Paul & Sabrina's cheap DIY 144v motor controller

[caspar21]

depends on the pull down strength you have setup.

[ga2500ev]

Originally Posted by squiggles

I have not seen the start of this discussion but it looks to me as though each of your IGBT will have to be rated for full motor current.....unless they are overlapping then the first that come on has to handle full current.
I am not seeing what the advantage would be?

I agree. Better would be to have all the IGBT's share the load. However, to do that you have to derate the max current each can carry because they will share unequally based on structure, temp, and gate characteristics. In short 4 400A IGBTs cannot carry 1600A.

A good paper on the subject is the Dynex application note AN5505. You can find it here:

Application Notes - Power - IGBT

Rules of thumb:

1) Derate the max current.
2) Use individual gate resistors for each module.
3) Put all the modules on the same heat sink.
4) If you cannot match the modules, then put a load sharing resistance on the emitters of each module.


ga2500ev

[asyx]

good information

[jmcginley]

Hey Paul,

I'm (finally!?) setting up to build my kit and I see there is no socket for U3. It seems like that would be a good idea. The other chips don't need one so much but the brain might have to come out some day. Are most people putting a 28 pin dip socket in there?

John

[ludiagsm]

Hi again today i made some tests with a 12volt controller with 6x irf540 mosfets and for diodes i use 3x irf 540

http://www.youtube.com/watch?v=wJr7BxCUcQE

that controller will be mounted on that

http://www.youtube.com/watch?v=bQ8EIZZef6M

electric BMX with starter dc motor till now it was controlled only with button and one 30Amp relay but hope soon to have new video with new controller

[LouMan]

ok guys,

hope I can explain but first I'm not good at explining so bare with me,
the the igbts parralled has to be match'd share the same heatsink as close as possible pluss the snubbers and gate resistors all need to match as close as possible then if you want to run the cheaper igbt's from ebay there most likely can top out at around 8khz.
pluss they have to switch every single pwm cycle at near the max rating,
and you'll still get a little squeel from your motor.(unless you can afford the new igbt's with a 16-20khz rating(don't know about you but I can afford them yet ))

now imagine breaking up the pwm clock to 4 600amp igbt's who only has one pwm clock out of 4 25% load, 4 600 amp igbt's is 2400 amp split in half for a desent nominal rating of 1200amp, witch will probley never go over 500 amps during taking off and a cruising amperage of 200 amps thats only 50 amps per igbt so I believe I have plenty of derating.
now picture this you have a line of people throwing sand bags to each other down the line one 50 lb bag behind the other imagine how hot you would get, now imagine 4 of you standing there and you only get one bag then get to rest and kool down while the other three get there turn before you get the next bag again.
switching the pwm clock to 2-4-6 so on devices allow the device to run kooler and eliminates a lot of the hassels of paralleling at lot of devices together at there near max speed.
I chose 4 600 amp igbt's for no other reason exept it was a neat size and they were free I just did a little yard work for my auntie so she got them for me
the numbers were easy to work with, they can handle 8khz time 4 would give 32khz at the motor but I dont see the reason, running them at 4 khz will let them run a little kooler and gives a better signal.
motherboard manufacturers found out by phase shifting the pwm to a lot of mosfets to the cpu they could use smaller cheaper mosfets and they would run kooler, remember the sand bags imagine having to wait thru 16 people till you get your bag again.
your not going to get a heavier bag but the same 50 lb bag not a 100 lb becuase you had to wait your turn.

well hope this helps, I think a 2400 amp power section that will probley never see over 450 amps during take off and only around 200 cruise will be very safe, and above that give me very nice bragging rights hehehehe. see yaw later

got to do a 200 mile poker run today with about 200 other bikers with a police escort had to put 3 gallons back in the bike when I got home (I GOT to do something about that milage hehehe)......

[MPaulHolmes]

Hi John, yes, that is a good idea. Some people have socketed the atmega, and some haven't. I've just been a cheapskate and not included a 28 pin socket. I know radio shack has them for like $1 or so.

-Paul

Originally Posted by jmcginley

Hey Paul,

I'm (finally!?) setting up to build my kit and I see there is no socket for U3. It seems like that would be a good idea. The other chips don't need one so much but the brain might have to come out some day. Are most people putting a 28 pin dip socket in there?

John

[MPaulHolmes]

Holy cow! CANbus is operational! I successfully had communication going between 2 nodes. I'm really really really really... really happy right now!

I sent 8 bytes as a single packet. They were 'a', 'b', 'c', ... , 'h'.

Originally, the 8 bytes of the receiving buffer had '1', '2', ... , '7', 'nothing' hehe.


I displayed it on the screen through the UART. So it can do CAN and RS-232 at the same time. hurray! noise immunity between everything that counts (and 500kBPS to 1 Mbps data rate), and slow crappy transmission for the user to see. haha

See the green and yellow wires connecting the 2 breadboards? They are the connecting wires for the canbus! You can't teach that, it's all instinct! hahaha


It worked on the first try. That never happens. Ever. Never.

[squiggles]

Originally Posted by MPaulHolmes

Holy cow! CANbus is operational! I successfully had communication going between 2 nodes. I'm really really really really... really happy right now!

We can see your excitement in the first photo, couldn't hold the camera still !

Well done by the way.

[jackbauer]

The 8khz "squeal" is nothing like the curtis. Its only noticable when reving the motor in neutral with the bonnet open. When driving its not audible at all. From either inside or outside the car. Ok in theory it'd be easier on the bus caps to run at 16khz but they have steadfastly remained stone cold in my controller despite all attempts to warm them up! The max recorded heatsink temp was 35c after blasting the car up and down a hill for 15 minutes. I used one igbt as the switch and one identical device with its gate and emitter tied together as the diode. No regrets so far and cost peanuts.

Now I am using a compound motor if that makes any difference?

Paul quick question. Would it be possible to program the controller to limit the motor voltage? Well done on the can bus stuff. I used it many years ago and it was a real pain to setup but once running impossible to glitch.