SMA Sunny Island not charging on sunny days (not solved)

[callmeburton]

Branching this off of my Generator post as I can see the thread getting too overwhelmed with replies. Original https://diysolarforum.com/threads/what-generator-works-with-sma-sunny-island.72963/post-1277033

Basics:
After getting my SI to work with a generator via manual mode it will sometimes not charge the batteries when there is full sun and the battery charge mode is boost or full (but I suspect all modes at this point) ... Changing #222.01 BatChgCurMax to anything other than what it is currently set to (90) will start the SI charging again. Changing back to 90 does not revert the state for that current "session" I am going to call it.

I call it a session as early morning I saw amps going into batter, SOC on SI 60%, I come back mid day and it is on BOOST and 1-2a max going into batteries. I change the parameter above and instantly get 50amps into both batteries for a while.

Likely FIX: 2024 12 21 UPDATE
Adding a 2k resistor to the temp sensor (since I never had one installed and technically VRLA is lead acid and "requires" one) has likely resolved this issue. Worked for 2-3 days, today 2024 12 22 stopped charging in full sun, no errors, no events to point to. Changing the single setting again allowed for charging.

Details
VRLA, 560a, 48v basic battery setup, open system. Have Batrium controlling pre-charge and main contractor, have JK BMS on both packs (another story).

SI has shunt configured as 50mV/500, previously worked this way.

Some settings I currently have, and I can't recall if they were set this way before this issue.
#222.12 AutoEquChrgEn is Disabled

#233 Grid Start section
03 GnSocTm1Str 80% (40% stock)
04 GnSocTm1Stp 95% (80% stock)
05 GnSocTm2Str 80% (40% stock)
06 GnStocTmStp 95% (80 stock)
09 GnPwrEna On (default Off)
10 GnPwrStr 13.0 kW (stock 4)
11 GnPwrStp 5.0 kW (stock 2)

#234 Generator Control section
07 GnStrMod Manual (I have to tell the SI to "start" gen on the SI for it to see it IIRC)
Lots of settings reduced for testing (aka to get faster failure points) but I don't think they are an issue
08 GnOpTmMin 1m
09 GnStpTmMin 1m
10 GnCoolTm 1m
11 GnErrStpTm 0h
12 GnWarmTm 10s
20 GnAiSns Low (stock is normal) Not sure if this was "required" to change to get things working with Gen

#231 section
10 TotPvPwr (is this DC or AC from the sunny boys?) Currently 7.5kW (assuming SB AC, might have previously been 11.5 but can't remember if I changed it)
11 ClstPwrNom 11.5kW (oddly this is my DC PV nominal IIRC)

 

[kommando]

When the full sun is up but no charge what are the reported battery V on the SI, Batrium and the JK BMS. What are the target V per cell for the Boost and Full (as you have VRLA then these are 6 (not 4) cells per 12V bat or 24 (not 16) for 48V battery.

 

[callmeburton]

Examples:
SOC in SI, 45%, charge mode FULL, JK BMS SOC 60% (they are not synced yet)
SOC SI 65% SOC, BOOST, JK BMS SOC 70%

The voltage / SOC of the packs and SI are not what is keeping the cells from getting amps. If they were changing the max charge amps setting wouldn't do anything IIRC.

I think these might be current targets, took the settings from here. https://bigbattery.com/2022/05/10/s...mbZrfxaCNAGVZxQ4OZoSNznZrgbfJzZDdtu2d63bgK3r0
That said the actual output voltage seems like it is 23 vs 24 count for the VRLA so I might have adjusted them slightly based on the voltage the SI was trying to get when looking at the compact meters.

#222
07 2.43 V
08 2.33 V
09 2.33 V
10 2.27 V

At one point I set these all so the final voltage from SI was 55.6 IIRC, didn't seem to affect anything in the pattern I am now noticing.

 

[kommando]

I got the cell count for VRLA wrong and corrected my post.

2.33 is 55.92V target V, but what is the Target V set in the Batrium and the JK, if its higher then it should be working, if the same or lower then the BMS's will stop the charge before the SI and the SI will be confused.

I run closed loop with off grid and the SI takes 1V off the BMS target V which I set to 56.2 and the SI works to 55.2 but in good sun the frequency shifting has a lag and I have seen 55.8V before the PV is reduced. The 1V reduction of the BMS target V seems to work so in VRLA I would be setting the Target V in the BMS to 56.92V. If you want a lower V on the BMS then take 1V off and give that to the SI.

In previous posts on running open loop on an SI in VRLA it was recommended to set the battery on initialisation as 47V which then allowed the 24 cell V to be set lower than with a 48V.

 

[callmeburton]

2.33 is 55.92V target V, but what is the Target V set in the Batrium and the JK, if its higher then it should be working, if the same or lower then the BMS's will stop the charge before the SI and the SI will be confused.

Just updated batrium to 3.56 to match JK, previous settings below but shouldn't have been affected by the voltages at the time of charging, and if it it was the only thing it can do is disconnect the contactor as it has no COMs with the SI.


For JK I have following, also higher end voltage then what I tell the SI to do


The SI is kinda odd with the voltages, so right now it is on BOOST, target is 2.39, in compact meters Bat2 page it is asking for 55.9 ... so the 24 multiplier doesn't work so well. I have been manually adjusting these numbers up / down to see what the target voltage is and then looking at the difference between what the SI thinks the bat voltage is vs what it really is to "set" them where I want them to be.

I haven't had issues with this setup prior to getting my gen working in the system. It is an open loop system for sure as the Batrium can drop coms with the SI as can the JK IIRC. That is why I look at Bat2 page on compact meters to see what the SI is trying to achieve

 

[callmeburton]

Real time update. Attached was one of the two JK BMS before I started this video

which shows the behavior I am highlighting.

 

[kommando]

222.13 BatChrgVtgMan is what the SI has as its target V, mine shows 55.2V when the BMS is set to 56.2V.

That should match 24 times boost or full.

Also have you calibrated the V on the JK, its not always right and needs to match the SI reading.

Note due to the steep incline in V to SOC at the top end of lithium the V will rise rapidly for a very small amount of charge. Once the cell hit 3.4V the V climbs steeply as you can see on this graph. The JK BMS will trip on one cell reading high if the pack is unbalanced but the battery V will not be 3.6x16 57.6V but lower.

 

[callmeburton]

222.13 BatChrgVtgMan is what the SI has as its target V, mine shows 55.2V when the BMS is set to 56.2V

222.13 does not exist in my settings likely because it is open loop not closed.

 

[kommando]

SI5048 running SI6048 firmware has a few quirks.

Try 226.01 BatChrgVtgMan
Manual setpoint of the battery charging voltage with disabled battery management in V
Range 41.0 V to 63.0 V
Default 54.0 V

 

[callmeburton]

#225 doesn't exist in my system, SI6048, GEN, with PV attached.

 

[kommando]

226 not 225

Its copy and pasted from the SI6048 manual for the later firmware that covers lithium.

 

[callmeburton]

Sorry 266, mistyped. just double checked

 

[callmeburton]

20 GnAiSns Low (stock is normal) Not sure if this was "required" to change to get things working with Gen

Changed this back to default (normal) will try on next cloudy day ... don't think it will affect this problem but it is one of those settings I don't think I "needed" to change to get gen working.

 

[kommando]

226 not 266

But I can only get to it using Sunny Data Control or Webbox, so why is it in the manual but editable only from outside the SI.

 

[callmeburton]

no web box, no data control. Again this 100% worked as expected before adding the gen.
I meant 226 but been going in / out to run backhoe and fat fingered and didn't read what I typed >__< Sorry.

Have you tried the reverse of what you are suggesting is the "fix" to see if it causes the issue on your end?
Change your 226 to out of range, then change max amps for battery and you see it suddenly charge?

 

[kommando]

In closed loop I have changed the 226.01 and it initially takes it but within 15 mins it reverts to the BMS -1V which must be a feature of the Lithium setting.

I can see what the 226.01 voltage is under meters 120.03 but its not editable.

I would be looking at the cell delta reported by the JK BMS at 55.2V, when my packs are balanced the delta is 0.008V but when I first set them up is was 0.120V on the worst pack.

There is a series of .evt files on the SI SD card which would contain the events the SI saw at the time the charging stopped prematurely including any error codes.

To look at the file use notepad++

https://notepad-plus-plus.org/

One of my typically uneventful evt file entries, look up the codes in the manual for more info

TimeStamp;Type;Number;Name;Add.Info;;
06.01.2024 00:35:44;W;755; LoBatMod1;0;;
06.01.2024 00:41:46;E;101; Standby;0;;
06.01.2024 06:00:00;E;102; Startup;0;;
06.01.2024 06:00:00;E;103; Run;0;;

 

[callmeburton]

If I get a chance today I will bring the system down to get the logs. Haven't done that since trying to figure out why the SI was loosing the signal from the batrium BMS Coms input :/ (One of the reasons the SI doesn't have eyes on any of the BMS's because being offgrid we can't afford surprise outages :D )

When I edit #222.07-10 if the SI is in the mode I am editing I can see it change the voltage target in the compact meters section. It takes a minute to do it but it does go up or down accordingly. When I have the system restarting I will check the #600 settings as I think you can direct edit stuff there when the system is down as well and look for the voltage setting you mentioned.

 

[1201]

If I get a chance today I will bring the system down to get the logs. Haven't done that since trying to figure out why the SI was loosing the signal from the batrium BMS Coms input :/ (One of the reasons the SI doesn't have eyes on any of the BMS's because being offgrid we can't afford surprise outages :D )

When I edit #222.07-10 if the SI is in the mode I am editing I can see it change the voltage target in the compact meters section. It takes a minute to do it but it does go up or down accordingly. When I have the system restarting I will check the #600 settings as I think you can direct edit stuff there when the system is down as well and look for the voltage setting you mentioned.

Do you see anything recently in the write or warning history on the island?

 

[kommando]

If I get a chance today I will bring the system down to get the logs.

No need to take the system down, just make sure before you pull the SD card the SI is not reading/writing, it tells you by saying do not remove the SD card. Once the SD card is out make a copy on your PC and reinsert the SD card into the SI. You can view all the info with the SD card back in the SI. If while the SD card is out the SI needs to write to the SD it just holds the info until its reinserted.

 

[callmeburton]

The evt logs are empty, IIRC I have to tell it to write them before pulling card so will do that later. BUT I did pull current params down to a CSV and the mentioned setting is in there at the default 54 value.

 

[callmeburton]

Will add this as I have to step away. My linked post first noting this behavior's logs. The highlighted record is when I changed the one setting for max char current for battery.

UPDATE: Don't see any events or failures for that day.

 

[callmeburton]

Still an issue :/
Since the above I have added my third pack to the mix. It was "good" for the day I added it then today, three days later and first partial sun day since, I come in to find the batteries getting 1 amp ... I have been trying to change other settings to see if it would "trigger" the correct behavior but so far only the one mentioned "fixes" the issue for at least a day ...

Debating somehow factory resetting both SMA devices and commissioning again. I feel like there is something really obvious I am missing here but can't figure it out. Things worked before adding the second pack ... so something I changed then, settings wise or physical wise, created this odd problem.

 

[1201]

@callmeburton

I noticed your video was reading battery temp at 71degC

Do you have battery temperature compensation 222.11 set to anything other than zero?

 

[callmeburton]

Do you have battery temperature compensation 222.11 set to anything other than zero?

It is 0 for sure as I remember default is 4 and I set it to 0 when I commission the battery as this setting gets reset :/
I don't think there is a "battery temp" probe hooked up to the SI at all. Never has been.

I don't think this is the issue though, that said I have some extra temp sensors at my other house we are moving out of I can grab maybe Friday. Scratch that, I see SMA has a "special" temp sensor so the ones I have likely wont work here :/ Update to the update, threw in a 2k resistor so the SI thinks it has a sensor connected and it is about 25C lol

Everything worked when I only had one battery in the system ... IIRC
no BMS COMs and SI set to VRLA

 

[callmeburton]

Update (day 1):
SI seems to be charging ... but IIRC it will be "ok" for up to a day ... so we will see tomorrow around noon when the clouds break a little if it is "fixed" by simply tricking the SI into thinking it has a temp sensor :/

Going to be odd if that hack works since I haven't had COMs hooked up between the batrium / SI for so long now.