16v Lithium battery Weber State Video

Just released is a detailed, but not overly technical, video about the 16v aux lithium battery for those who have it in newer Teslas.
The video has timestamps for easy access to pertinent information. Of most interest might be:

1:50 The jump-start terminals of the Model 3
2:18 Owner's manual jump-start cautions
4:18 Mislabeled 16V batteries
5:29 MUST SEE: Differences between 16V Li-Ion abs 12V Lead Acid battery
7:45 MUST SEE: The push-fit battery 16V electrical connector
8:55 The jump-start terminals of the Model Y
9:28 The Internal self-protection mode of the 16V battery
11:10 MUST SEE: Conditions for the MOSFET to open in protection mode
15:17 What alert is displayed with a bad 16V battery?
15:42 How to recover a 16V battery with an open MOSFET
16:08 The owner's manual method of recovery
17:47 The service manual method of recovery
18:15 Disconnecting the 16V battery
21:12 The Toolbox 3 and service mode methods of recovery

My model 3 is of the legacy 12v system, I found this video quite fascinating and informative.

 

[Gauss Guzzler]

Whoa dude, sorry if my combination of "maybe", "~", and "?" triggered your keyboard warrior instincts. Why don't you go ahead and tell us exactly how many kW the DC system is capable of and why it matters if my guess was off by a few kW. I'll wait here.

 

[n2mb_racing]

At least in the older cars, the DC/DC converter was rated at 2500W. 2012-2016 Tesla Model S MS DC-DC Converter Unit Module 2500W 15A-Max DAP-2500ABA | eBay

 

[MP3Mike]

I'm wondering if (while parked) the DC-DC runs continuously to supply requirements and the aux battery

The HV packs have had a continuous LV aux feed for a long time, to cover the power needs of keeping the system alive without draining the LV battery requiring closing the HV contactors and powering up the full-sized DC-DC converter multiple times per day. (It only provides enough power for when the car is asleep, if you are running Sentry mode the contactors are closed and the full-sized DC-DC is doing the work of keeping the full computer system operational.

 

[gearchruncher]

There must be trade offs between aux battery capacity and the frequency of cycling wear on the charging hardware.

Solid state hardware doesn't wear out from "cycling." That was never the calculation Tesla had to run. They already had to support all this hardware getting turned on and off every drive cycle, every charge cycle, etc.

The battery is there to support emergency loads when driving like the power steering and braking, and is sized for those.

The primary advantage I’m seeing of Tesla’s 16V system is that its BMS can request support as needed rather than having to rely on a timed query which may be too late to save an auxiliary battery.

This is false. The PCS already reads the lead acid battery voltage, and operates on the actual SoC of the battery, not on a fixed timer. This is why the car throws all sorts of errors if you replaced the 12V lead acid with a LiFePO4 battery. The car has a really, really good idea of how charged the lead acid battery is, and even the health of that battery.

The primary reason for the data bus on the Li-Ion battery is probably stuff like "AAAAAACCKKKK, please don't charge me!" when it's below freezing out since you can't charge Li-Ion below 0C and that battery has no heater. Li-Ion is also more sensitive to individual cell voltages than Lead Acid is, so it can report errors if a single cell is bad. But this doesn't mean the lead acid system doesn't have any sense of the battery state and only runs on a timer.

 

[MP3Mike]

since you can't charge Li-Ion below 0C and that battery has no heater.

Some people have said that Tesla draws a large amount of power from the battery when it is below freezing to heat the 16v battery and then allow it to be charged. (I'm not sure what they would use the power for, maybe they ramp down the amount supplied by the DC-DC so that the LV system is supported by the 16v battery temporarily.)

 

[gearchruncher]

Why don't you go ahead and tell us exactly how many kW the DC system is capable of and why it matters if my guess was off by a few kW. I'll wait here.

So you do admit it was just a guess, driven by no data? Nothing like the good old "I have no data, but YOU prove I'm wrong!"

The best data we have is that it's 1.5-2.5kW. Tesla doesn't appear to publish this. But the Model S was 2.5kW, and the size of the connector on the PCS doesn't support anywhere near 850A. It's only about 2-3X the size of the known 50A connectors on the PCS.

Why this matters? Because you're not off by "a few kW." You're off by 75-85%. That's huge. It also matters because you actually said all of this:

The DC-DC (in the penthouse) is capable of supplying many thousands of Watts (maybe ~10kW?) so it uses large transistors which require a lot of power to activate. This makes the system very inefficient for low power levels and is a major reason why the system disconnects and relies on repeatedly draining/recharging the LV battery when sleeping. It's possible they may have added a tiny secondary DC-DC to trickle charge the LV battery full-time. But I'd bet that they just made the sleep power consumption 3X lower to compensate.

Which is complete trash. You have NO DATA that the PCS is inefficient at converting 300-400W. Large transistors have lower RDS(on). Because that's exactly how they support huge currents. This makes them BETTER at all current loads. And you think large transistors "take more power to activate" on the scale of hundreds of watts? Tell me you don't know anything about MOSFETS without telling me you know nothing. And sorry, you didn't put "maybe" in front of this one, so you don't get out of saying it as if it was factual, your favorite defense when you get called on being completely wrong on technical details.

Yet, you seem completely unaware of the contactors, and the power they actually draw, and why you can't just leave the DC/DC on all the time.

It's "possible" they added a smaller DC/DC? Except they didn't, so you also just made this up. Go find this new DC/DC you created in your head in the parts catalog, which is public. Explain why this one can be hooked to the HV bus without contactors, but the main one needs isolation.

And what do you mean they made the sleep power 3X better? They didn't. As shown above, because you can drain a Li-Ion much deeper than lead acid, they have about the same usable energy. The architecture of the car barely changed for the Li-Ion battery. It's the same PCS. It's the same MCU, which is the primary draw on the system when it's sleeping.

How much we betting that they lowered sleep power 3X? I'll take that bet any day.

Just don't post when you don't know, and don't get defensive when you get called on it, especially when your only defense is "I said maybe" and "you find the source for my data."

 

[gearchruncher]

Some people have said that Tesla draws a large amount of power from the battery when it is below freezing to heat the 16v battery and then allow it to be charged.

The 16V battery only has about 500g of thermal mass. It would not take large amounts of power to heat it. But it's been proven it has no heater. Do you have a link to people describing this event?

The car takes large amounts of energy to heat the HV pack when it needs to charge that for sure, but I just don't see how it can heat the 16V. You'd have to draw tons of current from the 16V to try and use self heating, but this is generally pretty tough on the battery. It also would still only be maybe 1kW for a minute or two, which is not a lot of power.

 

[MP3Mike]

It's "possible" they added a smaller DC/DC? Except they didn't, so you also just made this up. Go find this new DC/DC you created in your head in the parts catalog, which is public. Explain why this one can be hooked to the HV bus without contactors, but the main one needs isolation.

@wk057 has reported about the small "support" DC/DC being added to HV packs long ago; it provides power to support sleep mode and reduce cycles on the LV battery.

Do you have a link to people describing this event?

I've just got what @Ingineer said in his video on the 16v battery and his guess on how they would heat it:

Followed up by another video which notes that the cells used likely have a 50C discharge rate, and can be charged down to -10C:

 

[gearchruncher]

I've just got what @Ingineer said in his video on the 16v battery and his guess on how they would heat it:

Thanks for the excellent links- but to be clear that is just him guessing, and isn't actually "Some people have said that Tesla draws a large amount of power from the battery when it is below freezing to heat the 16v battery and then allow it to be charged."

It's a reasonable guess- but it's also interesting that the cells can be charged to -10 but only discharged to -20. So there's only a 10C range where this method is useful, and in fact the car should technically brick at -20C, which really isn't that cold for the #1 car being sold in Norway. For all we know, the car just goes into DC/DC mode anytime the temp gets below -10C (the same as it does when sentry is on).

@wk057 has reported about the small "support" DC/DC being added to HV packs long ago; it provides power to support sleep mode and reduce cycles on the LV battery.

If we're discussing all Teslas, that's fair. I (possibly incorrectly) assumed we were discussing only the Model 3/Y, which is all the video covered. Do we have data that they changed the Model 3/Y battery when they added the Li-Ion to add this small DC/DC? Tesla's own parts catalog lists no changes to the PCS or battery that are tied to having a Li-Ion on a 3/Y.

But even for that, I can't find his post about this, much less it being tied to the change from Lead Acid to Lithium, which is the topic here.

Also, if this was true, you'd expect sentry mode to be much lower draw on newer cars, yet I have never seen that reported either.

 

[n2mb_racing]

I wonder how the DC/DC regulates charge current to the 16V battery, while also having a high-current pathway to the 16V bus in the discharge direction. I.e. The system is designed to allow 100+ amp discharge from the 16V battery to power the entire 16V system on the car, in the event of a HV battery failure. But, the charge current into the 16V battery needs to be limited to ~8A. Does the DC/DC converter adjust the voltage for the entire car in order to regulate the charging current?

 

[gearchruncher]

I wonder how the DC/DC regulates charge current to the 16V battery, while also having a high-current pathway to the 16V bus in the discharge direction. I.e. The system is designed to allow 100+ amp discharge from the 16V battery to power the entire 16V system on the car, in the event of a HV battery failure. But, the charge current into the 16V battery needs to be limited to ~8A. Does the DC/DC converter adjust the voltage for the entire car in order to regulate the charging current?

No. The Front Vehicle Controller (VCFRONT) is what hooks the battery to the LV bus. It has a FET, and current monitor for the battery. It can apply a lower voltage to the battery (via PWM) than the rest of the bus is at and limit charge current when the primary DC/DC is online.

Also note that the VCFRONT PN did not change when they introduced a Li-Ion battery....

 

[MP3Mike]

If we're discussing all Teslas, that's fair. I (possibly incorrectly) assumed we were discussing only the Model 3/Y, which is all the video covered. Do we have data that they changed the Model 3/Y battery when they added the Li-Ion to add this small DC/DC? Tesla's own parts catalog lists no changes to the PCS or battery that are tied to having a Li-Ion on a 3/Y.

The support DC-DC was added to the S&X when they still had lead acid batteries, and every Model 3/Y has probably come with this feature. (It was part of their work to make the 12v batteries last longer.) So yeah, there is no change, because they all have it.

 

[gearchruncher]

The support DC-DC was added to the S&X when they still had lead acid batteries, and every Model 3/Y has probably come with this feature. (It was part of their work to make the 12v batteries last longer.) So yeah, there is no change, because they all have it.

Gotcha. So like I said, they did not add a DC/DC when they went to Li-Ion as was suggested: " It's possible they may have added a tiny secondary DC-DC to trickle charge the LV battery full-time."

None of the DC/DC converters are hooked to the HV bus full time. They all go through contactors, and it's those contactors that draw hundreds of watts. So we know from observing older Model 3's, which have this "small DC/DC", that they go through cycles where they discharge the LV then charge it. They don't trickle charge it all the time, and they can't without having 150W+ of contactors engaged. It would absolutely take new HW tied to the Li-Ion LV battery to do this trickle charge, because if it was possible with previous HW, they would have already done it. But there's no new HW listed in the parts catalogs in the PCS of VCFRONT that is tied to this.

 

[MP3Mike]

None of the DC/DC converters are hooked to the HV bus full time. They all go through contactors, and it's those contactors that draw hundreds of watts. So we know from observing older Model 3's, which have this "small DC/DC", that they go through cycles where they discharge the LV then charge it. They don't trickle charge it all the time, and they can't without having 150W+ of contactors engaged.

The support DC/DC is powered 24x7, the goal being to power all of the 24x7 computers/LTE/WiFi without draining, and putting cycles on, the LV battery. It doesn't require contactors that use 150W. (I doubt even the main contactors require 150W to stay engaged. I'm pretty sure the full-in Sentry mode only uses ~220W, and that is powering the MCU/AP computer/cameras/coolant pump/main DC-DC/etc.)

 

[n2mb_racing]

The support DC/DC is powered 24x7, the goal being to power all of the 24x7 computers/LTE/WiFi without draining, and putting cycles on, the LV battery. It doesn't require contactors that use 150W. (I doubt even the main contactors require 150W to stay engaged. I'm pretty sure the full-in Sentry mode only uses ~220W, and that is powering the MCU/AP computer/cameras/coolant pump/main DC-DC/etc.)

Agreed, the contactors would melt if they used 150w. They might be 10w.

The new AMD computer uses about 220w on average in sentry mode, and that includes all the other consumption.

 

[gearchruncher]

The support DC/DC is powered 24x7, the goal being to power all of the 24x7 computers/LTE/WiFi without draining, and putting cycles on, the LV battery. It doesn't require contactors that use 150W.

This is trivial to disprove.
Pop the trunk on the car. Wait for it to go to sleep after an hour or so.
Now unplug the LV battery.
Now try to ping the car via the app.
Won't work.

Also, you can trivially see that when the car is idle, there is about 0.5A drain from the LV battery. All it takes is a current clamp around the wire going to it. And then when it switches to charging, you hear the HV contactors close, because they have to for the DC/DC to go online. Why is there any negative current from the battery if there is a maintenance DC/DC on the car? Why if you have a dead 12V is the car dead, if it is able to sustain itself off the HV with the contactors off? Why is Tesla's answer to detecting a bad 12V to leave the HV contactor on all the time and actually run off the DC/DC if they don't need the contactors?

You can do a bunch of things too, like reading the Vehicle Electrical Isolation Procedure in the Tesla service manual, which makes it clear that once the HV contactors are open, and the 12V battery is removed, there is no power in the car. Yet you are stating that there is a way for the car to generate enough 12V current to run the computers even with the contactors open.

All sorts of other questions caused by this too. Why does the car turn on the HV contactors any time it's awoken at all, even just to ask the car the position it was last at, if it's able to keep all that running off the HV through a "small" DC/DC?

The reality is that there's one PCS in the car, that's the only DC/DC, and for it to have HV, the HV contactors must be on.

And those contactors are 100W each, and there are two: Tesla Model 3 Contactors - openinverter.org wiki

Now yes, they can do smart things to reduce this. But all the support stuff to have the car be "on"- the contactors closed, the PCS running, etc, does add up to about 100W-200W of quiecent draw before anything in the car can do anything useful.

The new AMD computer uses about 220w on average in sentry mode, and that includes all the other consumption.

Correct. And about 50W of that is the computer. The rest is all wasted in just having the HV online. This isn't any different than the Intel computer BTW, and you also know that the AMD computer isn't drawing hundreds of watts given the AMD chip has a TDP of 45 watts....

 

[MP3Mike]

This is trivial to disprove.
Pop the trunk on the car. Wait for it to go to sleep after an hour or so.
Now unplug the LV battery.
Now try to ping the car via the app.
Won't work.

Because they have a safety inter-lock to turn off the support "DC-DC" when there is no battery hooked up?

You can do a bunch of things too, like reading the Vehicle Electrical Isolation Procedure in the Tesla service manual, which makes it clear that once the HV contactors are open, and the 12V battery is removed, there is no power in the car. Yet you are stating that there is a way for the car to generate enough 12V current to run the computers even with the contactors open.

I never said it supplied enough to completely run things, just to reduce the number of cycles on the LV battery.

All sorts of other questions caused by this too. Why does the car turn on the HV contactors any time it's awoken at all, even just to ask the car the position it was last at, if it's able to keep all that running off the HV through a "small" DC/DC?

Because the "support" DC-DC can't provide enough power to run the MCU when fully awake.

And those contactors are 100W each, and there are two: Tesla Model 3 Contactors - openinverter.org wiki

Yes, 100W to pull them, but not to maintain them closed. (As noted in your link.) You are trying to say that those use ~200W all the time, but with Sentry mode in use the total is only ~220W. Are you saying that the MCU/AP/cooling pump/DC-DC/etc. only use ~20W?

Correct. And about 50W of that is the computer. The rest is all wasted in just having the HV online. This isn't any different than the Intel computer BTW, and you also know that the AMD computer isn't drawing hundreds of watts given the AMD chip has a TDP of 45 watts....

I think the FSD computer takes about 72W on top of the 50W that the MCU uses. That accounts for more than half of the ~220W.

 

[gearchruncher]

I never said it supplied enough to completely run things, just to reduce the number of cycles on the LV battery.

You're claiming there is a DC/DC in the car that can handle about 5W, but not 12W?
What a weird engineering decision for Tesla to make. I mean, the engineering meeting would have to go like this:

<scene>

Ok, so we've got like 12W of idle power draw to deal with. We had all sorts of issues with cycling Lead Acid batteries in the Early S/X cars. So here's what we're gonna do.... We're gonna design a DC/DC that is super reliable and safe, such that it can be connected to the HV all the time. But we're gonna make it 6W. Not 12W. 12W is too much. 6W will halve how often we cycle the Lead Acid. But this way, we do still use up lead acids, and we do still get to cycle the contactors every day or so, and we still need to design a 12V charge controller since we do need to rapid charge the battery since we are so inefficient when we are charging it. But that all sounds fun, so let's do it.

Yep. That sounds great. You're right, making a 12-15W always on DC/DC that fixes all our problems would just be silly. Go with that 6W! And let's not just make the existing PCS more efficient at low loads with a different switching method. I mean, that's how OTHER EE's do that, but we're Tesla, what they do is stupid. We throw extra hardware at stuff for fun. So I for sure nominate both an inefficient PCS, and then an efficient second box.

Oh, and while you're at it, make sure we NEVER, EVER show this on any parts diagram. It would be awesome to hide it inside the PCS, but the PCS has no power when the contactors are off. So we'll put it in the penthouse, never mention it in any descriptions, and we'll also leave it off the wiring diagrams, even though it has dangerous high voltage going to it all the time even when the HV of the car is disabled. Nobody will ever need to service it, so it doesn't need to be in the service manual.

<end scene>

You can prove this isn't happening easily. You can unplug the wire under the rear seat to fully disconnect the HV. So check the idle current draw on the LV battery. It will be about 6W. Then unplug the interlock. You're saying the current from the LV battery should go way up? We taking bets on this?

Because they have a safety inter-lock to turn off the support "DC-DC" when there is no battery hooked up?

Yeah, those safety interlocks are called the contactors. You can hear them make a big CLUNK when they change state. You can hear them CLUNK on when you access the car via the app. You can hear them CLUNK when the car starts charging the LV battery in a CLEAR cycle where it goes from 0.5A discharge to 8A charge. Why are they ever off if there is a magic low power, high efficiency DC/DC that needs HV via an interlock?

Are you claiming there is a second set of contactors that interlock just this? That are small and you can't hear?

Because the "support" DC-DC can't provide enough power to run the MCU when fully awake.

We're not talking fully awake here. We're talking about a fully asleep car. We know for a fact the car still discharges the LV battery in this state and has to wake up every ~20 hours to recharge the LV battery.

At this point, someone really needs to point out where this special low power DC/DC converter is in the car. It's not in the PCS for a fact, since the PCS is well known to only have a single HV connection, and that HV connection is prov-ably only active when the contactors are closed. So please, what is the PN of this second DC/DC in the car?

I think the FSD computer takes about 72W on top of the 50W that the MCU uses. That accounts for more than half of the ~220W.

Neither the FSD computer or the MCU are running full bore, max TDP in sentry mode. Is sentry mode as computationally intensive as literally driving a car by itself? Is the MCU as busy as it is when it's showing a full FSD display, moving map, spotify, recording dashcams and playing fart noises? Even though the 10W of screen backlight is off?

Like people keep saying, the car is about 230W when in sentry mode. But there's no way more than about 50W of this is actual computational horsepower. Your phone could easily do sentry mode functions and it is only about 10W. So if it's 50W of processing, where is the other 180W of power going?

 

[n2mb_racing]

Like people keep saying, the car is about 230W when in sentry mode. But there's no way more than about 50W of this is actual computational horsepower. Your phone could easily do sentry mode functions and it is only about 10W. So if it's 50W of processing, where is the other 180W of power going?

I don't know. Get out the clamp ammeter and find out.... Or the data might already be on the CANBUS, since the car monitors the current on all of the 12V circuits.

 

[Gauss Guzzler]

This is trivial to disprove...
And those contactors are 100W each, and there are two: Tesla Model 3 Contactors - openinverter.org wiki

True. All I had to do was click on your link.

Pro tip: When making up garbage and pretending to know things that you clearly don't, don't use words like "are" in bold lettering. Instead, use words like "maybe" and symbols like "~" and "?" to indicate that you are just speculating about a possibility.