Project Plan: Making a Controller for EV
 

--------------------------------------------------------------------------------

Greetings everyone,

I am in college for electrical engineering, for my senior design project my group members and I have to design and create a Controller for a GEM e6 electric vehicle by Dec 2010. We have an idea about what we want to do but we don't know exactly were to start. If anyone have any ideas it would be apprieciated.

question
 

This vehicle has a 72v, shunt motor, 400AMP, motor has 7.0 horsepowers. We was wondering if we have to design the controller that on works for a shunt motor or can we design a controller that works for a shunt or series motor.

Series and shunt motor controllers in theory are interchangable.

To drive a shunt motor off a series wound motor controller you would only attach the controller to the motor armature and the batteries would attach directly to the motor field through a contactor, this would obviously cause some limitations but would work although speed/rpm and power selection would be very limited.

Also much of your design work has already been done on the open revolt controller (paul and sabrina have a thread on it) If you really want to plagurize :( Or you could look at their design and refine and modify it for your purposes then present it back here when done of coarse

For my sepex motor, 72 V continuous on the field would burn it out, so be careful about putting that much voltage directly on the field. Sepex controllers will very temporarily put higher voltage on the field, but just briefly for a burst of acceleration or regen. A rule of thumb from Lee Hart (EV god) was a sepex is often designed to run the field at about 1/3 of the armature voltage.

Hmm, sounds like you are running a 24v sepex at 72v, most sepex come out of forks and most antique sepex do indeed run straight off the pack voltage, of coarse if you up the ante it will not like more for long.

Of coarse modern sepex use PWM on both field and armature so you can similate different torque and rpm levels.

Right, make a control board with 2 separate mosfet drivers. Each of the 2 pwm channels from the microcontroller goes through its own hcpl-4504 (or 4506) optocoupler, which goes to the input of its own mosfet driver. Write your software such that the pwm of each is what you want it to be. For example, when just taking off, perhaps it's OK to have the field voltage be very high (so high pwm duty). Keep track of how long it was at a high voltage, and force the field pwm to throttle back at a certain point. For higher rpm, you can weaken the field (lower pwm duty). There are software ways to sense the approximate motor rpm on DC motors. The dsPIC30F family has 6 pwm channels, which gives a LOT of options. They have the humble and basic dsPIC30F2010, which can come in a 28 pin DIP, all the way up to the dsPIC 30f6011 (or maybe higher, I don't know! hahaha) which has like 60 pins. They are all functionally equivalent, but the higher pin micros just have more of the same stuff (and also has CAN bus and multiple UARTs).

Actually it's the opposite for my motor. Allegedly my Kostov motor is spec'd for 144 Volts, but at 1/3 that (48 V) the fields get pretty hot after 10 or 15 minutes. At 144 Volts I would expect the field to overheat in about 1 minute. One reason to run the field at less voltage is you can ramp it up for regen, down for accelerating, and have some overhead for higher or lower rpm.

I'll parrot what I've heard: There is some difference between shunt and sepex. The field and armature are run in parallel for both. The shunt field is wound with finer wire, and is designed to run at full voltage. This is what you'd do for a constant speed motor, like a conveyor belt or elevator. It would be harder to get regen on a shunt motor, practically speaking you could only do it by physically speeding up the motor as it would be hard to give it higher than pack voltage. A sepex field is wound with somewhat thicker wire, and the field meant to generally run at less than the full armature voltage -- this gives some voltage overhead for regen.

Hi,
I cannot quite remember which sepex you got but the reasons it is hot are:

1) Your motor is rated for triple the voltage you are using. Being at 48V means ventilation is not 3 times worse - it is 4-5 times worse as it varies exponentially with rpm. It is not a coincidence that we put a lower voltage spec on our motors...
2) Your motor is wound for much higher voltage. Operating as low as 24-48V results in lower efficiency which produces more heat.
3) Class H insulation allows the coils to get to 160-220C which feels quite hot even on the outside case. Even 60C feels quite hot to touch.
4) Running a sepex motor at 1/3 the voltage does not allow you to overload properly - your armature will witness increased amps, while the field will barely increase its amps. This is very bad for commutation and heats up the comm.

Why do you use 48V anyway?!?

Many thanks for the expert comments! Here are answers/questions:

I have an old 11 inch Kostov sepex. I bought it from Randy at Canadian Electric Vehicles Ltd. . I believe he upgraded the wiring, so there is a chance my field wiring is different than the original Kostov wiring. My field is 1.0 Ohm of resistance. He also upgraded the bearings and wrapped it (fiberglass? carbon fiber? not sure).

Given your 160 C to 220 C numbers, I think my motor is running much cooler than I realized even though it felt hot to the touch (I'd guess it was around 70 C).

48 V is just testing for short times, I'm going to much higher voltage. The car has not done serious hills or high speeds, and typically testing has been at 50 to 150 Amps.

Thanks for confirming the 144V spec voltage. What I'm not sure was clearly communicated, or maybe I'm not clear on your comments: I'd think you'd want to run the armature at 144 V, but you'd typically be running the field at much less than 144 V -- true or not? How long could I run the field on 144V? This page for the newer 11 inch shows field currents of around 6 to 12 Amps: , so that hints to me to the field is getting much less than 120V.

Thanks for the temperature numbers, I'll start using my IR thermometer instead of my hand. What would you recommend for monitoring temperature? I could do an IR measurement of the brushes... or measure the difference in temperature of intake air vs. exhausted air... It seems if one just monitored the motor case temperature it would be too late by the time one detected overheat.

In addition to keeping the stock fan and generally running higher rpm, I am also adding an external blower.

I did not realise you have such an old motor - it was produced 10-12 years ago. I though you have one of the new models.
Most of my comments are still valid though yours is class F (135-195C).

Your motor is most likely the ETH 20/96/40 so 20kW/96V/4000rpm.
Should be fine for higher voltage too with the modifications but generally do not exceed 120V.
The sepex coils are usually wound to work continuously at 1/3 of nominal voltage with 10-15A. Then you can overload by increasing sepex voltage within 33-100% of nominal. Armature voltage should be increased in proportion to field voltage. This is what the controller usually does.
Assuming you will go to 120V, you can run the armature at full voltage with continuous 190-200A. The field is rated to run at 32V so going to 120V would quadrupple the amps and it will overheat very quickly - probably within 2-3min. If field is seeing 3-4 times its nominal, the armature should too be overloaded 3-4 times. This is meant to be short term (racing?) overload condition not continuous.

Many thanks, that is awesome information for me. Got it about overloads need to be just briefly.