Ok, it sounds like i wasn’t reading you carefully enough, your concern is tripping the hard LVC, not just battery damage. My bad.
My personal strategy is to set my battery amps to the (independently) rated CDR of the cells, with the understanding that actually using that continuously is unhealthy because the ratings are generous, but that ESK8 can only use it on a pulsed basis. So that’s where I was coming from.
Maybe another strategy for you is to use a filtered voltage for LVC, similar to the Slow ABS cutoff setting? If the scenario is that a 60V pack can temporarily sag to 30V, it could wait a few seconds to bounce back, or ramp the current limit in gradually.
I think that would be the ideal solution, but I don’t think vesc tool supports this at the moment. Let me know if I am wrong though because I would really want to use this feature if it exists
In your experience how often has this happened? It definitely seems logical, since at low SOC’s and high discharges the cells can rapidly sag to well under the cutoff limits. If at 40A a cell at full charge sags to 3.5v, imagine what the same cell does at 3.5v rest…
That’s really good to know. Basically thermal runaway is very unlikely in our use case, and exceeding the 80°C cell rating drastically reduces the lifetime of the cell but isn’t a huge concern for thermal runaway? What immediate safety concerns are realistically at play when reaching or exceeding 80°C?
That’s the whole point. In this scenario we’re leaving it up to the rider to guess how hot everything is getting. With the battery amps set to an available 260A, perhaps a rider doing back to back burst draws of 80A+ per cell will rapidly exceed the 80°C cell rating, and with no system to tell the vesc that the pack is over its limit, cell damage may continue to accelerate. Ideally you want a temp sensor on the battery connected to the esc to act as a soft cuttoff, reducing amp draw as the cells approach 60-70°C. And there’s no guarantee that all cells IR’s are perfectly matched and thus overheat at the same rate.
These new cells are indeed amazing and it’s awesome to see @Skyart and @poastoast utilising them in racing. As @Jaws mentioned it would be really cool to see logs/graphs of the battery temperature during racing heats. (No puns intended).
I do the same as you, and assume that CDR is all I can use in practice because of sag. Sometimes slightly higher works, stepping up each time and monitoring the performance. For me, tabless cells are a great innovation because CDR is higher, not because PDR is higher. The risk is always there until an ESC has a way of managing voltage sag/pulse currents effectively.
I have not tested sag under load since years ago with p42a, but I would see sag from 4v down to 3v instantaneously for 2-4s in repeated hard pulls at 45a/cell. This would cut out. Even at 30a/cell I would see sag from 4v down to around 3.5-3.3v during a pull when pairing a small pack with large motors and having a large differential in phase:battery current, geared tall and with my weight. These conditions by design let me see what would happen. I couldn’t feel any power drop, until a cutout would happen, so going by feel doesn’t work. When using anything above CDR I would always exercise caution and take it in steps increasing each time. Also keeping in mind that your power output has to take into account voltage drop, it will only be as powerful as its voltage under load multiplied by the current.
@Skyart that’s why I mentioned using logs, not just for battery health but human health and the reputation of eskate as a whole. Stay safe.
Thermal runaway can easily happen with esk8, just not with very hard discharging IMO. Overcharging, charging at below 0°C, wildly unbalanced packs that result in reverse charging, physical damage, those can all cause thermal runaway.
Immediate safety concerns? Any of the above, at any temp.
Most li-ion “power” cells are rated to 80°C max (some 75°C) though so operation to that surface temp isn’t a big immediate safety issue to the manufacturers. But aging really speeds up when over about 60°C and things start breaking down internally up near 80°C.
Potentially that can eventually lead to failure of the cell. Probably not fire but that’s not impossible.
Personally, I’d be much more worried about pack balancing (to prevent reverse charging), proper charging, and retiring a pack if things get weird in any way. Temp just isn’t a huge issue IMO except for aging the pack.
However, this assumes a decently made pack that isn’t much hotter in the center than at the edges. That can unbalance a pack over time and require a very good BMS to protect the pack when it reaches the end of its life.
Good idea but….
The reading we get from the thermistors in our BMS’ can be pretty much useless in high current applications due to their incredibly slow response time. They’re great for ambient temps and tracking hours-long discharges but are too slow for anything else. Adding on the fact that they thermally couple very poorly to the cells we use (without a decent epoxy, properly placed) and that very few companies install them well and you’re left with having only a verrry rough idea of what your cell temps actually are.
Setting the temp cutoff, like you mentioned, a lot lower really helps with this though.
I agree, that would be great to see. But the data we get now would be pretty useless IMO…too low, maybe wildly so.
especially on that first section, they did exactly as i wanted and apart from lap 5 i don’t think i could be any more identical between laps in that section
I’ve got an oddly relevant chart. Coolwatts yt caught Bindingsmcgee’s eye, at least the title did, that long video.
100 minutes?! Appreciate effort, fumbles for easy button
Bindings doesn’t like ai tricks. I push AI shenanigans on strangers like pics of newborns I don’t have.
Vid transcript + all yt author included charts + prefer mooch cell stats over Oem. Distill all to a chart, relevant cells to appropriate column topics, 4-5 at max. Don’t hallucinate or guess (lol, I’m a prompt engineer), I use cells IRL.
If coolwatts stops by to this thread, I promise to support his patreon for that comment.
More to the topic at hand, I’ve got a strong bias to prefer known cell manufacturers. How established is EVE? How are they beating the multinational conglomerates to the punch? Eve rings a bell, barely.
That’s my thoughts exactly, even if a thermistor were to measure rapid changes accurately enough to be meaningful, it won’t account for each individual cell.
What can we do to increase the accuracy and response time of thermistors? Apart from the obvious. (Thermally adhereing the sensor to the cell itself)
It clearly gets more complicated the closer we use these cells near their limits. I found it crazy that in some of your testings some cells discharge to near their temperature limit of ~70°C, discharging stops and the temperature reading continues to rise to 80 to 90°C, even though current is no longer flowing. Which further illustrates your point about how delayed/innaccurate a temperature reading can be during rapid changes in discharge.
My understanding is that EVE is one of the most well known and established Chinese manufacturers, along with other reputable Chinese based companies BAK, EVE, Ampace (joint venture between CATL, ATL), Lishen etc.
Since we are talking about temp rise. My 21S4P P42A pack configured for 45A discharge per cell rises to about 55-58C after a hard racing session in 15C weather. I peak the full current nearly every good lap, but it’s only sustained for maybe a second or at most second and a half. Immediately after they see a similar pulse but for regen charging. But these temps are practice where I can do as many laps as I want, not actual race heats, where I need to wait half an hour between 2-3 minute heats.
I don’t think you need to worry about temp rise during an actual race, but during practice sessions it’s good to keep an eye on it if you push the cells very hard.
@Jaws I never got a cutoff like you mentioned pulling 45A out of P42A. I have a 0.3V difference between cutoff start and cutoff end for reference.
Do you know for sure that you are using 45a/cell? If it’s during T race under super short distances, that might be different since duty cycle could be lower and possibly the duration as well? Do you have any logs or even a screen recording of vesc tool?
My test battery was 12s3p and that allowed me to ensure it was easier to actually pull the power I set in the ESC. From memory the rest of the setup was v5 reachers and a spintend ESC in unlimited fw running bldc if you are curious (this was 2021-2022 times).
Back in the 2wd channel truck days I logged 160 battery amps at medium speed, 0.65V per cell sag. But I couldn’t push that setup nearly as hard, my battery to motor amps ratio along with the wobbly trucks and 9" tires meant I couldn’t reach full throttle at a large enough duty cycle to pull the full 180 amps.
On the new setup I hit full throttle at high speeds quite often.
Unfortunately I can’t show vesc logs because my Metr module is half broken, 5v keeps disconnecting all the time, I need to fix it before putting it back into the board. What I have is gps logs and it isn’t really evidence but I’m consistently pulling more acceleration than a dual hw max5 setup.
I can hit full throttle on T-race, reaching 57 km/h or 35.5mph on the straights.
That should definitely pull max battery amps, the duty cycle isn’t low.
I’ve got 520 motor amps total for acceleration, and 180 battery amps. Meaning about 35% duty cycle is already max battery amps. On the 160mm tires I am geared to about 70-75 km/h, so T-race should be about 70-75% duty cycle.
But I also raced some longer tracks on 9" tires, there I was also pulling full throttle at high speeds, reaching up to 70 km/h before I had to start braking.
This doesn’t really say anything, too many factors.
If what you say is true and you see such minimal sag, then I don’t know what to say. There must be some other variables or the duration is just too short. I don’t know, hard to say without hard logs.
Since this is a brick pack there’s less resistance coming from the series connections. I’ve got 0.2 nickel about 75-80 mm width and 30 mm length for all series connections minus the one in the back, that’s 3x12 awg. And the terminal wires are 8AWG soldered on in the full length of the nickel (this is done before spot welding of course).
The ESC is placed directly under the battery, keeping the terminal cables short as well.
But I agree that the time I was pulling this current in this log is very short. It would definitely be interesting to see logs for the current setup and possibly for longer tracks as well.
and the reputation of eskate as a whole. Stay safe.
