Range Extender - Any Updates? (MAR 2025)

[cybercricket]

You are ignoring cooling fluid connections.

Cooling loop is half of the problem. The other half is that there is no simple way to equalize the voltage between the two battery packs. I.e. equalization between the cell groups can be done by the BMS in each respective pack, but the cumulative voltages need to be equal before the two packs are paralleled to prevent a massive in-rush current. Tesla BMS uses active balancing, so it can't bleed off the voltage until the two packs are equal.

There could be some clever ways to integrate the two packs dynamically, like waiting until the main pack is at the same voltage as the extender during driving or charging cycle and then firing up the contactors to parallel the two. But any such approach requires more complexity that isn't necessarily justified given most customers won't be removing the extender once installed.

Sponsored

 

[mongo]

Cooling loop is half of the problem. The other half is that there is no simple way to equalize the voltage between the two battery packs. I.e. equalization between the cell groups can be done by the BMS in each respective pack, but the cumulative voltages need to be equal before the two packs are paralleled to prevent a massive in-rush current. Tesla BMS uses active balancing, so it can't bleed off the voltage until the two packs are equal.

There could be some clever ways to integrate the two packs dynamically, like waiting until the main pack is at the same voltage as the extender during driving or charging cycle and then firing up the contactors to parallel the two. But any such approach requires more complexity that isn't necessarily justified given most customers won't be removing the extender once installed.

The packs will stay at the same potential while the vehicle is off, so the matching only needs to occur once at installation.
Cybertruck already does something similar when switching from 800V mode to 400V.

 

[cybercricket]

The packs will stay at the same potential while the vehicle is off, so the matching only needs to occur once at installation.
Cybertruck already does something similar when switching from 800V mode to 400V.

It's not quite similar because the active balancing spans the entire pack regardless of whether it's in full 800V configuration or split into two. So when split&parallel operation happens the two halves are guaranteed to be at the same voltage and continue balancing across all cell groups equally - i.e. group X in half 1 will be able to equalize not just with a group Y in half 1, but also with a group Z in half 2.

 

[mongo]

It's not quite similar because the active balancing spans the entire pack regardless of whether it's in full 800V configuration or split into two. So when split&parallel operation happens the two halves are guaranteed to be at the same voltage and continue balancing across all cell groups equally - i.e. group X in half 1 will be able to equalize not just with a group Y in half 1, but also with a group Z in half 2.

So if modules in the same physical package can stay in sync, why can't modules in separate packages?
Total voltage of main and extender are the same due to the (expected) direct connection between them. Each parallel cell group is then total/numberSeriesGroups. There isn't an opportunity for the extender and main pack to develop a large offset unless the vehicle is kept off for a long time and they have very different self discharge rates (and no low current continuous balance path).

 

[cybercricket]

So if modules in the same physical package can stay in sync, why can't modules in separate packages?
Total voltage of main and extender are the same due to the (expected) direct connection between them. Each parallel cell group is then total/numberSeriesGroups. There isn't an opportunity for the extender and main pack to develop a large offset unless the vehicle is kept off for a long time and they have very different self discharge rates (and no low current continuous balance path).

It's just a design constraint - the battery pack has to be fully-contained without any exposed lines to the underlying cells. For example, besides HV (input/output), control (CAN/Ethernet/etc), and MV (input), what other lines are present on the main pack connector(s) ?

 

[mongo]

It's just a design constraint - the battery pack has to be fully-contained without any exposed lines to the underlying cells. For example, besides HV (input/output), control (CAN/Ethernet/etc), and MV (input), what other lines are present on the main pack connector(s) ?

Right, each module has it's own BMS that handles its cells and communicates with the vehicle/ other BMS. As long as the two units have the same voltage, the internals can be a black box though.

Extra effort is needed when first pairing the extender with the main pack to avoid excessive current. After that, there is no reason for them to get out of sync (to the level that direct connectionis an issue).

Regarding balancing, isn't it is the case that, while actively (software) controlled, the cell are kept balanced via dissipation, not energy transfer from high groups to low?

 

[cybercricket]

Extra effort is needed when first pairing the extender with the main pack to avoid excessive current. After that, there is no reason for them to get out of sync (to the level that direct connectionis an issue).

Theoretically speaking any variance in thermal regulation will result in different rates of self-discharge and also cell degradation. Even though the Extender would be integrated into the cooling loop, because of a different physical placement and its different thermal capacity it will likely be a bit off most of the time. Maybe they'll have a "precharge" style circuit to account for that. Won't know for sure what clever tricks they'll use until somebody takes one apart

Regarding balancing, isn't it is the case that, while actively (software) controlled, the cell are kept balanced via dissipation, not energy transfer from high groups to low?

That's the first thing I mentioned - as far as I know, Tesla BMS uses active balancing and not passive balancing. Active balancing is energy transfer, and is beneficial in the automotive applications because it has the potential to allow for a deeper pack discharge before the undervoltage condition occurs. Passive balancing relies on dumping some power into a discharge resistor typically during charging only.

 

[Pulaski]

Perhaps they removed it because they are releasing a long range version soon with a larger battery pack. I.e. it's not a range extender we'll be getting, but really a battery swap. It's wishful thinking, but gotta hope.

 

[mongo]

That's the first thing I mentioned - as far as I know, Tesla BMS uses active balancing and not passive balancing. Active balancing is energy transfer, and is beneficial in the automotive applications because it has the potential to allow for a deeper pack discharge before the undervoltage condition occurs. Passive balancing relies on dumping some power into a discharge resistor typically during charging only.

Yeah, I think the historical term active and passive (clipping) aren't quite applicable since Tesla's BMS (at least in the past) actively monitor voltages, but uses low cost bleed resistors, not active transfer. Especially since to do so across multiple modules requires an additional common rail.

Theoretically speaking any variance in thermal regulation will result in different rates of self-discharge and also cell degradation. Even though the Extender would be integrated into the cooling loop, because of a different physical placement and its different thermal capacity it will likely be a bit off most of the time. Maybe they'll have a "precharge" style circuit to account for that. Won't know for sure what clever tricks they'll use until somebody takes one apart

Yeah, total system performance will be limited by the hottest or cold section.
I think degradation is only an issue is if causes enough current imbalance to overload one of the packs. Otherwise, it's just a matter of how many usable kWh the unit has.

 

[cybercricket]

Yeah, I think the historical term active and passive (clipping) aren't quite applicable since Tesla's BMS (at least in the past) actively monitor voltages, but uses low cost bleed resistors, not active transfer. Especially since to do so across multiple modules requires an additional common rail.

Understood, so then I was incorrect in my assumption. As for the common rail that's not really true - energy transfer can be done on adjacent groups only, and that will be sufficient to propagate through the entire string. Even distributed BMSes don't typically implement per-cell control circuitry, instead specialized ICs are used to tap into a dozen or so cell groups with overlap to the adjacent groups.

 

[mongo]

Understood, so then I was incorrect in my assumption. As for the common rail that's not really true - energy transfer can be done on adjacent groups only, and that will be sufficient to propagate through the entire string. Even distributed BMSes don't typically implement per-cell control circuitry, instead specialized ICs are used to tap into a dozen or so cell groups with overlap to the adjacent groups.

Yeah, energy transfer within a module can bucket brigade or similar, but to energy transfer between the four 200V modules on Cybertruck would require some interlink.

 

[txtravwill]

I doubt the range extender will ever be released. Removed from configurator already. If lucky, they will provide refunds.

 

[Virtualarry]

Yeah, energy transfer within a module can bucket brigade or similar, but to energy transfer between the four 200V modules on Cybertruck would require some interlink.

Does the battery chemistry have to be the same in extender?
Can the extender have self contained cooling or double the parts so not feasible unless somehow it is a different type of self contained pack?

 

[mongo]

Does the battery chemistry have to be the same in extender?
Can the extender have self contained cooling or double the parts so not feasible unless somehow it is a different type of self contained pack?

It makes the most sense to use the same chemistry. Otherwise, you run into issues with SOC vs voltage. Adding a DC-DC converter to interface two different battery systems adds inefficiency and could limit charging speed.
Self contained thermal system duplicates a lot of stuff. I expect the installation includes new power cables and cooling lines to use the existing systems.

 

[cybercricket]

It makes the most sense to use the same chemistry. Otherwise, you run into issues with SOC vs voltage. Adding a DC-DC converter to interface two different battery systems adds inefficiency and could limit charging speed.
Self contained thermal system duplicates a lot of stuff. I expect the installation includes new power cables and cooling lines to use the existing systems.

The interesting thing is that if I recall correctly the Extender was supposed to be 50kWh or so, which makes me curious how that would work with the overall battery topology. Seems like if the main pack was two parallel strings of 61.5kWh, then it would be easy to parallel the Extender to them if it was also 61.5kWh.

Sponsored