Roadster Owner Based Study of Battery Pack Capacity Over Time

[lmore]

I earlier reported that the OVMS reports that the battery is warmer than the ambient temperature, usually 6-10 degrees C warmer (ambient temperature now is around 0 C). No problem has been found with the battery, Tesla Motors checked the log files and I got help here as well. Now I noticed that when I reach a quite low level of charge (below 30 %) and park the car without it being charged, the battery level drops real fast while the car is being parked. Yesterday night 22:00 the OVMS reported 28 %, 23:00 it reported 27 % and this morning it has dropped to 17 %. Now it has been stable on 17 % for two hours. It is not the first time I've seen this. Anyone experienced similar behavior?

I have not driven or charged the Roadster much the last weeks, only drive a few km each day and haven't fully charged it, perhaps that causes the battery pack not to be balanced?

 

[Timothy]

I just came across this thread tonight. Being a non-techy biology type, you guys are speaking a different language than I do. Anyhow, I just had my annual service at Tesla Menlo and asked them about the battery as I had noted the ideal standard charge had dropped from 184 to 176 since I bought the car.

I was told the best way to evaluate the battery is with the CAC (Calculated Amperage Capacity?) New cars are supposed to have 16 (who knows what units?) My car was manufactured 12/2010 and I bought it new 5/2011. At the car's service 12/2011 the CAC was 153. At the recent service 12/12 it was 147. I was told the 8% drop in 2 years since manufacture was 'normal.'

I drove from Menlo after the service to Santa Barbara (300 miles) and had about 25 ideal miles left on standard mode when I arrived. I charged in range mode--217 ideal miles--and unplugged the car and let it sit for a week to balance the battery. No change in the standard charge since. 175 ideal miles tonight after the standard charge completed.

So can any of you explain if this is a good way to evaluate the battery--and what it all means?

 

[JRP3]

For one the car needs to be plugged in to balance the battery. The BMS needs to bleed charge off the higher voltage cells while charging the lower voltage cells to bring them up. 8% drop in two years does sound fairly normal, maybe a bit on the high side.

 

[wiztecy]

For one the car needs to be plugged in to balance the battery. The BMS needs to bleed charge off the higher voltage cells while charging the lower voltage cells to bring them up.

I heard the opposite from Tesla. The car does not need to be plugged in to balance, but rather needs to be above 80% to kick off the balancing procedure. I verified this is true. Also I heard that in order to calibrate and bring up all the sheets to calibrate you have to drove on "one key turn" from above 80% SOC down to 30% SOC. If you turn the key off at any time the calibration will not occur.

As for the drop in range, I'd suggest seeing if you have a consistent lowest brick and possibly a sheet that's "lower" than the rest. You can run the "VMSparser -b" option
as well as the "DAY" option as Tom indicated a few threads up. If you see the same brick or the same set of bricks in a sheet always having the lowest Ah capacity that's most likely your issue. As for having Tesla fix that, that's another question. I don't know if 8% drop is for across the whole pack is normal or if 8% drop due to one bad sheet is also normal. I would say the latter should not be considered normal really.

 

[JRP3]

I heard the opposite from Tesla. The car does not need to be plugged in to balance, but rather needs to be above 80% to kick off the balancing procedure. I verified this is true. Also I heard that in order to calibrate and bring up all the sheets to calibrate you have to drove on "one key turn" from above 80% SOC down to 30% SOC. If you turn the key off at any time the calibration will not occur.

What mechanism is used to balance the pack when it's not plugged in?

 

[tomsax]

I'd like to encourage everyone to participate in a survey I'm conducting for Plug In America to help Roadster owners better understand what to expect for battery capacity over time and to better understand what "normal" is. Plug In America Tesla Roadster Survey

I've just added a question about CAC, so hopefully we can shed some light on how CAC changes over time and miles and also see how it correlates with ideal range after full standard and range mode charges.

Regarding extracting battery history from log files, in VMSParser, the -b option is for summarizing a bunch of the detailed data from many log files to participate in Rich's battery study (the topic of this thread). If you just want to look at the average and minimum brick amp-hours, you just need to run VMSParser on your most recent log file and use -pr DAY to dump out all of the daily records. It is believed that the brick average amp hours is the same as, or similar to, the CAC value.

Finally, I believe there are two types of battery pack balancing: during charging and after charging. After getting my battery pack replaced, I was told to do a full range mode charge and let the car sit for 2 or 3 weeks to let the modules balance. (I'm dubious that this does much, but I'm not a Tesla service tech.) I think there's also some module balancing that happens at the end of charging, at least sometimes.

 

[hcsharp]

I just came across this thread tonight. Being a non-techy biology type, you guys are speaking a different language than I do. Anyhow, I just had my annual service at Tesla Menlo and asked them about the battery as I had noted the ideal standard charge had dropped from 184 to 176 since I bought the car.

I was told the best way to evaluate the battery is with the CAC (Calculated Amperage Capacity?) New cars are supposed to have 16 (who knows what units?) My car was manufactured 12/2010 and I bought it new 5/2011. At the car's service 12/2011 the CAC was 153. At the recent service 12/12 it was 147. I was told the 8% drop in 2 years since manufacture was 'normal.'

I drove from Menlo after the service to Santa Barbara (300 miles) and had about 25 ideal miles left on standard mode when I arrived. I charged in range mode--217 ideal miles--and unplugged the car and let it sit for a week to balance the battery. No change in the standard charge since. 175 ideal miles tonight after the standard charge completed.

So can any of you explain if this is a good way to evaluate the battery--and what it all means?

It's true that CAC is a better measure of real capacity than post-charge ideal miles. It's also true that ideal miles displayed after charging has settled (10 min to 1 hr later) is usually fairly close (proportionally) to CAC. So if your car has gone from 184 ideal std miles when new to 176, then you've only lost about 5% which is good. I don't have the data that Tom Sax and Rich have but I've heard of new Roadsters with as much as 162 CAC and as low as 158. If you started with 160 CAC then you have lost 8%, but it's more likely that you started with about 157 or so if your car had 184 std mi when new, and 153 at the first annual. Personally I don't think you're losing range any faster than anybody else, but I wonder what Tesla did with your car for 5 months and what the CAC was when manufactured (12/2010). Keep in mind this is my .02 and probably worth about that much.

I heard the opposite from Tesla. The car does not need to be plugged in to balance, but rather needs to be above 80% to kick off the balancing procedure. I verified this is true. Also I heard that in order to calibrate and bring up all the sheets to calibrate you have to drove on "one key turn" from above 80% SOC down to 30% SOC. If you turn the key off at any time the calibration will not occur.
...

Based on my own experience, it will balance at the end of a charge whether plugged in or not, but there may be differences in how it's done. It doesn't sound like your (Timothy's) pack is out of balance. Sometimes it will recalculate the CAC without going below 30% SOC, and it doesn't have to be a one-key-turn session based on my own observations with my 2.5. Again, I'm not a Tesla engineer.

Like wiztecy said, I would recommend you pull your log file and have somebody look at it with the parser. I'm sure there's plenty of people who can do this in your So Cal owners' group or Green Speed. I'm curious what it reveals. Please post it here if you find out anything.

 

[tomsax]

What mechanism is used to balance the pack when it's not plugged in?

The car is always using power, especially at the top of a range mode charge when it keeps running the coolant pump and possibly the fans. It could be selectively drawing power from high-voltage bricks. I don't know enough to confirm that, but I know multiple owners who have been told to do a full range mode charge then let it sit for weeks to balance the pack.

 

[smorgasbord]

I highly recommend keeping the car plugged in whenever it's at full Range mode charge. That way the car won't hesitate to keep that battery as cool as possible. You don't want the car to sit at Range mode full and let the battery get warm.

 

[stopcrazypp]

What mechanism is used to balance the pack when it's not plugged in?

From the description it's pretty clear it's pretty much discharge only (I don't expect regen to do much). You have to reach a relatively low 30% SOC and I presume that is low enough that you don't have to bring up the lower voltage cells.

 

[wiztecy]

What mechanism is used to balance the pack when it's not plugged in?

Tesla's proprietary algorithm. My pack balances out within 10-15 mins typically no matter if I leave the charge socket plugged live in or have it disconnected. Tesla indicated to me this balancing algorithm gets fired off anytime the SOC is equal to or above 80% SOC and I have verified that. It is able to balance an "out of balance" pack faster when charged 100% range mode and I'm sure the Algorithm has to work harder and run longer to calculate and manage each brick and each sheets total capacity the closer the SOC is to 80.

As smorgasbord pointed out, I also agree that if plugged the thermal management has the opportunity to run if temps are reaching a critical temp. I don't need it in Santa Cruz but if you're in Vegas or Arizona I could see the benefit. But again I've never seen my pack temps rise just sitting there after it finished executing a charge cycle. The cooling system already managed what needed to be done where you'd have more control monitoring it yourself and re-executing a low amp charge cycle to bring the temps down again if needed.

As for the one-key / below 30% SOC advice I was given by the Tesla Service Manager, I tried it, didn't see any effect in my total CAC, it stayed the same. My pack is has always been a well balanced pack. I let it sit at 50%-60% SOC if its not going to be driven and when I do need to go out on a +30mile trip I do a full Std. mode charge then allow it to take 10-20 mins if needed to balance. Have all 0's in my balance grid, always have.

I have 148.6 to 149.8 Ah average capacity, charge to 181-182 Miles in Std. Mode and 232 miles in Range Mode. 14k on the clock.

Has anyone seen their Ah capacity rise, and if so what/how did it occur? Hotter Summer temps vs. Winter? Does cold diminish the Ah capacity?

 

[JRP3]

From the description it's pretty clear it's pretty much discharge only (I don't expect regen to do much). You have to reach a relatively low 30% SOC and I presume that is low enough that you don't have to bring up the lower voltage cells.

There seems as if there are two different suggestions, one is to charge the car and then let it sit, the other is to discharge it fairly deeply.

Tesla's proprietary algorithm.

An algorithm is an instruction set, not a mechanism. As far as I know there are two ways to balance, top or bottom. Top balance, which I assume Tesla uses, requires charge to be bled off of the highest voltage cells while lower voltage cells are kept charging to bring them up. This either happens while charging with resistors bleeding charge or it could happen by shuttling charge from higher voltage cells to lower voltage cells, which could happen even when not charging. Bottom balancing does the opposite, during a deep discharge it will balance all the cells so that they all hit the same voltage when near empty. That's how I manually balance my pack. Since top and bottom balancing are mutually exclusive I don't see how the one-key below 30% SOC makes any sense since I assume Tesla uses top balancing.

 

[dhrivnak]

I believe Tesla only top balances. So your batteries have to be above 80% for the voltage differences to show up to balance. You can balance faster with a range mode as it can more easily detect voltage differences.

Then I think the reason you occasionally want to do a deep discharge is so the software can better detect empty and full. If you only drive 10 miles a day and thus cycle your pack between 75% and 80% the software can loose track of the full range. Going to the bottom and top allows it to recalibrate.

Finally I have found I can vary my range (STD miles) by nearly 10 miles if I want. If I charge with 110V and check just after a charge it can be 183 miles. If I charge at 240V 30 amps at 4 am then when I leave for work at 7:30 I can have 178. If I charge say at 11 pm then my mileage may be down to 176 in the morning. If I charge at higher currents I can be in the low 170's. At one point I thought my mileage had dropped but most of the drop was chancing my charge time from 2 am to 11 pm.

So the time and rate of charge does have measurable effects.

 

[wiztecy]

There seems as if there are two different suggestions, one is to charge the car and then let it sit, the other is to discharge it fairly deeply.

As far as I know there are two ways to balance, top or bottom. Top balance, which I assume Tesla uses, requires charge to be bled off of the highest voltage cells while lower voltage cells are kept charging to bring them up. This either happens while charging with resistors bleeding charge or it could happen by shuttling charge from higher voltage cells to lower voltage cells, which could happen even when not charging. Bottom balancing does the opposite, during a deep discharge it will balance all the cells so that they all hit the same voltage when near empty. That's how I manually balance my pack. Since top and bottom balancing are mutually exclusive I don't see how the one-key below 30% SOC makes any sense since I assume Tesla uses top balancing.

From the behavior of my Std. Mode charge, it appears Tesla may use Top balancing? I don't know if this is seen across all Roadsters, a few, or mine. But now that you mentioned it... I tried three experiments to try and bring up my Ah capacity in sheet 4, its my weakest sheet with all the bricks lower than all the rest of the bricks in the pack. I think something went astray there.... After talking to Tesla about watching my range drop from 188 to 182 they suggested to swap out the BMB in sheet 4. I haven't seen it drop any further than 181/182, so possibly they addressed the degradation? I really don't know. At that time sheet 4's bricks were all around 148-149 Ah. The other sheets bricks were around 153, mostly at 155, and a few at 157 Ah. So in my experiment to recover sheet 4 I tried a couple of things:

Note that when I would charge in Std. Mode, the Pack would charge to 194 miles and then settle in at 182. This was seen before and a little after the BMB swap. 182 is 84% SOC for me.

1) Did a full Range mode charge after I was discharged down to 30%, multiple key turns. Pack charged up to 249 Miles then settled down to 232 miles. I allowed the pack to sit there bleeding off the charge for 3 days where the pack came down to 191 miles when put into Std Mode. I don't recall the range mode number.
-On my next Std mode charge, no change in std mode miles.

2) Did a slow 220V @ 12Amp charge from 30% to a full std. mode charge. Pack charged all the way to 212 ideal miles, and then after balancing came down to 183.
-Gained 1-2 miles and no significant Ah's, but only for that single charge. Multiple key turns. (Balancing took 20-30mins)

3) Did a 220V @ 30Amps, charged from 30% to full std mode. One single key turn from the previous full std. mode charge to 30% SOC. Pack charged up to 213 Ideal miles and then settled back in at 182 after balancing was completed (20-30mins). What is interesting is that I see the charger taking the pack all the way up to (213 miles) 94% SOC right when it shut off in Std. Mode. -No miles / Ah gained.

Possibly this high SOC is part of the Top balancing behavior you were describing to bleed off the voltages off the stronger cells while trying to bring up the weaker ones. I kicked something in/on with experiment 2 and three above but it didn't yield any positive results. Will see if my next few Std. Mode charges will be taken as high.

From talking to the Tesla Service Manager, I think when he described driving the SOC down to 30% this possibly could be Bottom balancing?

I know he mentioned two distinct balancing methods that occur, the one where it has to start charging below 30% SOC ( and would be nice to know if that single key turn theory is correct or not ) and the one above 80%.

Thanks for the clarification. Always intrigued, wanting to know more.

 

[JRP3]

Dhrivnak may be correct that the deeper discharge is helpful for the car to calculate actual ah capacity of the cells and/or better calibrate the range gauge. I can't imagine using both top and bottom balancing since they would be in conflict.

 

[wiztecy]

Dhrivnak may be correct that the deeper discharge is helpful for the car to calculate actual ah capacity of the cells and/or better calibrate the range gauge. I can't imagine using both top and bottom balancing since they would be in conflict.

Forgot one other experiment, #4 where I drove the SOC down to 9% SOC and then did a full range mode charge. No change in range. I have heard others who did get 5-8 miles back that way. Possibly people who's packs were out of balance in the 1st place unlike mine.

Wish Tesla could give an accredited run down of all these charging behaviors, to distinguish what's real and what's a myth.

 

[hcsharp]

Dhrivnak may be correct that the deeper discharge is helpful for the car to calculate actual ah capacity of the cells and/or better calibrate the range gauge. I can't imagine using both top and bottom balancing since they would be in conflict.

That's correct - it top balances and uses a deep discharge to recalibrate capacity. On my 2.5 it won't recalibrate after short trips (down to 70%+- SOC?) but it will recalculate your CAC (calculated A-hr capacity) if you go much lower than that. If you go down below 30% it does a better job.

The battery design and management is very sophisticated. The cells can be connected electronically in both series and parallel (not at the same time of course). They charge and discharge in series. From what I can tell there are 69 strings of 99 cells per string. When discharging, the 99 cells in a string are connected in series. Each of these strings is connected in parallel with the other strings. With 69 strings it's easier to manage the high currents involved.

When balancing, it connects the 69 cells in each brick in parallel to balance within the brick. Then the electronics monitor each brick and bleed them off until they accomplish two goals. First goal is to get them reasonably well balanced. But they also make some adjustments so that the lowest A-hr brick doesn't stay that way. Bleeding some bricks off more than others helps prevent failure runaway that would happen as soon as one brick lost one or two cells, taxing the remaining cells in the brick that much harder, causing that brick to fail sooner than the rest. Ultimately this will happen anyway, but this type of management slows the degradation way down.

Anybody who thinks the Roadster is just a giant battery with a fancy speed control and motor is very mistaken. It's a very sophisticated system and engineering accomplishment.

 

[wiztecy]

Finally I have found I can vary my range (STD miles) by nearly 10 miles if I want. If I charge with 110V and check just after a charge it can be 183 miles. If I charge at 240V 30 amps at 4 am then when I leave for work at 7:30 I can have 178. If I charge say at 11 pm then my mileage may be down to 176 in the morning. If I charge at higher currents I can be in the low 170's. At one point I thought my mileage had dropped but most of the drop was chancing my charge time from 2 am to 11 pm.

So the time and rate of charge does have measurable effects.

This is interesting, my Roadster doesn't lose that many miles sitting overnight, even when temps drop from 70 down to 30 overnight. If I charge to 182, at most I'll loose 1-2 miles overnight. It doesn't consume that many miles sitting idle. I may try a 110V charge one day to see if I can get my miles past 182/183. With my experiments, charging at 220v@12amps vs 30amps had very little effect. It appears others are seeing larger spreads like yours. I'm also curious if there's a relation with this spread and also if the how well the battery pack is doing balanced-wise.

 

[JRP3]

The battery design and management is very sophisticated. The cells can be connected electronically in both series and parallel (not at the same time of course).

This puzzles me. To do so the pack would have to physically disconnect all the high current series connections. If there are 69 strings that's 69 high current connections that have to be broken with large contactors, and then another set of connections in parallel have to be made, though at lower current levels for balancing. When the series connections are open pack voltage would drop to the level of a single string and the car could not be driven. I would think it more likely that Tesla simply bleeds off charge with resistors, or uses charge shuttling from high to low strings.

 

[dhrivnak]

This is interesting, my Roadster doesn't lose that many miles sitting overnight, even when temps drop from 70 down to 30 overnight. If I charge to 182, at most I'll loose 1-2 miles overnight. It doesn't consume that many miles sitting idle. I may try a 110V charge one day to see if I can get my miles past 182/183. With my experiments, charging at 220v@12amps vs 30amps had very little effect. It appears others are seeing larger spreads like yours. I'm also curious if there's a relation with this spread and also if the how well the battery pack is doing balanced-wise.

I do not think I lose more than 3 miles/day in "vampire load". Most of the other variation comes from different charging rates. Try charging at 120V and I suspect you will see a boost in range of a few miles. We are talking about 5% of the capacity so they are not large effects.