Still Going

A couple of folk have asked me about the Vogue since there hasn't been any blog entries for a while. Yes it's still going - and still waiting for the front end components to be fitted. The batteries have suffered a little from calendar life degradation but all 384 cells appear to be tracking each other pretty well. Range is down to around 55km. I'm always looking out for new cells to fit in the same place as the headways and if I can get the ones I like it would give me about 240km range. The pack weight would go from 140kg to just under 200kg - not bad. Hard to justify though as the range still gets me where I want to go and back.

Warmer Weather - better Range

I think I have mentioned before but with the warmer weather my range has increased by around 10km.
The Vogue measures remaining range by figuring how far you have travelled on what fraction of the battery pack's capacity - then calculating how far you can go based on the remaining capacity. All this is done assuming only 80% of the pack capacity is available. It does this 3 times a second so the remaining range moves around a lot based on how you are driving. This technique would not work well in a commercial EV where people got an initial reading of a 33 km range after travelling 600 meters to the top of their street's hill (as I do almost every morning) - then the range esimate changed with every take off (as it does).

At just under the 10 km mark on my way to work I get a very optimistic indication. I only get a few seconds to take this picture before the lights go green and I missed a great picture few days ago in this spot where it was 91 km remaining range - a total of just over 100 km with 80% DOD.

It doesn't last long though. About 2 km later, after a substantial climb at 70 to 80 km/h, it is usually down to 72 km remaining. By the time I get to work it's around 77 km remaining (Range) after 15 km travelled (TRP). The trip home is always worse and I'm arriving home lately with 31 km travelled and about 54 km remaining range. The Vogue has a 12 kw/hr pack (20 AH at 600 VDC) and all dashboard calculations are performed assuming only 9.6 kw/hr (actually 16 AH) is available.

A note or two about comments - I really like getting them. I have to moderate what comments get published - currently I publish everything. You can leave me comments privately or publicly though. If you like, you can leave a comment and just mention that you don't want it published, "don't publish this John", and I won't.

To leave a comment just click on the "No comments" (or "'n' comments") at the bottom of any post.
Next line down from this one...

Pack #5 Fixed and ready for Car.

I temporarily plugged the relevant parts of the BMS into the pack last night and charged it while the pack was still open and I could measure the cell voltages.
The cells were so perfectly balanced that the charger current died off at 59VDC with around 3.68V across each cell pair - so no balancing required.
I did  a quick test of the BMS monitoring (my opto-coupler mods) and it checked out.
The BMS is fine - no damage.
After now leaving the pack open for 24 hours or more to thoroughly dry out, I re-assembled the pack last night but it was too wet and rainy to bother putting it back in the car.

I'll put it back in tonight or tomorrow morning.

Astute readers my notice that this pack is identified as Pack #10. That's because, way back when I had problems with controller switching coupling from the motor leads into pack #10 in the engine bay, I swapped packs #5 and #10 - just in case.
It remains an annoying facit of always having to note which pack #5 or #10 I am dealing with in my pack notes.


If I ever have to pull the faraday cage off Pack #5 (engine bay - see), I'll shield Pack #10 (the one shown above) more elegantly and swap the packs back to their rightful positions.

Pack problem caused by Water

As soon as I got home from work last night I yanked Pack #5 from the boot (trunk), took it inside and pulled it apart. There was a little bit of water on it and a few drops on the tray but I kind of ignored them until I got the pack apart - a little water dribbled out but I still wasn't sure that it didn't get on the pack when I was removing it (boot lid was very wet). I checked all cells and all measured 3.27V per cell - no variation. So the cells were OK. I proceeded to remove the bottom of the pack enclosure - now it was obvious.

Most of the bottom cover was covered in a thin film of water with a patch about the size of a matchbox (50mm x 40mm) that had green tinged foam/slime around it - that's where the BMS's daughter board was sitting in the pool of water!
BMS daugher board circled in red. (Daughter board appears to be so they can use the same BMS for 12 or 16 cells systems. They add four more cell handlers on the little board.)

Unfortunately my sense of WRONGness was so great I instantly wiped it all up - then remembered that I should have taken a picture - oh well.

I removed the BMS board and cleaned it with genclean (PCB cleaning stuff) then resoldered any suspect looking pins (a couple of Molex pins looked like they had corroded slightly), then gave the board a coat of clear laquer.


OK - so where did the water come from.
Some while back I crowed about how the Vogue was fine charging in the rain because the charger inlet (located IMMEDIATELY above this water soaked pack) had a "moat" that went to a drain hose. Well the 48 year old drain hose had cracked even more than when I re-installed it and was leaking every time I charged in the rain. Also, any water that got under the unsealed cover while driving went down the drain hole/hose.

The charger inlet, "moat" and drain hole in question.

I cut the bad end off the hose and ran the hose in a straighter line (no fuel tank to go around).
View inside the boot.

This old picture shows where Pack #5 is mounted with respect to the charger inlet. It's the one sitting alone on the left under the contactor/fuse box.

I haven't tested the BMS board yet to see if it still works - fingers crossed.
I'll try to get the pack back together tonight - it's still in the state of the first picture in this post.

When I re-install the pack (assuming the BMS is OK) I will put a temporary cover over the pack and check it regularly for a while.
I am still amazed that the water had got into the pack enclosure so easily. The three packs under the front of the car (engine bay) have been sealed with waterproof gaffa tape but now I wonder about them as well. At least their orientation would not have the BMS board at the bottom of the pack like this one.

...and yes, I am very embarrassed.

Battery Pack Problem

I didn't drive the Vogue today. When I started it the battery monitoring system instantly sounded an alarm. A quick check showed Pack #5 in the boot (trunk) had it's indicator light off. I re-connected the charging lead but it's internal BMS switched off the charger path straight away. I took my daughter's car to work.
I'll check it out when I get home.
I have a sneaky suspicioun that I just had my first cell failure in active use (later edit - wrong - they are fine). Unfortunately my packs will tend to kill two cells at a time due the the two-parallel nature of the system.
At may be that my 65km on Friday night has pushed some week cells into failure (nope - see next post). I did have two cells fail while in storage - a couple of years ago now (and to date they are the ONLY two).
I'll post the results...

Pack #2 fixed

I built a new BMS monitor board, cleaned the underside with an aerosol cleaner for PCBs (at work) and sprayed it with a copper and brass laquer I had at home. Pack #2 is reassembled and back in the car.

Pack #2 BMS Monitor board

Well I'm not sure why the LED blows up but the board definitely has some issues. The low voltage side of the opto-coupled circuit (left side) appears fine but the higher voltage side has clear indications of electrolysis or corrosion of some kind.

When the car is off there is 50 VDC across two of the tracks that run side by side (third and fourth tracks up on the right hand side of the big gap).
Perhaps they had moisture on them when I heatshrunk the board.
Anyway - I have built a brand new one and I will clean and apply a conformal coating to it before heatshinking it.

It looks like I'll be looking at failures down the track but over the next couple of years I should have the packs out and I'll build replacements when required.

Pack #2 LED cooked again

Another minor issue has crept into the Vogue to-do list.
In this post and again in this post I note that the blue LED on Pack #2 in the boot had cooked. When I disassembled the pack and checked the LED was a dead short - both times.

Well it's happened again. It's probably no coincidence that the optically isolated connection to the five packs in the boot feed to the controller input that I blew up about a year ago (this post) - and at roughly the same time as the LED expired the second time. (The controller input fault was a surface mount 1nF ceramic capacitor that started believing that it was part 9V Zener diode - weird.)

So I will build another BMS Monitor Board and totally change out the one that appears to be blowing LEDs (and controller inputs).
I'll try to get that done this week after work.
I'll carefully visually inspect and test the one I remove and post the results.

Boot (Trunk) Covers and signs.

The boot covers and warning signs installed. The inertia cut-out switch is now also installed and wired in, as is the mains inlet plug. The inertia switch is mounted on the Vogue bodywork just behind the fuel tank cover (the black cover on the left). It's a bit difficult to actually see the polycarbonate panel in front of the four battery packs - the warning sign just kind of hangs in the air.

A view from the passenger side. There is plenty of room for shopping bags. The usable area of the boot hasn't really changed. The elevated section where the batteries are now was never much use.

 

Boot (Trunk) Partition

In order to satisfy regulations regarding not being able to touch the cables carrying EV traction battery voltages, I will need a plastic partition in the boot. The old fuel tank cover handles the switchgear on the left of the boot.

I will be using 3mm polycarbonate and it will be mounted just this side of the four packs along the rear firewall. It will be a bit fiddly as it has to be cut around the boot hinges.
I thought I'd get some quotes on cutting and drilling polycarbonate before I bought a piece at Bunnings (hardware store) and did it myself. The second quote was only just a bit more expensive than buying an uncut sheet. So, I dropped this drawing

into

Plastics for Industry P/L

13/189 Cheltenham Rd

Keysborough VIC 3173

today. Ashton was a pretty easy guy to deal with.

It should be ready early next week so I can instal it over the Christmas break - assuming I got all the measurements correct.

Oh, and my tailshaft is ready for pick-up.

Pack Balancing and Charger Modifications

With all the manual(ish) pack balancing I do, I had already designed a small circuit that would automatically deliver 3A until the pack's BMS cut out, then dial back to 180mA - the balancing current for my packs. I wasn't looking forward to making these, nor fitting them.
With that in mind I thought I would try again to find a schematic of the Kingpan chargers that I am using. I found a thread on Endless-sphere that (somewhat indirectly - I can't figure out now how I found it) pointed to a site in Europe where a guy (thanks Albert) had reversed engineered some other brand of charger based on a TL494 chip (my favourite PWM controller).

I grabbed the circuit, pulled the lid off one of my spare chargers and - guess what - yes they are practically identical. There are a few resistor values that are different but just about everything is the same. It was relatively easy to devise a small daughter board that used the Op-amp that was driving the transition of a bi-colour LED on the charger to trigger the changer to cut it's current from 3A down to 180mA. The challenge was to be able to do the modifications without having to remove the main PCB from the case - with all the complications of re-seating power devices on heatsinks. I found that the best was was to remove one resistor and solder on four wires to a daughter board. The daughter board is circled in red. The heatshrink gets cut down, shrunk, then glued to the nearest strong component. I'll improve that toroid's mounting too - while I have the Sikaflex out.

Now this charger has built-in balancing tailored to my pack's BMSs.
I don't have to do all the chargers at once, just as I get around to them.

Battery Packs all Good

I have been a bit slack with the blog this week. Last weekend, following dis-assembling pack #2, I  charged all packs as per normal then used my single balancing charger to "top off" each pack. I got the job done over the weekend - way better than the 2 weeks using two chargers last time. Some packs balanced up within 3 hours, maximum was 10 hours.
All the packs took 3 Amps for over 3 hours on the full charge so overall the cell balance is looking pretty good now.

I also got to the bottom of the AH difference from Vogue display to actual.
I stupidly was assuming that the current sensor put out 2.5 volts for no current (correct), then 5V for +100A and 0V for -100A - WRONG. Upon careful reading of the specs. I realised the the sensitivity spec. 20mv/A means exactly that. 20mv/A then 200mV/10A then 2000mV/100A or +4.5V for +100A and 0.5V for -100A. Curiously when we set up a 12A current flow to calibrate the sensor, it was only 17% out - not 25%. Anyway it's now calibrated to within 200mA - the resolution of the Digital to Analogue input on the dashboard PIC micro processor.
That also means that we really did pull 14.3AH from the pack which also means that all cell pairs are pretty much OK.
So the total range for 16AH that was projected as 80km is closer to 70km under those driving conditions. That's still good - old tyres, no front end alignment yet and it was a lot of stop/start driving (137Wh/km).

Another Battery Pack Check

Following the drive described in my previous post I need to iron out why the packs pulled up a little "short".
I pulled subpack #2 out of the boot, removed the lid,  and measured all the cell pairs. Fifteen of the sixteen cell pairs were between 3.19 and 3.21 VDC. One pair was 2.98 VDC.
I removed those two cells and marked them 11A and 11B (eleventh up from the -ve end of the pack) (the cells are 10AH).
I started by charging them individually at exactly 5A and noting the time and current every now and then. Both cells took 9.45AH - they were obviously quite discharged.

I then discharged the cells individually with a 0.8Ohm load down to 2.9VDC. I got just over 9.5AH. The white LED goes out at about 2.5 volts so lets me use a quick glance while I'm working to help prevent over-discharging the cell.


Obviously both cells are fine.

I recharged the cells to 4AH and placed them back in the pack.
I then charged the pack at 3A. The charger dropped out at 3 hours and 25 minutes. That was 10.25AH. The cells I had removed (#11) were the cause of the charge stopping. They had reached 3.9V and stopped the charge.
All remaining cell pairs were at 3.33 to 3.34VDC. I left the charge current at 170mA overnight - for 7 hours (1.2AH).
Next morning, I loaded cell pair #11 with 0.8 Ohms and kicked the charger back to 3A. It charged for 40 minutes before dropping back from another cell pair getting to 3.9VDC. Most cells pairs were now over 3.5VDC.

That's a total of 13.45AH for most of the cells. The Vogue had indicated 12.19AH which appears to be about 10% out. I will check the current sensor in the Vogue with a measured current, as that is something I have just "believed" - up until now.
I have also heavily integrated (smoothed) the current sensor value in both hardware and software. Perhaps I should not - and just smooth the display value.

I Have Regenerative Braking

Some help from AEVA members and regen. braking is now working (emoticon with red face). The Vogue is still on axle stands but a quick test this morning showed power going back into the batteries and the rear wheels coming to a pretty quick stop when I lifted my foot off the accelerator pedal - no thump either. I had to allow torque to go negative for regen. I went to a lot of trouble to NOT let that happen - picturing a ride on lawnmover with forward/reverse control on the one pedal.

I have kind of come to the conclusion that the differential probably has no oil in it. I don't mean low - I mean none. The car sat in the back yard for about 6 years, then 3 years getting restored and electrofied - and I know the diff leaks, and not just the seals, I think it leaks out a faulty seal in the drain plug as well....
The rumble of gears running in a hollow diff housing is unmistakable (this only became obvious when running the car in the garage on axle stands). Fortunately I haven't driven the car very far in the last decade - since I checked the diff oil. I won't be running it again until I fill it. I have to pull the rear axles at some stage in the next year so I won't go overboard at this stage - I'll just fill it.

Door Liners ready for fabric

I finished punching out the holes in the door liners and we laid one of the front and one of the rear door liners out to see how we would cover them. The rag stealer was nosing around.

Here is one without the black silhouette in the way.

The orginal liners were covered in red vinyl with vulcanizing in several strips under the armrest area.

We decided to do two or three strips of silver starting where the vulcanizing starts at the bottom but not going quite to the armrest (the two smaller holes through the vulcanized area). When we laid fabric on the liner it was apparent that, even though the fabric has a foam backing, that the backing was not thick enough - it felt cheap. I'll buy some 3mm foam backing to go behind the fabric.

At some point in the past two weeks I made the decision to use the constant current balancers on all the battery packs in the car, not just the ones that showed errors on the drive - I just didn't mention it here. I have been moving the balancer chargers from pack to pack as each pack got to 59 VDC (3.69 x 16). Every pack has taken more than two days to reach internal cell balance - thats about 9AH (for a 20AH cell pair). The packs were certainly not correctly charged. The last two battery packs are on the 180mA constant current balancer chargers now - nearly there!

Pack #9 Update (balancing the battery packs)

There's those who say that you can't tell a LiFePO4 cell's  State Of Charge (SOC) from the cell voltage.
They're right!

In my previous post I mentioned that cells around 3.3V appeared to only be about 2 to 4AH shy of fully charged.
Well in Pack #9, we have so far put 11AH into a cell that initially measured at 3.28V and it's only just up to 3.5V - so it's nearly charged. The critical voltage at 20 degrees C for reasonably new LiFePO4 Headway cells seems to be between 3.2 and 3.4 V.

I had ideas of logging pack voltage and current and coming up with a SOC indicator that didn't have to measure AH in and out of the pack. The idea was to compensate the pack voltage with the current being drawn from it and assume a particular SOC from the compensated voltage. I think it may be harder than that!


William is at home today and I'm at work, so I am getting cell voltage and current updates from him.

Battery Pack Balancing Update

The pack balancing is going well, albeit slowly. I have two 180mA current source chargers running all the time balancing the cells in packs that didn't show alerts during the drive. I move them to the next pack when the pack under charge gets to 59 VDC (3.69 * 16). As well as that, I am pulling each of the 5 alerting packs (minus one that was too difficult to remove and responded to 3.5 days of 180mA charging) out of the car, lifting the lid, and charging the low cells with an Individual Cell Charger (ICC).

The ICC is a small PC power supply that has an adjustable 5V, 5A output. It doesn't adjust all the way down to 3.7V so I have two 3A diodes in series and have it set to 5.2VDC. The diodes drop 1.6V at 3A and 1.5V at 180mA so it's reasonably safe to leave unattended. When I pull a pack from the car, I start by measuring all cell pairs then use the ICC on the low cells. I generally find that there are about 12 cell pairs over 3.6V and 4 at 3.3 to 3.4. The 3.3V seem to need about 4 AH put into them. Cells that have been 3.2V have needed up to 17AH. Considering the cell pairs are 20AH, that's pretty discharged.

So far all but one of the packs (the one that had a cell that took 17AH) I have pulled out have had the 180mA treatment before I decided to remove them, but pack #9 which get removed from the car tonight has only had the post-drive charge to 58.4V, (the 3A charger cut back after only an hour) so it will be interesting to see the SOC of the cells.
I only have one of the alerting packs left to pull out of the car, pack #9.

Battery Balancing and Steering column cowl

It's been a bit of a slow week with one battery pack opened up for individual cell balancing and others with a 0.01 C controversial balancing charge.
Pack opened for examination and individual cell charging.

I have also been experimenting with finishes for the steering wheel centre and steering cowl. I tried a leather dye but it would not soak in, then vinyl paint which peeled off in a very strange way. I have fallen back to prime and paint.
Primed.

 The top of the cowl.

Balancing my Battery Packs

I posted some thoughts on the AEVA forum so I could get some feedback from like-minded souls about my battery pack issues. It may be of interest to some who read this blog so here is the link. It's about the eighth post on the page.
http://forums.aeva.asn.au/forums/forum_posts.asp?TID=2707

Some notes on the Second drive

The drive on Sunday showed up some minor problems.

1. There was a "thump" from the driveline when I eased up on the accelerator too fast - even as very slow speeds. This was easy to avoid but it's a bug nevertheless.
2. Regen braking didn't work at all. Pretty clever for me to muck up something that is hard to prevent.
3. We only travelled about 5km before I got a battery low alert. Upon returning home, 5 out of the 12 packs were showing a low voltage event (blue LED off).

1.
The thump happens when the accelerator transitions through the point where I am requesting full speed (variable torque), to the point where I am requesting zero speed - at about 10% pedal depression.
I left a 100mS ramp up/ramp down in the speed section of the controller, thinking it did not effect torque control - but it may.
If that doesn't help, the easy fix is to implement a ramp function on the Analog input (a function block in the controller), and limit the ramp down speed to that which doesn't cause the thump - but I'd rather identify and take the correct action. I'll put it up on axle stands and have a fiddle.

2.
Weird. I will connect the laptop and monitor the internal controller signals and trace through while on the axle stands. This is pretty easy to do as the Lenze software lets you tap into any analog or digital code of function block output and display as a dial or digtal readout or both. Up to 8 at once.

3.
Seven packs were at 52.5V and 5 (with LEDs off) were below 51 V. After a full charge, I pulled the lowest voltage one (as measured after the fault event) from the car and have so far put 12AH into one of the 20AH cell pairs. There were 4 cells pairs under 3.25V - the rest were 3.6V or higher - keep in mind this pack just got fully charged!
They are obviously way out of balance. To use an Americanism - "my bad". I thought since all chargers had cut out and all packs were at 58V they were balanced - not at all! The remaining 11 packs are on charge for the 4 days it will take for 180mA to charge all cells. It took me a whole night just to charge one cell pair in one pack so that's not reasonable to do for the 5 packs. I will change the the pack I have out to the same "trickle" or CV charge and monitor it closely to ensure I'm kind of OK. I don't like doing this but it's untenable to pull every pack.