|
|
|
|
|
|
Spockie-Tech
Site Admin
Joined: 31 May 2004
Posts: 3160
Location: Melbourne, Australia
|
Not too bad for a first cut !
They should have added a simple O/C output or two for optional weapon or relay control instead just of the flip/reverse option and I wouldnt have stood the Electro Caps's up like that.. I reckon they will bend or possibly break off under high-G impacts.
I also would have designed the board to fit into an optional zippy box or some other standard sized ABS enclosure - Metal swarf has been the death of many an IBC.
The pins for the servo wires arent a great idea.. no retention of the wires means you will need to obtain double ended servo leads, and glue them into place and the BEC is a bit wimpy at 150ma.
The current ratings look reasonably realistic, nice to see some graphs, although Im a little sceptical of those single surface mount DPaks solder pads ability to pass 90amps for any length of time without melting the solder. I would have set the max current limit a bit lower. The 1405's TO-220 case legs are only rated for 75amps, and they look thinner than that with a much smaller contact area.
They dont say if its for one channel or both running at 30 amps, although given the Bridges are at opposite ends of the board, they probably arent dumping into the same heatsink, so that might not matter.
When load testing the IBC's we found they could survive a few 90-100amp x 1-2 second surges, but the legs and pads would very quickly get hot while the big heatsink blocks could soak up quite a few surges before it started to climb in temp. Of course, being a *sink* rather than a *radiator* (with lots of finned surface area), it needed a lot longer to cool down again after eating up the heat. This wasnt a problem for a controller designed for 3-5 minute matches with big downtime in between (low duty cycle) though.
Also, the TO220 heatsink tabs would get hot enough to melt standard-temperature isolation bushes pretty quick (we had to find high temp ones). 110Amps was really pushing it, and 120 amps pretty much meant near-instant destruction..
Whats with the heatsink on the *bottom* of the board while the fets are on the top ? I hope they have a ton of big plate-through holes under the Fets conducting the heat to the bottom, otherwise there is a big slab of temperature insulating fibreglass between the Fets and the heatsink !
The firmware has some entertaining bits and pieces in it.
code:
//Failsafe check routine. Are all channels reading?
if(micros() - fs_timer > DELTA_T_US_TO_FAILSAFE) {
...
oh_shit = false;
blink_furiously = false;
Be interesting to see how well it copes without TVS diodes or other protection circuitry on spikey motors.. *Theoretically*, the synchronous rectification method should help a bit there, although it offers no protection while the bridge is cycling between states
Overall, not a bad effort. but Im surprised at a few simple things that look wrong to me.. Time will tell how they survive under pressure.. _________________ Great minds discuss ideas. Average minds discuss events. Small minds discuss people
Last edited by Spockie-Tech on Fri Feb 08, 2013 3:52 pm; edited 1 time in total
|
Fri Feb 08, 2013 3:35 pm |
|
|
|
Spockie-Tech
Site Admin
Joined: 31 May 2004
Posts: 3160
Location: Melbourne, Australia
|
Heh, thanks for that Aaron. Interesting reading the development blog here
http://www.etotheipiplusone.net/?cat=85
Very Deja-Vu on the power supply design part.. we went through all the same stuff for IBCs switching regulator design years back. The requirement of Combat Bot controllers to run without brownouts or gate-drive problems on voltage sagging damn-near-flat batteries is a real PITA for designers..
Interesting, he *does* have lots of Plate-throughs for heat conduction - but *only* under 2 of the 4 Bridge Leg Fets (why is that ?),*and* that an early version that exploded in a pushy bot looks like it was the 2 Fets that *didnt* have the heat conduction channels that did the FET (Flame-Emitting-Transistor) thing.. Can I say "told you so" now ?
I dont see him talking about low-ESR specs on the main input filtering caps.. I hope they are for his sake.. Again, when you're running wildly varying loads on near-empty batteries, the input voltage will bounce around like a mofo, and the caps have to bear the brunt of that.. thats what gave us the biggest problems getting the IBCs Sepic switchmode to stay stable when switching from buck to boost mode and back again rapidly over and over again.
DC Motor Control design for battery powered is a fascinating but dangerous area, because its such a wildly variable environment, you have no idea what motors your controller will be asked to drive, under what load, for how long, with what power supply.. so youre really trying to hit a unpredictably moving target..
The number of emails we used to get from people going "Your controller blew up in my Robot, it sucks !".. further investigation would reveal they would be trying to drive a pair of EV Warriors (200amps+ each) from a 7ah Gel-cell brick in a 75Kg Robot with 18 gauge supply wires that they pushes spade connectors on and off the battery as a sparky power-switch or something stupid like that. (sigh).
I hope his controller does well, I personally wouldnt even consider venturing back into controller design until the mythical Silicon Carbide Fets or some new tech comes out that makes a re-spin worthwhile. _________________ Great minds discuss ideas. Average minds discuss events. Small minds discuss people
|
Fri Feb 08, 2013 4:21 pm |
|
|
|
|
|
|
|
|
|
|