Set MPPT/BMS to routinely chare to 90% (not 100%)

[joninca]

I can find a plethora of information on how to program an MPPT+BMS to routinely charge an LFP battery bank to 100%. But, what settings are optimal to routinely charge to 90% (to preserve the life of the battery bank)? Is it as simple as changing all of the charge settings (float, bulk) to 53.6V (for 16s)?
My system has been in service for 3 years, and I have always used "all-in-one" hybrid inverters from Growatt. But after my third all-in-one inverter failed after a year (this time, the MPPT broke, the previous one the output capacity was reduced by half, and my first one just stopped working entirely) I have decided to switch from "all-in-one" to individual components (see below) so that I can replace individual components as they break. Everything is going great with default settings, but this charges to 100% every day.
New MPPT: PowMr M60-Pro: Default "L15" program charges to 100%, but I would prefer 90%. There is a USE ("user") program that I would like to use to get to 90% everyday.
BMS: Sephlos 150A BMS x4
Batteries: 44 kWh LFP, unbranded from auction, 4x 16s 48V (51.2V) banks
Inverter: PowMR 10.2 kW (can't use the built in MPPT because its max input current is 10A)
Solar: 48V 10A panels 3s8p
Thanks!

 

[sunshine_eggo]

Why would you prefer 90%? This isn't necessarily going to "preserve the life of the battery bank."

How much degradation have you experienced in three years?

 

[joninca]

I was referring to charts like the one copied below, which suggest that charging daily to 100% reduces lifespan. In practice, I haven't noticed an appreciable drop in battery capacity over the past 3 years, but I was always using the "US2" setting on the Growatt inverter to stop charging at 90%. The separate MPPT I am now using doesn't have an option like this (nor a BMS interface), and if I check on the batteries around 4PM, they have been hitting 100% per the BMS.
So, my thinking was that if I could manually change the MPPT settings to a lower charging voltage for bulk/float (maybe 53.2V?), I could maximize the lifespan of the battery bank. I only use about one-half to two-thirds of my battery capacity overnight, so there is currently no practical need for me to fully charge every day.
Let me know if I am not thinking about this correctly. The only nice thing about the all-in-one hybrid inverters was that all of these issues were hidden to the user. Thanks for your help!

 

[Ampster]

Is it as simple as changing all of the charge settings (float, bulk) to 53.6V (for 16s)?

No, 53.6 V is only 3.35 volts per cell which is close to resting voltage and is at a fairly flat part of the curve. I charged my DIY pack to 3.45 per cell or 55.2 volts and after about an hour the pack rested at 53.6 volts. I chose that voltage because most charts suggested it was about 90%. Interestingly my new Pytes V5 batteries only charge to 3.50 volts per cell and they garantee them for ten years. To be clear, in both cases the SOC was reset to 100% so in a sense I had some built in cushion. Also my BMS balance started at 3.40 volts per cell.

 

[sunshine_eggo]

That chart is only relevant for resting voltages, and it's not horribly accurate.

You can't charge to SoC by voltage. Too many variables.

If you fail to regularly charge LFP to 100%, two things happen:

1) BMS SoC goes to shit.
2) Cell top balance goes to shit, which often reduces available capacity.

Fast charging:
3.55-3.65V/cell, 0.05C tail current (typically about 15 minutes absorption)
Slow charging:
3.45V/cell, 0.01C tail current (typically about 1-2 hours absorption).
3.40V/cell, tiny tail current (typically about 5 hours).

Fast charging is what you use when charging from generator.
Slow charging is what you use to minimize cell stress and extend cycle life.

LFP is subject to voltage depression where regular cycling that doesn't include periodic full charges will alter the voltage to SoC relationship and can affect charging and capacity.

 

[joninca]

Thanks, this is incredibly helpful! So, it seems like the advice is to just use the PowMr default lithium charging parameters and not overthink the "cycles vs lifespan" issue? If so, my options are programs L15 or L16 (see below or page 23 of https://cdn.shopify.com/s/files/1/0...60-PRO_User_Manual_V1.4.0327.pdf?v=1711526459). I'm guessing 15/16 refer to 15s and 16s, so I would want L16, which is closest to sunshine_eggo's settings?
L15: Boost,float=54.4V,51.8V
L16: Boost,float=58.0V,55.2V
(or the "use" manually-set program).

For my own conceptual understanding, why do LFP battery forums make such a big deal about "100% charge cycles vs lifespan" when there seems to be no easy way to control this (ie this thread, and the fact that most common MPPTs don't come with BMS communication options)?

 

[Ampster]

why do LFP battery forums make such a big deal about "100% charge cycles vs lifespan" when there seems to be no easy way to control this......

I disagree that that conclusion of this thread is that there "seems tp be no easy way to control this". Any voltage above 3.4 per cell will be at the steeper part of the charge curve and that is where voltage is a fairly acurate measure of SOC and a point where most BMSs begin balancing. If i sais anything that makes it seem difficult, let me know and I will clarify? i have been using LFP batteries in various systems for over ten years and made a few mistakes but learned from those mistakes and am happy to share my knowledge.

 

[sunshine_eggo]

Thanks, this is incredibly helpful! So, it seems like the advice is to just use the PowMr default lithium charging parameters and not overthink the "cycles vs lifespan" issue? If so, my options are programs L15 or L16 (see below or page 23 of https://cdn.shopify.com/s/files/1/0...60-PRO_User_Manual_V1.4.0327.pdf?v=1711526459). I'm guessing 15/16 refer to 15s and 16s, so I would want L16, which is closest to sunshine_eggo's settings?

Correct.

L15: Boost,float=54.4V,51.8V
L16: Boost,float=58.0V,55.2V
(or the "use" manually-set program).

This.

For my own conceptual understanding, why do LFP battery forums make such a big deal about "100% charge cycles vs lifespan" when there seems to be no easy way to control this (ie this thread, and the fact that most common MPPTs don't come with BMS communication options)?

Because there's a lot of parroting of theoretical information that's not actually practical.

I disagree that that conclusion of this thread is that there "seems tp be no easy way to control this". Any voltage above 3.4 per cell will be at the steeper part of the charge curve and that is where voltage is a fairly acurate measure of SOC and a point where most BMSs begin balancing. If i sais anything that makes it seem difficult, let me know and I will clarify? i have been using LFP batteries in various systems for over ten years and made a few mistakes but learned from those mistakes and am happy to share my knowledge.

This battery hit 3.4V/cell at 50% SoC at 0.2C and required a 5 hour absorption period (terminated at 0.02C @ 99.7% SoC):




This was influenced by a 0.08V drop at 20A and 0.05V drop at 12A due to wiring losses, but many chargers will be subject to these variables unless they're receiving battery OCV and net charge current.

Charge voltage and tail current are critically related as it pertains to SoC.

 

[Ampster]

This battery hit 3.4V/cell at 50% SoC at 0.2C

You are talking about an entirely different battery than the 16S battery that is the subject of this thread. That is an unusual charge curve that I have not seen in my over ten years of working with LFP. What was the basis to assume is was at 50% SOC? I do understand voltage calibration can affect those numbers.

My comment was based on the typical charge curve as seen below and that is why I picked 3.45 volts as my target charge voltage.

 

[sunshine_eggo]

You are talking about an entirely different battery than the 16S battery that is the subject of this thread.

Is it your position that cell characteristics change based on the number in series?

That is an unusual charge curve that I have not seen in my over ten years of working with LFP.

It's pretty typical for the many many dozens of cells I've tested.

What was the basis to assume is was at 50% SOC? I dounderstand voltage calibration can affect those numbers.

This charger measures within 0.01V of my Fluke.

My comment was based on the typical charge curve as seen below and that is why I picked 3.45 volts as my target charge voltage.
View attachment 216068

Pulled from Internet or graphed yourself based on measurements?

If you'll note the blue line on my chart has the same contours as the ones above up to about 50%.

My chart is much more indicative of what the OP's charger will see because OCV is not governing the charge. The charger will be working based on the current-influenced voltage measured by the charger itself.

Here's a plot using OCV (instrument correlates with my Fluke within 0.01V), 280Ah "Grade A" cell from Amy Wan:



This cell was 305Ah tested. Charged from the as-received condition (lower portion of voltage curve is missing) with about 100Ah in it. 3.4OCV was hit at 105 counted out of 204Ah total input. This leaves 99Ah before it was full, or at (305-99)/305 = 67.5% SoC when 3.40V was reached.

SoC @ 3.4V varies based on C rate as your own chart shows, though we show notably different values.

I think we ultimately agree, but you specifically mentioned anything above 3.4V. IMHO, 3.40V is too low. Even your chart shows ~40-50% @ 3.4V @ 1C, and my 0.23C charge rate shows about 90% SoC when 3.45V is hit.

IMHO, 3.45V is the magic number as anything below 1C is at or above 90% once it's attained.

 

[Ampster]

I think we are on the same page. I am not sure what I meant by the comment about different batteries. I also had not paid attention to the flatness of the curve at 3.4 volts per cell and how much capacity there was between that voltage and 3.45 Volts when charging.