mAh AAA's. At twice the weight this could be a problem, but it will get us going.
I spent the evening checking everything worked, then modifying the servos. The first one (with some trepidation and a great deal of care) took a while, but the second one literally only took ten minutes. I did not glue the feedback pots, so that the servos can be converted back to standard using a replacement gear set and reinstating the stop inside the case using a small pin (or a replacement case). The pots do move slightly during use, which means re-trimming every so often, but I will replace the pots with fixed resistors (see below) which will cure this.
One problem with the Hitec Ranger is that it has no channel mixing facility and the control sticks go up/down and left/right. It is impossible to make them both go up/down. So we need a mixer to go in the robot - more weight :o(
3 March 2002
James and I fitted the wheels to the circular servo horns by first filing the ridge off of the top surface of the horn so it fitted snugly against the wheel, then glueing the horn to the wheel with superglue. Our servos came with some longer screws, but they were not quite long enough. We milled away some of the wheel to allow the screw head to sit in a recess. This seemed to work well.
We found an old hard, clear plastic box which had once contained "something" from RS Components. Its lid measured about 70 x 90 mm - perfect! The lid has a small return all the way around, perfect for fixing the body to. Some of the return was cut away to clear the wheels.
|As this was to be a prototype/experimental robot, we wanted all the components to be easily recyclable so we didn't want to glue the servos to the base. We tried double-sided tape, but it didn't like the texture on the servo cases. We ended up with self-adhesive velcro which seems strong enough and certainly makes for easy changes.
I made up the battery pack and we stuck it to the top of the servos, then the on/off switch on top of that. The receiver (still in its case) was stuck in front of the servos, with the aerial wire bundled up at the front. The other wires were tidied up with a cable tie on top of the receiver.
It worked! :o) We switched it on and to our delight it worked, running around the kitchen floor quite happily with the rounded front of the smooth plastic base working well as a skid.
Time for the weigh-in ... with no armour or weapons, it tipped the scales at as close as we could measure to 150g! The battery pack accounted for about 50g, a lighter one is needed!
James made a cardboard bodyshell to experiment with - a wedge shape with a rear scoop which also makes it more resistant to being tipped over backwards.
Now for some performance testing and driving practice - the other half of the plastic box used for the base was loaded up with weights and pushed around. 300g was easy! 400g just about moved but we were running out of traction. Not too bad for a first attempt. Steering is awkward without a mixer, but James soon started to get the hang of it.
Having got ourselves started, we set about some much-needed improvements!
First, the pot's in the servos were replaced with fixed resistors. 0805-size surface-mount resistors are just the right size to fit across the pads where the pot' wires are soldered onto the PCB. We still suffer some variable creeping, this seems to be due to the rather crude mechanical trims in the transmitter. We can live with it.
Next, the wheels. The aeroplane wheels looked nice, but didn't give terrific grip, and were far too heavy. We saw on someone's web site mention of the lids from plastic milk bottles as wheels, so we quickly knocked some up, with elastic band tyres glued on for better grip. These were certainly lighter, and gave more grip, so a step in the right direction.
Meanwhile, the steering (or lack of it!) was being worked on. After figuring out how to do it in the robot and bemoaning the extra weight, I found mention of a transmitter mixer on the web. Some further research and a look inside the Hitec transmitter and our transmitter mixer was designed and built into a small enough board to fit in a handy space inside the transmitter. The mixer seems to work fine, it certainly makes steering easier!
Designs for a body and weapons also progressed, a flipper was decided on and various options were considered. A new chassis was made from 1mm polycarbonate and an experimental body also from polycarb' - a 1mm top and 0.5mm sides.
Another weigh-in, still just on 150g, still without a weapon - those heavy batteries have to go!
In the search for more traction, a variety of wheel designs were tried, and eventually some "scientific" traction test was needed. We finally ended up with milk-bottle-top wheels with foam-rubber (old mouse mat) tyres with elastic band tread.
The best traction was achieved with double-sided sticky tape (no surprise) but it stripped the gears in the servos!
We then turned our attention to weapons. James fancied an axe but I wasn't convinced we could get enough speed or mass to do any real damage to anything. After a while I came up with an idea for a spring-powered axe, which would give more speed.
|We eventually decided a flipper would be a better way to go (for our first attempt) and so we recycled the spring-powered idea into a flipper. After some searching, we decided on the Supertec Naro HP as a weapon servo as it has high torque and is also fast and only weighs 9g.
A prototype flipper was built up out of mainly scrap material, and it worked OK. At least it went up and down under remote control! However, tests showed that the (ex ball-point pen) spring was not strong enough to lift - let alone flip - 150g. After many fruitless searches for a stiffer spring we resorted to some different pen springs, this time two in parallel. This really did work well, flipping 150g with ease.
Next the transmitter came in for some more modifications, replacing the channel 3 pot' with a switch (to reset the flipper) and using the "racing start" switch to fire the flipper.
|Finally, five Varta PCB-mounting 150mAh cells were ordered from Farnell to make a 6V battery pack weighing a respectable 30g or so. The cells were soldered together side-by-side to make a thin battery pack which seemed to be the best shape to fit into our proposed designs. The photo doesn't show it, but the battery pack is only 9mm thick.|
The first two weeks of May were spent designing and building a battery charger for the new battery pack, using the Maxim MAX712 IC.
The prototype flipper design was re-worked so that it would fit around the drive servos, battery pack and radio receiver and three weeks careful design and building resulted in something that at least looked like a robot with a flipper weapon.
Disaster! When the robot was tested, the batteries lasted almost no time at all! For the weapon servo to hold the flipper down against the springs, the servo takes loads of current and the batteries don't last very long at all. The prototype was tested with 700mAh AAA's so I didn't notice the problem!
With AWS8 looming, the quick fix was to remove the springs and run with a servo-powered flipper. At least the weight was OK, somewhere between 140 and 145g. Without the springs, the flipper is slower, but will still easily lift 150g, and the batteries last about 10 minutes now.
8 June 2002
With AWS8 one week away, RONNY is complete and we have a week of driving practice! Pictures coming soon!
9 June 2002
A last-minute addition of some "mudguards" to try to protect the wheels a bit. They are a bit flimsy as only 0.5mm polycarbonate would fit in the weight limit - better than nothing though! Photo session organised for 12th June.
12 June 2002
The photo-shoot (thanks Roger!) produced some good pictures of RONNY in it's final form. You can see them here.
14 June 2002
Some Dycem non-slip mat was obtained today to replace the elastic band treads - much more grip, just in time for AWS8! Also, a solution for a low-current spring-powered flipper was devised (thanks again Roger for the initial idea!). This will be worked on after AWS8.
15 June 2002
AWS8 - for a brief report and pictures click here.
18 July 2002
I have (extensively) modified a servo to work in the spring-powered flipper, and obtained the all-important (0.3g !) micro-switch. Now I've finished the first stage of the Tech Talk part of this site and got all the bits I need, I will start trying to put R.O.N.N.Y back together again...
1 August 2002
Yesss! the spring-powered flipper mech' works! It always did work (see above) but is now guaranteed to take absolutely no battery power when holding the flipper down, so the batteries should last a good long time! Still some bodywork to rebuild...
Bodywork completed, with a more bulbous rear-end to prevent getting stuck with no wheels on the ground. R.O.N.N.Y. is once again battle-ready...
The flipper servo seems somewhat temperamental. Several people have reported interference probelms with Supertec Naro HP's. Added some more filtering to try to improve matters.
26 October 2002
AWS9 - R.O.N.N.Y's appearance was all-too-brief this time, going out in Round 1 after only 7 seconds! Pushed out of the arena by Vole, which later went on to come 4th overall. In the "super-melee" at the end, with 16 Ants starting in the arena, R.O.N.N.Y fared much better, lasting to the final 3 or 4 - c'est la vie!
Last updated 26 October 2002
Email us at email@example.com
Return to the Robots Page
Return to Gary's Home Page