Battery health: What is the ideal CAC ?

[smorgasbord]

I guess the background is that I understand how a low CAC can suggest a pack is worse than it really is (unbalanced, bad estimate, whatever), but can a CAC ever be higher than the pack condition really is?

This is a really good question. Without knowing the details of how CAC is calculated, or without knowing how to actually measure pack health, I don't think we can give a definitive answer. But, my guess would be "yes." If you do only shallow discharge/charge cycles for a period of time, the software probably doesn't know how the cells will behave when you eventually do perform a deeper discharge. So, it could overestimate as well as underestimate.

Could it be that Tesla designed the system to always underestimate? Maybe, but there certainly are failure modes that would break any such algorithm.

So, my question is: how do I bring my CAC back up? Tell me, and I'll try it for the next few months to see if we can see any difference.

Follow wiztecy's approach from post #15:
As for raising your CAC, its really experimental. Note that I have not range mode charged in a year and won't do it unless I really really need it. I know that 90-100% SOC and 0-15% SOC can be stress on the pack. I also try to keep my SOC sitting at around 55-65% over the weekends or when the Roadster is not in use. And I try not to punch the accelerator too hard when I'm below 50% SOC.

I would add that avoiding high battery temps is key, so running an OVMS COOLDOWN whenever the battery temp is above 35 degrees helps. And as people like slcasner have reported, one hot track day can ding your CAC permanently. In my case, leaving my car at Tesla service for a few weeks in the summer dinged my CAC permanently - they left it sitting outside in the hot sun without plugging it in.

 

[markwj]

I doubt that the CAC algorithm includes things like time sitting at temperature. It is Calculated Amp Hour (reportedly). That would be based on measurements of the capacity of the battery itself, perhaps averaged over time, and perhaps based on delta changes in charging and discharging cycles. That is what the ESS SOC screen on the VDS shows, anyway - all related to changes in current over time and no mention of temperature.

I can see how short cycles vs long cycles would make a difference. A long discharge then charge cycle would give a better estimate of capacity than a short one (or bunch of short ones).

I think I'm going to try a range mode charge, then long discharge cycle, as I haven't done that for several years. I'll let you know if there is any appreciable difference in CAC.

 

[djp]

I think I'm going to try a range mode charge, then long discharge cycle, as I haven't done that for several years. I'll let you know if there is any appreciable difference in CAC.

It would be interesting to double check the CAC calculation when you do the long discharge cycle. Make a note of the SOC LIM before and after the cycle, along with the #kWh used. With those three numbers you'll be able to estimate the total kWh capacity of your pack. The wider the SOC range the better. The #kWh is reset when you turn the key off, so doing the discharge on a single key turn would be ideal.

 

[wiztecy]

It would be interesting to double check the CAC calculation when you do the long discharge cycle. Make a note of the SOC LIM before and after the cycle, along with the #kWh used. With those three numbers you'll be able to estimate the total kWh capacity of your pack. The wider the SOC range the better. The #kWh is reset when you turn the key off, so doing the discharge on a single key turn would be ideal.

I'd also grab the ahr.log after the Range mode charge and also grab another snapshot of it when you've performed your discharge.

 

[m0rph]

It would be interesting to double check the CAC calculation when you do the long discharge cycle. Make a note of the SOC LIM before and after the cycle, along with the #kWh used. With those three numbers you'll be able to estimate the total kWh capacity of your pack. The wider the SOC range the better. The #kWh is reset when you turn the key off, so doing the discharge on a single key turn would be ideal.

I have noticed that 64A displayed on the VDS, is only 61.8 - 62.0A in real life... So my guess is that the displayed total kWh is also a bit higher than in reality.

 

[slcasner]

I have noticed that 64A displayed on the VDS, is only 61.8 - 62.0A in real life... So my guess is that the displayed total kWh is also a bit higher than in reality.

I have a dedicated kWh meter on the circuit used for charging the Roadster. In 2009 I complained to JB that the total kWh for a charge as shown on the VDS was consistently low by a ratio of real/VDS = 1.13 or so. Some part of the power was not being counted. They updated the firmware to fix that. Now I find that the kWh displayed on the VDS is typically less than 0.5 kWh, but it would be nice to have another digit in that display.

 

[smorgasbord]

I doubt that the CAC algorithm includes things like time sitting at temperature.

I agree, but having a hot battery does ding your CAC, as some have seen, including myself. So, there's some at least indirect effect.

 

[hcsharp]

I have noticed that 64A displayed on the VDS, is only 61.8 - 62.0A in real life... So my guess is that the displayed total kWh is also a bit higher than in reality.

How did you measure the 62A? Most of the times that I've measured the amps I've found it fluctuates up and down, although the average was always very close to what was displayed on the VDS.

 

[hcsharp]

I agree, but having a hot battery does ding your CAC, as some have seen, including myself. So, there's some at least indirect effect.

To clarify, I think you mean it doesn't ding your CAC directly because it's not included in the CAC algo. Time at high temp does, however, ding your battery health which in turn dings your CAC.

Most everyone experiences a decline in CAC in the winter, and a rise throughout the summer. I'm curious how the CAC algo takes pack temps into account. I assume it adjusts for pack temp when making the calculation, but perhaps not enough which results in seasonal variances? For example, a pack will have a lower real-life Ah capacity when it's colder, but the CAC doesn't change significantly on a daily basis with large pack temp swings. If I wake up to a 9C pack it won't have a significantly different CAC than the day before when I woke up to a 23C pack, but the 9C pack won't have as much capacity.

 

[tonybelding]

CAC is the Calculated Amp-hour Capacity, not the range. It's available in the log file - you'll need to download the log to a USB stick and run it through a parser to retrieve the CAC.

OK, I give up. Where-and-how does one find this information in the logs?

 

[djp]

OK, I give up. Where-and-how does one find this information in the logs?

Try using Tom Saxton's log parser tool:

VMS Log Parser for Tesla Roadster
Log Parsing tool available - Page 6

One section of the file has records labelled "timestamp, brickahmin, brickahave, bricknumber". The brickahave column is your daily CAC.

I did a long drive this afternoon to check my pack's capacity. Not exactly scientific, but the numbers beat my expectations:

Starting SOC LIM: 84%
Ending SOC LIM: 32%
#kWh used: 30.54kWh

Implied total capacity of the pack: 58.7kWh

Not bad for a five year old pack.

 

[wiztecy]

OK, I give up. Where-and-how does one find this information in the logs?

The (ahr.log) is in the "tar file" bundle once you've "untarred" the logs.

 

[djf]

When I click on the ahr.log file in the "flash" folder per Tom's written instructions, I just get this data set:

and this below it:

I can't find the CAC or "brickahave" that you smart guys refer to. Can anyone help me figure out how to see my CAC? If so, THANKS!

 

[djp]

I can't find the CAC or "brickahave" that you smart guys refer to. Can anyone help me figure out how to see my CAC? If so, THANKS!

The CAC isn't in the ahr.log. It's in the output file from Tom's parser. You'll need to run the parser against your log files and open up the results.

 

[Strib]

This thread has been valuable, not so much for how to improve the battery pack (because I don't think you can reverse any degradation) but for suggesting good discipline for preserving battery health as much as possible.Two, perhaps dumb, questions? - Can you do a cool-down without using OVMS? - Though doing a full Standard Charge, is it helpful to charge and lower current overnight if you have sufficient time (less thermal stress, etc., on battery)?

 

[hcsharp]

It's easy to do a manual cool-down. Just charge your car for 30 - 40 minutes. You can do it at any amperage although it will take a couple minutes longer to cool at 70A than 16A. Lower amps are better if you are on TOU rates.

Charging at about 40A is generally the most efficient rate although it depends on several factors. By efficient I mean the most energy stored in your pack per kWh consumed from the grid. Charging slower does not create significantly less thermal stress than charging faster. A lot of people dispute this but there is no empirical evidence that charging at 70A all the time has any impact on CAC.

This thread has been valuable, not so much for how to improve the battery pack (because I don't think you can reverse any degradation) but for suggesting good discipline for preserving battery health as much as possible.Two, perhaps dumb, questions? - Can you do a cool-down without using OVMS? - Though doing a full Standard Charge, is it helpful to charge and lower current overnight if you have sufficient time (less thermal stress, etc., on battery)?

 

[wiztecy]

I've been doing cool-downs manually myself, still can't get my OVMS going due to the Geo SIM card. I do my cool-downs at 220V@16 amps. Reason being is that I only want to cover the overhead of the HVAC system and don't want miles / a higher SOC. I've tried 220V@12amps but I've found many times the car never dropped the temp down fast enough. Could be just chance since if the ambient temps are higher, the cool-down can only do the best it can. So I settled in at 16AMPS.

Also I don't believe anyone is disputing that a "real" / "true" battery chemistry degradation cannot be be reversed. However the argument is more on what metric Tesla uses with the Roadster ESS to best guess and estimate the battery health over time, and that environmental changes can directly affect this metric depending on the environment conditions and / or how you care for your pack.

 

[hcsharp]

I've been doing cool-downs manually myself, still can't get my OVMS going due to the Geo SIM card. I do my cool-downs at 220V@16 amps. Reason being is that I only want to cover the overhead of the HVAC system and don't want miles / a higher SOC. I've tried 220V@12amps but I've found many times the car never dropped the temp down fast enough. Could be just chance since if the ambient temps are higher, the cool-down can only do the best it can. So I settled in at 16AMPS.

I've also found that 16 and 24A cool faster than 12A. Not sure why.

 

[spaceballs]

If my pack is already cool before I start charging (due to my location) I actually prefer 120v 12amp charging, as it never kicks on the HVAC, so it doesn't do any wear and tear on it i.e. cooling fans/pump/mechanical relays, etc.

 

[wiztecy]

If my pack is already cool before I start charging (due to my location) I actually prefer 120v 12amp charging, as it never kicks on the HVAC, so it doesn't do any wear and tear on it i.e. cooling fans/pump/mechanical relays, etc.

That's a good point and I'll try to keep that in mind.

Whenever the HVAC system is running I typically don't just shut it off which sounds/feels like it stresses it but rather allow the HVAC system to run its course before performing a shutdown. If you listen to the way the HVAC system shuts down by itself, its rather smooth. However if its up and running and you just hit "stop" to request the charging to stop. Everything comes to an abrupt halt which sounds really rough on the system.