DIY Open Source EV Charger

[MazdaMatt]

If we are going to be able to charge any type/size of battery pack from any voltage and amperage of outlet, i think we're venturing out of DIY into state of the art. The additional hardware required to accomodate EVERY battery is going to drive up complexity.

Could we select what is the most likely battery choice for a DIY EV and design the charger accordingly? I like the idea of 240/120 and selectable amperage draw.

[bennelson]

I was just thinking that if the charger used PWM controls or other fancy electronics (which could be designed as a kit, sort of) it would be a great feature to integrate.

I do think 240/120 selectable would be a great feature, as would adjustable amp draw. It would suck to not be able to charge AT ALL because your charger draws 16 amps, and you only have a 15 amp outlet available!

[lawoo]

Hello EVerybody,I want you opinion on this type of charger: Capacitive Battery Charger | The Fuelman

[NiHaoMike]

Originally Posted by lawoo

Hello EVerybody,I want you opinion on this type of charger: Capacitive Battery Charger | The Fuelman

That's actually a very simple circuit. It's a variation of the capacitive power supply commonly used in appliance controls. It's more or less just a motor run capacitor in series with the AC line and a bridge rectifier connected to the output along with a charge control circuit. Add a light dimmer in series and it becomes an effective desulfator.

However, capacitors large enough to supply enough current for EV batteries would be too expensive. A switching power supply or a series inductor would be a better choice.

[jyanof]

Originally Posted by MazdaMatt

Could we select what is the most likely battery choice for a DIY EV and design the charger accordingly? I like the idea of 240/120 and selectable amperage draw.

I agree that these would be the easiest things to implement. A lot of this would all be software related, i believe. A knob for peak voltage and for max current would be pretty easy to do.

Determining when to end charge could be tricky. Typically, you'd wait until the final current drops to a constant level. But depending on the final voltage, that current can range from 2 amps to 5 amps (if you're gassing a flooded LA pack). I think it'd be easy to have a knob for the amount of time (0 to 3 hours) to hold that final constant amperage. The controller could have a software algorithm to determine when the amps drop to a constant value and then hold that for the specified time.

Oh, maybe another knob for float voltage, unless you want to have a thermal probe to adjust the float voltage automatically.

So we're at 4 knobs, and maybe a switch for 120/240.

But, as I'm thinking, this is all software stuff. I've made some progress with the hardware - see below!

[MPaulHolmes]

See below!!!??? hehe. I think you forgot to attach something.

[jyanof]

This is my layout of all the components in the previous schematic.

Blue Caps: CORNELL DUBILIER 380LQ821M400A052
Diode on blue heatsink: IXYS SEMICONDUCTOR DSEI30-06A
Black Cap: RUBYCON 450USC560M35X50
Inductors: 150uh each E70340-015 from coilws.com
Mosfet: STMICROELECTRONICS STY60NM50
The shunt isn't being used at the moment...

It's pretty crude, and as you'll see, it presents some problems, but I think it's an ok test bed.

The small board in the bottom (half off the picture) is a 555 timer setup for PWM at 62 khz (according to the oscilloscope).

The DC clamp meter reads the load amps, 1.35 amps in this case. Off to the left of the picture, the yellow and black wire lead to a load; I started with some heating element wires and ended with a 12V battery.

The yellow wires up top leading away from the rectifier go to a 24V 10A transformer that I'm using just for testing. It's nice having an isolated low voltage power source.

[jyanof]

The first picture shows the gate-source and drain-source probes. The drain-source has all the ringing.

The first photo is at a lower current, maybe 2 amps, while the 2nd is at a higher current - around 6 amps. You can see the amplitude of the ringing increase with current.

[jyanof]

Two things that I thought weren't good with the previous scope readings were the high voltage spike and the high frequency oscillations. I suspect the high voltage spike is due to interrupting the current so quickly while having a large 300uH inductor in the low (which, btw, appears to do an excellent job of keeping a constant current through the load). My worry is that this voltage could spike above the rating of the MOSFET, which I suspect isn't good. The FET I selected is rated for 500A, but I don't know what'll happen with high voltage input. It seems benign for now...

One way to help this is to have the FET close a little slower by increasing the gate resistance. The gate resistance was increased from 6.2 ohms to 16.2 ohms and the test was re-run.

The first shows the scope at a moderate amp draw, maybe ~4 amps, and the 2nd at a higher amp draw at around 8 amps. There's still a large voltage spike, but it's much smaller, and the high frequency ringing is still present. Also, the spike does not increase drastically as current increases. Note also that the FET opens and closes slower which would lead to more heating and lower efficiency.

[jyanof]

Originally Posted by MPaulHolmes

See below!!!??? hehe. I think you forgot to attach something.

you're faster than I can type! I have a lot of stuff, so I'm spreading out the posts... just a few more now..

Did you notice these inductive spikes on your motor controller? I suspect there was much less inductance with the small bike motor...