DIY Open Source EV Charger
 

Hey Folks!

Since Paul is going on the whole Open Source 144V EV motor controller, we need to come up with an electric car battery charger to match.

One advantage of a higher system voltage on an EV is that it opens up the possibility of using a home-built 120V AC battery charger.

Does anyone here have experience building one of these?

Keep in mind that we want it to:

1) Be inexpensive
2) Safe
3) Actually work
4) Able to be built by a non-rocket scientist

Yes, Please! Someone do this! And make kits!!!

(My soldering looks WAY worse than the Forkenswift's welding ever could!!!) :D

For a 120v 6amp charger, the price ought to be lower than 10*20 = $200, as that is the cost of 10 12v 6amp Harbor Freight chargers when they are on sale, or for a 120v 10amp charger, the price ought to be lower than 10*30 = $300 (those go on sale for $30 each sometimes). Well, maybe a higher price is acceptable since the charger would be smaller than 10 12v chargers.

There are a few _very_ simple chargers at EV Technical Files.

I would personally only use the "Bonn Charger". To me, bad-boy charging is a rather dodgy affair and I would want to have as many safeties in it as possible.

Also, I think you could change the output voltage by adding a variac or transformer to the input (before the GFCI). Eventually your output voltage will decrease to the point where very little/no current will move between charger and battery pack.

To my mind, these particular circuits are very much sub-optimal for use in actual electric cars, but seem like a legitimate jumping-off point for adding additional controls and stuff to make a pretty acceptable charger.

I could make a black box that can read various inputs (signals about state of charge and whatever?), and have a microcontroller make decisions based on that information, putting out any voltage/current that you want, as long as it is less than or equal to the voltage of the input into the black box. Let's say it is a 144v charger. The black box would need to have an input of maybe 165v DC (I'm not sure how high the voltage gets when in fast charge mode. Is it around 13.7v per battery?).

Basically, that's all a motor controller is, is something that reads inputs, and puts out a voltage that is controlled by the user. You can control the voltage output by a micro-controller just as easily (with the human removed. No throttle!). You would only need a single 200v fancy mosfet, because the continuous current output of the charger wouldn't be any higher than 30 amps I would think, and 1 of those mosfets can put out 50 amps continuous.

It would be a PWM based charger. The PWM signal would be smoothed out by an inductor (like a motor) that would be inside of the black box. The inductor would just be a coil of maybe 12 gauge wire.

So, how do you make a steady 165v input source? All the charging algorithm stuff is just gravy with a micro-controller!

I believe that is how the cordless tool guys do their multi-voltage chargers...

Really? Hey! You are right about the multiple voltage. If you wanted to charge a lower voltage, you could just choose that setting. For example, if you wanted to charge 72v, the PWM inside the black box would only run at 50%. The average voltage would then be half of whatever the maximum DC that is being inputted to the black box. I think that could work. Thanks, Intrigued!

The smart chargers I have used in the past usually cut off at 14.4V on a 12V battery.

Dang! A 173v DC input to the black box then? The battery pack could be like the back EMF from the motor. Hmm... The black box could supply very high current to the batteries. You could vary the voltage to control the current going into the batteries (same as on a motor controller, but replace batteries with motor!), so people could choose the speed of charging. Some batteries have very low internal resistance, like my super fancy Sears AGM platinum editions! (Thank you Sabrina!) Those suckers can be charged at 40 amps each! I think current limiting might be a good feature. Imagine charging 144v at 40amps. You might want to check with the neighbors to see if they were going to use the power grid before doing that.

Paul, I think you missed your calling.

...or, maybe you just found it!!! :D

Nah, my calling is to be sworn at by middle schoolers. hahaha! But seriously, it sure is fun! I never knew electronics would be such a fun and creative process, and I get to use the programming experience! I love it!

The old K&W 20-amp chargers couldn't charge above 108V packs because they couldn't increase the line voltage. You needed a "line booster" to do that (and even then they could only do 120Vdc).

So here's a crazy (as in fairly loony) idea. Instead of rectifying, feed your AC line voltage through a full-wave voltage doubler. That should give you about 340Vdc to run through your PWM smart bits and charge whatever you want. It would also allow you to charge from a dc source if you put in such an input behind the doubler. I'd also definitely want an isolation transformer on the input of such a machine because those'd be 3 fairly massive electrolytic caps you'd have in the input stage.

Those Sears Diehards have reverse polarity, or dead cells, or some crazy relay. Started charging all 4 between 13.3V to 14.7V. 2 hours later 3 varied in the 13.5V range, but one dropped to 12.8.

Hauled the Diehard booster out and charged that low bat at 14.7 about 3 hours. Charge charge charge re charge. Pretty annoying.

I am getting to the point of buying 5 Diehard chargers, put them in 1 box (like an octopus with wires instead of legs; and charging all 4 (48V to 56V) at once, instead of spending a whole day charging. :mad:

That's what I do too! I have 6 12v 6amp chargers ($20 each). It keeps them balanced, but it's pretty slow. When each one is done, it's charger shuts itself off on it's own time.

I'm going to read up on transformers and voltage doublers and who knows what!? I think it's a great idea to start with a really high voltage so the charger could work for anything from an AC system to a DC system. You could have a little knob on the side that would select what voltage you want to charge at. You could have another knob that would select if it's Lead Acid, NiMH, or Lithium Ion. Each algorithm is well defined. It's just a matter of controlling voltage and current based on some sort of feedback from the batteries. Wouldn't that be cool!? It might be good for there to be 2 knobs for voltage, so there could be 10, 20, ..., 320 (or whatever), and the other 0.0 through 9.9. Lithium Ion are weird. Each cell is 3.7v. The Lithium Ion batteries have almost the same charging algorithm as Lead Acid, as far as I know. NiMH is totally different. None of them are very hard with a micro-controller though, and a little high school math! I love how you can use math in the real world! It actually makes real stuff work. It's weird!

Welcome to the World of Tomorrow!
http://i2.photobucket.com/albums/y42...k/Futurama.gif


Great work btw! :thumbup:

My personal experience is with Soneil chargers. They nominally charge constant-current to 14.4V and then float to 13.8V.

If you do that make SURE that the chargers are isolated. Check for continuity between the output and input leads, and if there is, it isn't isolated. Non-isolated chargers CANNOT be used in series or else you get blown fuses and general ugliness.

I did some reading on chargers and found the wikipedia page that explained that a "simple" charger either uses constant current or constant voltage, but may not be the most efficient... would it be reasonable to string 9 batteries in series and simply apply a wave-rectified, capacitor-smoothed wall outlet voltage across them? Sounds absolutely insane to me. I'm just thinking, if I had a 12 battery pack and had a dpdt switch that could make them either 2 parallel strings of 6 or one series string of 12 (not a difficult switch to imagine), then i could use a very gentle step-down transformer and a couple big capacitors to charge them...

again... sounds crazy to me, but that's what i gathered from a "simple charger" description.

The Sears Diehard charger goes to about 16V then slowly drops to 14.7V for 2 hours, then shuts off. After charging floats to 13.2 - tops! Yesterday 1t stayed at 15V with those confusing computer chargers on-board. Unusual!

Each battery post has about 6 bolt-ons, clips, loose wires, fuses.... Take me an hour to undo everything for a simple daily charge. Multiplied the problem!

I think that would be fine. You could have it shut off a tiny bit early when all the weirdness starts (the bulk charging is done, and hydrogen gassing might begin in earnest). Lead Acid is very forgiving during the bulk charging phase, or so I hear. I don't know what I'm talking about!

Actually, I think that's a really good idea. Charge 2 72v strings if you have a 144v pack. Then you could do without a transformer in the charger, and you could smart charge it with a micro-controller and PWM setup, just turned down a bit, since the rectified DC voltage would be around 160 or 165 (just a little bit less than the peak of about 170v.), and you could just run the pwm at 50% to get an average voltage in the 72v charging range.

I checked yesterday, and with 11 of my 200v ripple current capacitors, at 24v through the motor, there was a ripple voltage of 0.1v. I bet it wouldn't be that much ripple from the wall. I have no idea though. I need to test some of that crap out. haha!

I just stumbled across a fairly inexpensive 72V charger!

Schumacher Multi Battery Charger — 10 Amp, 72 Volt, Model# SE-1072 | Battery Chargers | Northern Tool + Equipment

http://www.northerntool.com/images/p.../167974_lg.gif

This charger is manually selectable up to 72V.

Thanks to Norm's YouTube video for the lead on this one.

This charger is about $220. It's not the exact same one as in the video, but manufacturers change the look of these products all the time.



http://www.youtube.com/watch?v=TONtT...eature=channel

Schumacher is a fairly well known name, that's been around a while. With a 144 volt system, maybe split it, and buy two chargers, to be able to charge it overnight???

Can we beat that building them ourselves, Paul???

The peak voltage of 120v AC is 170v, which should be plenty for most DIY EVs.

Directly rectify the AC, maybe with a voltage doubler for higher voltage batteries, and then use a buck regulator. Should be easily achievable using mostly off-the-shelf components. Start in current limit mode and switch to constant voltage once desired voltage is reached, then completely turn off after a while (or go into float charge).

I believe the most important thing to make a diy on is the CHARGE CONTROLLER

It is easy to make a dumb charger out of most anything, I even made a higher tech version of this thing to desulphate/controlled overcharge bad batteries

3 Dollar Battery Charger

As stated earlier, The trouble with any of the DIYs is we need a

1. SIMPLE
2. Cheap
3. Reliable
4. Safe will be a function of the charge controller

Device that can turn a contactor or mosfet on and off at preset voltage levels, example my pack voltage should be around 76.5v the max overcharge is usually around 81v if I had a simple, cheap and reliable device that could turn the juice on when voltage was below 76.5 and turn it off once it reached a preset max like 80/81v or whatever pleases me at the time (adjustable would be best) we would be a long way toward making existing cheap chargers safer and ones we build more effective and safe.

The next step would be to make a charge controller, I hear there are AC pwm dimmer switches around for lights and other household items (never knew that) if a controller could be setup to trim the pwm dimmer switch or equiv you could easily control charge current and follow a taper pattern. The beauty of any PWM system is that you end up with a pulsing pattern which helps desulphate batteries as well as charge them.

I think this is a more logical 1st step as suitable transformers are in almost every device I can think of, I just trashed out an audio system for someone to get their CD's out and got a boatanchor multi rail power supply with either of the above devices and a simple diode or bridge rectifier I would have a cheap reliable 8 amp multi volt charger.

Another type of charger I have been contemplating is a cap dump charger you need an input cap of high capacity and an output cap of very low capacity, 50/50 pwm oscillation speed determines charge rate (strap ac to large cap 1; unstrap ac and strap large cap to small cap2, cap 2 discharges into battery while cap 1 charges off ac line current through a bridge or diode)

Anyway I think a workable DIY charger is MORE important than a diy motor controller because it is simpler & cheaper to impliment on your own without lots of knowlege, it also seems to be an item that burns out more often as well. Not to mention a decent charge controller DIY could also spill into other things such as renewable solar and wind systems.

Cheers
Ryan

Ya, I was suprised just how powerful PWM is. I definitely think it should be a part of a charger. You should make one! That would be awesome. We need a good PWM based charger. Git' R' Done! ya!

Lol, I am probably just smart enough to cause lots of trouble :)

If we went buck it aint cheap
http://www.zahninc.com/sd1.html?gcli...FQxKGgodQCnydA

Geeze. That is expensive. That kid on the right looks like he is on a work release from prison. hehe. I'm not one to talk, though. I also look like I just escaped from Jail. Hey, I did just escape! hahaha! Don't tell anyone, though. It's a secret.

Did you get early release....for the long weekend?:D

I smile every time I see "early release" on school reader boards:p

Just found this thread...
 

... and i've been thinking of this project for a while now since I got my car working and am unsatisfied with the high power onboard chargers that are either massive in weight or massive in dollars.

The ultimate goal: I want something that'll pump 20+ amps into my 144V nominal pack when running on 220V, full 3 stage charging that is adjustable for the number of batteries I have or ambient temp, and will run on 110V without changing anything. Well, maybe a current knob to adjust the current down when on lower power lines.

I got so far as drawing up a design using the TI BQ2031 chip to run a power section as a buck converter with PWM, but have no EE basis to validate it (my work/job is all mechanical engineering).

But, I did order all the power section components and have laid them out. My plan is to test this with just a 555 pwm setup sending a signal to the mosfet driver, so essentially I'd have a fancy 'dumb' charger for now if it works. I attached the schematic below - I think it's virtually the same as a motor controller except that it has coil components for added inductance.

There's some big caps for input voltage smoothing from the bridge, another cap across the battery, some large inductors for current smoothing, a diode for clamping the voltage spikes when the FET closes, a large FET acting as a switch, and a shunt to measure current.

The components:

Big 400V caps: $10 each
Big 30A Inductors: $27 each
600V 37A diode: $5
500V 60A FET: $18
600V 40A bridge: $6 (from a surplus site)
Various heatsinks: Maybe $20 worth. I went overkill with just things I had around the house and found at a local surplus shop.

So that's about $140 worth of stuff. Everything else - the control board - shouldn't cost anywhere near that, so we're talking a fairly inexpensive high power charger if this works.

I went overboard with some of the specs since I see nothing wrong with having margin. Also, I can see this project running in stages:

1. Get the powerstage working with just a generic PWM signal (555 timer).

2. Get the control board to successfully control a charging algorithm. (the BQ2031 seems overly complicated - maybe Paul can help with the programing of a custom controller?)

3. (and this is where the high voltage components come in) Build a 300V (or so) Power Factor Corrected power supply that'll supply 300V (or so) to the charger from an AC input from 100 to 250V. This'll allow the charger to run from any typical voltage source AND maximize the power drawn from the power source.

But first things first. Anyone have any comments on this power section schematic? I read up a little on snubbers, but not sure if they're needed or how to implement/design one. At some point, I'll wire all of these up, borrow a scope, and see what happens (i'll start small, maybe just charge a 12V battery running off a 24V transformer).

But, any input is much appreciated!

Joe

I can help with that! There's no way we need a BQ2031! YA! Way to go Joe! I'll study your schematic more later. I have to do stupid school work right now! Makes me mad, way down deep. $140 for a 144v charger would be really really really nice.

jyanof, can you make a quick description of how the circuit operates? From what I gather, the battery and a 47k resistor are always in series with 240(ish)VDC. Then when the mosfet is on, the battery is then in series with the 2 coils straight to 240... so a pulsing of the mosfet causes a drop in voltage at the negative terminal of the battery (from the 100v or so that the 47k was holding up, down to near 0) and the coils push the current. (I'm no power electronics engineer!)

Excellent! I taught for two years... I know what you mean.

I can't imagine the control logic is that difficult. Maybe it's tough to have constant current and/or constant voltage control? I don't think the system would be dynamic enough to need fancy PID control... (like the current limiting suggested for your controller) I have a 12V power supply that I can hook up to a 6V battery to charge it and manually adjust to the desired current, so it seems like a computer could do it without missing a beat.

If there was a current sensor fed back to a cpu, that cpu could extremely easily control the pwm to make a constant current system. I'd use PI control... it isn't a difficult thing to do. Again, i really wish that i was not so busy, i'd jump on firmware development for this project and paul's motor control... all i can do now is yell stuff from the side lines...

Me neither!

But, I think the way you described the circuit matches my understanding of it.

The 47k resistor was left over from a sample circuit in the BQ2031 datasheet. Admittedly, I'm not sure of it's purpose, but I suspect it could help reduce initial voltage differences, or allow the cap to charge slowly when it's plugged in (though that would take forever). I can't see anything wrong with leaving it in, so I haven't taken it out...

But otherwise, the FET closes and the battery is connected to the DC powersource (rectified AC). A rapid rise in current is limited by the inductor. The FET then closes and the inductor releases it's energy through the diode and batteries. I don't know what to expect for voltage across the FET at this point. If my mental exercise is right, the voltage at the bottom of the diode must be greater than the top if current is going to flow through it, so that's something greater than about 180V near the end of charge. I suspect the 500V FET can handle whatever it is, but might be something to keep note of.

I would not remove that diode. I believe it serves the same purpose as the freewheels on the motor controller. Going back to basics, the voltage across an inductor is relative to di/dt... so when i goes from 10A to 0A in .000001seconds, you get a LOT of voltage. Again, synchronous rectification would be smart here, but i'm sure this is fine.
The resistor might be there to set up the Vds biasing... not sure (mostly slept through mosfet class :) )

Currently, that's what the 50mv shunt is for. It'd need amplification though. I suspect a hall effect sensor would work too.

I'm just thinking out loud now...
Perhaps an even better place to monitor current would be on the AC line. I think AC hall effect current sensors are pretty cheap and would allow you to set and max out the power source throughout the charge cycle. Otherwise, if you hold the battery current constant, your power draw will increase a little as the voltage rises, potentially tripping the breaker partway through charging (and likely unknowingly in the middle of the night).

Oh, I agree! I was referring to the 47k resistor...

I would definately keep the current sensing to the battery line. I don't understand what a 50mv shunt is? I have heard of 50mohm shunts in series with the power line and paralleled to op-amps as a current sensor (that's what is on the boards that i'm testing this morning, actually). So as, say, 10A passes by, it generates 0.5v, the op-amp has a gain of ten, so 5v goes to the cpu's ADC... that sort of thing.

Total system design should take into account a maximum power and the input ratings should be specified accordingly.

Oh, maybe better terminology would be a 50V 50mv shunt, which would be a shunt of 0.001 Ohm resistance. (My electrical engineering greenness is showing through)

As for input ratings, I think versatility of a charger is important. My goal would be able to plug in at home on my 240 30A line, then drive to my parent's place and charge off their 120V 20A line, and make a trip by the school and plug in to the 120V 10A outlet. My current problem (pun intended), is that the charger I have now only runs off of a 240V 30A outlet, so I can pretty much only charge at home.

Thus, the need to adjust the input power (current). I guess there's a number of ways to do this. (thinking out loud again...) Probably the most time efficient method I can think of would be:

1. Determine a max battery current based on the charger components. I'd probably conservatively set it at 20-25A for the ones in the schematic.
2. Set the max input current from the power source
3. Monitor both currents and maximize until one or the other reaches its respective limit.

From a 240V 30A source, the battery current would max out first at 20A or whatever, while the source current would max out first from a 120V 10A outlet.

This would of course only be during the bulk charge phase, but would allow for the quickest possible charge.

I guess it isn't necessary. The user could just select the battery amps knowing that the setting would have to be a little below the power source max. But, i think it'd be easy to do, so why not?

I don't know about the power factor (if you want, it should be able to be corrected by passive components), but this circuit should put out ~300V from a 120Vrms source.

If you want 240Vrms, just supply it to the leads marked "2Vpp" and "(-)".

I love the idea of a smart charger which would have some adjustments for amps and volts.

One down-side of the gel-cell batts. I am using right now, is that they require being charged at a little bit lower voltage. (higher voltage cooks the gel electrolyte!)

With a smarter open-source charger, the same charger design could be used no matter what style battery - AGM, Gel, Flooded, etc. is used.