I spoke to the engineers at N-Power Energy Limited (NPE, global distributor) and E-One Moli (Molicel) regarding high-current pulsed charging of the P42A cell. This information can be used to help set regen current limits for a P42A pack.
E-One Moli really stepped up to the plate and sent over the document below.
WARNING: I cannot recommend using any charging current level above the P42A’s rating of 8.4A per cell. The numbers below are as supplied by NPE and E-One Moli. Use of any of these numbers is done AT YOUR OWN RISK! I have not tested any of them and cannot guarantee anything regarding their (relatively) safe use.
The following numbers are pulsed charging and discharging power levels per cell. They are based on the charge voltage (voltage of the cell) and the charge current level and vary depending on the state-of-charge (SOC) of the cell and the temperature.
As you can see the forced charged current limit is significantly lower at high SOC’s. NPE and E-One Moli said that limiting the charge voltage to 4.0V-4.1V would help to reduce cell damage versus using 4.2V.
Yea, it shocked me at how much lower it is. Normal charging has very low current levels at high SOC’s but forced regen charging could significantly exceed normal charging levels.
I wonder if ESC manufacturers are aware of these high-SOC limits.
Makes sense, just look at a Model 3 charging curve, charging at 250 kW to around 20-25% state of charge in a 75 kWh battery, and tapering way down as it increase. As far as I understand lithium plating is the issue
What doesn’t makes much sense is they saying that higher temperature should decrease the pulsed charge rate. Higher temperature increase the chemical reactions speed, again, looks at Tesla, to those 250 kW charging rate to happen the cells have to be at 50-60°C
Would be cool to develop a standard skate cycle curve, and use a few version of it for long term cycling, for example, multiple maximum regen power to see it’s effects
I know, not feasible for anyone to do it, easily needing 100+ of cells and testing equipment to account for all possible variables
Probably yes, but I like the @b264 philosophy, braking as strong as needed and we can leave with short battery life due to it
E-One Moli sent over a detailed document for max per-cell power levels for pulsed charge and discharge current at different temperatures…incredible data. Thank you E-One Moli and NPE for your quick response to my request!
Not yet. They really busted a nut to get me this so quickly so I didn’t want to follow up with questions right away. I will ask about the criteria used to set these power levels though.
Just a few guesses, a really long term and complex study looking a cycle life, temperature rise over a bunch of these pulses back to back or the most likely, using the actual electro-chemical model for the cells and meeting a bunch of different criteria at the same time for a peak power to be considered ok to use
Thank you for this rare data, I don’t think it was ever discussed before bit worrisome to see how little regen can be taken at high soc
Maybe adding brake chopper to compensate close to full charge makes more sense now
Btw is it ok to share with us the doc? It mentions Confidential / Internal use only, maybe you don’t wanna post it as is and make an extract or a copy instead? (so you don’t get backfire from your contacts and stay tuned lol)
It is approved for sharing online. They requested the copyright text that I added but I did not delete the Confidential stamp as it is part of the original document.
An update…
E-One Moli, NPE, and I have been discussing the criteria used to set these pulsed charge and discharge power limits.
They are the max discharge/charge power levels and pulse lengths you can run at, based on state-of-charge, before the cell reaches its recommended limit of 4.2V (during charge) or 2.5V (during discharge).
According to E-One Moli, if the cell is kept within these voltage limits there is no expected impact on cycle life, capacity, or internal resistance when running at these power levels for their matching pulse length limits.
This makes the data even better than I thought for calculating max regen current levels for a pack.
I had missed this reply somehow, pretty simple actually, and can be limited by the vesc easily using the cutout values, not that I recommend, my life before cell life as always
Yea, interesting paper! LCO and LFP chemistry did a lot better but NMC (like what some of our cells are) could get a couple percent or a bit better cycle life.
The pulse charging in that paper was done at 100Hz to 2000HZ though with the higher frequencies being more effective. Regen is continuous (though short period) DC current but we could add another circuit to “chop it up” into pulses.
IMO the biggest issue would be that half (average) of the regen current, between the pulses fed to the pack, would need to be fed to resistors or another load. The circuit starts getting pretty big and hot for the bit of possible cycle life advantage.
Regular charging with high frequency pulsing could definitely be worth exploring if the circuit cost could be brought down low enough. We might only get a bit of benefit for the twice-as-long charge period (unless boosting the charging current 2x) but that’s still something!
bit worrisome to see how little regen can be taken at high soc
