Troubleshooting SMA and REC - Mostly (all?) Solved

[SilverbackMP]

— Most (maybe all?) kinks worked out thanks to forum members and Offgrid and Marine Energy Systems (OGM) —​

This crap is going to give me a stroke. I flew back home from Asia to try to get this over engineered German scheisse going.

We (me and my industrial/commercial electrician) have been trying to get this going for the better part of four days (when frustration levels are too high, we switch to wiring my shop).

While I was gone, the batteries ran down to 2.8 so we pulled them and charged them back up to 3.55 (56.8) using a chargeverter at my electrician's shop.

Believe I two issues going on. The Sunny Islands sometimes pull power from the batteries when they shouldn't be....i.e. when there's plenty of sunshine. I think this is what ran the batteries down.

The REC BMS refuses to budge from 0% - 3% SOC no matter the battery voltage. I can't find any way to reset them.

At this point I'd love to pull a Joker and load the powershed with ammonia nitrate.

God I wish I would have just went with Victron. Seems so much simpler. I have probably spent more on skilled labor than an entire Victron system would have cost.

Anyone know if the active power settings or reactive settings on Sunny Boys should be left to default "on" for offgrid? We think so, but given SMA's German bull schiesse of not speaking plainly, I am trying to find confirmation.

I am fairly certain that there is just one or two settings that need to be tweaked on the REC Master, the Sunny Islands, or Sunny Boys, but I'm going to go broke trying to figure this out.

Only meaningful tweaks we've done on this attempt to to take the SI SOC safety settings to 0% (since the REC controls this and based on old threads) and to install a temp sensor in the SIs to keep them from clogging up the error logs (again, this is primarily a REC BMS function with its own temp sensors).

 

[timselectric]

@Hedges is the forum SMA guru.

 

[SilverbackMP]

This was my install thread: https://diysolarforum.com/threads/my-sma-system-install-so-far-and-precharge-questions.54557/
And this was me acting as a conduit between my electrican and my mom and the forum's advice while I was in Korea (system is in MO): https://diysolarforum.com/threads/sunny-boy-7-7s-keep-going-off-line.71324/

I flew back to try to help get this thing working properly.

My mom's lead acid batteries are dying on her old Trace/Outback setup and we're having to run of a generator (too large of a one for the current system - 25 or 30kW powered by a BT4 Cummins) 3 times a day. I would rather not have to replace those LA batteries when I have $50-60k in equipment and labor tied up in a new setup.

 

[SilverbackMP]

Here I discuss my REC settings...which we haven't changed:

REC BMS Settings (M/S, SMA)

Hi all, Toms was kind enough to screen shot his REC settings and sent them to me. They seem to be on the very conservative side which is great. But they didn't include everything that my REC is showing. I think his is a slightly different version as I believe he uses a 3P16S battery whereas...

diysolarforum.com

 

[Hedges]

Other guys here use REC BMS. I have zero experience with that or anything else lithium related.

Sunny Island would supply current/voltage to battery as requested by BMS, if AC power available to it (either from grid or GT PV like Sunny Boy.) It would draw from battery to supply its internal consumption and loads. If configured to sell power to grid it might draw power from battery to backfeed grid (I don't know how that interacts with BMS; with lead-acid battery, target voltage is determined by Sunny Island, and if battery is at higher voltage, especially due to DC coupled charge controllers, then it can draw from battery to backfeed grid.)

While I was gone, the batteries ran down to 2.8 so we pulled them and charged them back up to 3.55 (56.8) using a chargeverter at my electrician's shop.

Believe I two issues going on. The Sunny Islands sometimes pull power from the batteries when they shouldn't be....i.e. when there's plenty of sunshine. I think this is what ran the batteries down.

Off grid with no generator, Sunny Island raises frequency to curtail production by Sunny Boy. If frequency is < 60.5 Hz, it is asking for 100% of what Sunny Boy has to give. Check wattage vs. apparent sun exposure. I've had bad string connections. But you have so many redundant, don't expect all failed.

Is there just too much AC load?
Load management, load shed.

What frequency do you observe?

If > 60.5 Hz (or > 61 Hz for Sunny Boy in offgrid mode), SI is requesting curtailment.

With my AGM, this frequency shift is done according to what SI things it should do to battery. With your lithium and REC BMS, this ought to be in response to voltage/current requests from REC.

Is communication with BMS occurring?

The REC BMS refuses to budge from 0% - 3% SOC no matter the battery voltage. I can't find any way to reset them.

I thought BMS normally reset to 100% after enough time at high enough (absorption) voltage.
SI similarly seems to take some time at elevated voltage, for systems running open-loop.

At this point I'd love to pull a Joker and load the powershed with ammonia nitrate.

I'd like to salvage everything before you do.

Anyone know if the active power settings or reactive settings on Sunny Boys should be left to default "on" for offgrid? We think so, but given SMA's German bull schiesse of not speaking plainly, I am trying to find confirmation.

I think reactive power settings are just for compensating a facility's loads on grid. I've read details in Sunny TriPower manual regarding that. Less sure about Sunny Island, which is grid-forming except when fed by generator or grid.

I am fairly certain that there is just one or two settings that need to be tweaked on the REC Master, the Sunny Islands, or Sunny Boys, but I'm going to go broke trying to figure this out.

Probably. The others who use REC may have an idea.

Did system work properly for a while?
What has changed, temperature? Anything else?

Only meaningful tweaks we've done on this attempt to to take the SI SOC safety settings to 0% (since the REC controls this and based on old threads) and to install a temp sensor in the SIs to keep them from clogging up the error logs (again, this is primarily a REC BMS function with its own temp sensors).

I don't know if SI does anything with temp sensor when BMS is connected. I would set temperature coefficient to zero so voltage doesn't get changed. The sensor could be used to enable an alarm or turn on a fan or air conditioner if battery gets hot.

 

[SilverbackMP]

Other guys here use REC BMS. I have zero experience with that or anything else lithium related.

Sunny Island would supply current/voltage to battery as requested by BMS, if AC power available to it (either from grid or GT PV like Sunny Boy.) It would draw from battery to supply its internal consumption and loads. If configured to sell power to grid it might draw power from battery to backfeed grid (I don't know how that interacts with BMS; with lead-acid battery, target voltage is determined by Sunny Island, and if battery is at higher voltage, especially due to DC coupled charge controllers, then it can draw from battery to backfeed grid.)



Off grid with no generator, Sunny Island raises frequency to curtail production by Sunny Boy. If frequency is < 60.5 Hz, it is asking for 100% of what Sunny Boy has to give. Check wattage vs. apparent sun exposure. I've had bad string connections. But you have so many redundant, don't expect all failed.

Is there just too much AC load?
Load management, load shed.

What frequency do you observe?

If > 60.5 Hz (or > 61 Hz for Sunny Boy in offgrid mode), SI is requesting curtailment.

With my AGM, this frequency shift is done according to what SI things it should do to battery. With your lithium and REC BMS, this ought to be in response to voltage/current requests from REC.

Is communication with BMS occurring?



I thought BMS normally reset to 100% after enough time at high enough (absorption) voltage.
SI similarly seems to take some time at elevated voltage, for systems running open-loop.



I'd like to salvage everything before you do.



I think reactive power settings are just for compensating a facility's loads on grid. I've read details in Sunny TriPower manual regarding that. Less sure about Sunny Island, which is grid-forming except when fed by generator or grid.



Probably. The others who use REC may have an idea.

Did system work properly for a while?
What has changed, temperature? Anything else?



I don't know if SI does anything with temp sensor when BMS is connected. I would set temperature coefficient to zero so voltage doesn't get changed. The sensor could be used to enable an alarm or turn on a fan or air conditioner if battery gets hot.

Hi Hedges....

At certain times, the Sunny Boys appear to be ramping up and down like they should. I.e. yesterday afternoon was really sunny and they were providing just 600 watts each powering a 1000 watt electric heater, the fan for a salimander style heater, and a few tool battery chargers....it stays online a TON longer when there is a load on the AC side. But take that load off, and the system shuts down in 10 minutes. REC displayed SOC did manage to go from around 2% to 3% during that time.

Rick at Offgrid and Marine (the US REC distributor) wanted more of a history, so I sent him this:


"We got this commissioned right before I flew at the end of April. Was showing the (I believe standard, as it "figures it out") 50% of charge. The system powered a freezer in my workshop for a couple of months. I have no idea what the SOC (real or displayed) was showing during this time as I didn't have the time to get remote connectivity going before I left.

After this, the system would stay on for a week or two and then my electrician would have to come out and restart it.

And then about two months ago, the system wouldn't stay on for more than 10 minutes....The displayed SOC would be in the single digits (and I think this was perhaps accurate). The cells got down to the 2.8v cut off limit, shutting down the BMS completely. And we/they quit trying to troubleshoot until I got home last week and we recharged the batteries to 3.55 but the SOC show over 3%. And the SIs sometimes show power going to the batteries and some being used by the AC side, And then sometimes it draws all the AC load from the battery even when there's AC available from the Sunny Boys."

 

[SilverbackMP]

REC BMS is showing 0% SOC now though, as that heater eventually started drawing from the batteries rather than available AC from Sunny Boy power. Voltage doesn't equal 0% SOC. I don't remember the exact voltage (I keep having to walk to the house from the powershed to use wifi), but it's somewhere around 3.3v per cell.

 

[SilverbackMP]

The coms seem to be working between the BMS. The Sunny Islands mirror the SOC that the REC BMS Master thinks it at.

 

[OGSS]

Hey Silverback, thanks for bringing this issue to our attention and although we don't always have all the answers we are here to help. In future feel free to contact us directly at support at ogm-energy.com or by mentioning us on this site.

Let's start with the simple things first and verify that all of your current shunts are reading correctly.

From what I can see it looks like you have a Master with 4 of 48V slaves and each slave has its own shunt so the first test is:

1) Disconnect all slave cell packs except #1
2) put a fixed load on the DC system like maybe a 1,500 watt heater and check your current reading. Assuming your pack is about 52V you should see a negative current on your BMS of about 28-30 amps
3) Repeat for each cell pack

Do all packs read about the same?
Do all currents read a negative value?

Rick

 

[SilverbackMP]

Hey Silverback, thanks for bringing this issue to our attention and although we don't always have all the answers we are here to help. In future feel free to contact us directly at support at ogm-energy.com or by mentioning us on this site.

Let's start with the simple things first and verify that all of your current shunts are reading correctly.

From what I can see it looks like you have a Master with 4 of 48V slaves and each slave has its own shunt so the first test is:

1) Disconnect all slave cell packs except #1
2) put a fixed load on the DC system like maybe a 1,500 watt heater and check your current reading. Assuming your pack is about 52V you should see a negative current on your BMS of about 28-30 amps
3) Repeat for each cell pack

Do all packs read about the same?
Do all currents read a negative value?

Rick

Hi Rick,

Thanks. We'll try it. But yes, all packs and cells were and have been reading within 0.01 volts of each other...even after going down to 2.8 and then charged back up without a fresh top balance. Those true grade A CALB cells from Current Connected are (perhaps) worth the premium.

 

[SilverbackMP]

Weird…..I don’t know if this affected things, but earlier this morning we logged into one of the Sunny Boys and resaved everything…we didn’t change any settings. Just ran through the setup as an installer. We have it running now producing about 1700 watts. The other two SBs are turned off.

The batteries are accepting a charge.

This was as of 10 minutes or so ago:




All about the same:

All about the same:

 

[SilverbackMP]

Currently at 2.4 SOC with 46A and 54.17V. And the time to 100% used to jump around a lot but now it’s remaining steady.

I’m thinking that the SB’s were the culprit…we’ll log into the other two and resave their settings.

 

[Hedges]

Thanks. We'll try it. But yes, all packs and cells were and have been reading within 0.01 volts of each other...even after going down to 2.8 and then charged back up without a fresh top balance. Those true grade A CALB cells from Current Connected are (perhaps) worth the premium.

I was going to ask what all individual cell voltages were. One bad connection of a balance wire could cause one cell to appear high voltage, then I think it would ask for zero charge current.

I think you're reporting all cells look the same, but confirm. The limited number you show are pretty uniform. Isn't 3.377V fairly close to full? NOT 1.2% SoC?

Does SI turn off due to "low battery" just because it sees an SOC that it thinks is low? Maybe BMS needs to see a sustained absorption voltage above some level before it resets its SoC counter to 100%. But I'd think it would ask for more current to get there? Maybe SoC counter limits are completely independent of target charge voltages.

0.750 mOhm - is that normal for these cells? I'm used to seeing 0.25 mOhm spec, 0.17 mOhm measured. All identical seems suspicious, unlikely - more like a configuration than a measurement.

What voltage and current does the master BMS request? each of the slaves?

Can you feed generator or other source to the system? Perhaps with SB disconnected so no unregulated power/voltages sources. Just to see if batteries get charged when power is available.

I’m thinking that the SB’s were the culprit…we’ll log into the other two and resave their settings.

Would want to know configuration (Rule-21 or offgrid).
Frequency from SI
Wattage from each inverter. And how much you think should be available at the moment according to sun on the panels.

 

[SilverbackMP]

I think the SOC is supposed to figure itself over the course of a few charging cycles based on an algo …at least that’s the conversation I remember having somewhere here.

We resaved settings in all three SBs….things seems to be going much more smoother now.

We noticed that the amps were ramping down to around 2 as cell voltage climbed towards 54.1v/3.38v…which makes sense.

What doesn’t make sense is that when we plugged in a 1500 watt heater, the SIs took the power from the batteries rather than the SBs.

Maybe it does that until the SBs ramp up?

After the heater was turned on for a few minutes…charging amps ramped back up:




I’m not liking that cell variation…but it might just have to live that way until I finish building the remaining three batteries (and have been thinking about four more for a total of 12).

They probably need a new top balance or at least a live cell by cell boost or resistance balance ala Andy’s off grid garage. But that ain’t easy due to the damned cabinet I have them shoved into (along with 12 ft of 4/0 wire each). Look neat…wouldn’t do it again though.

 

[SilverbackMP]

I got all excited again…I knew better. The f$&&n system just shut itself off again. Every time we think we possibly have it fixed….

And then wait 15 minutes between restarts because SMA is German and that’s the way it is period, no questioning. I should send them a bill from my electrician.

 

[Hedges]

Frequency shift and ramp up of power from SB takes about 3 seconds.

Unless they are knocked offline, more likely with Rule 21 than offgrid, which do you have?
SI makes an excursion past the target frequency or frequency for target watts, possibly Rule-21 parameters cause 5 minute disconnect when driven to/past frequency for 0% watts.

Your cell voltages may not be identical, but still fairly full.

In some SCC, there is "rebulk" voltage, where battery is used to supply power for a while before it decides to harvest more PV. That doesn't fit the idea of keeping at float.

Does your battery drop below float voltage for a while before charging resumes?
or is voltage still above float, having dropped below absorption, so heater gets powered from battery for a while?

"Shuts off" - what shuts off and why?
BMS can disconnect.
Testing old AGM batteries, I have run them down until LVD or low SoC disconnect. Not sure if I then had to turn off breaker, let caps drain, and wake SI back up, or something else.

If you have load-shed relay, default from AGM had been shed load at 70% DoD, keep SI operating down to 80% DoD (at which point it shuts off), don't reconnect loads until back up to 50% DoD.

 

[SilverbackMP]

Frequency shift and ramp up of power from SB takes about 3 seconds.

Unless they are knocked offline, more likely with Rule 21 than offgrid, which do you have?
SI makes an excursion past the target frequency or frequency for target watts, possibly Rule-21 parameters cause 5 minute disconnect when driven to/past frequency for 0% watts.

Your cell voltages may not be identical, but still fairly full.

In some SCC, there is "rebulk" voltage, where battery is used to supply power for a while before it decides to harvest more PV. That doesn't fit the idea of keeping at float.

Does your battery drop below float voltage for a while before charging resumes?
or is voltage still above float, having dropped below absorption, so heater gets powered from battery for a while?

"Shuts off" - what shuts off and why?
BMS can disconnect.
Testing old AGM batteries, I have run them down until LVD or low SoC disconnect. Not sure if I then had to turn off breaker, let caps drain, and wake SI back up, or something else.

If you have load-shed relay, default from AGM had been shed load at 70% DoD, keep SI operating down to 80% DoD (at which point it shuts off), don't reconnect loads until back up to 50% DoD.

The Sunny Islands shut down.

We selected Island 60 hz when we commissioned these, but I noticed that two of them weren’t on the setting. No idea about the first one as I thought to check on two and three when we went back through.

I am thinking the settings didn’t stick again and that the first one is good to go. I just restarted the SI and number one SB and it seems to be working again.

I will deal with the other two tomorrow.

Curious to see what happens with the SOC as we are now pushing past 54.4 / 3.4 and should be in balancing territory.

 

[SilverbackMP]

Hmmm…we are going backwards now with zero external load…I suspect this is maybe the REC resistors burning off excess voltage as part of the balancing?

 

[SilverbackMP]

Looks like that one online Sunny Boy ramped down completely but no error mode so that’s a good thing.

I think the BMS is still working on balancing…still slightly worried about the SOC but I think (?) it will work itself out.



Anyhow…will work on it it tomorrow. Electrician went to a couple of small jobs and I’m sleep deprived (jet leg and then waking up in the middle of the night to a CPAP without power and then a trip out to the generator to let in run for a cycle (no autostart…yet)).

 

[Hedges]

So you think balancing is going on, therefore current has been requested to be low?
Reported SoC is 4%, even though lowest cell voltage is 3.354V?

If BMS wants low charge rate to prevent runner cells while balancing the low one, I'm wondering why max cell voltage 3.391V. Is this targeting 3.4V max? I'd think driving higher like 3.55V or 3.60V could accelerate charging.

And what voltage (and time) is used to reset SoC to 100%?

Active balancer, so able to put maybe 2A into one low cell? Rather than bleeding off 15 or 47 high cells?

Is this home built battery? What if you disconnected the (only one?) battery with a low cell - would the other batteries then permit higher charge rate?