Replace Stock BMS with Remote Mounted Active Balancer (RS19)

[0000]

It was some really nice weather for riding in the PNW, and coincidentally, I got to experience tilt-back at the LVC for the first time. It was actually pretty wild, you get so dialed in to moving in tune with the wheel (especially after a couple hours ride).. then you noticed all of the sudden the wheel is very subtly not responding normally... are my pedals dipping back... is my wheel beeping or something... holy shit it's really tilting back. Wasn't expecting it at all, it was kind of startling at first, thinking is something broken! Anyway, so I walked 2 miles today...

On 4/17/2022 at 8:29 PM, goatman said:

yes, but if youre riding under load and the lvc trips at 3.0v, the at rest voltage of the cells will bounce back up to 3.3v.

Yeah I get that, today I definitely would've want to dial down my LVC to at least 3.2V so I could make it to the ****ing car, would've saved me a long walk and busted trolley handle! I also get why each EUC design team specifically chose these LVCs to limit voltage drift due to their crappy top-balancing solution and extend cycle life.

I rode the hell out of the tilt-back and got back on multiple times until it wasn't worth it. Ended up at 80V or 3.33V/cell when I checked it at home so their LVC is definitely working as intended.

On 4/17/2022 at 8:29 PM, goatman said:

with the cheap blinky blink active cell balancers, i would balance the pack down around 2.7v to 2.5v

I wouldn't even want to take my packs down to this level, let alone spend time balancing them there as it accelerates capacity fade unnecessarily. Here, I'll do you a solid - take a look at section 6.6 and 7.2 of this study on Aging of Lithium-Ion Batteries in Electric Vehicles (or you can parse the whole thing for the raw data). I'll quote the easy parts:

Quote

Cycle depth has been identified as a major influencing factor for battery degradation. With higher
cycle depths, capacity fade and resistance increase are aggravated. In particular, the resistance
increase has revealed a strong dependency on cycle depth and only a small dependency on SoC.
Moreover, the resistance increase is mainly caused by rising charge transfer resistances of the NCA
cathode. The increasing resistances lead to higher losses and a lower energy efficiency. This has to
be considered when designing cooling system for EV batteries. (Page 139)

and

Quote

The resistance increase of the lithium-ion cells examined in this thesis has originated largely from
rising charge transfer resistances of the NCA cathode. Particularly for large cycle depths and deep
discharging, the resistances have increased markedly. When reducing the typical cycle depth to 20–
40%, the resistance rise of an EV battery can be minimized. This maintains a higher power capability
of the battery also for an aged battery. (page 143)

edit: wrong emphasis, 20% and 40% DoD is pretty unrealistic for an EUC.

On 4/17/2022 at 8:29 PM, goatman said:

i dont need to balance those salvaged cell battery packs every time i charge, maybe every 50 charges id do it when id notice the at rest pack voltage had changed 0.2v from 69.7v to 69.5v (4.1v/17s packs) and if they had a programmable controller lvc that you could set for doing pack maintenance it would be nice and easy to do, like on my bikes

All I can say to that is you're doing some crazy shit with salvaged cells and manually balancing man. I wouldn't be too keen on busting open my case every time I wanted to ensure pack balance.

Edited April 19, 2022 by Vanturion

[Freestyler]

Something to keep in mind, is that 3.3v is not a universal empty cell (tiltback) value for all Begode wheels.

You can find the minimum voltage for each wheel (black motherboards) in the source code of my app here: https://github.com/freestyl3r/euc-dash/blob/main/js/begode.js#L7-L23

 

The minimum voltage should be easy to change in firmware, but I have not bothered in mine because I feel safer knowing that I have all this "juice" even when approaching low volt beeps.

Edited April 18, 2022 by Freestyler

[mrelwood]

 

11 hours ago, Vanturion said:

I wouldn't even want to take my packs down to this level, let alone spend time balancing them there as it accelerates capacity fade unnecessarily. Here, I'll do you a solid - take a look at section 6.6 and 7.2 of this study on Aging of Lithium-Ion Batteries in Electric Vehicles

Look at the graph scales though: at 200 cycles the capacity will decrease to 91% instead of 94%. 200 cycles equates to roughly 10’000 miles on a 1800Wh EUC. I’d call that a very small difference.

 Still, I don’t see the point in deflating the cells to 2.5-2.7V when the same gain can be achieved within the operating voltage range.

Balancing every 50th cycle though… you’d have the unbalanced packs burdened for quite a long time. Besides, 50 cycles equates to roughly 2500 miles per 1800Wh. I’d be sure to have the cells balanced at the very minimum every 10th cycle, and every single cycle for top balancing systems.

[0000]

10 hours ago, Freestyler said:

The minimum voltage should be easy to change in firmware, but I have not bothered in mine because I feel safer knowing that I have all this "juice" even when approaching low volt beeps.

Wait, do any of the apps/dashboards have the ability to change the LVC then??? I'd actually really like to set mine to 3.2V.

To mrelwood's point earlier, a 3.3 LVC really leaves a lot on the table with high-energy cells like the 50E. Just going from 3.3 to 3.2 LVC yields + 16% more capacity at 5A * 4P * 80V = 1600W, which is still a relatively high amount of cruising power. Not to mention, we're still in the linear part of the discharge curve, so I would say that the threshold is overly conservative even if it protects better against long-term capacity fade, which itself is hard to characterize outside of controlled bench tests.

[0000]

2 hours ago, mrelwood said:

Look at the graph scales though: at 200 cycles the capacity will decrease to 91% instead of 94%. 200 cycles equates to roughly 10’000 miles on a 1800Wh EUC. I’d call that a very small difference.

Hmm, which graph/page are you looking at?

2 hours ago, mrelwood said:

Balancing every 50th cycle though… you’d have the unbalanced packs burdened for quite a long time. Besides, 50 cycles equates to roughly 2500 miles per 1800Wh. I’d be sure to have the cells balanced at the very minimum every 10th cycle, and every single cycle for top balancing systems.

I know it's been said before in other threads, but totally agree with this. Resistor bleed balancing current is much to slow to not ensure balancing is working every charge cycle.

[Freestyler]

20 hours ago, Vanturion said:

Wait, do any of the apps/dashboards have the ability to change the LVC then??? I'd actually really like to set mine to 3.2V.

If you have an RS with a black controller, then the limit is at 3.25v already.

If you want a 3.2v limit and have a black controller, I can cook you a custom firmware if you'd like.

Edited April 19, 2022 by Freestyler

[0000]

9 hours ago, Freestyler said:

If you have an RS with a black controller, then the limit is at 3.25v already.

If you want a 3.2v limit and have a black controller, I can cook you a custom firmware of you'd like.

I do have the black controller, and that's a kind offer, thanks Freestyler. Thinking on it some more, if my wheel's LVC is actually at 3.25V vs the 3.3V I thought--given voltage sag, I think I'd want just a little bit more out of it with LVC set at 3.18V for range emergencies. Hopefully with going that much lower it wouldn't induce much more voltage drift over time for the top-balancing BMS to patch up.

In any case, I'm about to take my wheel on a bit of a touring road trip, so I'm not keen on messing with the firmware beforehand in case anything goes wrong so if you don't get to it anytime soon, no worries there.

[mrelwood]

On 4/18/2022 at 9:49 PM, Vanturion said:

Hmm, which graph/page are you looking at?

It was in the mentioned section 6.6 in the (very long) document you linked to. I’ve actually seen similar results and graph scales in another battery life test as well.

Edited April 19, 2022 by mrelwood

[0000]

9 hours ago, mrelwood said:

It was in the mentioned section 6.6 in the (very long) document you linked to

Haha yeah, I hope you didn't read the whole thing! Unless you're a researcher, I find it best to just skip to the results and conclusions. Anyway, I think you're talking about the graph on page 127 -- if you don't mind, I'll point out a few things as I understand them and maybe they'll be helpful to some in gaining some insight on the nuances of lithium-ion battery performance:

In this comparison, they're only looking at up to 61% Depth of Discharge (curve in black) whereas if we were to consider EUC LVC, with a 3.3V cutoff threshold we're probably at 80% depth of discharge or higher when discharged from the full 4.2V. Notice the exponential rise in battery internal resistance on graph (b) of the 61% curve, the deeper you get into the depth of discharge, the greater the capacity fade too as the relationship is not linear. This is what they were pointing out in section 6.6 and 7.2.

I've exaggerated a "guesstimated" curve in red as to what an 80% DoD might look while including the higher SoC effect too.

In each of their tests above, they are not charging to the full 4.2V which also affects relative capacity remaining over EFC. Example:

Here you can see both how the state of charge and charge current effects the capacity fade the cycle life. In other words, at an appropriate charge rate of C/2 or less, just charging to 4.1V instead of 4.2V yields about an extra 150 cycles alone before hitting the 75% relative capacity remaining lower limit.

To add another compounding factor, you could also examine differences in the rate of discharge:

I'm not sure what they used in the first chart comparing depth of discharge, but here you can see some pretty big differences in 1A vs 2A rates for these cells.

Now all that said, none of the specifics in this study should be used to make direct comparisons with our EUC batteries unless they are specifically using the cells that they used in the study, Panasonic NCR18650PDs. While there are general principles that hold true among all lithium-ion batteries in order to limit or slow capacity degradation over cycles and time, as I understand it, we can't make accurate direct comparisons without performing equivalent tests as the blend of battery chemistries between manufacturers and models tends to yield unpredictable comparative performance.

In other words, you can't predict the exact behavior or capacity fade of your Samsung 50Es, Molicel P42As, or whatever other cells without testing and measuring for those specific cells. However, the non-linear relationship between cycling with higher DoD having a greater deleterious effect to cycle life than charging to full holds true between all lithium-ions AFAIK. So, don't let your cell-phone run to zero...

 

[mrelwood]

 @Vanturion, you are of course correct in everything you say, but what is still missing is the relevance.

19 minutes ago, Vanturion said:

just charging to 4.1V instead of 4.2V yields about an extra 150 cycles alone before hitting the 75%

Again, only after reaching the amount of charge cycles that no EUC rider has seemed to reach yet, anywhere in the world.

 There aren’t many riders in my country who have ridden further than I have on a single battery, and even I only reached to roughly 200 charge cycles (14000km @1600Wh). And the wheel before that had it’s first battery failure at roughly 80 cycles, due to insufficient balancing as I generally only charged it up to 80% and balanced only every 10th charge.

 What happens at around 500 cycles is simply not relevant to us.

[0000]

15 minutes ago, mrelwood said:

 @Vanturion, you are of course correct in everything you say, but what is still missing is the relevance.

Again, only after reaching the amount of charge cycles that no EUC rider has seemed to reach yet, anywhere in the world.

 There aren’t many riders in my country who have ridden further than I have on a single battery, and even I only reached to roughly 200 charge cycles (14000km @1600Wh). And the wheel before that had it’s first battery failure at roughly 80 cycles, due to insufficient balancing as I generally only charged it up to 80% and balanced only every 10th charge.

 What happens at around 500 cycles is simply not relevant to us.

What part of battery enthusiast and min/maxing personality trait *^deficiency did you not understand! J/K.

Yeah agree, default LVC and 4.2V top-end is good enough for basically all riders as long as they're practicing other good battery management techniques like not leaving batteries at 100% for long periods of time unused, not charging (or possibly even heavy on braking/regen) when batteries are cold, and not braking down a hill fresh off the charger. Many cell lines have improved over time and cycle life is pretty good for many of these type of performance cells like the 50E, P42A, and 40T.

[0000]

Extracted my damaged trolley handle today and figured I'd measure the RS empty cavity dimensions while it was apart:

So the 4A Active Balancer dimensions are listed as 130 x 140 x 25 mm and I measured roughly 127 x 130 x 35 mm in red. So pretty much off by a half inch or so like I originally thought, go figure. It looks like one might be able to make it fit by cutting into the inside shell wall and shaving the top corner circled on the right, but it'd be a tight fit.

I'm going to hold off on this unless I hear of any issues with the LiTech packs in the future, but it'd be an interesting project for someone who was making or already planning on modifying their own battery packs for whatever reason.