Printable View
I was thinking an arduino would be the easier way to control the IGBT and relays for the h-bridge. May get one of those IGBTs just to experiment with, can anyone make enough sense of the datasheet to figure out what voltage/current it takes to turn on the IGBT/activate the gate?
Are you still going for a single PWM transistor and a relay H-bridge? IMHO, that's doomed to failure unless you can find an absolutely massive relay. What was the reasoning behind selecting that IGBT rather than a FET? IGBTs are generally used when you need to switch high voltages and FETs do a better job switching high current at lower voltages. That particular IGBT is not going to switch much more current than a TZ85a in a real-world situation; the large current figures are for 1 millisecond at 25 degrees.
The relays should be alright as the pwm transistor would shut off the current whilst the relay switched then turn it back on once it has switched. This avoids sparking/arcing which is how relays normally fail, once closed they can take a lot more than their rated current.
I had noticed that a MOSFET was more suitable for low voltages than an IGBT but my issue is with connecting the mosfets, they all seem to have such tiny pins which are hard to connect a large wire to, IGBTs seem to be larger packages which are easier to connect to and you seem to only need one rather than having to connect a bunch in parallel. I am after simplicity but high current handling. Maybe the advantages of the IGBT aren't worth it though.
There are plenty of FETs in large packages like the one you found - if you have some spare gold bars to pay for them :). I did some detailed research and concluded that unless you need to switch thousands of amps, FETs in a smaller package like the TO-247 has the best combination of current capability, heat transfer and size. I picked the IRFP4004 FET and mounted four of them to a small heatsink with copper strips to connect the source and drain pins. Its fairly compact and will comfortably switch 120A continuously in a real-world situation - more if I increased the heatsink or used a fan.
If you want to DIY a single direction FET switch,have a look at the OSMC circuit: http://www.robotpower.com/downloads/...2sch-clean.pdf. you just need to duplicate one of the low-side FET legs and then find a suitable driver. The HIP driver chip isn't necessary as you are only driving one string of FETs and you don't need the voltage pump as all the FETs are on the low side of the circuit.
I guess this is essentially a package of mosfets: http://docs-europe.electrocomponents...6b800bbecc.pdf
Could be used with a switching frequency of up to 3kHZ but I fear that could be a bit too low. Not sure how well it would perform though and imagine when used in this sort of application as with mosfets the 100a rating would drop down a lot.
I have lost count of how many of those I have blown up! I used two 100A SSRs to switch a short Mag motor (no PWM) and they regularly died if the motor came close to stalling. Internally, they have two FETs in TO-220 cases. The other problem with these SSRs is impact resistance; mine had a thin a brittle plastic shell that quickly shattered. I had to cut the top of the shell off and inject potting epoxy to make them rugged enough for combat use.
On the other hand, they are super easy to control and the screw-on terminals make wiring easy too.
I tried something similar to this had a tendency to blow up although I think it could have been due to doing pwm in two micros. It might work but I couldn't be sure. Big heatsink block helps a lot and you might just get away with running a single 85A on a reasonably geared motor.
Steve
PityQuote:
Originally Posted by marto
Ain't the browning out of the control circuits solvable with a seperate powersource for those?Quote:
It all works fine until one of the ESCs browns out. And if you happen to be using very low inductance motors, like I am, this can happen even at very low PWM duty cycle. If the motors are stalled even at low throttle this can cause the input voltage to drop very low and the ESC to reset. As there is communication between the two paired ESCs to ensure they are synced and running properly this means that if they don't reset at exactly the same time or if one powers back up first they can enter a situation where there is shoot through in the ESC causing it to blow up.
A not uncommon trick. Over here there is a machine running astro's on XXL's in Mirror mode.Quote:
There is no way to solve this with software. It could be solved in hardware but its easier to just use a different ESC. For the middle they are 2 85As running on separate brushes.
Lets go on with the test pleaseQuote:
I do not know how much current is coming out of the batteries but I think it is a lot. The 8AWG wire from the batteries is too hot to touch after just a few minutes of driving. So it will be interesting to see what happens.
Steve