Paul & Sabrina's cheap DIY 144v motor controller

[NiHaoMike]

Use an instrumentation amplifier to amplify the small voltage coming off the shunt.

[magicsmoke]

Hi blackpanther-st, have a look at alliedelec.com (we here know them as RS). They do one of the LEM range .. HTFS 800-P/SP2 which is an 800A nominal, +/- 1200A beast for USD23.83

Rob

[TheSGC]

How's is Ben's controller coming and when can we expect some dirty driving with it?

I am itching to see some videos of these beta units getting some road time, and I am interested in getting data on Hill climbs and highway driving.

[blackpanther-st]

Thanks both NiHaoMike and NiHaoMike for your suggestions. I had thought of using an op amp, but had wondered about about it's accuracy going from 75mv to something that would trigger a low volt zenner or reference diode, but I suppose they are made to be exceptionally linear.

That HTFS 800-P/SP2 looks really good for the price. A few complications I may have to resolve to use it, but I think it may be the answer I was looking for.

My buss bars are 3/4” square so I will have to round the corners a bit to get the 22mm aperture to fit over. I may not have to amplify if I use a 1.2v reference diode. My aim here is not to do throttle to current measurements but just a simple but quick and sure fire hardware level current limit and protection.

[MPaulHolmes]

I FINALLY tested Ben's controller with the latchup problem fixed. I just did a sort of quick and dirty program and emailed it to Joe for him to try his out, and to test Ben's. It was awesome! I floored it from a stop, and I could hear the frequency change just a bit when it hit the hardware overcurrent shutoff point, because in this version, I turn off the mosfets for 4 cycles if an overcurrent trip happens. It had a bit of a jerk at 0 rpm, but that's really easy to fix. Most of my throttle code has not had any jerk at 0 rpm. I just did a quick and dirty throttle control that I knew would work, so Joe could try it out.

I want to do a PI loop next. I'll do a video of Ben's controller, and when the laptop gets here, I'll do some PI loop tuning to get the throttle perfect, and then I'll ship it off to Ben, for him to do his worst! Then I'll email the new code to Adrian and Joe (or put it on the svn if I ever take the 10 minutes that I need to read up on it).

I also need to write up some directions for Adrian to assemble his. I'll try to get to that this weekend.

[MPaulHolmes]

An interesting thing to consider for all those clutchless people in the house. If you do throttle proportional to current only, it can cause slight problems under certain circumstances.

For example, let's say you took the car out of gear, and you are coasting at maybe 25 mph. Now let's say you want to put it in gear before stopping. Well, you have to rev up the unloaded motor a bit to be able to put it in gear. However, throttle proportional to current means that an unloaded motor is VERY sensitive.

[blowncurtice]

I can blame it on age or rusty brains however I got it wrong last time, I mixed up the prices Mouser charge for AU$ orders free shipping and their price if you elect to pay US$ . Here is my revised list prices. Of course the more people to share the cost of shipping the cheaper per person.
John

[MPaulHolmes]

The new laptop from the SGC is on it's way, but a friend of mine (evrag) let me borrow his laptop, so I decided to get to work loop tuning! I installed all the software, and tried it out.

Here's a graph of the current response when throttle suddenly jumps from 0 to 50 (on a scale of 0 to about 500). This isn't a PI loop, but instead was just some ramping code, like this:

if (throttle < current)
pwmDuty--;
else if (throttle > current)
pwmDuty++;

Also, pwmDuty was in the range 0 to 4088. Throttle feels pretty good on this setting. Not TERRIBLY responsive, but not bad.

The current is blue, and the throttle is red. Each tick mark on the x-axis represents 1/4000 of a second, so it took the current 0.04 seconds to get up to where the throttle was (50 amps about):



I was just trying it out. I wrote the code for the PI loops, and now I'm going to test that. It's almost like having a digital storage oscilloscope! hehe.

[reggiewatson]

When do you guys think we will be able to buy the boards for the controller?
Thanks,
Reggie

(Keep up the great work)

[blackpanther-st]

Paul that must make it so much easer to adjust your code when you can see the reaction data like that.

It would be great if you could get your Bata testers to send this type of data to you from different motor / vehicle weight applications. Then you could write some presets or a better general setting. Even better yet find a high and low range and program a simple user accessible setting to vary the response to anywhere between the high and low points. The user just picks their motor and vehicle weight from a chart and enters an appropriate value for their combination and they would have a controller tuned for their particular application simple as that.