DIY Open Source EV Charger

[MPaulHolmes]

I was reading about your questions on the EVTech list. You are doing an awesome job!!! This is very close to the way I was wanting to try it. I really appreciate you putting all that work into this! Down the road if you want to switch out the 555 timer for a micro-controller, I just bought 10 ATMega8's awhile ago, and I could help with the programming if you want. Awesome Job!

[MPaulHolmes]

I did get voltage spikes and ringing on the rise and falls. They were pretty small, like 3-4 volts or so. The ringing didn't seem to increase in magnitude as the voltage increased if I remember right. I'm about to test the car's version in a couple days, so I'll double check that. One thing that helped with the spikes was putting small (like 0.1uF, 0.01uF, 1uF) MLCC capacitors from the input of the gate driver to ground.

I actually did it in real time just to see how it affected things. It was pretty cool. You could actually see the spike shrink each time I inserted a capacitor, and it would grow when I removed it.

Is the freewheel diode responding too slowly? What's the response time for it? I guess I could look it up. Never mind. hehe. I think you gave the part numbers. Hmm... Could you get away with just a single 150 uH inductor? Have you seen what it looks like with just one of them in the circuit? (are you using 2 of them? I forgot. I guess I can go back and read it in a second. hehe)

[jyanof]

I did some research about the high frequency ringing and determined that it was due to the parasitic inductance and capacitance in the wires/FET etc... The frequency is around 3.3 MHz, so I think this is reasonable.

Anyway, I also read about RCD snubbers and hacked one together out of components laying around. It's pretty ugly, so I didn't even take a picture of it. But it took care of the high frequency oscillations, but left the high voltage spike.

The first picture shows the new readout at low power, and the 2nd at high power. The high frequency oscillation is damped away, but the voltage spike remains.

The last photo is actually taken while charging a battery at 20 amps (I used the 12V SLA accessory battery from my EV as a guinea pig). The FET was starting to get warm, but still touchable, but I didn't let it go long since I didn't want to push it.

So, given that my purpose of this stage was to demo the basic power circuit by actually charging a battery, I'd say this is a success so far!

Things to do:

Determine if the voltage spike is a risk. Will it increase at higher input and pack voltages? Is it only dependent on current? I plan to do some tests at 120V, but have to get my hands on an isolation transformer in order to use the scope (I tried - apparently it doesn't like grounding the scope to a rectified rail).

Get some more input caps. At 20 amps, they were draining to zero quickly on each AC cycle. I don't think this is good, but I really don't know.

Then, further into the future, implement a microcontroller to fully automate charging...

[jyanof]

I do plan to switch out the 555 timer at some point, and I have zero experience with micro-controllers. But, your success so far with the motor controller gives me hope! Were you able to learn just from the application notes? What materials would you recommend to help learn?

I'll have to try the small caps on the driver input - you're referring to the signal input, right?

As for the diode, it appears that the spike takes about 100 ns to build, but I couldn't find the forward-recovery time (is that right?) in the data sheet and I have no idea whether a typical diode can handle that. I haven't tried removing one of the two inductors, but that could at least verify the source of the spike.

Lotsa stuff to try! Will have to save it for another day...

[bennelson]

Jyanof,

What range of voltages does your current charger setup allow for?

Lower voltage chargers are not that expensive, but above 72V is. An open source charger that could do 72-144V would be ideal.

[esoneson]

Me too, me too! I am also looking for a reliable high voltage 144 volt battery charge system. I love the idea of the open source solution. There are plenty of people who would love to have that option.

Paul let me know about the development system for the AtMegaXX microprocessors and I now have one on order. Complete setup including several processors for under $100. Develop code in Windows and load to the SDK board. Couldn't be easier. Software CAN be a viable solution.

Now I just need to learn more about hardware.

Eric

[jyanof]

Originally Posted by bennelson

Jyanof,

What range of voltages does your current charger setup allow for?

Lower voltage chargers are not that expensive, but above 72V is. An open source charger that could do 72-144V would be ideal.

My current charger is a quickcharge model - 20A and set for 144V. I find it adequate as a stand-alone charger that sits in the garage. I think it lacks input flexibility (240 only), accurate charge termination (doesn't detect a minimum constant current at end of charge), and is outrageously bulky and heavy at 60 lbs and the size of a large over-stuffed backpack.

Obviously, I'm moving forward with the intent of being able to charge higher voltage packs like mine, but that doesn't mean the charger would be limited to such high voltage packs. It'd be easy to adjust the settings - either in software or externally with knobs - to allow the controller to sense different pack voltages. I think the only question would be if the system is stable enough at different pack voltages to allow the controller to maintain current control during bulk charge or voltage control during the constant voltage phase.

I suspect that 72-144V would be doable, but we'll see!

[MPaulHolmes]

Originally Posted by jyanof

I do plan to switch out the 555 timer at some point, and I have zero experience with micro-controllers. But, your success so far with the motor controller gives me hope! Were you able to learn just from the application notes? What materials would you recommend to help learn?

If you go with AVR, there's AVR Freaks

They are very helpful. I always had an answer to any question I had before the end of the day. I learned mostly how to do stuff from the data sheet. It was one horrible week.

Originally Posted by jyanof

I'll have to try the small caps on the driver input - you're referring to the signal input, right?

Ya. In addition to the small capacitors, I also use two 4.7uF tantalum capacitors in parallel (which I guess are still small haha) connected from the signal input to the gate driver to ground. If you use electrolytic capacitors, they will get really really hot, in fact, any capacitor with too large of ESR, it will burn up (at least it did in my case).

Originally Posted by jyanof

As for the diode, it appears that the spike takes about 100 ns to build, but I couldn't find the forward-recovery time (is that right?) in the data sheet and I have no idea whether a typical diode can handle that. I haven't tried removing one of the two inductors, but that could at least verify the source of the spike.

Lotsa stuff to try! Will have to save it for another day...

The diodes I'm using have a reverse recovery time of 22ns and forward recovery time of 220ns. It's sort of weird. 220 ns seems like a long time, but I didn't have any problems with spikes. I'm switching at 16 khz.

What switching frequency are you using with the 555 timer? Higher switching frequency can prevent the capacitors from draining completely, and might allow you to use a smaller inductor for smooth current, but I think you can overcome the voltage spike with the right diode. It must be possible, dang it! haha!

[MPaulHolmes]

Originally Posted by jyanof

I think the only question would be if the system is stable enough at different pack voltages to allow the controller to maintain current control during bulk charge or voltage control during the constant voltage phase.

I suspect that 72-144V would be doable, but we'll see!

Current control was a little tricky, but turned out not to be that bad. You just need a hall effect current sensor, and then a control loop to modify the PWM duty until you get the current that you want. That was the last 2 months of my life. I sort of wish I had it back. haha! just kidding. I think it would be the same idea for charging batteries, of course I have no idea what I'm talking about.

[lectruck]

Is there any update on the Chargers progress, now that Ben is trying out the Cougar Controller in could come in handy?