Troubleshooting SMA and REC - Mostly (all?) Solved

[SilverbackMP]

Tried it at 500 A, thought I was cooking with bacon grease:


That lasted for about two minutes until it the contactors for the Sunny Boys disconnected (no errors) and the SIs disconnected from the generator with loads switching to battery.

Same deal at 400 A but the Sunny Boys stayed on. SIs disconnected from the generator.

Lasted maybe 10 minutes at 300 A, 15 minutes at 250A

Took it down to 224 A (doubled the 112 A setting but I don’t think it has anything to do with it). Lasted maybe 20 minutes be before SIs disconnect from gen.

Tied changing BMS setting to charge at 0.5 C…again lasted maybe 20 minutes. No change in charging watts either.

Tried it a second time, lasted until the last Sunny Boy went offline due to dusk. When it when offline the generator disconnected. Gave it it 3-4 minutes to see if would reconnect - no dice.

Note that I can been letting the system run for a bit on most of these to see if the generator would reconnect. Maybe I am not waiting long enough?

Also discovered that flipping the generator’s main breaker off on will prompt the Sunny Islands to reconnect. So it knows when it’s available at all times.

Also I think one or more of Sunny Boys were flipping a contactor ever time the Sunny Islands disconnected from the generator. Will pay more attention to this tomorrow.

 

[1201]

Tried it at 500 A, thought I was cooking with bacon grease:

View attachment 188971
That lasted for about two minutes until it the contactors for the Sunny Boys disconnected (no errors) and the SIs disconnected from the generator with loads switching to battery.

Same deal at 400 A but the Sunny Boys stayed on. SIs disconnected from the generator.

Lasted maybe 10 minutes at 300 A, 15 minutes at 250A

Took it down to 224 A (doubled the 112 A setting but I don’t think it has anything to do with it). Lasted maybe 20 minutes be before SIs disconnect from gen.

Tied changing BMS setting to charge at 0.5 C…again lasted maybe 20 minutes. No change in charging watts either.

Tried it a second time, lasted until the last Sunny Boy went offline due to dusk. When it when offline the generator disconnected. Gave it it 3-4 minutes to see if would reconnect - no dice.

Note that I can been letting the system run for a bit on most of these to see if the generator would reconnect. Maybe I am not waiting long enough?

Also discovered that flipping the generator’s main breaker off on will prompt the Sunny Islands to reconnect. So it knows when it’s available at all times.

Also I think one or more of Sunny Boys were flipping a contactor ever time the Sunny Islands disconnected from the generator. Will pay more attention to this tomorrow.

More progress. That's great. Now it just needs to stay charging.

If you go to the menu for warnings, events and failures, and look at the history it should tell you why it disconnected, with time stamps

 

[SilverbackMP]

I just check with no input from the Sunny Boys (I could turn them off in the day time but this testing also needs to put some level of power in my batters).

The SIs are still disconnecting from the generator after 20 or so minutes with no Sunny Boy activity. So that’s one element that might be out of the equation.

 

[1201]

I just check with no input from the Sunny Boys (I could turn them off in the day time but this testing also needs to put some level of power in my batters).

The SIs are still disconnecting from the generator after 20 or so minutes with no Sunny Boy activity. So that’s one element that might be out of the equation.

What does the most recent warning say?

 

[SilverbackMP]

What does the most recent warning say?

Didn’t check it. Will look in the morning.

 

[SilverbackMP]

Well I started reading manuals from the SMA approved (closed loop) battery list…every one of them wants you to manually set 221 02 BatCapNom. Mine is showing 230 AH (which is my individual battery sized). But this should be 230 AH x 5 (number of batteries). It won’t let me change it. Even with level 1 access. I thought…well maybe the BMS is feeding it after all. So I changed the AH capacity in the REC BMS from 230 AH to 1150 (Note, this entry is supposed to be per battery, not total). It showed 5750 AHs (1150 x 5) on the home screen, which is erroneous but I tried it out. Restarted the master BMS and the Sunny Islands several times, but the 230 AH figure in the BatCapNom never changed.

I think this has a lot to do with it and the BatCapNom needs to be adjusted, but I’m still trying to figure out how to do so.

 

[1201]

Well I started reading manuals from the SMA approved (closed loop) battery list…every one of them wants you to manually set 221 02 BatCapNom. Mine is showing 230 AH (which is my individual battery sized). But this should be 230 AH x 5 (number of batteries. It won’t let me change it. Even with level 1 access. I thought…well maybe the BMS is feeding it after all. So I changed the AH capacity in the REC BMS from 230 AH to 1150 (Note, this entry is supposed to be per battery, not total). It showed 5750 AHs (1150 x 5) on the home screen, which is erroneous but I tried it out. Restarted the master BMS and the Sunny Islands several times, but the 230 AH figure in the BatCapNom never changed.

I think this has a lot to do with it and the BatCapNom needs to be adjusted, but I’m still trying to figure out how to do so.

You have to
-stop the sunny island
-flip the DC breakers to off
- wait a minute or so and turn them back on(be careful about startup surge)
-on the init screen hold the button to enter the quick start guide and choose "new battery".

Then you'll be able to set the nominal battery capacity.

I think @Hedges said you may also be able to do this using the direct access menu but I've never tried that.

Before you do all that though- check the warning history. It may not be related to that at all

 

[kommando]

On my SI5048 BatCapNom is set when you Initialise a new battery at the start of the setup routine. Its then fixed and cannot be altered only displayed even with level 1 access.

 

[SilverbackMP]

Before you do all that though- check the warning history. It may not be related to that at all

Yep working on that now:

Mostly

W332 S1 Warning - AntiIslSlv1
W332 S2 Warning - AntiIslSv2
W331 AntiIsl

And a few others like:

W326 S3 Warning - ExtFrqLosSlv3
W324 S1 Warning - ExtFrqLosSlv1

 

[1201]

Yep working on that now:

Mostly

W332 S1 Warning - AntiIslSlv1
W332 S2 Warning - AntiIslSv2
W331 AntiIsl

And a few others like:

W326 S3 Warning - ExtFrqLosSlv3
W324 S1 Warning - ExtFrqLosSlv1

W332 1 Disconnection from utility grid/generator due to islanding (slave 1)

Ok it's definitely this but I wonder why it's happening.

My wag is it has to do with these settings
Can't you try turning anti islanding sensitivity to low.


Also is your source set to gen only or gengrid?

 

[kommando]

Yep working on that now:

Mostly

W332 S1 Warning - AntiIslSlv1
W332 S2 Warning - AntiIslSv2
W331 AntiIsl

And a few others like:

W326 S3 Warning - ExtFrqLosSlv3
W324 S1 Warning - ExtFrqLosSlv1

Those are related to anti islanding on S2 and S1 plus Grid/gen disconnections due to insufficient external frequency on S1 and S3.

That could be intermittent comms issues between master and the slaves or real issues with some slaves with grid and frequency which means connections need looking at.

More debugging needed.

 

[SilverbackMP]

W332 1 Disconnection from utility grid/generator due to islanding (slave 1)

Ok it's definitely this but I wonder why it's happening.

My wag is it has to do with these settings
Can't you try turning anti islanding sensitivity to low.


Also is your source set to gen only or gengrid?

Gen only.

Going to eat, then play with the battery size and also adjust that sensitivity. Probably try the sensitivity first.

 

[SilverbackMP]

We put a clamp meter on the generator the other day and it was dead on nuts for frequency and voltage and I’ve been letting it get plenty warm up and smooth out. Still, could be some stuff going on that would just show up on an oscilloscope.

It’s an older Ingersol Rand unit with fairly low hours (3xxx) with a Cummins 4BT, so I doubt they cheaped out on the generator part.

 

[Hedges]

When SI disconnects from grid input, I see SB drop off. With SI carrying the load, SB reconnects later (mine are commanded to switch between on grid UL-1741/off grid island, rather than using Rule 21. Your may be island all the time because generator.)

With generator, I'd expect SI to disconnect as soon as loads + charging are less than SB produces. That disconnect hopefully don't glitch, and SB remains connected.

To set some parameters, I have to first navigate to password. Manual explains how to calculate a password from something like date or runtime hours. Then additional parameters can be adjusted. I just did that to change some VRLA settings.

 

[SilverbackMP]

When SI disconnects from grid input, I see SB drop off. With SI carrying the load, SB reconnects later (mine are commanded to switch between on grid UL-1741/off grid island, rather than using Rule 21. Your may be island all the time because generator.)

With generator, I'd expect SI to disconnect as soon as loads + charging are less than SB produces. That disconnect hopefully don't glitch, and SB remains connected.

To set some parameters, I have to first navigate to password. Manual explains how to calculate a password from something like date or runtime hours. Then additional parameters can be adjusted. I just did that to change some VRLA settings.

Yep, been doing the password thing, going from “Lvl 0” access to “Lvl 1.”

 

[kommando]

My understanding of BatCapNom is that it is used in the SOC calc when the SI is set to non Lithium, so for Lithium it should not come into play as the SOC comes from the BMS. Mine is on lead and this needed setting correctly plus a shunt calibration so the SOC calcs would be accurate. I initially got it wrong and the SOC reported did not make sense after a few days, now its set closer to reality its a lot more accurate.

 

[SilverbackMP]

My understanding of BatCapNom is that it is used in the SOC calc when the SI is set to non Lithium, so for Lithium it should not come into play as the SOC comes from the BMS. Mine is on lead and this needed setting correctly plus a shunt calibration so the SOC calcs would be accurate. I initially got it wrong and the SOC reported did not make sense after a few days, now its set closer to reality its a lot more accurate.

I dunno. It prompts it to be set when on the Lithium mode and Discover Batteries (which is not a SMA partner as I later figured out) has you set it. Their system is Master/Slave setup couched in marketing terms (calling their master the “Lynk II”). From their manual.




My BatCapNom currently shows 230 AH.

Then reading from the SI manual (page 72 in my physical copy):

4, The following parameter must be set been “New Battery” is selected:

- Nominal capacity of the battery (100 Ah to 10,000 Ah for VLRA, FLA and NiCd. 50 Ah to 10,000 Ah for Lion_Ext-BMS), default setting: “100 Ah.”

You’d think the BMS would pass this information to the SI, but it apparently doesn’t since it prompts you for it. And since there is no place to enter the number of batteries, what they are calling “nominal” must mean all combined. More lack of SMA clarity. “You must attend our academy, anything less is verboten.”

 

[SilverbackMP]

And this is why I wish I had went Victron…besides the ability to stack more of them, there are many more home brews with specific recipes…and pieces designed to jigsaw together. I am sure SMA’s tech (even though it’s older) is great and probably better…it’s a more complex system and complex systems have more failure points. And getting It dialed in is a bit of chore. Victron seems a little more Apple like.

 

[kommando]

You have to change the setting in case it does matter, but the manual was rewritten for Lithium late on in the life of the SI5048/SI6048 so likely its just a section not edited correctly and with a lead only requirement reading as universal.

 

[1201]

And this is why I wish I had went Victron…besides the ability to stack more of them, there are many more home brews with specific recipes…and pieces designed to jigsaw together. I am sure SMA’s tech (even though it’s older) is great and probably better…it’s a more complex system and complex systems have more failure points. And getting It dialed in is a bit of chore. Victron seems a little more Apple like.

Once its all running smoothly everything will be forgotten. I'm shocked out how well my sunny islands work.
Don't get me wrong- there are some things I wish were different- for example the sunny island can't be set to export to grid without also importing from grid. So I have to wait till the battery is full, to turn on the grid otherwise it will charge the battery from the grid

But, there are some things that absolutely amaze me.

How well it controls my growatt pv inverter. I tried one or two settings on the growatt and the sunny island can clamp the growatt output from 11k down to zero and anywhere in between. I'm not sure victron can do this. I haven't seen how they control pv inverters except fronius

Grid as generator. I love that I can have it seamlessly turn on the grid at my set soc.

I also like that it can turn on the grid at high loads.

I'm using vrla settings but I've found it to be highly accurate once I added enough batteries. Si recommends 100ah per kw so for two si there should be 1200 ah. When I only had 560 ah I noticed that it would turn on the grid at 35% soc in the morning under high loads. My setting was 20% so this was wrong. Apparently the sunny island resets the soc to 20% based on voltage(it doesn't say what voltage) so in the morning when I had everything running, the voltage would drop and sunny island would reset soc to 20% and kick in the grid.

Once I got to 1120ah I noticed the voltage didn't sag so badly and the soc doesn't reset. Even down to 25% soc its within 2% of what my smart shunt says and it even tells you the amount of error in it's soc calculation.

So before I was looking to run bms comms but I don't feel the need to any longer since I believe the internal soc is very accurate with enough battery capacity