Pics/Info: Inside the battery pack

[spaceballs]

wk057 this is a Awesome thread!

BTW there are gid tie inverters that work at ~400v that might work so you shouldn't have to break apart the existing pack. Also there are cheap china grid tie inverters that you can put in series that would be capable of handing the ~400v, but you would have to build a way to disconnect them when the battery pack got too low.

Saw a few chats on the battery balancing, is anything like the Roadster battery balancing the BMB (which I've been fighting against for a few days, but won). There is a power darlington transistors connected to an surface mount bleed resistor ~40 ohms (1210 size I think), across every brick of cells.

Again great post!

 

[magnet]

Can you get a picture of the negative terminal of an individual cell?

The Rav4EV uses the 2900 mAh NCR18650PD with a C shaped vent designed to break the bond wire on the terminal when it opens. The C shaped vent is pictured in the Tesla patent but more recent patents also describe a circle shaped vent.

Pansonic recently released the NCR18650BE cell which is also an NCA cell rated at 2C max just like the PD cell, but rated at 3300 mAh! This cell has the same triangular postive terminal as the PD cell but has a circular vent. No other Panasonic cells besides the PD and BE have negative terminal vents! The nominal voltage of a NCA cell is 3.65V so a 3.3 Ah cell would result in a 85.5 kWh pack

The 3400 mAh NCR18650B is a high energy cell not suitable for EV use. The latest high energy cell is the NCR18650G rated at 3600 mAh. Neither of these cells has the Tesla vent on the neg terminal.

Tesla does not use the BMS to balance the pack. They use active impedance control to shuttle charge between bricks. It is always working and it much safer than a typical BMS (battery murder system)

 

[arg]

BTW there are gid tie inverters that work at ~400v that might work so you shouldn't have to break apart the existing pack. Also there are cheap china grid tie inverters that you can put in series that would be capable of handing the ~400v, but you would have to build a way to disconnect them when the battery pack got too low.

Grid tie inverters probably won't do the job since they aren't designed to regulate their output (they are designed to deliver everything they can get from the input to the output and rely on the grid to regulate voltage).
Maybe a strategy could be to buy one and replace the control part of the electronics - keeping essentially just the mechanical arrangements and power transistors, but that's the hardest part to prototype safely.
Is there a separate thread for this off-grid stuff?

Saw a few chats on the battery balancing, is anything like the Roadster battery balancing the BMB (which I've been fighting against for a few days, but won). There is a power darlington transistors connected to an surface mount bleed resistor ~400 ohms (1210 size I think), across every brick of cells.

That's only 10mA or so, which seems insufficient even given the slow balancing we see.

Using the Model S numbers, a 1% imbalance on the worst brick (which equates to 1% capacity loss on the pack as a whole, quite a commonly seen amount of imbalance) is 1% of (74 * 3400 mAh) = 2516mAh, which would take 250 hours to rebalance at 10mA.

1210 resistors are typically rated 1/3 W, which in this case would allow about 80mA. Are you sure that resistor wasn't 47R (marked 470) rather than 470R (marked 471)?

 

[spaceballs]

Grid tie inverters probably won't do the job since they aren't designed to regulate their output ...

Depends on model some of them can be variable. Also the cheap Chinese ones (I have two) will ramp on it's current draw up depending on the input voltage.
Opps looks like he doing off grid, I thought it was on grid solar system, so yeah no go on Grid tie. I would just use HV solar inverter then.

... 1210 resistors are typically rated 1/3 W, which in this case would allow about 80mA. Are you sure that resistor wasn't 47R (marked 470) rather than 470R (marked 471)?

Well I'll let you decide, here is actual resistor next to SOT-223.

 

[arg]

Well I'll let you decide, here is actual resistor next to SOT-223.

That's 39R, so about 100mA while balancing. Using Model S numbers, that's then 25 hours to correct a 1% imbalance, somewhat less for a Roadster (for which I can't remember the cell counts & capacities).

Rather slow, but consistent with what people have reported (taking several days to cure severe imbalance).

 

[scaesare]

Tesla does not use the BMS to balance the pack. They use active impedance control to shuttle charge between bricks. It is always working and it much safer than a typical BMS (battery murder system)

Now that's interesting. It does however seem to be at odds with the reports of folks suggesting that unless they do a full charge (or several) the packs seem out of balance... and the final 1% of charging can take up to several hours. The assumptions has been that's when balancing was occurring... but again that was just an assumption based on some increased range afterward.

What's your source for the active-impedance balancing scheme on The S?

 

[djp]

The 3400 mAh NCR18650B is a high energy cell not suitable for EV use. The latest high energy cell is the NCR18650G rated at 3600 mAh. Neither of these cells has the Tesla vent on the neg terminal.

Tesla isn't using a stock Panasonic cell, they've heavily customized it for EV use. It's very possible that they're using a 3400 mAh cell with custom packaging and custom electrolyte additives.

 

[deonb]

Tesla isn't using a stock Panasonic cell, they've heavily customized it for EV use. It's very possible that they're using a 3400 mAh cell with custom packaging and custom electrolyte additives.

I guess there's no chance that someone left a big fat Panasonic NCR18650 label on one of the cells in the pack, wk057?

 

[tom66]

I think one of the cost savings Tesla are making is no wrapping or printing on the cells at all.

 

[wk057]

I think one of the cost savings Tesla are making is no wrapping or printing on the cells at all.

Definitely no printing on the cells that I could see. They *appear* to just be metal shelled with no wrapping at all, but don't hold me to that... could be a plastic coating but it wouldn't really make sense since they're locked in place anyway. And since the majority are in parallel electrically, it wouldn't even matter if the housings touched, electrically, due to some kind of damage.

 

[JRP3]

Would be interesting to do a full capacity charge/discharge cycle to answer once and for all the actual cell capacity. Kind of a pain in the butt of course. One possibility to explain the balancing issue, there are chemical cell additives that allow cells to self balance, I believe through a redox action, and bleed off excess charge over time without resistors. I have no idea if Tesla/Panasonic are using them.

 

[tom66]

In the Nat Geo Megafactories video, there is a red ring around the cells. This might be insulation or perhaps just an early design, as they were using early cells in pre production vehicles maybe?
https://www.youtube.com/watch?v=Y7V5DLl3L0o&noredirect=1

 

[magnet]

Now that's interesting. It does however seem to be at odds with the reports of folks suggesting that unless they do a full charge (or several) the packs seem out of balance... and the final 1% of charging can take up to several hours. The assumptions has been that's when balancing was occurring... but again that was just an assumption based on some increased range afterward.

What's your source for the active-impedance balancing scheme on The S?

Rob Ferber. He built the 18650 pack for the T-Zero at AC propulsion and then went on to found Tesla. Now he builds 18650 packs with no welds (flex pcb with foam cusion also like Neal Saiki). You will note Tesla is trying to patent flex PCB bus bars but with cooling on the neg terminal. Both the positive and negative connections will be made on the postive side of the cell on the next pack so the other side can be baseplate cooled. Much simpler than the current system with curvy aluminum pipes wrapped in Kapton. Also the current bond wire technique is senstive to vibration induced cracking of the bond wire. This means the pack is very heavy with all the cells glued in place. The Model 3 pack will be much more elegant and lighter than this over engineered tank pack.

 

[magnet]

Would be interesting to do a full capacity charge/discharge cycle to answer once and for all the actual cell capacity. Kind of a pain in the butt of course. One possibility to explain the balancing issue, there are chemical cell additives that allow cells to self balance, I believe through a redox action, and bleed off excess charge over time without resistors. I have no idea if Tesla/Panasonic are using them.

What you are describing is a redox shuttle additive not compatible with NCA cells.

 

[magnet]

Tesla isn't using a stock Panasonic cell, they've heavily customized it for EV use. It's very possible that they're using a 3400 mAh cell with custom packaging and custom electrolyte additives.

They are not heavily customized other than removing the wrapper and adding a vent to the negative terminal. My bet is Tesla uses the NCR18650BE cells in the Model S with a minimum capacity of 3300 mAh (closer to 3650). The grade B cells are resold as the BE cell with 3300 nominal rated capacity, 3200 min guaranteed. The grade A cells also have lower impedance so can easily handle the 10.5 A peak they see in the P85 (over the 2C max rating but Tesla has active cooling). You can even see a little dot on the BE cell negative in the exact location the bond wire would go.

 

[magnet]

In the Nat Geo Megafactories video, there is a red ring around the cells. This might be insulation or perhaps just an early design, as they were using early cells in pre production vehicles maybe?
https://www.youtube.com/watch?v=Y7V5DLl3L0o&noredirect=1

Yeah that was JB's idea. The whole case of a cell is connected to negative which is why you wrap the whole cell for safety but it kills heat tranfer. With proper mounting you only need a ring



The present invention eliminates the adverse effects of the dielectric material covering the cylindrical case of a conventional 18650 cell by eliminating this covering and replacing it with a ring-shaped dielectric material, the ring-shaped dielectric material not extending down or otherwise covering the cylindrical outer surface of the cell's casing. Accordingly, the ring-shaped dielectric material provides access to the battery terminal while preventing shorting between the terminal and the edge of the cell casing. This design significantly improves cell heat transfer efficiency while providing a better surface, i.e., the bare cell casing, to which to bond, clamp, or otherwise attach to during cell integration within a battery pack or other package.

 

[qwk]

The highest cell voltage during charging that I've seen is 4100mv, but it definitely goes below 3000mv. I ran them down to 2800, with voltage sags under power to 2600mv.

I suspect contactor opening is when any cell group hits 2500mv.

This is all data from the BMS in the Rav4 EV with Tesla 41.8kWh usable capacity battery. Another interesting observation is that I have not seen the cell voltages go above 4200mv during fully charged regen.

I can't speak for the compliance cars that Tesla had a part in, but the Roadster, and Model S BMS both follow the voltage limits I mentioned.

Roadster Technology - Power Control | Tesla Motors

 

[yobigd20]

theres an 85kW salvage up on ebay for $24.5k with a working battery pack. Tesla Model s 85 | eBay this might be good for someone else trying to do the same thing as you.

 

[tom66]

theres an 85kW salvage up on ebay for $24.5k with a working battery pack. Tesla Model s 85 | eBay this might be good for someone else trying to do the same thing as you.

If the car says "pull over safely" it's a fair bet in the very least that the main pack fuse is out. You can't be sure pack is ok unless it's connected to the car. The range readout is just an indication of the cell capacity at the time of the accident.

 

[yobigd20]

If the car says "pull over safely" it's a fair bet in the very least that the main pack fuse is out. You can't be sure pack is ok unless it's connected to the car. The range readout is just an indication of the cell capacity at the time of the accident.

Well if you looked at the eBay listing it is connected and powers on fine.