PHEV conversion using CAN-View and BMS+

[jawshoeaw]

I'm starting my thread again in the proper category (I hope)
I'm slowly converting a 2008 Prius to PHEV using the CAN-View and BMS+ with two extra (total of three) Prius traction batteries. Both batteries came from low mileage 2008 Prius.

Charger is simple capacitive doubler which sags under load, i.e. when charging the batteries, to some V lower than 2x the mains peak of 170V. So open circuit is 340V but when I plug it in to one of the battery packs for testing, it just shows the voltage of that battery, but slowly rising. It's about $40 in parts including two caps, two diodes, and two resistors to bleed the caps down when you unplug it. Very simple. Also dangerous, since I'm not sure at what voltage it would stop charging. So several safeguards will be added such as temp, voltage, current (fuse).

Progress is slow, but charger prototype is built, tested to put out about 3.5 A at 220V. It gets a little warm but not bad. The Prius battery doesn't change temperature at all (that I can tell) when charging from 213V up to 230V then discharged back to 213V. I haven't pushed it much - I'm load testing it on an old 220V baseboard heater which draws by coincidence about 3.5A DC at 230V.

I'm wiring in EV mode button tonight (homemade) and will begin testing the car as a plugin with just one battery (stock) to make sure the BMS+ is working OK. WIll post pictures soon.

[vertex]

Originally Posted by jawshoeaw

I'm starting my thread again in the proper category (I hope)
I'm slowly converting a 2008 Prius to PHEV using the CAN-View and BMS+ with two extra (total of three) Prius traction batteries. Both batteries came from low mileage 2008 Prius.

Charger is simple capacitive doubler which sags under load, i.e. when charging the batteries, to some V lower than 2x the mains peak of 170V. So open circuit is 340V but when I plug it in to one of the battery packs for testing, it just shows the voltage of that battery, but slowly rising. It's about $40 in parts including two caps, two diodes, and two resistors to bleed the caps down when you unplug it. Very simple. Also dangerous, since I'm not sure at what voltage it would stop charging. So several safeguards will be added such as temp, voltage, current (fuse).

Progress is slow, but charger prototype is built, tested to put out about 3.5 A at 220V. It gets a little warm but not bad. The Prius battery doesn't change temperature at all (that I can tell) when charging from 213V up to 230V then discharged back to 213V. I haven't pushed it much - I'm load testing it on an old 220V baseboard heater which draws by coincidence about 3.5A DC at 230V.

I'm wiring in EV mode button tonight (homemade) and will begin testing the car as a plugin with just one battery (stock) to make sure the BMS+ is working OK. WIll post pictures soon.

I am working on basically the same thing. I have the extra pack, and I designed a charger with the same voltage doubler, but I use a FET to control the peak current and turn it off at max voltage. I have not built it yet, but I was planning on doing that when I figure out the whole BMS system. I don't just want to put the batteries in parallel, and have them smoke.

[jawshoeaw]

Originally Posted by vertex

I am working on basically the same thing. I have the extra pack, and I designed a charger with the same voltage doubler, but I use a FET to control the peak current and turn it off at max voltage. I have not built it yet, but I was planning on doing that when I figure out the whole BMS system. I don't just want to put the batteries in parallel, and have them smoke.

are you using the fet to pump high frequency pulses to drive a transformer (as opposed to mains frequency)? I'm curious how the FET can limit current. In simulation software, the cap doubler has low average current but some pretty nasty spikes. the battery seems to soak these up. i hope.

[vertex]

Originally Posted by jawshoeaw

are you using the fet to pump high frequency pulses to drive a transformer (as opposed to mains frequency)? I'm curious how the FET can limit current. In simulation software, the cap doubler has low average current but some pretty nasty spikes. the battery seems to soak these up. i hope.

I am using the doubler, it is easy to do at these currents/voltages. The diodes have huge current spikes, you need 80 amp diodes. The average current of course is lower, 12 amps in for 6 out. I used a 6 amp resistive load in my simulations. I found you need a lot of capacitance, 2 miliF to get a low enough ripple to have an average voltage over 250V. (I don't remember the exact #s, but this is close). Of course the battery will act as a big capacitor, but I wanted the current to flow continuously and not be modulated by the AC ripple. I a syncronous buck controller so with current and voltage feedback, so it works in costant current mode until reaching the preset charging voltage, and goes into constant voltage automatically. I'd be glad to post the schematic, but I would like to test it before I say it works.

[jawshoeaw]

Originally Posted by vertex

I am using the doubler, it is easy to do at these currents/voltages. The diodes have huge current spikes, you need 80 amp diodes. The average current of course is lower, 12 amps in for 6 out. I used a 6 amp resistive load in my simulations. I found you need a lot of capacitance, 2 miliF to get a low enough ripple to have an average voltage over 250V. (I don't remember the exact #s, but this is close). Of course the battery will act as a big capacitor, but I wanted the current to flow continuously and not be modulated by the AC ripple. I a syncronous buck controller so with current and voltage feedback, so it works in costant current mode until reaching the preset charging voltage, and goes into constant voltage automatically. I'd be glad to post the schematic, but I would like to test it before I say it works.

It's interesting - I found the same problem in simulation, needing high capacitance. Probably because you can't easily simulate a battery. But in the real world, I'm using two 60amp diodes and two 390uF caps in a typical doubler arrangement (which I think means the capacitance is additive to 780uF) and in this arrangement, the output "follows" the battery's voltage right up to 238V. All that ripple is just swallowed up and obviously even if average voltage is low, a lot of current is going to the battery. efficiency on this setup is about 70% as measure by kilowatt meter and amps measured going to battery. I don't have a scope but someone else with the same rig as mine had a scope put on during charging and said the spikes at least (if not the ripple), were gone during charging.

Using 390uF caps, I'm getting average current of 3.5 A, and that's measuring through my little $10 multimeter, which seems quite happy to be part of the current path. If you use 680uF caps instead of 390, current jumps up to more like 8 amps. Prius battery doesn't even get noticeably warm after 5 minutes of that.

It bugs me though that I can't get the simulation software to more closely follow the bench results. it's an awfully simple circuit, except of course the battery.

last thought - my understanding is that these batteries will go up in smoke at constant voltage charging, no? Above about 240V, you get a lot of heat generated even after charger is removed (from self discharge I assume) but even if you held at 238V, I think you can still in some sense be charging. In an idealized flat line DC current maybe not since in theory the voltages would "cancel out" but for safety I think it's better to just kill the charging current at some set point in time, plus temp cutout of course. I wanted to do voltage based kill switch but the voltage reads 242V then you turn off the charger and it sags down to 236V or something. Too imprecise. Might still do it as a safety device.

[jawshoeaw]

Update: Slow progress, I know, I know. I charged the OEM pack up to 238V with my prototype charger and did some test runs. Man, one pack even charged beyond what Toyota allows doesn't get you far. I drove about a half mile up a gentle hill and the voltage was 213V (after I stopped and let the car rest a minute). Drove home which was of course mostly downhill. Voltage back at home was 208V. 25 minutes on the charger put it back to 238V.

So, success! My harness is in place, wired for one extra pack at this point, will test two then add the third. Or maybe I'll just throw both in there. I'm using 6 AWG wire from Lowes. It's not very fine-stranded (not like OEM wires) but it's OK. I sort of half-crimped on the connectors, enough so that they wouldn't budge with a hard pull, then soldered them. I know it's not ideal but I don't have a heavy guage crimper and the environment is mostly isolated, minor vibrations, etc. so I'm not worried about metal fatigue around the solder/copper. Needless to say I will periodically inspect.

To connect the HV lines, I'm using 1/4" stainless steel nuts, bolts, washers, lockwashers and lead free solder on silver coated crimp-terminals. Not too worried about galvanic corrosion since the environment is dry and lead-free solder is close to 300 series stainless, silver, and copper in reactivity.

Power cord to my charger gets warm so I need to upgrade that - it was some salvaged power cord to a computer monitor I think, probably 18 or 20 ga wire and the charger according to kill-o-watt meter is pulling just about 1kW or 14A at 117V. This is more than I thought I would get based on other's similar charger, but he runs his through a cheap multimeter to monitor current, and I'm suspicious now that the multimeter in his setup is slowing things down. Either that or this charger is incredibly inefficient. I'm assuming 112V x 14A = 230V x X so X=7A but when I place my multimeter in line with the current flow it read 3.5A, about what the other guy got doing the same thing. I don't think 3.5A of current is being wasted as heat. The heat sinks on the diodes get warm and the caps get warm, but not 500W worth. I'm thinking more like 100-200W wasted heat. Will do more testing and post pictures soon.

Oh, final bits of data, difference in V between subpacks is very good after a couple charge/discharge cycle, less than .06V (high-low). Temp of battery on charge cycle from 208 to 238V raised battery from 75 to 85 degrees F in ambient temp of 68 degrees F. I checked ten minutes later for any temp. elevation but the temp had actually gone down.

The Prius does some very strange thing if the 12V gets low (like when you leave the traction pack disconnected by mistake - oops). First it warns you that your parking brake pawl is not engaged anymore and to use emergency brake. That was the first hint. Then CAN-View turned off. Finally light on dash get dim (duh). I charged back up the gel-cell for a few minutes because at that point i couldn't even get the car to start.

[jawshoeaw]

Wow, I needed to update this a long time ago. I want to post my progress properly but for the impatient, I'll summarize with good news. I have had no problems with my three total packs but I find myself constantly a little short on volts (you know what I mean) so I've added a fourth battery!!! Yeah! This one is from 2007 with 30K miles and I got for much less money. And after selling off some of the unneeded parts (like the battery ECU) the cost was brought down to under $200. Now I'm getting greedy and want to add more packs under front driver and passenger seat, but that will have to wait.

I built an upgraded charger which can switch between two "banks" of capacitors and is continuously cooled. Fast speed is 9A, and the slow speed is about 4 Amps, which is really slow on 4 packs, less than an Amp per pack. Only problem with fast speed is it pulls about 24Amps from 120V outlet. (Power factor of ~75%) That's fine for 30A dedicated breaker at home on a short run of 12 gauge wire to an unhappy GFCI then 25 feet of 12 gauge extension cord. It's not fine when plugging into a public charger, most of which are 20A. So I use the slow setting in town Even worse, power factor on 'slow' drops to 60%, so still close to 15A draw at 120V

At home I almost always use the fast setting, which uses a cluster of 4 440 microFarad caps. I decided just for the fun of it to use some 1 Oz copper PCB and etched two stripes off with FeCL (I used packing tape as a resist ) So much easier than using punchboard!!!! OMG, I can't believe how much easier, smaller and tidier this was.So here are some pictures, 1st is my splice into OEM traction battery. Note that what looks like a short between + and - is actually just a little spacer i taped in to keep them apart:

Next picture is my silly punchboard charger prototype. The nonsense on the left is just a 120V to usb adapter to power a little 5V fan I had lying around. It didn't do much in the way of cooling but I wasnted to be sure my components would survive a fan failure without cooking. They did! 105 deg C rating helps on the caps.

This next picture shows my priorities! It's a hack to replace my dome lights and map lights in the front with LEDs. Total cost, about $8. There not quite bright enough,but the auto dimming still works and they cast an eerie bluish glow at night. Fun stuff - 12V seems so comfortable after 238V!!!

Here's the backside
And here's the LED modules, they are 12V waterproof with built in current limiting resistor. They come in a string, as many as you order, and cost about $2 each. If you look closely, you can see the little plugs i added were salvaged out of the guts of some of the traction batteries, i think they energized the contactors, now they have the much easier duty of energizing LEDs.

If anyone's interested I can add more pics and info. I'll add pictures of my newer charger soon. Here's a teaser of my first PCB etching Capacitor board thingamawhatsit. If you look closely, you can see the lighter stripes through the PCB where the copper was removed, so it's just three stripes of copper, width based on an online calculator I found for 1 Oz copper, 10A current, 170V each lane.
Here's the full charger's guts spread out on our new rug (Fancy! From Target!! Made somewhere else besides China!!!) You can see the big DPDT switch that toggles between slow/fast charging and the new punchboard housing the the "slow" 390microF caps and the two diodes. I swapped in a different diode which takes a more conventional heatsink and mica spacer. Heatsinks were a $1 at Radio Shack, diodes are Vishay, rated at 30A continuous and were like $2 each or something. I like them better than the stud mount, even the stud diodes were rated for 60A. Without fan cooling, the smaller 30A ones don't even get that hot.