What is a normal cell delta V?

[evan]

IMO 3.3V open circuit (after resting) is going a bit emptier than you’d want to for most applications. It shows the state of charge is really <5%. I know we want to get the most flight time out of the pack but that’s not great for the lifespan, and it means you had really no reserve.
At 3.0V I’d start to consider the cells are at risk of damage.
That said I don’t know exactly what cells are used for the SP140 - anyone know?

Why is 1 group lower? It’s probably just the one with the lowest capacity, and the cells are “top balanced”. The BMS should have cut off discharge based on this group. Generally you do see more spread in voltage at “empty”, although the rest are really tight, as you can see the voltage takes a knee at that level of SoC so a small difference could explain it.

What Bratwurst said is worth checking though - if you did have one cell disconnected it would show up like this, although it should not be able to happen and would be a major issue by itself.

From ShieldSquare Captcha

[skduncan560 (Scott)]

Good question; hadn’t thought to check that before. With a full charge, my delta is 0.05V, with cell 5 being nearly exactly the same as everything else.

I haven’t experienced just the motor cutout; only a full cutout of the entire system on 3 of 3 flights where I intended to find the limit. You’re saying that you’ve had the motor cutout but the system is re-engagable just by reapplying throttle maybe at a lower power setting? That seems like the ideal failure mode — each time mine has cut the motor (reapplying throttle does nothing), then the screen says “BATTERY DEAD” and then within a second or two blanks out.

Yes, I’m sure I could get an hour on a properly sized non-reflex glider. Ironically I sold my Rush 5 to get the funds for the SP140, haha. The Viper is just as efficient as any other appropriately sized reflex glider I’ve flown, but the Rush was a full 30% better on fuel consumption on my 2-stroke from an endurance perspective.

[skduncan560 (Scott)]

Thanks for chiming in, @Pdwhite ! As suggested, I adjusted the “charging start limit” to 3.0V from 3.8V and flew to cutoff again yesterday with pretty much exactly the same result. Prior to flight, the maximum delta between cells when fully charged was 0.05V with cell 5 being nearly the same as everything else, but when I hooked it back up to the charger post-flight, cell 5 was again 0.3V lower than everything else, with cell 5 being the sole outlier. As @evan mentioned, if the cell delta V increases as the battery gets closer to dying and if no one’s actually flying to cutout, then maybe this is normal?

That being said, I was watching one of your videos where you said that you usually don’t fly below 70V and that the battery protects itself with a cutout at 60V but I was well above each of those limits when my battery cutout. Has the cutout limit changed since you made that video (the 50 min typical flight one)? Would flying a more efficient glider let you dip deeper into that voltage range before cutout? Also, what exactly did I change by adjusting the “charging start limit”? I really know nearly nothing about these batteries and presumably the settings are there for a reason (?). And finally, is it actually that bad for the battery to be run to cutout? I assumed that the cutout was set to protect the battery before I could hurt it?

Thanks for putting together such a cool project! Im sure you’re swamped with current orders and such, but it would be very cool to get a video or something on what different BMS settings mean and what we can safely do with it at some point.

[evan]

SK, from your description, the BMS shut the power off, and that will have been based on individual cell group voltage (probably your group 5) rather than total pack voltage.

That’s the whole reason to have a BMS, because otherwise you wouldn’t know that one group was empty, and the ESC would just keep on pulling current - it will quickly go down below 2V (killing it) or even be forced to a negative voltage by the other cells in series. That’s when you get a “thermal event”.

So the good news is the BMS is doing its job although, the way it cuts off power without warning isn’t ideal.

What you need to figure out is, is one of your cell groups significantly lower capacity, or less charged than the rest, which is shortening your run time?

To check that I would do a ground test. Secure the frame to something, and hold the current steady at (say) 50A. Time how long it runs, watch the BMS app and verify the reason it’s shutting off, and record how long you get.

For example if you got 30 minutes, you know you’re getting 25Ah at that load.

If that’s below the expected range, then the question is, is that cell group weaker, or just not fully charged for some reason? That’s what the balancing function in the BMS is for. However it can only compensate by bleeding off a very small amount of power from the highest cells (all the rest in this case!) to bring them down to match the lowest one, while charging (and above a set voltage as Paul explained).

That can take a very long time to balance out, and you’re hammering that low cell group every time you use it, which will actually cause it to lose capacity.

One remedy is to manually balance the cells (just charge the affected group with a PSU when the normal charge is complete), and if it really was just unbalanced then the problem will be solved.

[GliderPilot (Paul Butler)]

@evan

To check that I would do a ground test. Secure the frame to something, and hold the current steady at (say) 50A. Time how long it runs, watch the BMS app and verify the reason it’s shutting off

I would not recommend a ground test… That can be dangerous if not properly secured. It’s a lot easier to attach your phone somewhere to monitor the app during flight than to attach your motor to the ground. The rest of what you said is accurate and very good advice.

@evan

What you need to figure out is, is one of your cell groups significantly lower capacity, or less charged than the rest, which is shortening your run time?

His description of has last flight answers this question. When the pack was fully charged it was balanced within 0.05V. When he landed he had one cell group 0.3V lower than the rest. That is not a normal delta at all and since it started balanced it points to a capacity issue.

@skduncan560

if the cell delta V increases as the battery gets closer to dying and if no one’s actually flying to cutout, then maybe this is normal?

When I fly to cutout or near cut out my cell groups maintain their balance which means all of my cells are equally healthy and have equal capacity. It is not normal to be 0.3V different at the end of a flight. That’s quite excessive, specially if you start with all cells within 0.05V.

@skduncan560

Also, what exactly did I change by adjusting the “charging start limit”?

The BMS only balances the cells during charge and normally towards at the end of the charge. What you did was initiate the balance earlier so that it had more time to bring the low cell group up. It just helped to guarantee that your delta was good at the end of the charge and the start of your next flight. Unfortunately it got off significantly by the end of the flight. Therefore it’s not a balance issue… you have a bad cell group. The BMS will continue to protect that cell group by shutting down the pack. The other 23 groups aren’t getting fully utilized.

[GliderPilot (Paul Butler)]

@skduncan560

You’re saying that you’ve had the motor cutout but the system is re-engagable just by reapplying throttle maybe at a lower power setting?

Yes. The voltage of my pack got low enough that the ESC cut because of low pack voltage. Definitely a more ideal warning, however the pack was so empty that I couldn’t maintain level flight.

What you experienced was the BMS shutting the battery down. It’s basically like unplugging the battery.

I recommend flying with your phone mounted so you can see the lowest cell group with the app. Then you will have that visual warning before it cuts. That’s the best “fuel gage”.

[evan]

GP, two comments to add:

It would be near impossible to maintain an exact current setting while flying. Surely it’s possible to secure the thing properly with a bit of care and some 2x4’s?

Second, as you say it probably isn’t a balance issue as they seem close enough to start with. But the other cells are almost empty at end of charge too, going by the resting voltage, from my experience. He could just be using more energy than expected.

[GliderPilot (Paul Butler)]

@evan

It would be near impossible to maintain an exact current setting while flying

We do have cruise control.

I agree that analyzing on the ground can have its benefits but the app is a pretty powerful tool that can be used in the air and there aren’t other instruments needed which would require being grounded.

Yes, there are safe ways to secure the motor on the ground but there are a lot of unsafe ways to do it as well and I’m sure those ways would be discovered by someone if we go around recommending that. With gas motors, ground starts are highly frowned upon.

I also agree that if the cells are close to empty they will drop off faster and potentially make a bigger delta at the end if one group hits that steep part of the curve first. A simple check would be to fly with the app and see what pack voltage the delta starts to become excessive. Or he could fly till the pack is low but stop earlier than he has been stopping and see what the resting voltages are. I bet in his case it will start pretty early based on how severe the delta has been.

[evan]

Fair enough, I guess what I’m getting at is it would be useful to record exactly how much Ah he is getting from the pack, if it’s in question at all.

The instrumentation supplied can’t really do that so it’s why I suggest a fixed current run down test. Obviously to be done safely! I think I will do that when I get mine anyway (and/or add instrumentation)

[GliderPilot (Paul Butler)]

That’s cool… I’m sure you understand more about it than most people. What instrumentation are you wanting to add?

The throttle controller displays kWh from the ESC. Not sure how accurate it is but it should be pretty accurate if you operate the throttle smoothly. Also, the ESC records a couple hours of data which can be downloaded.

[evan]

Building batteries and BMSs, and integrating them is my day job so please forgive the nerdery!

I would like to have an accurate Ah-based percentage display on the controller. Plus a warning and few seconds grace before you hit empty. As you said it’s already displaying kW and kWh so the ESC data contains enough to do this without adding any hardware.

The BMS app looks good for diagnosing things, but ideally end users shouldn’t need to be in there worrying about balancing and cut off voltages.

Maybe I’m making too much out of this but I read a lot of stories with people having cut-outs in the air and kind of guessing where the limits are. SP140 will be my first paramotor after paragliding and I won’t be as competent a pilot as some of you guys, so I’d rather work around the avoidable issues like that. If I can contribute to the project usefully, I will.

[skduncan560 (Scott)]

@evan , thank you for the very detailed explanation. Let me just make sure I understand what you’re saying: By my math, the battery should have a capacity for 40Ah, and the implication of what you’re saying is that Ah is the best measure of the battery’s capacity (do I understand that part right?). If I can run it at 50A for an 48 minutes (40Ah/50A x 60 min/hr= 48 min), then my battery capacity is fine and my expectations are off. If I can’t get the full 40Ah, then there’s something wrong with the battery. Is that right? Should I expect to get the full rated Ah of the battery? I’ve been getting about 3.3-3.4 kWh of battery but maybe I souldn’t expect to get 4 and Ah is a better predictor of capacity?

As far as testing, for the sake of control, it would be ideal to ground test, but I’d have to go out into the middle of nowhere to not annoy the neighbors, and at that point I might as well fly, haha! So maybe I could just do a 500’ climb, then set cruise control and run it out which I imagine would be close enough to get an idea of where I’m at. And as far as losing power, as long as I know it’s coming, I’m pretty comfortable running it to cutout from a flying perspective. The last two flights, I suspected cutout would happen sometime after 30 minutes, so I made sure to just loiter a few hundred feet over the LZ until it did. Even on my gas unit, I always do power off landings, sometimes with a dry tank.

The next thing you mention is that it’s possible to rebalance the lower cell group just by a series of discharge-recharge cycles? This would be great if I could just have the pack “self-repair” rather than have to send it in or whatever I’d have to do to get it fixed, but that seems kind of like magic to me that that could do it.

However, it’s still curious to me that the “bad” cell group delta is ~0.05V from everything else when fully charged and even when I brought the battery up to a storage voltage of 86.4V after my last flight running to cutout. Although, as you guys seem to be saying, it’s possible that the BMS is able to even it out when charging but not when discharging, if I understand correctly. In that case, I think I’ll try what @GliderPilot suggests and monitor the BMS in flight and see what that cell is doing. Obviously, I don’t want to be watching that app every flight, but for a test flight to see what’s going on, it’s easy enough. Is there anything I’d be watching for other than seeing whether that cell group is low throughout the flight or just at the end?

[skduncan560 (Scott)]

@GliderPilot

The BMS will continue to protect that cell group by shutting down the pack. The other 23 groups aren’t getting fully utilized.

This is the thing I’m worried about from a flying perspective. Obviously, I want the BMS to keep me from hurting anything, but it would be a shame to have one parasitic cell dragging down my flight time when there’s still juice left in every other cell group.

@GliderPilot

Yes. The voltage of my pack got low enough that the ESC cut because of low pack voltage. Definitely a more ideal warning, however the pack was so empty that I couldn’t maintain level flight.

Sure it would be nice to maintain level flight, but even if it helps you hobble to a better LZ I’d prefer that to just shutting off altogether. Interesting to know that that’s what it’s supposed to do; I wouldn’t have known otherwise! This suggests that there is a genuine issue with that cell group. Good suggestion on flying with the app. I’m going to give that a try on my next flight and see what I find!

Thanks @GliderPilot and @evan for your guys’ expertise and help! I thought that coming to the world of electric would make for a simpler time, but I’m finding out that it’s a lot more complicated (or at least different) than just rebuilding a carb or replacing parts on a 2 stroke!

[evan]

@skduncan560

Let me just make sure I understand what you’re saying: By my math, the battery should have a capacity for 40Ah, and the implication of what you’re saying is that Ah is the best measure of the battery’s capacity (do I understand that part right?). If I can run it at 50A for an 48 minutes (40Ah/50A x 60 min/hr= 48 min), then my battery capacity is fine and my expectations are off. If I can’t get the full 40Ah, then there’s something wrong with the battery. Is that right? Should I expect to get the full rated Ah of the battery?

Basically correct, I think. It would be good to have a benchmark from a known good pack to compare against - you won’t get exactly the rated capacity at higher C rates but it shouldn’t be far off.

You do have a kWh reading but that includes the voltage which is heavily affected by resistance, load, temperature, etc, so it will give a less clear comparison.

On closer reading of the info you’ve posted so far, I’d probably agree with Gliderpilot that this isn’t a balance problem, but first thing to do is establish if you’re really lacking capacity or not, and depending on that we could try to determine the cause.
My guess is that you actually are getting close to full usable capacity.

Another clue will be, if you can watch that cell voltage throughout the run - if it’s consistently sagging below the rest that would indicate a problem too.

In the end, the electric system definitely can be fly and forget, and I’m sure it will get there. I think at the moment it’s still slightly early days so it’s a bit more hands on, and more like the R/C flying experience.

[bratwurst (Thomas)]

before i told you about my point of view regarding error analysis. now i would like to say something about batteries in general. batteries in this performance class have a relatively high value and also danger potential in general. which of course can be very well mastered. but what surprises me in comparison to european batteries from companies is:

during the performance check that a new battery usually has to pass under full load and thermo check before delivery, a higher cell drift would be noticed immediately.

a bms programmed in this performance class would have to immediately go to fail with such a high drift of a cell group and prohibit a restart. only the manufacturer can “unlock” the battery again. or a service company that checks the battery.

since this is of course an opensource project where customers are allowed to control the bms themselves, it will probably not be possible for the manufacturer to give a warranty because he does not know what the customer is doing with his product. so of course i understand that the manufacturer does not immediately write with such a posting as here: please send immediately, we must analyze the battery for safety reasons and warranty immediately in the factory.

but i think that openppg will surely exchange the battery if indeed a mistake happened during the manufacturing process. this is always possible, no matter which manufacturer.

[skduncan560 (Scott)]

Yesterday I did some data collection. As suggested, the bad cell seems to be what’s driving the cutout, not the total pack voltage. As soon as the lowest cell hit 2.5V, the system completely shut down. I’m guessing that this bad cell cost me 5-10 minutes of flight time, but not really sure of all the factors at play for that. It was also suggested to measure Ah instead whatever the percentage scale measures or kWh displayed to determine what capacity I’m actually getting out of the battery. Taking an amperage measure every minute, I came up with a consumption of 35 Ah. I have no idea how close to 40 Ah I should actually expect to get, but from the first chart it looks like I’m missing at least a little bit of juice due to the low cell.

Methodology:

My flight profile was a reduced power climb to 500’, then set the cruise control to maintain level flight until battery cutout. I screen recorded the Smart BMS app for the duration of the flight and then manually recorded that data post-flight every minute from the screen recorded video.

Known data anomalies:

• The zero time data point is with no load, immediately after plugging in the battery before flight.

• I didn’t realize I wasn’t actually screen recording until in flight, so the first real data point starts at 4 minutes.

• At the 53 minute mark, a battery fault message pops up on the Smart BMS app, moving voltage data for half of the cells out of view, so no data for those cells for the last 2 minutes.

[evan]

Great work Scott. Glad you’re getting a decent flight time first of all!
Seems that the low cell group has the lowest voltage for the full duration, so it does look like it’s got a lightly higher resistance as well as lower capacity.

However. Supposing you shut off at 3.0V which would have been a bit easier on the cell, you can see the other cells aren’t far behind at all - really only a couple of minutes in it, or like 3% of duration.

It might be more apparent because due to that difference you’re getting a BMS lock out instead of an ESC power reduction or restartable cut-off, which would feel less dramatic.

[skduncan560 (Scott)]

I was skeptical, so I extrapolated the data from the next 5 lowest cells and you’re 100% right. My “5-10 minutes” is actually closer to 2-3 minutes. I still want those 2-3 minutes, but it’s not nearly as big of a loss as I thought due to how rapid the drop-off is.

So my next question is: is there anything I can do to get the ESC power reduction before the BMS lockout in this case? Some setting to tweak?

Also, now that I have a good idea what’s going on at the end, I don’t really feel the need to run all the way to cutout. However, I still need to know what the limit is to maximize flight time (which was the point of running to cutout on these early flights in the first place). So suppose I want to end the flight by the time the low cell hits 3V. Could I monitor the BMS app at the end of the flight and see what controller voltage corresponds to the low cell reaching 3V? Does that correlate? Would the value wander over time? It would be great to be able to see the parameter closest to causing cutout (in this case the lowest cell voltage) on the controller; as it is now I feel like I’m kind of just guessing.

[Allex]

Yes, we are discussing it here:

SP140 - Battery low power cut off mechanism & thresold

For those who uses ant bms on their new packs you can go with this display where it will show you the lowest cell voltage during flight.

[evan]

@skduncan560

Could I monitor the BMS app at the end of the flight and see what controller voltage corresponds to the low cell reaching 3V? Does that correlate? Would the value wander over time?

Yes you can definitely do that. You can also see from your chart that when the total reaches (3.3V * 23 = 75V) you only have a few minutes left. At that point it would be great to get a warning, and you can look at the voltage on the handle (reported by the ESC) at least.

Will that change over time? Yes, somewhat, and it will vary with temperature and of course load. And if your weakest cell gets worse, it could still catch you out.
My intention would be to get an individual cell voltage based “elegant shut down” - i.e. a vibration or sound warning followed by a gradual cut of motor power. But also not need to hit that limit, instead plan to be back in your LZ based on an accurate % readout.
As I mention in that other thread that Allex linked, that is how other EVs work, so it’s a reasonable aim.