30 day garage queen -> battery badly out of balance

[markwj]

I didn't drive my car for about 30 days. It sat 'dark', in standard mode, with SOC falling from around 84% to 27%. It really hasn't been driven much this year (weather exceptionally wet, and too many kids activities requiring >2seats).

So, yesterday I plugged it in and gave it a full charge. It stopped at 77%, and wouldn't top up greater than that.

I tried a range mode charge, and that didn't improve it.

ESS SOC screen is showing SOC LIM 77% MIN 77% MAX 97% (with a CAC of 145.55Ah). I tried a drive, and car behaves fine. After about 20km drive to work, ESS SOC showed LIM 71% MIN 71% MAX 92%.

So, my issue is that one (or more) bricks are low (77%) and one (or more) are high (97%); things are badly out of balance. It could be a bad brick, I guess.

Looking at VMSparser -b, I see the following limiting bricknumber changes (cutting out all the bits where the brick didn't change):

timestamp, brickahmin, brickahave, bricknumber
12/04/2012 17:53:45, 157.05, 158.31, 5
04/22/2013 22:07:22, 156.48, 157.85, 5
04/23/2013 22:07:22, 156.37, 157.68, 9
05/09/2013 16:30:43, 155.17, 156.54, 9
05/10/2013 16:30:43, 155.23, 156.60, 5
05/24/2013 16:47:55, 155.40, 156.71, 5
05/25/2013 16:47:56, 155.46, 156.88, 9
08/01/2013 22:21:06, 154.83, 156.48, 9
08/02/2013 22:21:06, 154.20, 155.86, 81
10/03/2013 22:44:45, 153.92, 155.74, 81
10/04/2013 22:44:45, 153.69, 155.57, 5
12/09/2013 15:13:06, 152.15, 154.03, 5
12/10/2013 15:13:07, 152.38, 154.26, 85
01/13/2014 15:55:46, 152.38, 154.26, 85
01/14/2014 15:55:46, 152.27, 154.09, 5
03/10/2014 21:02:41, 152.15, 153.98, 5
03/11/2014 21:02:41, 152.09, 153.92, 85
03/18/2014 21:02:42, 152.09, 153.92, 85
03/19/2014 21:02:42, 152.21, 154.03, 5
06/29/2014 00:15:46, 152.15, 153.92, 5
06/30/2014 00:15:47, 152.04, 153.86, 81
07/31/2014 00:26:15, 151.87, 153.58, 81
08/01/2014 00:26:16, 151.47, 153.41, 5
09/15/2014 00:26:23, 151.24, 153.06, 5
09/16/2014 00:26:22, 150.84, 152.95, 54
01/11/2016 21:20:14, 147.82, 150.27, 54
01/12/2016 21:20:14, 147.19, 149.41, 9
01/19/2016 21:20:15, 147.19, 149.41, 9
01/20/2016 21:20:16, 147.19, 149.41, 54
01/26/2016 22:23:31, 147.25, 149.47, 54
01/27/2016 22:23:31, 146.96, 149.19, 9
02/24/2016 22:23:33, 147.13, 149.36, 9
02/25/2016 22:23:33, 147.25, 149.53, 81
03/03/2016 22:23:34, 147.25, 149.59, 81
03/04/2016 22:23:34, 146.91, 149.24, 9
04/25/2016 11:02:16, 146.22, 148.45, 9
04/26/2016 11:02:17, 146.22, 148.39, 81
04/27/2016 11:02:17, 145.65, 147.82, 9
10/11/2016 11:51:15, 146.22, 148.50, 9
10/12/2016 11:51:16, 146.34, 148.62, 81
10/18/2016 14:23:19, 146.34, 148.62, 81
10/19/2016 14:23:20, 146.39, 148.73, 9
11/10/2016 14:23:25, 145.65, 147.88, 9
11/11/2016 14:23:26, 145.37, 147.70, 81
11/24/2016 14:23:29, 145.37, 147.76, 81
11/25/2016 14:23:30, 145.54, 147.99, 9
03/31/2017 14:13:20, 145.42, 147.82, 9
04/01/2017 14:13:19, 145.08, 147.76, 54
07/02/2017 17:51:54, 145.25, 147.82, 54

So, seems that #54 has been the limiting one all year.

A sorted ahr.log shows:

        brick   soc     v
    74    739196    32289
    73    916631    33366
    0    916631    33368
    54    916631    33368
    72    916631    33368
    55    922448    33406
    1    922448    33408
    63    928199    33448
    64    933899    33486
    80    933899    33486
...
    33    950686    33608
    34    950686    33608
    35    950686    33608
    43    950686    33608
    44    950686    33608
    53    950686    33608
    62    950686    33608
    8    950686    33608
    88    950686    33608
    98    950686    33608

So, that #74 doesn't look good (but can't really understand why #54 shows up as the limiting brick where #74 looks worse.

Lowest/Highest
32289/33608 = 0.9607
739196/950686 = 0.7775

I'm guessing the car is getting upset about the SOC guess (22%) vs the voltage (4%) difference. The SOC guess on #74 is scary; 20% off from it's nearest neighbour (#73).

So, I could either have a bad brick (#74), things are just out of balance, or the SOC calculation on that brick is wrong (sensor, etc).

Anybody got any good suggestions? Options are (a) drive it to give it some exercise, (b) charge it to as much as I can put in and leave it there to sit, or (c) drive it to Tesla. Given that the voltage looks ok, and it looks like a SOC estimate thing, perhaps the best is to drive it down to 20% or so then try to charge back up as much as possible.

 

[markwj]

Oh, and one other piece of information:

ESS V&T is showing Vave 4.08V, Vmin 3.93V (#75), Vmax 4.10V (#27). That is about 4% diff min/max, but average is close to max so definitely just one (or a few) bricks pulling it down.

 

[dhrivnak]

Did you check for balancing? I would range charge and let it set to bleed off the high bricks. This could take days.

But I am afraid you have a weak brick.

 

[supersnoop]

Did you check for balancing? I would range charge and let it set to bleed off the high bricks. This could take days.

But I am afraid you have a weak brick.

You just need to get it above 80% (maybe 82%) for it to balance. Letting it sit at a full range charge for more than a couple hours would be bad for the battery.

 

[wiztecy]

What scares me is your Vmin and Vavg spread. And with that I also am getting a bad feeling as you are with the notion there's a bad brick. My Roadster has been sitting longer than yours, over 1 year and driven a couple of times. I've been commuting and using my electric skateboard so I haven't taken the Roadster nor my other cars out nor had the need. I keep my SOC around 50-65%, I usually disable APS so its less of a vampire, but I can't do that at this moment due to the driver's side manual lock failing to function and unlock the car. I need to take some time to pull the panel, investigate, and fix. The car's been good however and barely loses any miles sitting, so its not too bad. It stays 65-85F normally here by the coast and the marine layer in the morning prevents high heat during the day.

Anyways, getting back. So my pack is way out of balance, I know that and its fine since I can re-balance it once I charge and drive regularly to allow it to balance. I looked at my numbers and I'm about right at 106 ideal miles standard, my numbers are:

CAC: 147.65
SOC LIM: 54% SOC MIN: 54% SOC MAX: 57%
Vmin: 3.82 Vavg: 3.84 Vmax:3.85

 

[markwj]

Did you check for balancing? I would range charge and let it set to bleed off the high bricks. This could take days.

Yep. All zeros.

You just need to get it above 80% (maybe 82%) for it to balance.

Yep. That was my concern as well.

What scares me is your Vmin and Vavg spread. And with that I also am getting a bad feeling as you are with the notion there's a bad brick.

Vmin to Vmax is "only" 4%. It is the SOC estimate for brick #74 that scares me.

I'm guessing 32289 is 3.2289v? With the next brick at 3.3366v, that seems a pretty large drop, but not too bad considering the entire spread of voltages. Balancing the rest of the pack down to that would be very hard (even if I could get balancing to even start).

I guess I'll try driving it for a few days, then pull logs, to see what happens. For the moment, I'm going to continue charging it each night (unless anybody has got any better ideas).

 

[markwj]

40+km drive yesterday, and charge, got me ... 77%. No change.

I'll try driving it down to 10% or so over the next couple of days, but this is seemingly a faulty brick and I doubt it will come back.

This whole 'balancing' thing does seem very kludgy. 98 out of 99 bricks are fine, but just 1 brings the whole thing down 20%.

 

[wiztecy]

Mark, have you tried pulling up the low brick by applying the lowest amps to your charger? Here in the US I'd put it on the 110v / 10amp charger and see if I could coax it back up.

And yes, that value in the ahr.log file does indeed appear to be the brick's voltage.

 

[dhrivnak]

It is really weird that that the high brick is not bleeding off. So since there is no balancing I fear you have a brick with a few bad cells in it. You may need to get in line for the 3.0 battery or resign yourself to 150 miles of range.

 

[hcsharp]

You just need to get it above 80% (maybe 82%) for it to balance. Letting it sit at a full range charge for more than a couple hours would be bad for the battery.

I would take this advice except I wouldn't worry about letting it sit after a range charge unless you did it every day for 2 weeks. It's better for the battery to have it in balance so doing a range charge will have an overall positive impact.

32289 is not 3.2289v. If your cells got to 3.23v it wouldn't be able to charge at all unless you take it apart and do it manually. The number represents volts but it's not a power of 10.

 

[markwj]

32289 is not 3.2289v. If your cells got to 3.23v it wouldn't be able to charge at all unless you take it apart and do it manually. The number represents volts but it's not a power of 10.

32289 in the logs was at around the same time as ESS V&T showing Vmin 3.93V.
Similarly, 33608 had a Vmax 4.10V.

I'm guessing 8192 is the factor? That's pretty close, and would give Vmin 3.9415V, and Vmax 4.1025V.

 

[markwj]

Mark, have you tried pulling up the low brick by applying the lowest amps to your charger? Here in the US I'd put it on the 110v / 10amp charger and see if I could coax it back up.

I'm going to drive it down to perhaps 20%, then I'll try your advise and do a long slow charge. That said, I've now reported it to Tesla so we'll see what they say.

 

[markwj]

Drove down to 40%, then slow (16A) charged back up to full. Charging stopped at 61%

SOC LIM 61% MIN 61% MAX 87%. Vmin 3.87v (#75-1), Vmax 4.07v, and Vave 4.06v.

Logs now showing bricks #0, #1, #54, #55, #63, #64, #72 and #73 all being pulled down now. #74 is still the worst.

Definitively a bad brick. Car goes back to Tesla next week. I suspect that it will be a while...

 

[ML Auto]

If it's out of warranty, I'm sure Tesla will want to sell you a whole new battery instead of just replacing one sheet. Before spending any money, I would definitely try to charge brick #75 externally thru the BMS connector to bring the voltage up enough to balance. That is a big drop in just 30 days, but if you can get it to balance and keep it plugged in, it might be enough to keep it on the road for awhile.

 

[markwj]

To 'end' this story: I got the car back from Tesla service today. They had it for about a month (I was away on holiday anyway, so wasn't in a hurry).

Diagnosis was faulty sheet replaced, plus what looks like a general clean up and service inside the battery itself.

SOC after last night's full standard mode charge was LIM 87%, MIN 85%, MAX 88%, which seems much more reasonable. Car showed 100% SOC, and got to 286km @48A charge (was 302km when vehicle was new). Battery was quite unbalanced (all 1's, except for one sheet with three 0's, and two sheets all 0's) so I left it plugged in and this morning I see nicely balanced and all 0's. CAC is now 145.70 (was 144.75 at the time of the issue). I guess it will take a while for things to settle down.

After two weeks of lousy weather, including two typhoons, the sun came out this morning and blue skies greeted my morning drive to work. I had a blast back in the driver's seat of my roadster. I've made a promise to the 'roadster gods' to drive it more.

Very happy with Tesla service, and the great job they did getting this back to me in such great shape. The car is coming up to 7 years old, 60,000km on the clock.

 

[ML Auto]

Glad to see they didn't just sell you another ESS. I guess with all the 3.0 upgrades they now have a bunch of spare parts to service the original ones, and are allowing the service centers to do so. How much did they charge you for that?

 

[Rotarypower]

Yep curious what the cost was?

 

[markwj]

Yep curious what the cost was?

I'd rather not say on an open forum; things get mis-quoted/mis-interpreted (especially by the press). Just glad to say that Tesla is still working hard to support these ageing Roadsters and keep them on the road.