|
DavidM
Joined: 07 Jul 2004
Posts: 41
Location: Victoria, Australia, Earth, Milky Way Galaxy
|
Werewolf - Machina La Mort - Victoria
I'm having a go at two bots, so far all I have is two frames and one controller. The controller I have is barely running (but running still the same), it is about the size of three IBCs, and has an ATMega8535 running it and two H-bridges of BUK9535 and IRF5210 Mosfets, I'm using P types in stead of a Hipchip with charge pumps, thus enabling higher voltages, but thats the not the point, the extra control lines are the point.
I'll see how it goes, theres a lot of work in the IBC and I'm not sure that I will be able to get a mandatory fail safe integrated and I know there's been lots of subtle stuff done with the IBC to make it a good performer in battle. _________________ "Limitation shows the Master."
|
Sun Aug 22, 2004 9:55 pm |
|
|
|
|
|
|
|
DavidM
Joined: 07 Jul 2004
Posts: 41
Location: Victoria, Australia, Earth, Milky Way Galaxy
|
Well finally I will spill the beans, here is the test frame.
http://robowars.org/forum/album_pic.php?pic_id=224
It still isn't performing how I'd like to yet, the simulation was great, the 1/3 scale version works fine, but the full scale version with compliance isn't working that great (yet), it moves but its scary, imagine making a dog eat 20 no-doze tablets then get it to smoke a packet of Peter Stuyvesant and give it some LSD and yell 'watch out for the spiders on your tail', thats how it moves. _________________ "Limitation shows the Master."
|
Thu Sep 09, 2004 11:52 pm |
|
|
|
DavidM
Joined: 07 Jul 2004
Posts: 41
Location: Victoria, Australia, Earth, Milky Way Galaxy
|
Well spotted Brett.
At work I'm designing a medical table that uses linear actuators, this board drives two actuators (2000N each), the board measures drive voltages, currents, motor continuity, actuator position pots, failure modes, plus the normal PWM stuff, as the other thread says its not hard to make a controller but its not trivial either, a bit of time went into protection and even to the level 'what if someone removed this chip doing a repair'. The board uses garden variety devices. I'm using digital logic to prevent any deliberate attempts by the micro to do something it shouldn't like a shoot-thru condition, but the micro still has to do the 'miller time' (sorry bad pun) for a polarity changeover. The board has a few spare IRQs and I/O lines plus an RS-232 port/drivers. The cost is fairly reasonable, but not squeezed for real estate like the IBC the layout was designed for easy reliability, maintenance, failure tolerance/warning and simple assembly and no surface mount. You could place 3 IBCs side by side on the board, so its not diminutive.
So when the first protype boards were made for the medical project, I thought one PCB wouldn't mind a change of software and change of scenery to the rough and tumble world of robot combat.
And the question, did I blow any FETs up, sure did, noise into the bias circuit, partially driving the opposite P type, sort of a partial shoot-thru got warm and took out the bias circuit, a bit more filtering on the bias and its fine now, but not what I expected to happen, of course Spice didn't see it either.
Why, because Werewolf needs both motors 180 degrees synchronised, if not the two motors effectively 'beat' and the bot gyrates until it comes out of a synchronicity (not a bad album). _________________ "Limitation shows the Master."
|
Mon Sep 13, 2004 5:51 pm |
|
|
|