diy solar

diy solar

Why do the batteries just keep charging?

Madcodger

Solar Enthusiast
Joined
Aug 17, 2022
Messages
474
Hi Folks. Hoping our battery gurus can shed some light on this situation. The system is comprised of six 210W panels (1260 watts max total, 3S2P) feeding a Victron MPPT 150/45. Inverter is a Victron 24V single phase (120V) 3000 KVA Multiplus. A Victron SmartShunt monitors charge/discharge and was calibrated at install, then again about a week ago, when the system was at 100% and stopped charging the batteries (i.e., it seemed to be "topped off". The batteries here are three 100AH Amperetime LiFePO4 24V batteries, in parallel (via a bus bar) with factory-crimped 1/0 cables (Battery Cables USA), all the same length, ~ 18" long. All connections are nice and tight and no cables become even slightly warm regardless of the load on the batteries. Two of these batteries were installed about a year ago, and a third added about six months ago. All appear fine, physically. All parameters are set according to the specs supplied by Amperetime for the batteries, with bulk charge being 28.4 V, Float 27V. Float is set to the Victron minimum of 0.25 hours. The Amperetime batteries apparently have a built-in BMS, but there is no way to interface with that (no app, etc.). There is no third party/external BMS.

The system has been working well, but there are a couple of things about which I'm curious, and that make me wonder if one of these batteries is "bad".

1) As shown below, the batteries continue to charge at a reasonably high current (usualy 26-30A, with 28.20 A showing in the pic below), and this continues for hours after the batteries show 100% SOC. They were full at 9:30 this morning, and they're still accepting much solar input at 12:23. Load has been fairly constant at 70-100W all morning. So, where is all that energy going, if the batteries were full at 9:30?

2) A few days ago, before I got my propane heater installed, i was using an oil-fired radiator to take the chill off the room, with about a 900W draw. Combined with the regular load, the system was supplying a fairly constant 1000W. The batteries were showing ~ 30% SOC, but Voltage was down below 22V early in the evening. When it hit the inverter low cutoff of 21.6V, the inverter shut down. I turned on the AC (generator) input and recharged, and all was fine. The next day I took the batteries down to about 33% but then removed the radiator load, and allowed the system to continue to discharge, but with a load of only about 30-50W plus the inverter draw of about 30W. When it got down to about 29% SOC overnight, the inverter again shut down because the batteries got down to 21.6 volts. But it got to that low voltage with under 100W of load total. Highly suspect, I think.

So, before I go disconnecting batteries and testing things, what's the group consensus? I suspect one of the three batteries is bad, but what do you think? Any advice on next steps? Thanks in advance.

1698854210109.png
1698854238487.png
 
It looks like your SoC has drifted and is out of sync with battery voltage.
Your charger must see 26.88V and continue charging to 28.4V regardless of the 100% SoC reading.

Combined with the regular load, the system was supplying a fairly constant 1000W.
For how many hours? Cannot do meaningful capacity and usage calcs with just watts.

3x 25.6V x 100Ah = 7680Wh
 
Can you take a photo of the shunt and the battery wires?
 
What other cable connections do you have between the battery negative and the shunt?
 
Post your setting of the smart shunt, almost looks like the 100% SOC is set for 26.xxv
 
All AmpereTime (now LiTime) LiFePO4 batteries have a built-in BMS. The only problem I've ever had with mine is where one of those bms's acted flaky, and exhibited some weird voltage numbers ... this is how you know it's a problem battery (bms). You can usually catch this when the inverter shuts down, and the batteries are in a state where one (or more) of them needs "waking up".

In your spare time, go ahead and disconnect all batteries from inverter and each other ... then:
- check each with multimeter, record numbers
- ensure each is topped off, using appropriate standalone charger for their voltage
- hook them all in parallel for 24 hours, and let them balance with each other
- check each with multimeter again, record numbers & verify they are close to each other

Then put them back into your battery-bank arrangement. At this point, you've balanced them as well as AmpereTime (now LiTime) wants, per their docs. If the inverter shuts down again, disconnect all the batteries, test each with multimeter, and if one of them shows weird/high voltages (like in the range of 500+ volts), then you know it's bms is acting up.

This happened to me one time, and with some pics of the multimeter reading, LiTime swapped the battery out on their nickel ... I saved the battery boxes, they provided a free ups shipping label, and when the battery showed as in UPS's hands, they cross-shipped a new one to me.

So far, great batteries, excellent company and customer service ... after several years now.

I can't explain why lots of power being "consumed" in your charging scenarios ... perhaps LiTime can ... send 'em an email to their support line, after you do the above, as you can tell them all batteries are balanced per their docs.

Hope this helps ...
 
If all batteries read 3.30v at a resting state, the SOC can be anywhere from 30% to 70%.

What would be nice to whiteness is while charging is the current going into each battery, and compare across the three.

While all cable lengths are equal I think each SOC has changed. I’d wager one may be close to 100% but the another maybe at 60% and is taking in all that current.

Seeing the shut is looking at voltage setpoint to reach 100%, it is very easy a wrong setting that gives the bank an apperance of 100% but is far from it.
 
Looks like shunt drift and some mixed up settings. 26.88v on an 8 lifepo4 cell 24v nominal system isn't fully charged.

My SmartShunt drifted too, but it went the other way, showed low even when the batteries were fully charged. Very frustrating at first, WHY are my batteries only accepting 15w from a 1600w panel array, when the shunt says they're only at 83% SOC?!? My BMSs were saying both batteries were at 100% SOC and 27.2v; the batteries were full, no matter what the shunt said. Finally figured it out, my AIOs were set to float at 27.2v, shunt set to read 27.2v as full, but the shunt was only seeing 27.186v or whatever where it was at. Might be cable losses(doubtful with half an amp of current and 1/0 cables), or(more likely) might be some small variance in calibration between the AIOs and the shunt. Set the inverters to float at 27.3v, now the shunt gets to 27.2v and resets SOC to 100%, no issues since. Could have taken shunt "full" voltage down to 27.1v as well.
 
This is why I enjoy VE.smart network, the shunt talks to the SCC such that the SCC knows exactly bank voltage and controls bull/float.

The BMS setting and SOC % are just there as a ride along.

I think if the OP sets his float for 2-3hr this will keep that voltage higher throughout the day and will allow some balancing between batteries.
 
Post your setting of the smart shunt, almost looks like the 100% SOC is set for 26.xxv
Well, there's a good chance I was just in idiot mode. For whatever reason, I completely missed this. Yeah, the shunt was set to 26.88 as 100%. Corrected it to 28.4 but not early enough to fully recharge with solar today. Will charge with AC tomorrow and then let the sun top it all off, and reset the shunt. Will report back. Thanks to everyone for the input and ideas - greatly appreciated. No idea where my head was at on that. I'll blame work distraction.
 
Update: Battery charged to 28.4, stayed there in absorption for a while, and finally stopped charging. Currently at 87.1% an hour or so before sunrise and showing 26.5V so not unreasonable. Battery mfr (Amperetime) manual isn't very helpful as they say 28.4 OR 29.2 is fully charged. Huh?? That's either 3.55 or 3.65 V per cell, so WHICH? I think I might revisit some Off Grid Garage videos but am open to input. What's the disadvatage of charging to 3.65V per cell?
 
Update: Battery charged to 28.4, stayed there in absorption for a while, and finally stopped charging. Currently at 87.1% an hour or so before sunrise and showing 26.5V so not unreasonable. Battery mfr (Amperetime) manual isn't very helpful as they say 28.4 OR 29.2 is fully charged. Huh?? That's either 3.55 or 3.65 V per cell, so WHICH? I think I might revisit some Off Grid Garage videos but am open to input. What's the disadvatage of charging to 3.65V per cell?
I read somewhere that an optimum charge voltage is 5 to 13% of nominal battery voltage. But I'm just getting educated on solar so don't go by my advice. All I have right now is 20,480 kwh of batteries and one harbor freight 1.5w solar panel lol
I ordered the batteries yesterday but I think I need a bigger panel ?
 
a few minor details...
When charging/discharging the shunt has a BIAS.
Charge mode will always show a different voltage than discharge. It is usually only minimal but can be up to 0.2V difference. This requires you to correct for that in your settings. It's called Calibrating the system ! You must use a Multimeter and look at the voltage on the SCC, Inverter and of course at the Battery Terminals. and you need to do this during a normal charge, normal discharge & when completely idle (no loads other than just inverter). Believe it or not, even a 0.1 Volt difference with Lithium based Batteries can make a difference.

Adjust your Charging settings UP to compensate for the slight loss.
Correct your disconnect voltage, also compensating for that loss so if the inverter reads as 24.0 but the batteries read 23.9 then change the inverter to 24.1.

REMEMBER, your Shunt is "in between", so the software settings for the shunt must also account for the loss bias from charging vs discharging so that it is "reasonably" accurate... they are not 100% perfect, not much is.

Because you are using 100AH Batteries, your TailCurrent is 5A. That is the point where Absorb/Bulk needs to switch to Float (Constant Voltage Variable Current). At 5A Charge taken the batteries are FULL, float servioces any load that it can while trickle topping the LFP packs allowing them to balance calmly. DO NOT LIMIT FLOAT TIME ! If you have enough solar incoming, float can & will service the Inverter demands FIRST and them pull balance from battery IF required. Properly setup this will enhance the batteries NOT harm them... (old wives tales need not apply)

Charge your battery packs to 28.0V (3.500V per cell)
Float at 27.6V (3.450V per cell)
LFP will Naturally SETTLE to 3.400 Volts per cell as that is the top of the Working Voltage Range. AKA 100% SOC !
Remember, that LFP Nominal Voltage is 3.200 Volts per cell or 25.6V
The Chemistries Working Voltage Range is from 3.000V per cell (0% SOC) to 3.400V per cell (100% SOC) because this chemistry has a very flat voltage curve. NB: This also makes software that reads voltage with a single decimal point not that accurate due to the tight spread.


quick-voltage-chart-lfp-jpg.150247
 
Actually the smart shunt had voltage sensing leads that should be directly on the battery terminals. This allows the shunt to tell the Victron SCC actual battery terminals (independent) of small voltage drops from SCC to battery. The SCC compensates and adjusts its output voltage as needed. another reason why victron units are worth the few extra $.
 
GREAT info, folks. Thank you! I think the thing I'm trying to wrap my head around now is float. I always thought float was essentially unnecessary with LiFePO4, but the info from @Steve_S and some Off Grid Garage viewing now have me convinced I shouldn't be limiting it so much, and that I probably need to tweak my float voltage up to 27.6 instead of 27.

This is a small "starter" system that I built for my "outbuildings" to gain experience before we build another house next year, that we may keep off grid. So, while I have Victron equipment I went with Amperetime batteries. I think they're a good value for the money, but I now wish I had spent a little more for batteries that might benefit from a "real" BMS, where I could see the voltage of each cell. I drool every time I watch Off Grid Garage. Anyway, you folks have been a big help, and I appreciate it. Thank you!
 
Back
Top