[Harmon20]

Hello All,

I've been toying with the idea of a DIY EV for a while. It's been all words and ideas but no action for quite a while. I finally got motivated by George Karellas on his YouTube channel and series he's producing called SOUP. It's all about "get up and get to it" and him documenting as he does just that. If your motivation is waning or never was, you should totally watch that series.

Then I found this forum and it prompted me to start on the EV project I've always thought about. So I'm going to document the project here so maybe I'll keep motivated (or shamed) into keeping at it. I might even get some useful feedback once in a while, eh?

Here is the project - a 1966 VW Beetle patient and JAC Motors HFC7000AEV (rolls off the tongue, don't it?) organ donor.

The Beetle was acquired in the mid 90's. I drove it for a while, handed it off to one of my brothers who drove it a while and then handed it off to the next brother. He built modded the single carb 1600 to a dual carb 1914, welded some Pontiac Sunfire bucket seats to the pan, and welded a barn door hinge to the top of the passenger door to make some kind of janky gull wing that's held closed by a bungee from the window crank to the shifter. He gave it back to me and it has sat in a barn for about 20 years. Mostly fine but the motor doesn't turn.

The JAC was wrecked and abandoned around July this year. The state recovered it and when no one claimed it sent it to auction. The guy that bought it works on cars for fun to make interesting and functional hotrod-style vehicles. (He's got a project under way now putting a 1927 GMC truck on an early Silverado chassis/motor.) He didn't know what to do with this EV and sold it for the $600 he had in it.

I can't find any info about this car at all and no idea who sells them or parts. Frankly, I have no idea how it even got in the country. JAC apparently only distributes to countries outside the US. I can't imagine this car is DOT approved for street use. Given the lack of info and parts I'm going to have to reverse engineer this thing as I pull it apart to try to figure out what everything does.

I thought it was going to be a cheap-as-dirt mass produced piece of carp, but as I started taking it apart I grew to love it. The car is light weight and not put together very tightly so it would never hold up on the country roads I drive, but I think it would actually make a dandy city car, which is where an EV shines anyway. It seems thoughtfully designed and the workmanship is pretty decent. As I started getting into it I decided I'm going to transfer every system I can over to the Beetle - power windows, ABS power brakes, electric A/C, drive electrics, everything. Maybe even this doodad:

Not sure what it is, but looks useful. Tracker maybe?

The battery pack was a suitcase in the trunk under the deck. There is a steel tube at the back of the pack, passes through a box with wires to it, then disappears into the RR fender. A couple smaller tubes exit on the other side of the pack and go to a steel member under the seat pan. Cell cooling maybe? The charge port appears to be a GB20234.2-2011 standard.

Not going to find a charger like that laying around so I'm not sure what I'm going to do. I think I'm going to see about maybe pulling the bus bars/cables off the LiFePO4 cells and charge them individually or in small series banks with an off-the-shelf charger. I just want to see if they'll take a charge and if I can get this beast rolling under its own power before I start digging into this project too much. From what I can tell it's only been a few months since it was operating so I'm optimistic all will be well once I get a charge into this thing. I popped the top of the battery pack to see what it would take to get to the cells. As a former industrial/commercial electrician I was impressed with the neatness and layout of the battery control cabinet.

Moving up front, this is the engine bay:

I think this is the onboard charger:

I'm not sure what the big heat sink down by the LF wheel was:

Oooo, electric A/C. That is definitely going in the Beetle.

Finally, the drive motor. It appears to be an air cooled 3ph AC. I opened up the controller just to look around. Woof, those caps make me real nervous. I suddenly got a lot more careful with my tools and screws when that top was off.

I'll have another installment when I get the battery cells exposed and try to charge them up somehow. With any luck I'll get the JAC rolling, at least enought to let me know it's OK to proceed with the project as planned.

Harmon

 

[Harmon20]

Had just enough time last night to get the meter out and check the battery pack. The LFP cell battery is putting out voltage to the upper deck control section. My analog meter jumped to around 430v for a while then slowly settled down to 360v over the course of about 10 seconds. I checked calibration on the 12v accessory battery and it looked good. So it looks like the cells are probably fine and I just need to figure out a way to get a charge in them.

Kind of odd that the pack rating is 304v but the meter is reading 360-430v, no? Why would that be?

I hooked up the accessory battery to see what would happen. Everything seems to work fine (other than the fact the drive has no go juice) despite the wreck. No warning lights on the instrument cluster aside from the fuel gauge, power lock and windows work, shift selector works as expected. Contacts in the battery pulled in as expected when turning on. Looks like a good donor.

I've got a 56v lithium charger for cordless tools. I'm considering splitting the cell battery up into sections and seeing if I can charge them with the cordless tool lithium charger.

Harmon

 

[Harmon20]

I got the battery case open to see what was going on inside. I was kind of hoping things would be more exposed and easier to work on, but I guess covered and safer is good too.

The plastic is covering what looks like bricks about 3x4x10 or thereabouts. The packs are connected by means of some thin plates. One plate goes down on the plastic, the tabs of the bricks fold down over the plate, the next plate goes down, hex phillips screws go through it all and include one BMS monitor ring terminal and safety clips to prevent the screws from backing out.

The BMS is ordered in four sets with an extra cable going to each 10th monitored "bat".

There are therefore only 40 BMS monitor cable assemblies.

Even though there are only 40 BMS leads the physical dimensions and layout suggest there are 60 bricks under this top. The plastic terminal covers are stamped with alphanumeric info - "A3-1.5", "B4-1.5", etc - and 60*1.5 = 40 so...I'm not sure what that means. Apparently there is a parrallel-serial arrangement rather than the straight serial I'd assumed. I think. I'd really hoped not to pull the entire array apart, but it looks like that's what I need to do to fully understand what's going on.

Harmon

 

[Harmon20]

I'm a little more informed now, but no less confused. I opened up the cell assembly and it turns out there are 40 packs. The reason I couldn't figure out the arrangement is that there are actually two different kinds of cell packs: opposing tabs and adjacent tabs. They are placed in the assembly to make the serial connections short.

I'm confuse by the voltages here. The array is "rated" at 304v. With 40 packs that means each pack is rated at 7.6v.

Huh? Where does that number come from? A standard LiFePO4 cell produces 3-3.65v, minimum working level to max charge. That's no lfp cell I've ever heard of and nothing I can divide into 7.6v gets me in that range.

7.6v/2 = 3.8v
7.6v/3 = 2.533v

What's more, if I look at the spot welded tabs and looking through the bottom of the pack it appears that there are 5 cells in this brick.

The only battery I can find that is anywhere in the neighborhood of 7.6v is a SLA, and this doesn't appear to be anything like that.

So I'm not sure how to charge these right now. I've got to get a new meter and see what each of these packs is actually putting out.

Any ideas what the charge voltage should be for these cells?

Harmon

 

[brian_]

I assume that the JAC is the previous version of their current IEV5 (which is the HFC7001AEV). Auto-CHE.com has a page of information about a bunch of variants ("batches") with links to more detailed pages for each of them. I suppose you already had that...

The Wikipedia article for JAC says they assembled Mitsubishi vehicles at one time, so it seem reasonable to guess that this car is largely a copy of some Mitsubishi model... although this might not matter since those would be the parts which you will not be using.

 

[brian_]

Perhaps most importantly, there are just empty spaces where the text suggests that there are supposed to be images.

Update: I copied the URL for one of the images to a separate window, got a security warning from my browser (Chrome), told it to go ahead and risk problems, and now I see the images.

 

[brian_]

Maybe even this doodad:

Not sure what it is, but looks useful. Tracker maybe?

The marking appears to say "GPS GSM CAN"... so a combined GPS receiver and mobile data (the old GSM standard) transceiver, communicating with the rest of the car on the CAN network?

 

[brian_]

I think this is the onboard charger:

Presumably, based on the markings, that's the DC-to-DC converter, with 200+ VDC input.

 

[Harmon20]

Auto-CHE.com has a page of information about a bunch of variants ("batches") with links to more detailed pages for each of them. I suppose you already had that...

Yes, I had looked over that site I was, by far, the most informative of any site I'd found but it didn't make things any clearer for me at the time.

Looking back over all the JAC electric vehicles they all say 3.2v cells, but the clue that got me to re-thinking things was that some of the vehicles (not necessarily cars) had battery rattings noted as things like "320v/304v". That 320v being the voltage of 100 3.2v LiFePO4 batteries got me to thinking...I count 40 packs with 5 cells each = 200 cells, a multiple of 100. Rather than being straight series as I thought, I need to check polarity on the cells. Quit assuming based on sight and verify. It could be there is a combination of series and parallel, even though it appears at a glance to be pure series.

I already know that I have at least two tab configurations - opposing and adjacent - so I might also have two different polarity configurations. Two tabs on the left and the positive is on top, two tabs on the left and positive is on the bottom. That could allow for combo connection that appears to be a serial connection. I think.

If I actually have two parallel banks of 20 five-cell packs, or some other combo, then I can get to 320v. That's not the name plate rating, or even what my meter showed, but it is one of the figures listed on a handful of the JAC ev pages at auto-che.com

What it comes down to is I need to quit looking and assuming I know what's going on and put the meter to things to verify everything.

Harmon

 

[Harmon20]

Presumably, based on the markings, that's the DC-to-DC converter, with 200+ VDC input.

Yeah, my ignorance showing. I realized as I did more reading what I was looking at but never went back to change the post. I'll chalk it up to recording my enlightenment as I get into this project.

Harmon

 

[Harmon20]

Perhaps most importantly, there are just empty spaces where the text suggests that there are supposed to be images.

Update: I copied the URL for one of the images to a separate window, got a security warning from my browser (Chrome), told it to go ahead and risk problems, and now I see the images.

I put the image links in incorrectly, using https instead of http. I use a self-signed security certificate and forgot that the rest of the world's browsers might have a problem with the cert my browser was specifically told to trust. Should be fixed now. Thanks for the heads-up.

Harmon

 

[Harmon20]

OMG, I'm an idiot. Those bricks have to be 5 parallel cells because of the way they're tabbed.

Harmon

 

[Harmon20]

OK, I bought a new meter and confirmed everything. All 40 packs are five paralleled cells for 3.2V (currently) and the packs are all connected in series.

What I don't get is why the placard rating on the battery cabinet and the specifications listed at auto-che.com say the battery voltage is 304v.

40 * 3.2v = 128v
40 * 3.6v = 144v

I don't get where the 304v is coming from. Maybe that's the amount of voltage you have to push through it to charge?

At least I know any LiFePo4 charger will work, just link cells in series to the charger's capacity. Any recommendations?

Harmon

 

[brian_]

I put the image links in incorrectly, using https instead of http. I use a self-signed security certificate and forgot that the rest of the world's browsers might have a problem with the cert my browser was specifically told to trust. Should be fixed now. Thanks for the heads-up.

That matches what I saw.

 

[brian_]

OK, I bought a new meter and confirmed everything. All 40 packs are five paralleled cells for 3.2V (currently) and the packs are all connected in series.

What I don't get is why the placard rating on the battery cabinet and the specifications listed at auto-che.com say the battery voltage is 304v.

40 * 3.2v = 128v
40 * 3.6v = 144v

Each 5-cell pack must be either all in parallel or all in series; both a visual assessment and the pack voltage confirm that it is parallel.

That makes sense, but...

I don't get where the 304v is coming from. Maybe that's the amount of voltage you have to push through it to charge?

That sounds like an impossibly high charging voltage for that configuration, and an unreasonably low operating voltage for a modern production EV. I note that the input voltage range on the DC-to-DC converter is from 200 V to something... presumably higher, but you could check. And you already measured over 300 volts from the fully connected set, right?

I'm confused.

 

[Harmon20]

Yes, I'm confused as well. I'm thinking my meter is suspect so maybe the readings are not actually in the high 300v range. I've acquired a new meter but the pack was dismantled by the time I got it so I haven't confirmed the total series voltage yet.

I finally found a charger on AliExpress. I'll have to do some modification to it to work on US power, but at least I'll have a proper charging system for the car's systems.

I might even see if I can use the control system in the charger and convert it from a Mode 2 to a Mode 3 charger. I found a detailed spec sheet for the GB20234 standard. The spec in combination with a working device communicating according to the GB standards I might be able to reduce my charge times.

When the charger gets in I'll connect the car and see if I can get this thing going and use the new meter to sort out what's going on.


Josh

 

[Harmon20]

I obtained some CC-CV 6A chargers from Battery Space to try to get some power in the cells while I wait for the slow boat from China.

They don't seem to act as advertised, but maybe I don't know what I'm talking about. Set me straight if I'm confused.

They seem to be letting current and voltage float and never reach the rated 6A. They do limit voltage to 3.6v and shut off at [email protected], so maybe I don't know what is supposed to happen.

I thought a charge cycle was supposed to look like this:

(Yes, I know the numbers are off for a single LFP cell. It's just the curves I'm interested in.)


But instead I get this:

The voltage curves are 3.1v to 3.6v; I just left off the labels. The graph covers 25:20hr.

My data is HERE. (3kb CSV file)

It takes 24hr to charge when my 62Ah cells should be taking around 14-15hr @ 6A.

Do those curves look like they should for a CCCV charger?

Harmon

 

[rev0]

They seem to be letting current and voltage float and never reach the rated 6A. They do limit voltage to 3.6v and shut off at [email protected], so maybe I don't know what is supposed to happen.

I thought a charge cycle was supposed to look like this:


(Yes, I know the numbers are off for a single LFP cell. It's just the curves I'm interested in.)


But instead I get this:


The voltage curves are 3.1v to 3.6v; I just left off the labels. The graph covers 25:20hr.

It takes 24hr to charge when my 62Ah cells should be taking around 14-15hr @ 6A.

Do those curves look like they should for a CCCV charger?

Harmon

What are you using to measure the voltage? I highly doubt the current would have a smooth curve if the voltage were really abruptly changing like that. I have a feeling there's some rounding going on with the measurement tool.

 

[Harmon20]

What are you using to measure the voltage? I highly doubt the current would have a smooth curve if the voltage were really abruptly changing like that. I have a feeling there's some rounding going on with the measurement tool.

The "curves" look like abrupt changes because there is only a .5v difference between the high and low voltage during the entire 25hr period and the digital meter is only capable of .1v resolution, or 20% of the total change.

My flat data file (and therefore the graphs derived from the data) can't reflect the periods where the numbers on the voltage and amperage jumped between high and low before finally settling down. The transition from 3.1v to 3.2v wasn't an instantaneous jump - the 1 & 2 jumped back and forth for a good 10-20 seconds before finally settling on a 2. The abruptness in voltage changes is an artifact of representing digital measurements with analog visualization. The little blips at the transition points is an artifact of the line drawing algorithms in the charting utility.

When the voltage is plotted with the same scale as the amperage then it looks like a smooth line with a very slight incline. I blew up the scale of the voltage graph (perhaps wrongly) to exaggerate the changes so it was possible to correlate changes in voltage with changes in current.

Harmon

 

[Harmon20]

My car charger got here quicker than I expected. I'm only 10 cells in to charging the pack with individual cell chargers.

Opinions? Assemble the pack and let the on-board BMS do its thing or continue topping off the cells individually for the next month?

Also - anybody read Chinese?