What to build?

[maddox]

With the idea Jim brings, using a Plant drive, fairly cheap and useable, but not top of the bill(no Storm II like you mention), but for starters fine, and even has some resell value, if you are experienced enough to go for the expensive stuff.
So for the budget stuff, I have a foolproof on/of drive system, but as the relais, Albright SW88B 24 are reasonably expensive- £50 each over here, brandnew- no improvement over the Curtis actualy, but a fair basis for a homebrew speedo, needing only 1 mosfetbank.Just an idea I add.

Personaly I wont spend time on speedos. About every Dutch electronics wiz that tried it went over the budget needed to buy a Vantec.Only Kees Sicking, the builder of the Mythras speedos favoured in the Netherlands and Germany succeded to now.(Now Hammerhead used also hi powered homebuild speedos, but cheaper?)

Pneumatics. The cheap safe route is to go the compressor way. For a heavy, team up several of the small carcompressors. Im working on systems like that(for live event machines), and they seem to be damn cheap ,and if the batteriepower is on board, even lighter and longer lasting than CO2 powered low pressure systems.With the added benefits of cheap recharging, easy acces to replacement parts, and on top of that all, no liquid CO2 problems.
Now for more info about these things, my email, and even living room is available for a lot of questions(and the bar behind the corner has lots of good stuff)
MSN is also available.

[roger_plant]

Rather than the Reliance motor and £15 type S axle for your heavy weight, use my CIM2 motor and my type TE axle.
The CIM2s cost £75 each, cheaper than a Bosch and more powerful.
The Reliance motor is fine for fun robots but can€™t compete when pushing against the average heavy weight.
On 36 volts the CIM2€™s give 3.3 KWatts output each when driven by a 150 amp current limited speed controller.
They are very reliable too, I have been using them, at 36V and 150A, on my 170 Kg off road robot for 3 years, with no problems or overheating.
It does 15 MPH across ploughed fields and powers up steep slopes with ease.

The components to build your own chain drive with axles and bearings will probably cost more than TE gearboxes.
My gearboxes and motors just fix into the chassis as one unit, simply put your wheels onto the output shafts, you have an instant drive system without the alignment problems you will have with a chain system.
They are sealed from dirt, well lubricated and long lasting units that will out last several of your robot design changes.
I have sold over 500 units, many of them have made it onto RW on TV.
I now have adapter sleeves to take 25 mm bore Go-kart hubs directly onto the ¾ inch axle shafts on my TE gearboxes.
Two motors and axles cost £270, 4QD 120/160 amp 36 volt speed controllers £280 the pair.

I will be selling my own dual motor speed controllers soon for a similar price. They are especially developed for robots with all the right facilities built in.
One model covers 12 to 60 volts.

That is about half your £1000 budget so leaving plenty for the rest of the robot.
CIM2 plus TE axle weighs 8 Kg, 16 Kg for the pair, not bad for 6.6 KWatts of drive power.
The Reliance and type S axle will give you 900 watts output and weigh more than the CIM2 + type TE. The Reliance and type S have a much shorter life expectancy too.

I would advise on using the KISS principle for you robot design, I have gone the other route with The Mule and The Big Cheese.
The Mule was entirely home brew from aerial through speed controllers and gearboxes.
Fine if you don€™t mind fault finding to component level and replacing chips, between bouts, in your own electronics.
Much easier and quicker to change a complete standard speed controller, even borrowing one from an eliminated robot if need be.
All your robot building time comes to nought if you get unreliability problems from your own electronics.
I know it is more satisfying to build your own but very frustrating when the whole thing stops, just because you used an unwise choice of component in some vital part of it.

[leorcc]

Big Nipper uses Cim2s, and with success.

I have the reliance motors in RCC2. They are fine motors, but as Roger stated in his posting not that powerfull. At the time though, i bought them keeping the cost factor to a minimum. RCC2 still drives and handles beautifully with the Reliance, but it is a very bad power to weight ratio.

So when i do a refit of RCC2, they will more than likely be with AEGs, or with CIM2/CIM4 motors.

[fwsorg]

CIMs sound like a good option. But your arguments for Bosches also ring particularly true.

I think you are right to go with the heavy, having read your reasoning above. And I certainly think that flippers need powerful drive trains. Your objective as a robot builder is to deliver your weapon to your opponent. And the better drive you have, the better you will be able to do that. Id be very much inclined to go for 36V if you can. Bosches on 36V are a fantastic drive (as Tornado demonstrates so eloquently (in as far as fighting robots can be considered eloquent ). Geared down reasonably for something like 12mph (a very good speed for most live event arenas). If you get some decent current limiting speedos, then youre laughing.

As for orders of magnitude, well, yes conceivably, but there is absolutely no reason why (apart from money obviously) you shouldnt have a 36V LEM powered drive train- LEMs weigh as much as Bosches, and whilst you need a slightly more resilient drive train to cope with them, that is made up for by having considerably lower RPMs, so not so much reduction required.

That brings me on to Speedos. You can build them, yes, but they will probably end up being more expensive, but certainly a fantastic learning curve, and when they do blow up, youll be able to fix them. The low power side is basically a PIC, which, being a software person ) you should have no problems with. Im not a software person, but I can cope with PICs, and that is the more fun side of the current speedos Im building. The high power side is more difficult for me- I can cope with MOSFETs, their drivers, high and low side, charge pumps and so on, but current limiting is giving me enormous headaches, to the point that Im tempted to get some logic level MOSFETs with built in current limiting. While their on resistence is fractionally higher, we can get round that problem by paralleling them. The choice between H-Bridge and Relays is largely a software one. H-Bridges are harder to program for, but are nicer in terms of control. I would go for a relay controller however, as it uses fewer MOSFETs (cheaper and considerably less bulk when you consider how banks of MOSFETs in an H-bridge need to be arranged, plus you dont need a complex system of copper buss bars), is easier to program for, you dont have to worry about high and low side driving, and a number of other reasons which are randomly entangling themselves in my head. If you wont to get something on the arena, and having fun, buy some 4qds or similar (or perhaps a roboteq- though that€™s probably a little on the expensive side) then you can build speedos on the side, and incorporate features into them as you begin to get an idea of what you want in a speedo. As I said, theyre fun, satisfying projects, but not really enormously practical for a first robot.

Ground clearance- the likes of Cassius and firestorm produce rather elegant wood shavings as their fronts plane across the floor. Storm 2 has more ground clearance than I expected, however it still seems to get under everything because its front is so bloody hard, so delivers its KE more effectively into the opposing robot such that the opposing robot jumps up perpendicular to storms wedge, and so storm gets under it. At least that€™s what seems to be happening last time I saw it.

No need to apologise for long ramblings, theyre well punctuated and readable, which makes a refreshing change. Although far too much time spent on MSN has done no favours for my typed English, Im still very much from the €˜Babeth€™ school of thought when it comes to forum postings, as the way you/one writes speaks volumes.

[jim]

Cayenne has 459MHz radio as does Serrano, but they use the LPRS modules which work a bit like a one-way radio modem. I dont have the equipment to do the actual RF stuff myself.

[fluppet]

Thank you for your feedback, everyone. Excuse the delayed reply - needed time to sit down for a good witter! Here goes, and I hope some of this is still of use to other beginners who read this thread...

Babeth: Meant to say before, your rule 1 is very well phrased (as I would expect of you!) The complex robot designs which I have on paper are complex because of a high degree of redundancy. I include a long list of failure modes in the designs, and how they would cope with them. This thread is about avoiding having more failures than I can design around. :-)

Thank you, everyone, for the advice on motors. I remember Rogers site of old (i.e. about a year ago - Ill get my zimmer frame out) and the Cim2s sound like a good option. Big Nipper did indeed seem to have good performance (nippy, as it were). Im trusting that Im not going to have too much trouble with replacements if I go down the Cim2 route - I remember Roger being involved in another thread last week about problems acquiring replacement magnets for Bosches, so my perceived benefit to using them may not in fact be valid. Roger - should I be doing something on your web page to get a more complete catalogue? Ive never been quite clear from your site precisely how much you sell - you seem to mention things Ive not seen listed - should I be contacting you directly for more details? I know you have better things to do than be a webmaster, so this is just an inquiry, not a criticism!

The gearbox combination also sounds like a good thing for a novice. Strangely, many of my designs seem to have such odd drive mechanisms that the stick a wheel on an axle advice may not be as much use, long term, as youd think. But still, I could be better off starting with a managable RPM and hooking that up to something odd in a single stage, rather than trying to build the gearing into the mechanism myself. Give me a year or twos practice with a welder and I hope Ill be showing the community what I mean!

My only concern about a single piece unit is the possibility of repair. If I snap a chain or shred a sprocket, its not so much to replace, but Im not sure what can be done to a knackered gearbox. Obviously Im planning to shock mount the drive anyway, but am I going to have to buy them in bulk for every tournament, or is there a chance of repair?

I might take Eddys advice when it comes to speedos - buy a couple, then start building some. My biggest concerns are over paying for unwanted facilities, and the one size fits all problem. I doubt Ill ever have a use for mixing - even when I start Ill probably have a PIC doing some work to add some failsafe and add a bit of autonomous control. It seems a waste (and less reliable) to generate PWM speed control from a PIC and have to shove it into an analogue speed controller input. Anyone got a speed controller which, essentially, just does the current limiting and drives lots of FETs? What do the - apparently single channel - Curtises do?

Also the number of motors and power required from them is likely to vary vastly between my robots: Ive expressed an interest in innovation, and that means that whatever I build, and however ineffective it will be, Im unlikely to end up with a box on 2/4 wheels very much. My most developed design uses 16 independent motors, each needing speed control... (but it would keep functioning, after a fashion, with only a couple intact - and Jim, dont worry, the bleeding-edge technology is all my own, and therefore a lot cheaper than using anyone elses!) Ive had a cunning plan for a first design which is much simpler (Roger should be proud of me for applying KISS), but the big money needs to be spent while thinking of the long term.

On the other hand, the prices people have been stating for controllers are lower than I was led to believe a year or so back; possibly the market has moved on, or maybe I just memorized the Vantec cost (I remember the £4-600 bracket). While I still dont want to throw away more unnecessary speedos than I have to, it sounds like Im not going to blow the whole budget on something I can only use once.

Jim - thanks for the radio components tip; actually I could use bidirectional data, so I may well be better off just buying a radio modem and being done with it. Ill have to investigate - someone gave me a pointer on the old forum, and I can dig it out.

Mario - I realize you should know, but about the compressors: really? I was under the impression the power to weight was far inferior to a CO2 cylinder, but Ill believe you, and I guess the choice of low pressure would help. Obviously it would make my life easier, consumables-wise, in terms of testing at home, and might avoid me venting an entire tank in a closed room and knocking myself out...

[Ironically, having bought my new camera for taking snaps at Debenham, I asked about cleaning the sensor. You cant use propellant-based air dusters, youll have to use carbon dioxide; I can give you the name of a manufacturer said the bloke in Jessops. I advised him Id probably cope. :-)]

Also, thank you, Mario, for the offer of help - you may come to regret it as I get my plans together!

Finally, thanks, Eddy, for the reassuring news about Storms ground clearance. Now I just have to find a way to exploit it! Also, thanks for the reassurance about my heavyweight reasoning - very welcome to hear someone tell me Im not spouting utter nonsense - and Im glad Ive fooled you into thinking Im coherent. :-)

(The tiddlywinks crowd I hang around with are the most pedantic people I know, and include a fellow of classical linguistics. Babeth holds few fears for me when I spend my time with people who use pusillanimous and callipygous in regular conversation, and using less when you mean fewer brings months of scorn.)

Oh, one last thing. I was looking up the FERM MIG 100 TURBO MIG WELDER on Screwfix Direct. You might also be interested in... Chainsaw. Theyve dealt with us before, havent they? (Who else would need both a welder and a chainsaw? How do you weld the tree back together again? Unless youve just lumberjacked a cellphone mast, obviously.) :-)

Thanks again, everyone. I hope to be able to enter something within the next few paycheques!

--
Fluppet
[Britney mode: Yargh - I did it again.]

[jim]

The Curtis controllers are single channel, they take a variable resistance or variable voltage input. Annnoyingly, it has 2.5V as the centre (off) position, so you have to be careful you dont switch them on with 0V applied as this is full reverse. The PICs I use have a single PWM output to generate a variable voltage, so I have four PICs running the drive motors on Cayenne, and another two running the weapons. The only speed controller Im aware of with more flexible input is the rather pricey RoboteQ which accepts RS232 input.

If you find a radio modem which works on 459MHz or 40Mhz please let me know. If you want to use other frequencies, prepare for a bit of debating with the FRA.

[fluppet]

Thanks, Jim. Are you actively smoothing the PIC output for the Curtis, or is there enough capacitance in it anyway? Maybe it doesnt matter, but if its point sampling then I could see some weird beating effect could happen. Just being paranoid. And thanks for the 2.5V warning!

Ill certainly publicise any radio modem I find, but it sounds like - if youve been looking for one and not found it - Ill be doing some soldering. Im not planning to be a pain regarding odd frequencies (if I was going to play those games Id stick an 802.11b card in the thing and make much of my life simpler) so if I can only find something inappropriate then Ill roll my own.

Cheers for the feedback,

--
Fluppet
(Woohoo - short post.)