diy solar

diy solar

Sol-Ark 12K - Batteries Drop from 75% to 0%

1) I do not have washers nor dielectric grease....will correct.

2) I have maintained water levels.

3) Recent hydrometer and multimeter voltage readings attached.

4) Photos of charge settings attached....possibly set at 70A?

5) Several current photos attached...I'm new to this stuff.

Under Battery Setup on the BATT tab I would set your MAX A CHARGE to 70 amps which works out to 35 amps per string. Run it there for a few months and see how it affects your water levels. After a couple months, you can increase to 80 (40 per string) and see if you notice any excessive water losses. I don't think I would go higher than 80 though as you're probably going to be pouring water into those cells frequently at that rate.

Under Battery Setup on the CHARGE tab, I would set that for 50 amps, but you can set it for whatever you want as I suspect the inverter will not allow a charge rate higher than what you set in the BATT tab at 70 amps.

With lead acid cells, if you charge them too slow, they won't get charged fast enough and the sulfates can stick to the plates and harm the battery. But if you charge them too fast, that's just as bad as it can blow apart your plates and boil them dry.

I think some of your problem is in your battery connections. Those connections look corroded and, (guessing), look like they aren't mated to the terminals well. This doesn't really make much of a difference when you're passing 30 or 40 amps back and forth, but when you start drawing real current and sucking up 70+ amps, those connections are going to start heating up and the resistance is going to cause a problem.

Clean them up, shine up the terminals and remove all corrosion. Polish those connections clean, then coat the crap out of everything with dielectric grease and use a torque wrench to tighten them down to spec if you're not accustomed to doing it. Too loose and bad things happen (corrosion, heat, etc).. too tight and you can physically damage them (stripped threads, broken terminals, etc)

I didn't read everything in this thread.. Did you use a hydrometer? If not, then you need to get one. Owning a lead acid battery bank without owning a hydrometer is like owning a car that doesn't have any fuel, temp, or speedometer gauges. Take the guesswork away with a hydrometer.
And you are using distilled water right? Not tap water or bottled water?
 
Does the SolArk have a shunt built into it? Can you enable that and make it count amp hours to determine battery SOC instead of using the voltage?
They do have self calculations but if he switches back to SOC instead of voltage it will turn the charger off if it thinks the battery is full.

That’s why I told him to switch to Voltage.
There are no voltage charts in Sol-ark.
You just input what voltage you want.


1672722542136.jpeg

If you give him an Equalization voltage and let him do that it will work.

I wouldn’t use SOC until sure he has full capacity.

He has no BMS.
 
Last edited:
There are no voltage charts in Sol-ark.
Sorry, I meant a datalogged graph. I'm trying to get a graph of the actual voltage what the SolArk thinks the batteries are when the SOC suddenly goes to 0%. That might give us an idea of what's going on. I've seen bad crimps/connections act just fine until they warm up and then the oddness starts happening.

I don't think any lead acid battery has a BMS.

If you give him an Equalization voltage and let him do that it will work.
That's a good voltage for the EQ cycle to start with. He may have to change the 30 day interval to 1 or 2 days until the SG responds on the lower cells and all cells are within ~0005 of each other. Still the first thing I'd do is check the voltage drop across every connection and repair any connections that stood out.

Here's the spec sheet that I just found on those batteries. It recommends max amps available <100 and 2.7 VPC (64.8) when used in EV service but makes no mention of RE service. Could be worth reaching out the manufacturer to get some clarity on for RE. Still I would have no problem with a C/5 charge rate of ~86 amps per string and 2.6 VPC (63.6) for equalization.

I haven't seen if this is off grid site or grid tied?
 
Last edited:
Sorry, I'm meant a datalogged graph.
I’m not aware of one on the inverter itself.
PV pro should have the data if he knows how to access it.
I don't think any lead acid battery has a BMS.
Sorry My experience with FLA and RE is nil.

He may have to change the 30 day interval to 1 or 2 days until the SG responds on the lower cells and all cells are within ~0005 of each other.
He can easily do that just in that setting.

I haven't seen if this is off grid site or grid tied?
I honestly do not know. Have to wait on OP to respond to that.
 
Last edited:
Sorry, I meant a datalogged graph. I'm trying to get a graph of the actual voltage what the SolArk thinks the batteries are when the SOC suddenly goes to 0%. That might give us an idea of what's going on. I've seen bad crimps/connections act just fine until they warm up and then the oddness starts happening.

I don't think any lead acid battery has a BMS.


That's a good voltage for the EQ cycle to start with. He may have to change the 30 day interval to 1 or 2 days until the SG responds on the lower cells and all cells are within ~0005 of each other. Still the first thing I'd do is check the voltage drop across every connection and repair any connections that stood out.

Here's the spec sheet that I just found on those batteries. It recommends max amps available <100 and 2.7 VPC (64.8) when used in EV service but makes no mention of RE service. Could be worth reaching out the manufacturer to get some clarity on for RE. Still I would have no problem with a C/5 charge rate of ~86 amps per string and 2.6 VPC (63.6) for equalization.

I haven't seen if this is off grid site or grid tied?
Off grid
 
Nice work on that battery chart. Putting the small load on there was smart.

In my experience a sulfated battery will have a resting voltage that indicates it's full but the SG will indicate it's anything but. It does not appear that yours are severely sulphated.

1250 vs 1275: Is that all the resolution your hydrometer has? I would expect readings more like 1245, 1255, 1270, etc. Still I think it's time to do an equalization charge until you can get all cells to 1275 but I don't think that will totally solve this issue.

Have you checked voltage drop from the battery neg to the SolArk battery neg when it's under load? (do the same for the positive as well) Put one lead of your meter on each end of the same cable. A zero reading would be perfect. More than a few volts could be bad. EG: ~50v x 3% = 1.5v

Does the SolArk have a voltage chart that you can share with us? I see you've got "use battery V charged" check in one of those pictures. One theory is that your voltage dropping due to a bad connection or cell telling which makes the SolArk think that the batteries are empty.

Does the SolArk have a shunt built into it? Can you enable that and make it count amp hours to determine battery SOC instead of using the voltage?
Does this chart shed any more light? SOC in percent and battery voltage.
 

Attachments

  • IMG_BDB5B4B4FF2A-1.jpeg
    IMG_BDB5B4B4FF2A-1.jpeg
    108.3 KB · Views: 13
Yea, wow that was bad. Huge fire hazard!

8S2P

Have you tested the rest of the batteries? Specific gravity test and water addition has to be part of regular maintenance, right?
How old are those batteries? It is not recommended to mix and match lead acid batteries, especially old and new.
Batteries have been in service for 10 months.
 
I'm at a loss to explain why the SolArk thinks that.

Is there one of the OP's pictures of his SolArk menus that showed what the battery AH are set at?
430AH is what is set in the inverter.
This is just one of the reasons I like running Voltage control over SOC for open loop communication.
 
Looking at the image the op loaded, immediately before the dip in SOC we see the battery voltage falling. When the voltage reaches the "battery empty" that's set, the Sol-Ark will correct the SOC estimate to zero. That mystery is solved, maybe.
Over the next several hours the SOC rises as we'd expect while it's charged, until it reaches around 40%, where it soars back to 100.
The resolution of the Powerview graphs is too low to see what's going on there.
There's no way the battery has been charged the 430 amp hours the op has programmed in just those 4 hours at the 70 amp charge.
Try lowering battery capacity to 200 or something and see what happens.
I have no idea what the correct settings are for your battteries, but they don't seem to be delivering the 22kWh the Sol-Ark is expecting.
 
430AH is what is set in the inverter.
This is just one of the reasons I like running Voltage control over SOC for open loop communication.
I'm have no clue how "voltage control" works in a SolArk since I've not done one yet.

In my experience it's hard to beat using a shunt and counting AH. Even that tends to get a little lost so programming in the finishing amps and volts allows the AH meter to reset to full.

Looking at the image the op loaded, immediately before the dip in SOC we see the battery voltage falling. When the voltage reaches the "battery empty" that's set, the Sol-Ark will correct the SOC estimate to zero.
That's what I've been trying to say..... LOL
 
To be clear, I think the Sol-Ark is performing as it should when the SOC dips.
I think the battery settings need to be reviewed by some lead-acid expert and changed. I know that there's no way the internal resistance is even in the ballpark, and the 99% efficiency? That's too high for any chemistry, arguably.
 
I don't know if this ever got resolved, as this thread is over a year old now, but the battery empty volts setting should be set closer to probably 40 volts, or something like that. Generally speaking on FLA batteries, 46V will be approximately 50% SOC, which tends to be the sweat spot to get a decent amount of battery life, while not needing to go way oversized on capacity.

But....... this SOC jumping thing is most definitely a Sol-Ark issue!!! I am dealing with this on a handful of systems, although these all have LFP batteries, but also no closed loop comm. And just an FYI for anyone who may read this: closed loop comm is not the answer to all problems in any way, shape or form! Many brands of LFP batteries have issues with accurately tracking SOC, especially under very low load scenarios, and when they don't often get completely fully charged.

I have worked with quite a few different battery monitors over the years, and Sol-Ark's "built-in" battery monitoring takes the cake on worst ever! I will probably start a thread on the issues I'm seeing, to try to get some attention from Sol-Ark, as tech support has been of no help, and I have not managed to get help from anyone further up the chain...

(Oh yes, someone wondered if they have a shunt inside the Sol-Arks, yes they do!)
 
Back
Top