Wiki Sudden Loss Of Range With 2019.16.x Software

[mymagiccarpet]

The pumps always run at full speed (or near full) during high rate charging, high rate discharging, and above a threshold of SoC vs Temp. If you look at CAN data (or the diag screen) you can see that usually this is done with the battery loop isolated (as in, not attempting to cool or heat the cells). This is simply to help equalize the temperature of the cells throughout the pack, and this mechanism is present in even the earliest firmware versions I've examined (6+ years old now). Nothing new.

Basically, the more thermal variation there is between the cells, the more variation there is in how much power each cell will output (load) or accept (charge). This translate to a higher voltage imbalance over time. Equalizing the temperatures of the cells slows any voltage/capacity imbalance between cell groups that could be caused by thermal variations.

This is also why in my solar setup I put the pairs of modules that were in series on the same horizontal plane, since this would work out to roughly the same air temperature since they were at the same level... and it actually worked out quite nicely, as I've never had any major imbalances in any strings despite using no liquid cooling. (My highest charge/discharge rate is about 1/3C, so definitely not a huge factor). (Edit: Just to note, the system just reached 5.5 years of continuous operation, with over 200 MWh produced.)

There's some links and a graphic in a post above that are completely irrelevant. As always, take such things with a grain of salt... or a pile. The above info is referring to cells charged at freezing temperatures at high rates... which is known to cause damage, degradation, and otherwise be unsafe. The temperatures above freezing that they are referring to appear to be the permitted self-heating targets for the cells under charge/discharge when starting at freezing. Fortunately, since day 1, Tesla has included protections to prevent the battery from ever being subject to those conditions. The cells are never charged when below an unsafe temp, and the max charge current is monitored vs temperature at the millisecond level... so, as noted and as usual, irrelevant information.

Are you referring to discussion about pump draining battery with SOC above 80%?
That sure is new, and not since May'19 update, it was later, at least for me, I want to say later in the summer of '19. My car was not making any noise whatsoever above 80% charge, hot or cold.
Maybe what you are saying is it's not the pump that's draining the battery?

 

[wk057]

My car was not making any noise whatsoever above 80% charge, hot or cold.

The pumps definitely get louder with age, but I digress.

Are you referring to discussion about pump draining battery with SOC above 80%?
That sure is new, and not since May'19 update, it was later, at least for me, I want to say later in the summer of '19. My car was not making any noise whatsoever above 80% charge, hot or cold.
Maybe what you are saying is it's not the pump that's draining the battery?

If a pre-refresh car doesn't sleep, the pack will drain at a minimum of about 0.5 MPH of range. This is because the pre-refresh cars (and some early VIN Model X) don't have a standby power supply inside the pack (basically a mini isolated DCDC converter to keep the MCU and body controller powered with minimal power). On pre-refresh cars, if there is any 12V drain above a threshold, the main contactors have to close to power the DCDC. The DCDC is liquid cooled, so needs the powertrain pump running. If the pack pump is running at 100%, that means the pack contactors have to be closed, too, thus introducing more range loss. Most likely the gateway and/or MCU won't sleep either.

I haven't done a lot of testing on it, but it initial impression is that the connectivity changes Tesla made last year adversely affect vampire drain on pre-refresh cars also.

Tons of reasons why a car would lose range overnight.

But the pack pump running at full would tend to mean that a temp target was trying to be reached, or a temp imbalance was trying to be corrected... both of which are more important for the pack's overall health, especially if there is any existing imbalance (on the electrical side of things), than the loss of some range overnight.

 

[Droschke]

My car was not making any noise whatsoever above 80% charge, hot or cold.

The same for lots of other owners after 2019.16.x update. As @DJRas showed above, the battery temp actually shows relatively cooler above 80%, nevertheless the pumps run (not because they are older and louder as they age), but because the battery is presumably unsafe due to, as it might be, the internal heat buildup. Don't get distracted.

 

[Chaserr]

@Droschke I recently linked several published studies showing the hows and whys of internal cell heat build up in cells exactly like ours. This is the most plausible and scientifically backed explanation for why Tesla has illegally capped volts, perhaps reasonably if not popularly throttled charging, lowered regen limit temperatures, and runs the pumps constantly. These all combine to what looks like attempts to cease internal structural changes that lead to batteries self-combusting from a slow buildup of internal heat generation. Those studies not only corroborate Tesla's actions, they also fit the real world evidence that led to them. Batteries simply don't burn for no reason, and parked sleeping Teslas should be the least prone to any kind of accident except internal heating which is more probably in a battery that is slowly building up heat but is sleeping and therefore not running cooling at full draingate. I've already shown how a number of post-batterygate studies have been proven this will become a problem in cells exactly like ours at temperatures as low as 95 degrees F when subjected to situations exactly like we were encouraged to treat our batteries. Scientific phenomena like this was one of the risks of being an early adopter - and the reason were all enticed by legal promises of the worlds best battery and safest cars. When we bought our cars Tesla would have stood behind those promises. It will again.

 

[wk057]

the pumps run because the battery is presumably unsafe due to, as it might be, the internal heat buildup.

Evidence? Because this is contrary to all data I've seen on this, including CAN logs from some cars affected by the range loss issue.

The pack pump generally runs in a closed loop, which wouldn't remove any "internal heat buildup"... instead, if a cell or group of cells were internally heating, the loop temp would rise as a result. However, you yourself note the following:

the battery temp actually shows relatively cooler above 80%

So unless you've got pumps AND fans running the car is NOT cooling the battery. Due to their positioning behind the A/C condensors, which are to the sides of the small coolant radiator, the fans don't have a lot of effect on the coolant radiator when the car is stationary. So to remove heat from it while stationary they have to run at high speeds to have any effect on the coolant radiator at all... and thus be pretty noticeable. Alternatively, the A/C compressor and chiller can run, which is quieter than the fans+radiator method at the expensive of power... so this generally is only done while charging.

Long story short, the pumps are not running to cool the battery, and generally can't do this anyway, without the help of the A/C system, while stationary.

 

[wk057]

Oh, I forgot to mention a major cause of hearing pumps run at 100%, especially on older cars: Pump failure. If one or more pumps fail, or otherwise can not properly have their run speed detected, the car runs all pumps at 100% for any pump command at all. So, if the car just needs 5% speed out of a single pump for what it wants at the time, and can't detect that pump's speed to know that it is in fact working, it commands all pumps to run at 100% as a fail-safe.

A negative side effect of this is that the pumps then consume enough power to prevent sleep, possibly causing a vicious vampire drain cycle (need pump to cool DCDC, cant shutdown pump because DCDC is running, cant sleep because DCDC is running, etc).

 

[MP3Mike]

So, if the car just needs 5% speed out of a single pump for what it wants at the time, and can't detect that pump's speed to know that it is in fact working, it commands all pumps to run at 100% as a fail-safe.

That is probably a good fail-safe, but they should really display an alert so that the owner can get the car serviced and the problem resolved.

Is there something people can see in the CAN data that would indicate that it has detected a pump issue and is running in fail-safe mode?

 

[mymagiccarpet]

The pumps definitely get louder with age, but I digress.



If a pre-refresh car doesn't sleep, the pack will drain at a minimum of about 0.5 MPH of range. This is because the pre-refresh cars (and some early VIN Model X) don't have a standby power supply inside the pack (basically a mini isolated DCDC converter to keep the MCU and body controller powered with minimal power). On pre-refresh cars, if there is any 12V drain above a threshold, the main contactors have to close to power the DCDC. The DCDC is liquid cooled, so needs the powertrain pump running. If the pack pump is running at 100%, that means the pack contactors have to be closed, too, thus introducing more range loss. Most likely the gateway and/or MCU won't sleep either.

I haven't done a lot of testing on it, but it initial impression is that the connectivity changes Tesla made last year adversely affect vampire drain on pre-refresh cars also.

Tons of reasons why a car would lose range overnight.

But the pack pump running at full would tend to mean that a temp target was trying to be reached, or a temp imbalance was trying to be corrected... both of which are more important for the pack's overall health, especially if there is any existing imbalance (on the electrical side of things), than the loss of some range overnight.

This is informative. Thanks. Seems that you are saying (i don't like to interpret other's words), if pump is running, it's purely b/c of temp imbalance?
A couple of things that come to mind are, I'm not even sure it is the pump that's running, but something is, and it's not a fan.
The other thought is, it seems to always drain specifically to ~78%. In reality, maybe it drains to 80 and the rest is vampire drain, who knows.
So for example, these are vampire drain at different SOCs:
If I charge the car to 70%, and park shortly after, my vampire overnight drain is 1-2mi. You are correct that there is more variation now after they changed connectivity settings/code, but lately, seems to have settled back to 1-2mi.
In case of not charging recently:
If car's SOC is 20ish% to 78%, vampire drain is also 1-2mi overnight.
If car's SOCis <20%, vampire drain is significantly higher, eg, it will go from 40->32mi overnight
If car's 12V is low, car's vampire drain will significantly increase, eg. mine was 8-10mi overnight (seems like we understand and expect this one)
If, however, car's SOC is above 80% (whether charging recently or not) it will continue to drain at a much faster pace until it gets to 78%. This, on surface at least, does not seem to be related to temp, but rather SOC, unless temp and SOC are 100% correlated at 78% SOC.
That's where it stops and goes back to normal vampire drain. It is a very consistent and specific behavior.
Do you have any insight or theories on this?

Not specifically to wk057, but all:
Someone mentioned earlier that 78% correlates to 3.99V? Can anyone confirm this with their own car?
Maybe first of all, how many people see this same behavior with their car? Is it everyone or not?

 

[mymagiccarpet]

Oh, I forgot to mention a major cause of hearing pumps run at 100%, especially on older cars: Pump failure. If one or more pumps fail, or otherwise can not properly have their run speed detected, the car runs all pumps at 100% for any pump command at all. So, if the car just needs 5% speed out of a single pump for what it wants at the time, and can't detect that pump's speed to know that it is in fact working, it commands all pumps to run at 100% as a fail-safe.

A negative side effect of this is that the pumps then consume enough power to prevent sleep, possibly causing a vicious vampire drain cycle (need pump to cool DCDC, cant shutdown pump because DCDC is running, cant sleep because DCDC is running, etc).

Thanks for this info too. To your knowledge (and anyone else's), is there a sensor that would throw error if a pump has gone bad?

 

[wk057]

That is probably a good fail-safe, but they should really display an alert so that the owner can get the car serviced and the problem resolved.

Is there something people can see in the CAN data that would indicate that it has detected a pump issue and is running in fail-safe mode?

Unfortunately seems to only trigger hidden alerts.

This is informative. Thanks. Seems that you are saying (i don't like to interpret other's words), if pump is running, it's purely b/c of temp imbalance?
A couple of things that come to mind are, I'm not even sure it is the pump that's running, but something is, and it's not a fan.
The other thought is, it seems to always drain specifically to ~78%. In reality, maybe it drains to 80 and the rest is vampire drain, who knows.
So for example, these are vampire drain at different SOCs:
If I charge the car to 70%, and park shortly after, my vampire overnight drain is 1-2mi. You are correct that there is more variation now after they changed connectivity settings/code, but lately, seems to have settled back to 1-2mi.
In case of not charging recently:
If car's SOC is 20ish% to 78%, vampire drain is also 1-2mi overnight.
If car's SOCis <20%, vampire drain is significantly higher, eg, it will go from 40->32mi overnight
If car's 12V is low, car's vampire drain will significantly increase, eg. mine was 8-10mi overnight (seems like we understand and expect this one)
If, however, car's SOC is above 80% (whether charging recently or not) it will continue to drain at a much faster pace until it gets to 78%. This, on surface at least, does not seem to be related to temp, but rather SOC, unless temp and SOC are 100% correlated at 78% SOC.
That's where it stops and goes back to normal vampire drain. It is a very consistent and specific behavior.
Do you have any insight or theories on this?

Not specifically to wk057, but all:
Someone mentioned earlier that 78% correlates to 3.99V? Can anyone confirm this with their own car?
Maybe first of all, how many people see this same behavior with their car? Is it everyone or not?

Would need to see some CAN logs of this behavior (in the various states noted) to check, but I know that older packs tend to have more thermal and electrical deltas, especially after charging. The permitted deltas adjust based on SoC, also, so what _appears_ to be happening is that your pack is unable to climb out of the allowed thermal delta for a particular SoC range, and dies trying.

This is somewhat expected of older packs, although Tesla should probably fix their algos to account for this at some point.

Thanks for this info too. To your knowledge (and anyone else's), is there a sensor that would throw error if a pump has gone bad?

No sure what you mean, but the pumps each have a sensor that reports to the thermal controller. If the THC can't correctly correlate commanded flow with measured flow from the pump, then it assumes the pump is faulty.

If I recall correctly, one pump failure is not enough to trigger a user facing alert, but two can if one is the battery loop pump.

 

[Droschke]

Someone mentioned earlier that 78% correlates to 3.99V? Can anyone confirm this with their own car?
Maybe first of all, how many people see this same behavior with their car? Is it everyone or not?

Many owners in this thread and other threads. It's not uncommon for the S85 cars.

Have you been wondering what happens if these pumps do not run for hours on and what the adverse consequences are, including the fire risks?

 

[Chaserr]

Have you been wondering what happens if these pumps do not run for hours on and what the adverse consequences are, including the fire risks?

There is an incredibly irresponsible way to find out. Supercharge to "100%" and immediately pull forward to a parking spot in the lot right in front of the supercharger. Disconnect both the 12V and HV. None of the cooling pumps will be able to run and battery won't be able to drain. Let it sit for a week like that and see if temperatures rise. If you disassemble the frunk first, you should be able to see and touch the number nine brick under the Microwave. Tesla officially announced "one or two modules at the front of the pack" were the cause of the fires and the 350v 85 replacement doesn't have those, so measure temps on the top middle module and see if they stabilize at ambient. If enough people do this we would probably see a few confirmations. If everyone does thsi we would probably see photos of a car in flames in front of a Supercharger.

@David99 I don't recommend doing this with your car's confirmed short circuiting battery. You would probably wind up in a fancy new Tesla and wind up being a catalyst for us all finally receiving the transparency we've been begging and pleading for, but knowingly starting public fires is even more irresponsible than Tesla's secretive responses to them.

 

[Fierofiend]

Another interesting thing to note is that my pumps will run at 100% until I open a door, then they drop to 23-25%. As soon as I close the door, they ramp back to 100%. Not sure if others experience this too.

 

[wk057]

Another interesting thing to note is that my pumps will run at 100% until I open a door, then they drop to 23-25%. As soon as I close the door, they ramp back to 100%. Not sure if others experience this too.

Believe this is intended.

There were a lot of patches designed to improve the perception of things a while back that all happened about the same time. One of them quiets fans/pumps/compressor/air suspension/brake booster/etc for a bit during ingress/egress.

 

[mymagiccarpet]

No sure what you mean, but the pumps each have a sensor that reports to the thermal controller. If the THC can't correctly correlate commanded flow with measured flow from the pump, then it assumes the pump is faulty.

If I recall correctly, one pump failure is not enough to trigger a user facing alert, but two can if one is the battery loop pump.

I think you answered it in the last sentence. Do you know or think that SC can run a check and find an issue if there is an actual issue with one pump? Meaning, if I report that pumps are running every time i charge over 80% until SOC hits 78%. if they run a "pump check" they could see that one is bad? or they dont' have that level of granularity

 

[JRP3]

Have you been wondering what happens if these pumps do not run for hours on and what the adverse consequences are, including the fire risks?

So unless you've got pumps AND fans running the car is NOT cooling the battery.

 

[mymagiccarpet]

Many owners in this thread and other threads. It's not uncommon for the S85 cars.

Have you been wondering what happens if these pumps do not run for hours on and what the adverse consequences are, including the fire risks?

Funny you should say that, bc I did write something about it, and then in the interest of not adding to unvalidated theories, I deleted it and posted
Yes I definitely do. Why is it so critical that my battery is not above 80%?, but again, don't have all the details, so my perception may not be what's actually happening

 

[Droschke]

Funny you should say that, bc I did write something about it, and then in the interest of not adding to unvalidated theories, I deleted it and posted
Yes I definitely do. Why is it so critical that my battery is not above 80%?, but again, don't have all the details, so my perception may not be what's actually happening

Appreciate that. The validation has been right in front us for over a year. Just remember how the batterygate, chargegate and draingate started:

Tesla gives updates on cause of a battery fire, says single module is responsible - Electrek
Tesla is updating its battery software following a car fire, claims improve longevity - Electrek

There are more of course.

 

[ran349]

Funny you should say that, bc I did write something about it, and then in the interest of not adding to unvalidated theories, I deleted it and posted
Yes I definitely do. Why is it so critical that my battery is not above 80%?, but again, don't have all the details, so my perception may not be what's actually happening

If I charge to 75%, the pumps do not come on at all after charging. If I charge to 80%, they come on for exactly 3 hours. If I charge to 85% or 90%, or anything higher than that, they still come on for exactly 3 hours. Are you observing something different than that?

 

[trayloader]

Unfortunately seems to only trigger hidden alerts.



Would need to see some CAN logs of this behavior (in the various states noted) to check, but I know that older packs tend to have more thermal and electrical deltas, especially after charging. The permitted deltas adjust based on SoC, also, so what _appears_ to be happening is that your pack is unable to climb out of the allowed thermal delta for a particular SoC range, and dies trying.

This is somewhat expected of older packs, although Tesla should probably fix their algos to account for this at some point.



No sure what you mean, but the pumps each have a sensor that reports to the thermal controller. If the THC can't correctly correlate commanded flow with measured flow from the pump, then it assumes the pump is faulty.

If I recall correctly, one pump failure is not enough to trigger a user facing alert, but two can if one is the battery loop pump.

Evidence? Because this is contrary to all data I've seen on this, including CAN logs from some cars affected by the range loss issue.

The pack pump generally runs in a closed loop, which wouldn't remove any "internal heat buildup"... instead, if a cell or group of cells were internally heating, the loop temp would rise as a result. However, you yourself note the following:



So unless you've got pumps AND fans running the car is NOT cooling the battery. Due to their positioning behind the A/C condensors, which are to the sides of the small coolant radiator, the fans don't have a lot of effect on the coolant radiator when the car is stationary. So to remove heat from it while stationary they have to run at high speeds to have any effect on the coolant radiator at all... and thus be pretty noticeable. Alternatively, the A/C compressor and chiller can run, which is quieter than the fans+radiator method at the expensive of power... so this generally is only done while charging.

Long story short, the pumps are not running to cool the battery, and generally can't do this anyway, without the help of the A/C system, while stationary.

Hi Guru Expert Enigma

so basically what you are saying is that older packs run hotter?