Sunny Island SI5048 - disable battery management- driving me crazy

[cattledog]

Hi everyone,
I have a Sunny Island SI5048. Its a nice inverter and works fine for the most part, but ever since I switched to Lithium based batteries the SOC tracking the inverter tries to do is very inaccurate. I go through the usual process to commission the new batteries (4 ampere time 200ah batteries in series, 48v) and the inverter works well afterwards - starting with a 50% SOC - even though the batteries were fully charged when connected.

my solar array is producing - and putting energy back in the batteries - the SOC according to the SI5048 steadily goes down, until "battery protection" activates and the whole inverter shuts down - even though plenty of power remains in the batteries. The solar arrays are connected to Victron MPPT's with their own separate connection to the batteries not involving the SI5048, except that all power in and out does go to a shunt that is connected to the SI5048 of course.

I can see that the SI5048 is able to see current going into the batteries from solar with the TotBatCur parameter on the front panel. Regardless the SOC just keeps heading down until the inverter shuts down - not a nice situation in the middle of the night lol.

I can probably troubleshoot this further, but what I really want to do is disable the whole battery protection mess and just let the BMS's in my batteries do the job of protecting the batteries. I'm used to Victron inverters that don't have this over complicated SOC tracking crap to deal with that SMA seems to require.

Anyone know how to disable battery protection with the SMA Sunny Island units? I called SMA tech support and the guy didn't really have clue on what was causing my issues, and said the battery protection cannot be disabled or "we can't tell you how" - was one or the other.

Thank you!
Cattledog

 

[Hedges]

There is a correct and an incorrect order when connecting shunt and charge controller to batteries. Presumably you got the correct one, for the case of SCC which does not communicate with Sunny Island.

Do your lithium batteries have communication with Sunny Island or not?
It seemed to me that VRLA settings in Sunny Island would not line up with desired max/min voltages of a 16s LiFePO4 battery using 48V nominal (24s) lead-acid battery. A different nominal representing different number of lead-acid cells will get adjustable parametres within range.

I don't think you should have BMS protect batteries on a normal basis. You should have SI shed loads and shut down if necessary before BMS steps in.

We understand that SI will do even worse estimating SoC for lithium than it does for lead-acid. However, I think voltage limits override. What voltage limits have you set and at what voltages does SI shut down? What SoC according to BMS? What lithium capacity do you have (200 Ah) and what capacity have you told SI?

 

[sunshine_eggo]

This statement:

my solar array is producing - and putting energy back in the batteries - the SOC according to the SI5048 steadily goes down, until "battery protection" activates and the whole inverter shuts down - even though plenty of power remains in the batteries. The solar arrays are connected to Victron MPPT's with their own separate connection to the batteries not involving the SI5048, except that all power in and out does go to a shunt that is connected to the SI5048 of course.

Are you certain the MPPT are sending the charge through the shunt used by the SI? If the Soc is continually decreasing, it sounds like the shunt isn't seeing the input current.

 

[Hedges]

Do you have loads? How many watts?
Each SI consumes 25W
How much power is being produced by PV?

You could lower the SoC for shut-off from default 30% SoC, maybe to zero or just above. But presumably it would still shut down.

Is polarity of shunt sense wires and orientation of shunt correct?

I could understand lithium SoC being estimated wrong, but I don't expect it to drop if battery is full and voltage remains high. Wrong polarity signal from shunt would explain it. Or SCC charging battery and only SI's draw going through shunt, as Sunshines suggests.

Other people do use this configuration successfully, with and without BMS communication. Not sure if they have DC coupled SCC. TrukinBear used SCC with shunt and FLA, but also had communication and SoC was incorrectly calculated due to connection on wrong side of shunt.

 

[cattledog]

Hi there, I am using two Victron MPPT controllers, one for each of my solar arrays. I don't think Victron MPPT units support any sort of communication with SI's. My batteries don't support communication either, so I am reliant on a shunt to keep the SI informed of battery status.

LOL I guess it isn't possible to disable this troublesome SOC "feature" on the SI. No one seemed to engage that question. The Victron inverters I use elsewhere use voltage readings to initiate battery protection shutdown, and are not reliant on a SOC "calculation" using a shunt or communications etc. I much prefer the simpler voltage based setup Victron follows! If anyone knows how to turn off or defeat SOC tracking on the SI I'd still like to hear about it.

As I stated in the original post, if I view the TotBatCur information on the front panel of the inverter (120.06) I can see that number jumping around when the arrays are producing energy.

I have the shunt setup correctly in terms of loads and charging sources. All the loads and charging currents are connected to the shunt terminal that is more distant from the batteries.

I also wondered if the polarity on the wires from the shunt to the SI5048 might be reversed - I think that could be possible. In watching the TotBatCur diagnostic on the SI5048 today, I noticed that number showed a positive value when the arrays were energized and charging the batteries. Could anyone look at their SI unit, and tell me if TotBatCur should show positive or negative, when charging or discharging? If TotBatCur should show a negative number when the batteries are charging then I bet the wires from the shunt to the SI 5048 are reversed.

The site with this setup is a bit remote so I'll have to take these ideas and return to troubleshoot further.

Chris

 

[cattledog]

Do you have loads? How many watts?
Each SI consumes 25W
How much power is being produced by PV?

You could lower the SoC for shut-off from default 30% SoC, maybe to zero or just above. But presumably it would still shut down.

Is polarity of shunt sense wires and orientation of shunt correct?

I could understand lithium SoC being estimated wrong, but I don't expect it to drop if battery is full and voltage remains high. Wrong polarity signal from shunt would explain it. Or SCC charging battery and only SI's draw going through shunt, as Sunshines suggests.

Other people do use this configuration successfully, with and without BMS communication. Not sure if they have DC coupled SCC. TrukinBear used SCC with shunt and FLA, but also had communication and SoC was incorrectly calculated due to connection on wrong side of shunt.

A few weeks ago I did lower the battery protection threshhold for all three battery protection levels to "0" - zero - thinking that would disable this dumb SOC problem. No such luck - when the SOC goes to zero the SI shuts down and doesn't care that the battery voltage was still showing around 54 volts! lol

 

[sunshine_eggo]

A few weeks ago I did lower the battery protection threshhold for all three battery protection levels to "0" - zero - thinking that would disable this dumb SOC problem. No such luck - when the SOC goes to zero the SI shuts down and doesn't care that the battery voltage was still showing around 54 volts! lol

Sorry I didn't mention it. I looked it up in the manual, and I saw no way to disable it; however, the battery monitor does count current to establish SoC. It sounds like it's only counting down which implies it's not seeing charge (wrong side of shunt or shunt direction reversed), or it's registering charge as discharge (shunt direction or sensing wires reversed).

FWIW, SoC based protections are better than voltage. With voltage based, a heavy load may trigger cut-off even if the battery is at a high state of charge. It happens all the time around here where people see 12.1-12.2V as "50%", so they set their disconnect there and wonder why their inverter has tripped off in the morning.

I bleed blue, so it pains me to admit that another manufacturer does it better natively.

 

[toms]

Not sure if the SI5048 is the same as the SI8.0, but with mine I use a REC BMS and the SOC is provided to the SI via the REC shunt. The SI doesn’t have a shunt.

The only thing is you need to reset the SI critical SOC to 1%, because if it does disconnect it uses those numbers and will not remain on until the BMS reports SOC is over it’s thresholds.

I use both SMA and Victron. The Victron is cheaper, that is the only thing better about it.

 

[sunshine_eggo]

Not sure if the SI5048 is the same as the SI8.0, but with mine I use a REC BMS and the SOC is provided to the SI via the REC shunt. The SI doesn’t have a shunt.

The only thing is you need to reset the SI critical SOC to 1%, because if it does disconnect it uses those numbers and will not remain on until the BMS reports SOC is over it’s thresholds.

His batteries are not in communication. 4S 12V "dumb" Amperetime. He's being forced to use the in-built battery monitoring designed for lead-acid because it can't be disabled.

I use both SMA and Victron. The Victron is cheaper, that is the only thing better about it.

You cut me deep bro...

 

[cattledog]

Sorry I didn't mention it. I looked it up in the manual, and I saw no way to disable it; however, the battery monitor does count current to establish SoC. It sounds like it's only counting down which implies it's not seeing charge (wrong side of shunt or shunt direction reversed), or it's registering charge as discharge (shunt direction or sensing wires reversed).

FWIW, SoC based protections are better than voltage. With voltage based, a heavy load may trigger cut-off even if the battery is at a high state of charge. It happens all the time around here where people see 12.1-12.2V as "50%", so they set their disconnect there and wonder why their inverter has tripped off in the morning.

I bleed blue, so it pains me to admit that another manufacturer does it better natively.

That makes sense - that you could get unwanted shutdowns from temporary voltage sag when under a heavy load- and soc based shutdown could avoid this. Still prefer being able to turn off soc based “protection” but doesn’t seem possible.

 

[cattledog]

Check this out - looks like my shunt wires are probably reversed

https://www.altestore.com/static/datafiles/Others/SI5048U_Manual.pdf#page45

See page 46, bullet #3

I am not seeing negative numbers under charging conditions so my shunt wires may be reversed - that would be an easy fix!

Will try it soon

Cattledog

 

[Hedges]

Once you fix the immediate problem ...

Also review your absorption, float, low-voltage disconnect settings.
Determine if the Sunny Island settings are able to hit the lithium battery voltages you want.
Default is 48V nominal, and I don't believe the desired settings are possible.
You can set VRLA nominal battery from 42V to 52V in steps of 2V, and one of those should make such settings possible.

 

[SilverbackMP]

His batteries are not in communication. 4S 12V "dumb" Amperetime. He's being forced to use the in-built battery monitoring designed for lead-acid because it can't be disabled.



You cut me deep bro...

Could just take em apart, buy more cells, and make 16S. I am truly impressed with the REC. I went in thinking I’d have to fight things and f$ck with it. A lot of work putting the batteries together but the actual commissioning was about as plug n play as one would hope for. It just works. Put the Sunny Island(s) on lithium mode and the Rec just takes over.

Edit: I think some of the REC units support 4S. The M/S units might even do this.

 

[cattledog]

Once you fix the immediate problem ...

Also review your absorption, float, low-voltage disconnect settings.
Determine if the Sunny Island settings are able to hit the lithium battery voltages you want.
Default is 48V nominal, and I don't believe the desired settings are possible.
You can set VRLA nominal battery from 42V to 52V in steps of 2V, and one of those should make such settings possible.

Thank you- my si5048 is older and does not have a lithium option, only VRLA. I will look into changing the nominal voltage setting as you suggest.

It might also be possible to upgrade my si firmware to gain access to a lithium option. The si says “2006” when it boots so it’s an older one!

Seems like the nominal voltage with my batteries should be 13.2 x 4 (52.8)

Chris

 

[Hedges]

SI 5048 is firmware compatible with new models, so that can be updated.
I have not figured out how to back up previous version, and SMA only posts most recent version.

Firmware update only beneficial if you have a compatible BMS. REC is. Current Connected says they now have SOC pretty much working closed-loop.

13.2 x 4 = 52.8 would not be nominal voltage.
12V x 4 = 48V is a typical nominal voltage.

For SI, nominal lead-acid settings can be 42V to 52V, in 2V increments.
This was so a failed cell could be removed.

You would not be setting SI's VRLA parameters for actual nominal voltage.
You would be lying to SI about how many (2V nominal) lead-acid cells you have, such that low-voltage disconnect of that number of lead-acid cells matches desired 48V lithium low voltage shutdown, and such that absorption and float voltage of that number of lead-acid cells matches desired lithium voltage.

The allowed parameter ranges in SI for low-voltage disconnect, absorption, float multiplied by 24 (lead-acid) cells do not all line up. Something like 46V nominal (23 lead-acid cells) did, if I remember correctly. But double check the parameter ranges and figure it out yourself. I haven't used it but another member did.

It may also be useful to lie about the Ah capacity, depending on how its SoC estimation works, to get load-shed, shutdown, and other features working.

You should have a precharge scheme to avoid burning SI breaker and any relay contacts.

Max wattage and surge capability of SI may or may not be limited by your batteries' BMS current limit.

 

[HamzaH]

Thank you- my si5048 is older and does not have a lithium option, only VRLA. I will look into changing the nominal voltage setting as you suggest.

It might also be possible to upgrade my si firmware to gain access to a lithium option. The si says “2006” when it boots so it’s an older one!

Seems like the nominal voltage with my batteries should be 13.2 x 4 (52.8)

Chris

Hello, I am new to this forum, and I have recently ( one month ago) hooked up an si5048 with lifepo4 batteries with 250ah Daly bms, I have been also having problems with my si5048. especially when the bms cuts the charging due to full battery. have you managed to find a fix to this issue ?
also I would like to point that I have just found a firmware for the si5048-us that has the option to have li-ion bms in battery options. i still haven't tried to install the firmware on my SI5048, as i am in Australia and the SI5048 that I have is not a US model .

does anyone know if the firmware would work on mine. or if there is a fix for the battery issues .

thanks all

 

[HamzaH]

Hello, I am new to this forum, and I have recently ( one month ago) hooked up an si5048 with lifepo4 batteries with 250ah Daly bms, I have been also having problems with my si5048. especially when the bms cuts the charging due to full battery. have you managed to find a fix to this issue ?
also I would like to point that I have just found a firmware for the si5048-us that has the option to have li-ion bms in battery options. i still haven't tried to install the firmware on my SI5048, as i am in Australia and the SI5048 that I have is not a US model .

does anyone know if the firmware would work on mine. or if there is a fix for the battery issues .

thanks all

this is the link for the setup of the battery and down in the video, there is a learn more that takes to a download page that has the US firmware that I downloaded .

 

[Hedges]

I thought someone mentioned an SI firmware revision that was for lithium batteries themselves, but all I'm aware of is one to communicate with lithium BMS.

DIY users can get REC-SI BMS for their builds, which is supported by the US version SI 4548, 6048, and 5048 inverters. There is also a list of supported batteries. The newer 8.0 H series of European model SI have a different battery list. I see that is the latest offering in your market.

Battery Inverters | SMA Australia

The Sunny Islands offer you first class possibilities in the installation of self-sufficient energy systems. In conjunction with a battery pack the devices form a self-sufficient AC voltage grid, which meets with the highest of quality standards.

www.sma-australia.com.au

Yours is a 230V 50Hz market, so the US firmware almost certainly won't work.

Check your Australian SMA site for downloads of firmware matching your model. See if there is a way to back up your existing firmware in case you want to revert to it; I don't think SMA makes older versions available.

Using search feature I found:

Search Results | SMA Australia

www.sma-australia.com.au

https://files.sma.de/downloads/SUNNY-ISLAND5048-7210-7200E.zip

If you unzip that, there should be release notes.


Without a BMS that is compatible, use VRLA but change voltage and time settings to taste.
As I recall, nominal 48V 24s lead-acid battery and per cell voltage settings do not match 16s LiFePO4 voltage. I think nominal 46V 23s lead-acid was the one that did. But put the numbers in a spreadsheet and determine for yourself.

 

[HamzaH]

I thought someone mentioned an SI firmware revision that was for lithium batteries themselves, but all I'm aware of is one to communicate with lithium BMS.

DIY users can get REC-SI BMS for their builds, which is supported by the US version SI 4548, 6048, and 5048 inverters. There is also a list of supported batteries. The newer 8.0 H series of European model SI have a different battery list. I see that is the latest offering in your market.

Battery Inverters | SMA Australia

The Sunny Islands offer you first class possibilities in the installation of self-sufficient energy systems. In conjunction with a battery pack the devices form a self-sufficient AC voltage grid, which meets with the highest of quality standards.

www.sma-australia.com.au

Yours is a 230V 50Hz market, so the US firmware almost certainly won't work.

Check your Australian SMA site for downloads of firmware matching your model. See if there is a way to back up your existing firmware in case you want to revert to it; I don't think SMA makes older versions available.

Using search feature I found:

Search Results | SMA Australia

www.sma-australia.com.au

https://files.sma.de/downloads/SUNNY-ISLAND5048-7210-7200E.zip

If you unzip that, there should be release notes.


Without a BMS that is compatible, use VRLA but change voltage and time settings to taste.
As I recall, nominal 48V 24s lead-acid battery and per cell voltage settings do not match 16s LiFePO4 voltage. I think nominal 46V 23s lead-acid was the one that did. But put the numbers in a spreadsheet and determine for yourself.

I appreciate your time and effort to reply to me.
I already have the latest firmware , but li-ion is disabled in this please check below comments :

Firmware release v7.203/7.200 (only for SI 5048)
========================================================
- Bug Fixing: Disabling the option to set “Li-Ion” as the battery
type in the Quick Configuration Guide and in the menu “Battery”.
This is a failure and it is still not allowed to use Li-Ion batteries
in combination with the SI5048!! Please do not support or apply this function!


Also please note that I already have the VRLA set at 23 cells and this is what I am using now, but occasionally I get a failure (f710 several system restart) on the system, when the battery gets full and the bms blocks the charging of battery.

I have lowered the cell voltage settings on the SI5048 for all settings (boost, float...) to 2.43v/cell = 2.43V x23 cells=55.89v /16cells Lifepo4 = 3.5v/cell which is lower than the 3.6v/cell for the lifepo4 when full.
I have also tries to set the battery capacity amp lower (I have 400ah cells and tried setting it at 350ah so that the SI would think it is full before the bms and would turn off the charging before the bms, but also didn't work.

can you please advise what recommended settings to have? am I missing anything here ?



I have started havethe doubt that my SI is not changing the frequency from 50HZ when full to have the sunnyboys disconnect when battery is full, and

 

[Hedges]

Sounds like you're doing the right things. I haven't used SI with lithium, but others here have.

How many kW of PV and Sunny Boy do you have?
How many kWh of battery? (400 Ah)

I have oversize PV array compared to battery, and default charge rate of about 0.6C would never occur if sized according to SMA recommendations.
I reduced charge rate to 0.2C for my AGM.

Reduced charge current may help with the issue.

Does BMS report what it disconnected for? Maybe a cell that was a runner?
How much "absorption" time is it set for? I think holding within a voltage range allows BMS to rebalance. Maybe yours are drifting apart? Check cell voltages while fully charged.

Your SI should raise frequency when battery isn't accepting all the power at voltage setpoint. But there is a delay, some seconds to frequency shift.
Reduced charge current could help with that, except in the "load dump" case when a large load suddenly turns off.
Is ramp-down supposed to be 100% at 51 Hz to 0% at 52 Hz? Could reduce that closer to 50 Hz & 51 Hz to save time on ramping down.

Are you sure SB ramps down voltage in response to frequency? That takes a setting, either something like our newer "Rule 21" frequency-watts, or "Island", or "backup" for grid connected systems. Do you have a DMM with "Hz" range? I can see the frequency shift. When it works, my frequency hovers between 61 and 62 Hz. When it doesn't, runs to 64 Hz and disconnects.

What model SB? Some don't do frequency-watts as intended.