*Update again, replaced* EG4 Wall Mount battery cell not reaching absorb voltage.

[immortl]

Huh, interesting thread. I am having a similar issue with a Midnite Powerflo16. Cell imbalance of 150-250mv when BMS reports 100% SOC and cell high voltage alarm as well if I try to bump up the charging voltage a bit. Does your SOC suddenly drop to 0% when the lowest cell voltage gets low but overall battery voltage is still up in 52v range?

Past month and a half and getting a similar run around from Midnite tech support. Supposed to try new firmware and put the battery in parallel with the other good batteries and 'it'll sort itself out, it's just a calibration for SOC issue" according to their support and engineers. I'm skeptical but am going through the steps or will once they send the updated firmware my way. They've been 'getting it tonight' for 2 weeks now.

So the tie in with this thread is similar not making much sense to me either tech support advice. EG4 vs Midnite - on battery side doesn't seem much different. Sigh... I'm definitely going DIY battery for the next additions.

 

[Zapper77]

@Zapper77 I know it's easy to lose track of queries, but this one is fairly important, can you answer the question?

I've never tested COV to see if that alone would trigger the balancing logic. From my personal experience testing BMS firmware, a cell imbalance was the most common trigger.

 

[wpns]

I've never tested COV to see if that alone would trigger the balancing logic. From my personal experience testing BMS firmware, a cell imbalance was the most common trigger.

Basically that appears to be their solution for any top end imbalance is to set off the over voltage protection and that is supposed to prompt the BMS to begin balancing on future charging.

So no? You can't answer the question?

 

[rogerheflin]

Huh, interesting thread. I am having a similar issue with a Midnite Powerflo16. Cell imbalance of 150-250mv when BMS reports 100% SOC and cell high voltage alarm as well if I try to bump up the charging voltage a bit. Does your SOC suddenly drop to 0% when the lowest cell voltage gets low but overall battery voltage is still up in 52v range?

Past month and a half and getting a similar run around from Midnite tech support. Supposed to try new firmware and put the battery in parallel with the other good batteries and 'it'll sort itself out, it's just a calibration for SOC issue" according to their support and engineers. I'm skeptical but am going through the steps or will once they send the updated firmware my way. They've been 'getting it tonight' for 2 weeks now.

So the tie in with this thread is similar not making much sense to me either tech support advice. EG4 vs Midnite - on battery side doesn't seem much different. Sigh... I'm definitely going DIY battery for the next additions.

There is a generic BMS issue that can happen when one cell has a higher resistance connection to a cell, and is allowed to balance when below 3.42 (when charging) that causes typically one or 2 cells with the condition to have charge removed from then (thinking the voltage is higher because of the resistance in the charge path) and the cells that have that condition have to have that charge added back (by the balancer) when charging stops. See.

Post in thread 'Determining when a cell is going bad?'

Aug 5, 2025

This is not the resistance shown by the bms, this is the path length between the 2 wires measuring the cell voltage. The other 3 cells all have to go through the length of a bus bar and cell01 does not. And it causes an increase in voltage between the 2 measurement points when current is flowing across the bus bar. All other cells are cell voltage + current * busbarresistance. With 30A through that style of bus bar I get 1mV from cell- to next cell+. Double the current and you are at roughly 2mv difference causing cell01 to APPEAR to be behind the other 3 cells. When you see...

• HRTKD

 

[IAm_Not_Lost]

Ok, so after however many months of one cell declining and not fully charging, EG4 is sending me a replacement battery. Should be here tomorrow.

So, my question is, what’s the best way to insert the new battery into the charge cycle with the other two I have? Discharge the two I have down to a very close voltage and estimated percentage as what the new one is at and then turn the new one on and get it going?

EG4’s initiation instructions are very lacking in general, based on the manual I have.

Thoughts?

 

[wpns]

Get the voltage (not the SOC) close, within maybe 1/10 volt, and initial current will be low.

 

[fnnwizard]

Ok, so after however many months of one cell declining and not fully charging, EG4 is sending me a replacement battery. Should be here tomorrow.

So, my question is, what’s the best way to insert the new battery into the charge cycle with the other two I have? Discharge the two I have down to a very close voltage and estimated percentage as what the new one is at and then turn the new one on and get it going?

EG4’s initiation instructions are very lacking in general, based on the manual I have.

Thoughts?

If you are within 2V between packs you will be fine. Here's the math to prove it. There is a minimum .050Ω between the 16cells, connections, cables, etc. So if you are 2V diff, you'll see the single pack hit 40A either charge or discharge within 1 sec into however many packs you currently have. (2V/.050Ω)

So say you have 3 packs now, you hook up new pack and it is 52.656V ( it should be close to this if mfg knows their stuff) while existing packs are 54.6Voc. You'll see the new pack charging at ~40A and other 3 dischrg at ~13.xxA or so for a sec or so and current will quickly get lower over the next 5 minutes and slowly decreases after until the single pack gets to the same V as the other 3.

There's 3 things causing the various current deltas, ohmic resistance, Ionic resistance, actual soc changes.

I just happened to post a chrt of cell overpotential slump curve today.

F

Post in thread 'Can't tell SOC by voltage! So many different answers.'

Sep 30, 2025

I was able to get into the office today to pull a couple charts. Still need to pull rest of data for the Slump V's but these charts give a good example of the time it takes for V slump to stabilize.

It's 2 Eve 280LFK V3 cells, top caps out at 300.07Ah, bottom caps out at 299.94Ah, so true .25C rate is 75A. Top one is 20% soc, bottom close 40%.

Full and magnified view. Note: Y Scales are diff for top and bottom curves of same graph, also each box of the bottom chart is 1/100 of a sec.

To the unexperienced observer, (s)he may say the V slump hasn't stabilized, but it has at this C rate...

• fnnwizard

 

[immortl]

Went through similar with a MNPowerflo16. Same as you - took several months. After finally getting the BMS utility from them, it showed the individual cell voltages and 4 were dangerously low (like 1.xxx V) while 2 others were above 3.6v. They quickly said they'd replace the battery.

When adding in batteries to the system, I find it easiest to do it around full charge. I use the BMS utility to view the cell voltages and match the new battery within .010 of the existing battery. I'll have the 2nd battery close to full via a chargeverter directly hooked up to it. I then shut the system down, and charge the new battery if necessary a tick more while monitoring the cell voltages and when they match the existing battery I turn off the chargeverter and hook up the additional battery to the system. Then I bring everything back up again. It's worked for me a few times as I currently only run 2 batteries at a time on my 'temporary' install, but have rotated all 4 into service for testing.

Added on edit - this is just what I did and it worked well for me. The math in the post above me by @fnnwizard makes sense to me. The .010 I used was arbitrary. They raise a good point.

 

[fnnwizard]

Just to add, of course the closer the V's are the better.
Extrapolating to single cell, for the same ~40A current, you need to be within .125V (16*.0125V = 2V).

Cell ohmic Ir with connections is closer to .003Ω, while ionic+Ohmic is closer to .062Ω at 40A. So in an instant you are at 41A and then tapers lower following closely to its Vslump/current curve.

The real life issue tends to be that most users won't have an easy way to get the individual pack charged up in a timely fashion so it can join its brethen, so likely easiest just to hook it up "in situ" using guide above.

Or, wait til other packs get down in V.

 

[IAm_Not_Lost]

Just to add, of course the closer the V's are the better.
Extrapolating to single cell, for the same ~40A current, you need to be within .125V (16*.0125V = 2V).

Cell ohmic Ir with connections is closer to .003Ω, while ionic+Ohmic is closer to .062Ω at 40A. So in an instant you are at 41A and then tapers lower following closely to its Vslump/current curve.

The real life issue tends to be that most users won't have an easy way to get the individual pack charged up in a timely fashion so it can join its brethen, so likely easiest just to hook it up "in situ" using guide above.

Or, wait til other packs get down in V.

yeah I was planning to just hook it up when the other batteries were as close to the new batteries voltage as possible, accounting for voltage drop under load.

 

[Quattrohead]

If you are within 2V between packs you will be fine. Here's the math to prove it.

That is way too big of a voltage difference and your math proves nothing. There will be a massive current flow and the BMS's will probably trip.
My practical experience says I'm okay with half a volt difference and I'm a bit of a rebel, the closer the better. As wpns says you can't trust the SOC reading until the battery has cycled a couple of times with its mates.

 

[teal95]

Since the voltage is pretty flat in the middle (30-70%???) I've hooked them up with SOC in the middle. The only disturbing part about this is when you do get them towards full charge whichever is closer to full will stop charging and all the current will go to the others. So if your old batteries are at 60% and the new ones at 30%, they will keep close to that same difference until they get to the top. Then all the current will be going to the new battery until it catches up.

 

[Fisherguy66]

Since the voltage is pretty flat in the middle (30-70%???) I've hooked them up with SOC in the middle. The only disturbing part about this is when you do get them towards full charge whichever is closer to full will stop charging and all the current will go to the others. So if your old batteries are at 60% and the new ones at 30%, they will keep close to that same difference until they get to the top. Then all the current will be going to the new battery until it catches up.

I've noticed this too, I just bought a 2nd 280ah wallmount 1 year after the first. The older batt has 182cycles and the newer one is at 5. When I hooked up the new one I put it as the 2nd in line connected to the old ones bus bars and left the old one as the master. I got hold of a 48v charger and brought both batts to 58.4 initial charge and then let them settle. When I hooked them up they were less than 0.2v out. The system has been working fine but I find the soc is all over the place and they charge and discharge at different rates. This snap is in the morning just as the sun is coming up. almost identical v but pretty good soc discrepancy. Eventually they even out. They recommended I put the newer batt as the master and connect it to the inv first. I will be trying this when I get back home.

 

[wpns]

I've noticed this too, I just bought a 2nd 280ah wallmount 1 year after the first. The older batt has 182cycles and the newer one is at 5. When I hooked up the new one I put it as the 2nd in line connected to the old ones bus bars and left the old one as the master. I got hold of a 48v charger and brought both batts to 58.4 initial charge and then let them settle. When I hooked them up they were less than 0.2v out. The system has been working fine but I find the soc is all over the place and they charge and discharge at different rates. This snap is in the morning just as the sun is coming up. almost identical v but pretty good soc discrepancy. Eventually they even out. They recommended I put the newer batt as the master and connect it to the inv first. I will be trying this when I get back home.

View attachment 334834

That seems like a lot, did they both get to full charge and spend some time balancing the previous day? My (six) PowerProOutDoor are the same age, and they generally stay within 5% of each other:

And I'd like to add another pair without rearranging the order of them.

Maybe you could add another set of parallel cables between them?

 

[Fisherguy66]

That seems like a lot, did they both get to full charge and spend some time balancing the previous day? My (six) PowerProOutDoor are the same age, and they generally stay within 5% of each other:
View attachment 334843
And I'd like to add another pair without rearranging the order of them.

Maybe you could add another set of parallel cables between them?

Yea that looks like how they should operate. I haven't been able to get mine to stay close. I dunno how the charging/discharging algorithm works for these batts but they definitely play catchup with each all the time.

I did use lead acid mode to boost em one day for a few hours at 57v but I don't know if that was long enough to balance. Honestly I don't even know if they are balancing there's no way to tell. Maybe I need to keep em at 57v for a whole week. I'm still in discovery mode

 

[Fisherguy66]

When using closed loop, the eg4 protocol deems about 55v to 55.7 or 55.8 as 100% soc and shuts down the charging.


Then it floats around 54.5 for the rest of the day. Is it balancing then? Who knows? Both batts show 100%. It would be nice if the closed loop comms could incorporate an absorb cycle at a slightly higher v for a couple hours each day if there's enough sun or even during a grid charge. This is how they set it to work. Maybe its to extend battery life keeping it a little lower?

 

[wpns]

Honestly I don't even know if they are balancing there's no way to tell.

You need to talk to them directly over a serial connection, I use Modbus with some custom Python code I wrote, but maybe the EG4 battery monitor software would work? You should be able to tell by looking at the individual cell voltages using the front panel display, though that's clunky at best. Here are my cell deltas, if that helps:



Once they are all above 3450mv and the delta's below 40mv they'll stop balancing, so that's as good as it gets.

 

[Fisherguy66]

You need to talk to them directly over a serial connection, I use Modbus with some custom Python code I wrote, but maybe the EG4 battery monitor software would work? You should be able to tell by looking at the individual cell voltages using the front panel display, though that's clunky at best. Here are my cell deltas, if that helps:
View attachment 334850
View attachment 334860
View attachment 334861
Once they are all above 3450mv and the delta's below 40mv they'll stop balancing, so that's as good as it gets.

Yea nice monitoring. I have looked at the front panel ind. voltages but it's way to tedious to sit there and watch. The screen goes to sleep after about 10s without input. Are you using closed loop or manual mode?

 

[fnnwizard]

That is way too big of a voltage difference and your math proves nothing. There will be a massive current flow and the BMS's will probably trip.
My practical experience says I'm okay with half a volt difference and I'm a bit of a rebel, the closer the better. As wpns says you can't trust the SOC reading until the battery has cycled a couple of times with its mates.

What do you consider massive? It better not be more than 100A.

I'll bet you no one will have data that shows greater than 100A on 2V diff, on a 16S pack for the tyical solar ess setup ( I can actually think of a way to do it tho). The best packs with large copper buss and 4' total of 4/0AWG, non Degson connectors might get to 60A's.

I also have test data to the cell level @ about .141V diff so would be ~2.25V at 16S. Max current, measured down at the the millisec level was 69.3A and within 1 minute down near 20A's.

I'll reiterate, up to 2V diff for 51.2V packs (~300Ahr Cells, in typical solar ess) is fine, Your BMS won't shutdown from OCP. And, as long as there is current going through the packs, they will all eventually balance out.

Edit:
I just realized that my tests involved 4 total packs in P. It could theoretically get to >100A on 6 or > P packs (if all the buss bars and cables were sized properly and evenly distributed).

I might be eating my words. Will have a test on 6P in a couple days.
My excuse is that I've had a rough few days. Uncle passed away Sunday morning suddenly from heart attack... Ive been assigned to help settle his estate... Just got back from final funeral arrangements and now onto next item on the list..

 

[wpns]

Yea nice monitoring. I have looked at the front panel ind. voltages but it's way to tedious to sit there and watch. The screen goes to sleep after about 10s without input. Are you using closed loop or manual mode?

Mine’s closed loop with the Primary 18Kpv inverter over CAN and my own stuff over RS485

 

[Fisherguy66]

Mine’s closed loop with the Primary 18Kpv inverter over CAN and my own stuff over RS485

Interesting, the 6000xp is limited in its comms, so limited to what you can do. I prefer the closed loop, kind of a set it and forget it type thing and I just glance at SA or the eg4 site once in a while to see what's going on. I spend all my time at camp till November so I don't play with it much when I'm not home but when I get back I will try a few things like switching the new battery to master.

 

[fnnwizard]

Here's a worst case scenario on connecting imbalanced packs in P.

Close to fully charged 3P packs (R = 1/3 of the 1P) into a semi charged 1P ( 280Ah with Seplos 10E bms).

Using brand new Tyco contactor for switching connection to keep OvrPot bleedover from slow connections to a min.

.2% accuracy shunt = good enough for this test.

The 1st 2 charts shows the Current/time curve with the zoomed in one showing the first few millisec = ~178A. this is with ~3.35V diff between a 1P and a 3P, and in addition to charger providing up 70A of current into the mix as shown in the seplos logger.

The attached Seplos logger is ~1second per line, = response time is not the greatest, so it didn't capture the peak. It still caught a big chunk at 171A tho..

 

[fnnwizard]

Here's a worst case scenario on connecting imbalanced packs in P...

... in addition to charger providing up 70A of current into the mix as shown in the seplos logger.

Final details with 2 examples of A and B:

Wanted to explain further why the additional 70A of current from the charger makes this a worst case scenario because little details like this is what can add confusion and what can make me "eat my words..."

A: Adding on a pack with lower V, = do so when existing packs should be in idle or discharging mode.

In the above example, there is an additional 70A of current (~4kW) that's added to the single pack's charge current that really has no Impedance effect to the 3Ps, where as if the 3P packs were to provide that current, their V's of course would sag ( based on their Impedance curve) so it would not be close to additional 70A = total peak current would be way less than 178A.

The table below, top example at near peak current, the 3P pack is providing ~94A while the charger adds another ~70A. ( The math is a bit off only due to time delays in the reporting of, but it's close enough to get the point.)

So, iow, if we remove the "free" current provided by the charger; to reach near peak of 171A, the 3P has to provide that diff. It can't provide all of it due to its total impedance curve. Hopefully readers get what I'm trying to convey.



B: when adding on a pack that has higher V than existing, the load to existing would be opposite of above = do it while existing batts are idle, or should be charging.

This whole adding on P packs is not as critical for most DIY persons since they can wait for the existing packs' V to get close enough to the new addition.

But for professionals, we don't always have the time to wait so need to know these details to make it work 100% of the time.

 

[wpns]

Final details with 2 examples of A and B:

Wanted to explain further why the additional 70A of current from the charger makes this a worst case scenario because little details like this is what can add confusion and what can make me "eat my words..."

A: Adding on a pack with lower V, = do so when existing packs should be in idle or discharging mode.

In the above example, there is an additional 70A of current (~4kW) that's added to the single pack's charge current that really has no Impedance effect to the 3Ps, where as if the 3P packs were to provide that current, their V's of course would sag ( based on their Impedance curve) so it would not be close to additional 70A = total peak current would be way less than 178A.

The table below, top example at near peak current, the 3P pack is providing ~94A while the charger adds another ~70A. ( The math is a bit off only due to time delays in the reporting of, but it's close enough to get the point.)

So, iow, if we remove the "free" current provided by the charger; to reach near peak of 171A, the 3P has to provide that diff. It can't provide all of it due to its total impedance curve. Hopefully readers get what I'm trying to convey.

View attachment 335380

B: when adding on a pack that has higher V than existing, the load to existing would be opposite of above = do it while existing batts are idle, or should be charging.

This whole adding on P packs is not as critical for most DIY persons since they can wait for the existing packs' V to get close enough to the new addition.

But for professionals, we don't always have the time to wait so need to know these details to make it work 100% of the time.

So my previous paradigm (guidance) of 1/10, 1/4, 1/2 volt delta before connecting them should really be 'under a volt'?

 

[fnnwizard]

So my previous paradigm (guidance) of 1/10, 1/4, 1/2 volt delta before connecting them should really be 'under a volt'?

Under 2V is fine, but under 1V is of course better.
For your specific scenario, the actual current would depend on if you are adding a 1P pack to the 3P bank or adding 1P to 2P and, how the parallel cables are connected = not all the packs share equal load duties.