Paul & Sabrina's cheap DIY 144v motor controller

[MPaulHolmes]

80 MHz would be nice, but I'm OK with the ATMega8 for now. I can add a crystal oscillator to crank the speed up to 16 Mhz if I need to, but I don't think I will. The only major concern I had was keeping current under control, and I now that is not a problem, because I fully understand how to do a very robust hardware current limit. I'm using Fran's design. It's actually very clean. On a pulse by pulse basis, the 393 comparator and flip-flop doesn't let ANY current pulses get through to the mosfets (not even 1!), and the nand gates I'm using for the flip-flop (to remember the overcurrent condition) are very fast. Each time a pulse tries to get too big, the hardware shuts down the mosfet driver within NANOSECONDS, and keeps it off until the microcontroller gets around to turning it back on when the next pulse starts.

I'm very excited that I understand how to do this now. I worked on it the whole weekend, and had lots of chats with Fran. Fran is awesome. He is from some eastern bloc country, and recently made a controller that makes it so NO MECHANICAL BRAKING IS NEEDED! All of it is regen! He was testing it out on a very fast go-cart. It doesn't even have regular brakes! How awesome is that! He also just finished a controller that is VERY high power on a "tank" that's using like 300v of Lithium Iron Phosphate batteries. It's using 6 IGBTs. He sent me a video of it shoving back a 10,000 pound forklift, when the forklift was trying to fight back. It was really sad if you are a fan of really big forklifts. The forklift was whooped like a red-headed stepchild.

Also, the hardware current limiting costs around $1.00. Ya!

[motor_control]

Originally Posted by MPaulHolmes

shuts down the mosfet driver within NANOSECONDS.

You need notice current sensor delay. For IR2127 all loop take about 800 nS and befor 750 nS current protection is no active.

Originally Posted by MPaulHolmes

Fran is awesome.

He is from some eastern bloc country, and recently made a controller that makes it so NO MECHANICAL BRAKING IS NEEDED! All of it is regen!

Very interesting. Does Fran shared some infomation refer this project ?

[MPaulHolmes]

It's a LEM 500amp (900 amp peak) current transducer. It responds to a 90% of max current step in under 5 us. However, there will be no 90% current jump because I'm having the throttle control current rather than PWM duty, and I'll add sufficient ramping to slow di/dt to reasonable levels, and the mosfets I'll be using will be able to handle a current pulse that is much larger than 500 amps. More like 500 amps per mosfet for a single 100 us pulse according to the datasheet, which is a 5000 amp pulse.

Fran is the chief engineer at a company that designs brushless DC motors. He just volunteers his time on the EVTech list. I've never seen him receive any sort of help on there. He's very helpful though.

-Paul

[motor_control]

Originally Posted by MPaulHolmes

Fran is the chief engineer at a company that designs brushless DC motors.

I'm too developing BLDC from auto alternator, but w/o permanent magnets.
Now I'm working on controller power stage.

My opto sensors HLC1395 (insted hall sensors) pcb on altarnator - photos
electroauto.ru • ПроÑ�мотр темы - Самодельные Ð*лектродвигатели

and pcb toner transfer for SOIC14, SOT23

[MPaulHolmes]

A BLDC motor from an alternator! That's awesome!

The copper bus bars (well, bus bar until I cut it) and heat spreader came yesterday. Commercially pure copper. 99.9%. Very high conductivity. 101% on the conductivity scale. Also about 390 W/(m*K) thermal conductivity, compared to 167 W/(m*K) for 6061 Aluminum, which is all the welding shop near my house sells. Diamond can have over 2000 W/(m*K), but that would be sort of expensive.

I'm sort of angry. I tried to order some of the controller parts. I ordered 25 of the new fancy 200v mosfets (to get the bulk discount), and they only had 6 available, so 19 are on back order, and will be shipped MARCH 21! I am glad I was wearing flame retardant material, cause I was real mad, way down deep! I found the perfect (at long last!) dc-dc converter for the inside of the controller. Otmar (the Zilla Otmar) helped me out with that. He had used quite a few of the same ones that I had just ordered, and told me they kept failing because the isolation voltage was too low, so he suggested I use the DC-DC part that he's going to use in a new High voltage project he's working on.

[almightybmw]

Wicked cool. This is beyond the skills I have, but I'm still curious how much it has cost up till now? I couldn't find a price list in the 345 posts, but I was only skimming.

[MPaulHolmes]

I now have a pretty good idea about what parts to use. There was a bit of trial and error, but that is gratefully now mostly over! There are several possible variations, but here is just one. This doesn't include the cost of shipping. Buying in bulk would really help to eliminate the cost of shipping, and would make the parts cheaper. Some of the component prices assume buying in modest bulk (say minimum of 10 or whatever).

Total cost of 144v 600amp controller without synchronous rectification:

MOSFETS: $5.59*12
DIODES: $2.36*12
RIPPLE CAPACITORS: $2.00*24
MOSFET DRIVERS: $1.95*2
POWER PCB: $0.80
CONTROL PCB: $2.00 (unsure at this point)
COPPER HEAT SPREADER: $28
ALUMINUM HEAT SINK: $25 (I don't know! Ben sent me his!)
COPPER BUS BARS: $9
ISOLATED 12V-12V DC TO DC CONVERTER: $15
BRASS NUTS AND BOLTS: $8
CURRENT SENSOR: $27 (You could get a $15 sensor if you had a band saw.)
MOSFET CLAMPS: $0.27*24
MISCELLANEOUS SMALL ELECTRONIC COMPONENTS: $20
SOLDER: $3
BOX FOR CONTROLLER: $20 (I have no idea! I still don't have one!)

ATMEGA8: $2

Total: $313.58



With Synchronous Rectification, you would need to add about $40, because you would need 24 mosfets, and no freewheel diodes, and would need a couple extra mosfet drivers and resistors.

[MPaulHolmes]

Well, I found another place after lots of searching that had the fancy mosfets. They were like $8.50 each, but I just had to have at least a total of 12 for experimenting until the other 19 come at the end of March (which are only $5.50 each). So, if I don't destroy any, we will get a full 500 or 600 amp controller with those 12! No synchronous rectification, though. We'll save that for later. Mazda matt says K.I.S.S.!!! And I agree!!!

[bennelson]

ALUMINUM HEAT SINK: $25 (I don't know! Ben sent me his!)
[/QUOTE]

It's $30.

$20 to buy it, $10 to ship it!

Bought it from a local guy I know who is also working on a controller (although at his rate he will be done 2 years from now...)

[brihoo2k]

I know that I haven’t really posted in here but I have read until my eyes have bled, and this is the sweetest thread. I'm true ally interest in a diy controller as I would like to do a home bred hybrid using waste vegetable oil as well as biodiesel. I've spent a good portion of my life as an auto mechanic, a maintenance mechanic, a company electrician and a programmer. I have a perfect candidate vehicle an 81 vw rabbit diesel unfortunately i need to tow it to my house and it so freaking cold right now I'm not going to attempt it just yet. You see its been sitting in a field with the cylinder head removed about 30 miles from my house. I have a whole other engine and trans and another head. My plan is to keep the fwd as diesel and do a rwd as electric I have studied extensively in doing the wvo conversion as well as the making of biodiesel. I also have a slight outside chance of aquiring a komatsu 48v forklift which would supply the motor that I would mate to one of my tansmissions then scronge my local junk yard to get another front suspension and use my sawsall die grinder welder etc to graft the two together. Completely experimental but I beleive to be feasible. Your Idea of a diy controller is true ally a godsend as money is tight and nothing is free and anything to help trim the cost as well as get an education is great