Under those aluminium bars and copper tabs.Eva said:You are crazy
Where are the MOSFETs?
http://www.ee.oulu.fi/~tech/kart/6_8_07/IMG_3860.jpgdjQUAN said:are there pics of the vehicle, motor and batts?
Previous version with only one motor.
http://www.youtube.com/watch?v=I5pw14BhCv4
http://www.youtube.com/watch?v=NsWyoqgYblI
60Ah car starter batteries, but better Optima batts are on their way. High-current Li-ion cells would be nice if we can get enough $$$$$phase_accurate said:
Looks like a fun project with lots of challenges.
For paralleling lots of MOSFETs for high-current applications, I’ve used 5” pieces of 14AWG wire to form a very low R “source resistor.” Naturally, length and AWG will be dependant on your current-per-FET and matching requirements. It can also radiate a lot of noise, so a shielded enclosure is recommended.
Once you get the controller figured out, I’d look into designing a FET-shorting circuit using some very high current contactors (like several Ford starter solenoids paralleled or MIG-welder relays). The way it would work is that once the FETs’ duty cycle hits 100% “on,” the contactor would close shorting the Drain-Source-junction eliminating your FET I*R losses. This would maximize full-throttle power and greatly reduce loses. I would suggest caution and use an appropriate acceleration ramp-up to avoid the contactors instantly closing and going from off to full-on and the associated currents. Then again, what a ride!
For paralleling lots of MOSFETs for high-current applications, I’ve used 5” pieces of 14AWG wire to form a very low R “source resistor.” Naturally, length and AWG will be dependant on your current-per-FET and matching requirements. It can also radiate a lot of noise, so a shielded enclosure is recommended.
Once you get the controller figured out, I’d look into designing a FET-shorting circuit using some very high current contactors (like several Ford starter solenoids paralleled or MIG-welder relays). The way it would work is that once the FETs’ duty cycle hits 100% “on,” the contactor would close shorting the Drain-Source-junction eliminating your FET I*R losses. This would maximize full-throttle power and greatly reduce loses. I would suggest caution and use an appropriate acceleration ramp-up to avoid the contactors instantly closing and going from off to full-on and the associated currents. Then again, what a ride!
Motors? what rating? Probably about 2kW continous. We have rebuilt them 4-5 times already, now half of the stuff inside is teflon to stand against the heat. More like drag-racer than a go-kart.Eva said:I'm really impressed
Are you the driver mzzj?
At how much power are rated those motors? How about 415V three phase stuff?
Those are my Friends who are driving in waterpools, I am that long-haired hippie on second video
Just give us batts for 415v and we are ready to go
We have been half-seriously playing with an idea of making our own permanent magnet synchronous motors, something like this but bigger: http://www.plettenberg-motoren.com/UK/Motoren/aussen/Predator37/Motor.htm
Thats something I would like, 15kW(60sec peak I guess) and weights mere 2kg!
star882 said:
Not only a bit easier, but 400V system would do with single! igbt.mzzj said:
DCPreamp said:
We were planning to use contactors, but in a different way. Starting with motors in series and then switching to parallei connection once we have got some speed.
Once you have a few batteries, it's a matter of connecting them in series to get 48V, 60V, 96V etc... and feeding a push-pull converter with them. Then you get 400V to 500V DC. Note that 500V electrolytic capacitors are very compact for the amount of energy that they can store... Then comes the three-phase industrial motor control part with some 100A IGBTs (or maybe a single module including IGBTs and clamping diodes).
Or you could consider building a motor intended to work with the resulting voltage of all your batteries in series...
Concerning current sharing, how about sensing the Vds drop of each MOSFET with diff. amps.? Then you could adjust Vgs to either limit current to safe values in case some device wants to conduct too much or to prevent too much deviation from the average (although this may be troublesome to control). Also, you would gain a lot of reliability with cycle-by-cycle current limiting.
Edit:
Uhm... Why don't you have implemented average current control of the PWM to start with? Isn't motor PAR proportional to average current?? But average current is not proportional at all to duty cycle because we have this damn discontinuous-inductor-current to continuous-inductor-current boundary... There is not even a constant relationship, the boundary moves depending on the crossing conditions!! And we have back EMF too...
Then average current limiting would come for free and you can have short and long term limits to protect the motor, as winding heating should be proportional to i^2...
Umm... And you are going to need a high power "pocket" charger for all those batteries...
with some 100A IGBTs (or maybe a single module including IGBTs and clamping diodes).Or you could consider building a motor intended to work with the resulting voltage of all your batteries in series...
Concerning current sharing, how about sensing the Vds drop of each MOSFET with diff. amps.? Then you could adjust Vgs to either limit current to safe values in case some device wants to conduct too much or to prevent too much deviation from the average (although this may be troublesome to control). Also, you would gain a lot of reliability with cycle-by-cycle current limiting.
Edit:
Uhm... Why don't you have implemented average current control of the PWM to start with? Isn't motor PAR proportional to average current?? But average current is not proportional at all to duty cycle because we have this damn discontinuous-inductor-current to continuous-inductor-current boundary... There is not even a constant relationship, the boundary moves depending on the crossing conditions!! And we have back EMF too...
Then average current limiting would come for free and you can have short and long term limits to protect the motor, as winding heating should be proportional to i^2...
Umm... And you are going to need a high power "pocket" charger for all those batteries...
Eva said:Once you have a few batteries, it's a matter of connecting them in series to get 48V, 60V, 96V etc... and feeding a push-pull converter with them. Then you get 400V to 500V DC. Note that 500V electrolytic capacitors are very compact for the amount of energy that they can store... Then comes the three-phase industrial motor control part
Or you could consider building a motor intended to work with the resulting voltage of all your batteries in series...
Hey, why dont you suggest feeding 50kW push-pull converter from 12v battery?
And yes, single large 450V electrolytic could probably handle the required 25A ripple current. I like more about your idea of 400v battery pack.
We already got some 100Ah li-ion cells/batts (EV/traction), but 2C max dischage current makes them total trash. Better li-ions are costly and we havent found willing sponsor for them. 400pcs of ANR26650M1 cells and go-kart would be ready to fly
(downsides: cost: 5000eur, charger and battery manager with 100 cells in series)
this beast http://www.killacycle.com/ uses 990pcs of ANR26650M1 cells. 375V battery pack with 1600A discharge rating
Eva could wire them all in parallei to get some challenge making 3V, 150kA input push-pull converter.
With 350 rwhp they must pump those motors with half a megawatt. Our motors have easy and peacefull life compared to that
Re: FET Drivers
2x ucc27322
It is bit underrated for driving 3500nC gate charge, but didnt bother to use anything else as extreme current prevents switching much faster anyways. 80-120A per to-220 makes every nH count. Either your mosfets enter avalance region or gate voltage rings like insane.
TO-220-7 would be better as it has 6 source leads.(also cost 6 times more
TechGuy said:
2x ucc27322
It is bit underrated for driving 3500nC gate charge, but didnt bother to use anything else as extreme current prevents switching much faster anyways. 80-120A per to-220 makes every nH count. Either your mosfets enter avalance region or gate voltage rings like insane.
TO-220-7 would be better as it has 6 source leads.(also cost 6 times more
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.