SMA sunny island settings for signature solar EG4ll batteries

[carlsonultimate]

Hi,
I recently replaced my old lead acid batteries with new signature solar eg4ll batteries.

System specs
Two Sunny boys AC coupled with Two Sunny islands
10,000W PV running two the sunny boys
1 outback flexmax80 with 3000W PV
12 (48V) signature Solar EG4ll LFP Batteries (60KW)(1200AHr)
Propane backup geni

I went through the new battery set up on the SI and set to VLRA as suggested by tech support sense the BMS on the EG4's is not compatible with the the SI's.
set as
48 V
VLRA
1200AHr

System started up fine but the SOC on the SI has been significantly lower than what the EG4 read. Example; the SI read 1% and the EG4's read 78% at 53.3V. So this is my first problem I am not able to use the whole capacity of the batteries. The second problem is during charging the EG4's go into Over voltage protection around 54V. The specs for the BMS say that OV protection occurs at 60V so maybe I just haven't been there to observe it. I've lowered all of the charging values to 2.3V. Once the EG4's go into OV the SI will start to give a F201 Malfunction VBATMAX, this happens 3 times with the system turning off for about 30 seconds and coming back on until I get a F710 Failure autostrcnt which puts it in standby.

Can anyone recommend different settings? Any help would be greatly appreciated, I've talked to SMA tech support but there not much help, they just tell me that the EG4's are not a supported battery...

 

[SignatureSolarJames]

do you have a shunt from SMA?

 

[carlsonultimate]

do you have a shunt from SMA?

I do, it was working properly with the FLA before. The SOC corrected itself in the right direction today, not sure how but it is only a 20% difference now. Still confused why the BMS on the EG4's are going into OV protection, I have not observed any voltage over 57.5V.

 

[Hedges]

With AC coupled Sunny Boys, if an AC load goes away then Sunny Island has to shove the power into batteries for a few seconds.
LiFePO4 has a curve, shoots up in voltage near full.
54V / 16 cells = 3.375V/cell, seems reasonably low (compared to 60V = 3.75V/cell) but try setting lower max "VRLA" voltage in SI.

SI would reset SoC when full. Maybe low SoC issue will go away, and maybe you'll need to play with voltage interpretation for that too.

 

[carlsonultimate]

With AC coupled Sunny Boys, if an AC load goes away then Sunny Island has to shove the power into batteries for a few seconds.
LiFePO4 has a curve, shoots up in voltage near full.
54V / 16 cells = 3.375V/cell, seems reasonably low (compared to 60V = 3.75V/cell) but try setting lower max "VRLA" voltage in SI.

SI would reset SoC when full. Maybe low SoC issue will go away, and maybe you'll need to play with voltage interpretation for that too.

Thanks for the reply, all my charge values are pretty much at the minimum allowed. Here are what I have my values set at
BatrChrgCurMax 1000A
AptTmBoost 60min
AptTmFul 60Min
CycTmFul 30Days
CycTmEqu 180days
ChrgVtgBoost 2.26V
ChrgVtgFul 2.3V
ChrgVtgEqu 2.3V
ChrgVtgFlo 2.2V
BatChrgVtgMan 54.5V
Where do you set the Max Bat voltage in SI? I have not seen a setting for this.

 

[Hedges]

Thanks for the reply, all my charge values are pretty much at the minimum allowed. Here are what I have my values set at
BatrChrgCurMax 1000A

"12 (48V) signature Solar EG4ll LFP Batteries (60KW)(1200AHr)"

What is maximum charge current/rate for the battery? I would expect 0.5C or 600A. Which is still 30kW, 3x the 10kW of PV you have.
So shouldn't be a problem, but you could set BatrChrgCurMax accordingly. Not likely to make a difference; 10kW of PV is just 0.16C or 200A.

I think 0.16C is pretty ideal, low but still plenty to recharge battery during a single day. A cold battery, approaching freezing, shouldn't be given 0.5C but can take 0.16C relatively cold.

AptTmBoost 60min
AptTmFul 60Min
CycTmFul 30Days
CycTmEqu 180days
ChrgVtgBoost 2.26V
ChrgVtgFul 2.3V
ChrgVtgEqu 2.3V
ChrgVtgFlo 2.2V
BatChrgVtgMan 54.5V
Where do you set the Max Bat voltage in SI? I have not seen a setting for this.

Those various (lead-acid) cell voltages are the battery voltage. Because VRLA no FLA, ChrgVtgEqu is about same as the others.

From SI manual, "ChrgVtgFul Cell voltage setpoint for full charge 2.3 V to 2.7 V VRLA 2.40 V", you've already got the minimum, can't lower further.
2.3V/cell x 6 cells = 13.8V for a 12V battery, 55.2V for 48V battery.

Another option is to raise the over-voltage setpoint of BMS, but we would rather not do that. 3.75V is reasonable. However, it could be an option.
Maybe during some conditions voltage briefly goes higher, not long enough to cause any problems.

AC split-phase is 60 Hz, with currents going to a peak and returning to zero. I've found that at high wattage, the ripple current comes almost entirely from my lead-acid battery, little from the capacitors. Lithium being lower internal resistance, even less would come from capacitors (at flat part of voltage). Well up the charging curve, however, voltage rises more quickly. Possibly supercapacitors at SI would help, together with enough cable resistance to decouple from battery.

We don't know if your problem occurs when SB is producing power and an AC load suddenly shuts off, or if it is just the AC ripple doing this.

Oh, does BMS report exactly what the fault was? One cell over-voltage while the rest were OK? Could be not fully top-balanced. Could be poor contact of one cell busbar (that would show up as cell voltage rising above others when charging, dropping below when discharging, even at a middle range SoC where voltage curve is flat.)

The ripple current may produce peaks exceeding cutoff voltage even though average is lower. Any chance you've got a scope, could see waveform? Can you make a diode-capacitor circuit, with high value bleed resistor, to sample and hold peak ripple voltage (less one diode drop)?
Use AC range of DMM to read ripple voltage on battery. Assuming sine wave, multiply by square root of 2 to get peak AC voltage. Add to DC voltage to calculate peak voltage on battery.
Maybe if sense wires of BMS went through an RC filter they would ignore AC component, just see DC. If this was DIY with discrete BMS you could add that, but it is a packaged rack mount battery. Digital sampling has its problems. usually filtered at around sampling frequency, but we would like low-pass around 10 Hz.

What model Sunny Boy, with what settings?
If this is a case of slow frequency ramp, SB continues delivering power for a few seconds after a load is removed, a couple settings affect that.
Sunny Island by default swings frequency below 59 Hz to make up for mechanical clocks running fast at 62 Hz, so curtailment takes longer. You could change that to 59.5 Hz minimum.
UL-1741-SA frequency-watts of SB ramps down from 100% at 60.5 Hz to 0% at 61 Hz, while off-grid ramps down 61 Hz to 62 Hz (and is adjustable). Ramping down from a lower frequency would let it finish sooner.
Off-grid allows larger voltage excursions (adjustable), while UL-1741-SA allows less, UL-1741 even less. During a load-dump, this could have it drop off sooner.

Try turning off one Sunny Boy so only half as much power has to be curtailed, and peaks are half as high.

So I'm guessing it could be the 60 Hz ripple voltage on battery I'm familiar with.
You also have DC coupled Midnight. What are its settings?
Maybe if you set Midnight to charge up to 58V, Sunny Island will see that higher DC voltage and curtail Sunny Boys sooner, so their 60 Hz stops getting imposed sooner.

Turning off one Sunny Boy, then both, would be a first step to see if curtailment time and/or ripple are the problem.
Using Midnight to hit same ripple-free voltage would be next, see if issue goes away (inverter drawing ripple current might bounce it higher.)
Then setting Midnight voltage higher, causing SB curtailment at lower voltage.
Alternate settings of SB, and of SI frequencies.

 

[Hedges]

Got an idea!

Although 2.3V per cell is the minimum setting, you can lie to SI and say you have a 46V nominal VRLA battery, not 48V.
SI supports a few extra cells, and removing a few. Nominal 42V to 52V, I think.

This would have SI stop charging further down the curve, maybe almost the flat area. Needs to be high enough that slight variations don't cause it to stop charging at low SoC, no readjust cell voltages if needed to ensure just starting up the knee.

Midnight can then pull voltage a bit higher to allow balancing (only works if not too much AC load, otherwise SI will draw from battery to keep voltage down.)

 

[carlsonultimate]

Thanks for all the ideas!
My BatrChrgCurMax is set at 1000A, each EG4ll is rated at 100A, I have 12 so that would be 1200A max I set it slightly lower. I doubt that limit has been approached, haven't run any big loads and charging doesn't put out that much.
My thought is that it is the AC ripple that you are talking about, as the error happens both from solar and from generator charging, but only when the batteries reach full charge. If the SOC matched up I wonder if there would if there even be any issue. As of now the batteries say there 23% higher then the SI.
The BMS dose not report an exact error just states (OV) protection, which according to the values is 60V (See attached BMS parameters). To check packs BMS values with Signature Solar's software you must plug into each back they haven't figured out how to have them all populate. I have not checked 3 packs so far and all the cells within the those packs seem balanced.
Attached is the SB nameplate.
So this 60Hz ripple voltage would create a voltage over 60V?
I have an Outback flexmax 80 solar charger that takes about 3000 W of solar. I have it set at 57.5V for bulk and 56.5 Float.
I'll try turning off one SB then two and see where that gets me. If that doesn't work I'll try lowering the nominal V.


Thanks!

 

[Hedges]

Sunny Island reports average battery voltage, as will a DMM. It also reports current, or maybe power.
I used a current transformer to read ripple current on battery, and it was quite significant when drawing 40% of inverter rating.
If battery is near full the charge current should be reduced, but I think Sunny Island will impose a 60 Hz ripple current, which will cause a 60 Hz ripple voltage. Just use AC Volts range of your DMM. That will report the RMS voltage superimposed on DC, going above and below.
So I figure having Sunny Island back off earlier and let Midnight take over should help.

 

[BigPines]

@carlsonultimate did you ever resolve this issue? I also have two SMA Sunny Islands and have the EG4-LL batteries on pre-order. I have since learned SOK batteries support closed loop CANBus communication with the Sunny Islands.

SK48V100 Server Rack Battery Owner & Operator's Manual | Current Connected

Full instruction manual for the SOK SK48V100 server rack battery. Everything you need to know is contained within this document!

www.currentconnected.com

I like the Signature Solar rack which I also ordered and the EG4-LLs are cheaper so I would rather stick with my Signature Solar order but if I won't be able to get the EG4-LL to communicate with the Sunny Islands closed loop, I will cancel my Signature Solar order and go with the SOKs.

Any information is appreciated.

 

[Hedges]

I'd say for sure get communication with Sunny Island.
Otherwise, you have to enter dummy VRLA parameters for voltage and zero temperature coefficient. Still won't get the feature of reducing charge to the rate balancer can keep up with. Or SoC knowledge for load control.

I am presuming Sunny Island will obtain SoC values from BMS, so it can manage load-shed, generator start, etc., but haven't used anything other than AGM.

 

[BigPines]

Thanks for the quick response!

Well crap, I called Signature Solar and they said my batteries shipped today - a full 10 days early! Coincidence?

So now it looks like I have two EG4-LLs to get rid of.

 

[carlsonultimate]

@carlsonultimate did you ever resolve this issue? I also have two SMA Sunny Islands and have the EG4-LL batteries on pre-order. I have since learned SOK batteries support closed loop CANBus communication with the Sunny Islands.

SK48V100 Server Rack Battery Owner & Operator's Manual | Current Connected

Full instruction manual for the SOK SK48V100 server rack battery. Everything you need to know is contained within this document!

www.currentconnected.com

I like the Signature Solar rack which I also ordered and the EG4-LLs are cheaper so I would rather stick with my Signature Solar order but if I won't be able to get the EG4-LL to communicate with the Sunny Islands closed loop, I will cancel my Signature Solar order and go with the SOKs.

Any information is appreciated.

I have got my system working now. I had to put the SI battery settings as a 46V system instead of 48V. I haven't had a trip on the BMS sense and the SOC finally is close to the same on BMS and SI. The BMS SOC drops faster than the SI SOC which is better than the otter way around. As for communication with SI and BMS I have found no way to get that to work, both companies say they are not compatible yet but may offer firmware updates in the future to offer compatibility.

 

[BigPines]

Glad to hear you got your system working. I ended up canceling my Signature Solar order and buying some SOK batteries because I really wanted the closed loop communication. From what I understand, the BMS in both batteries is similar and may in fact be the same so a future firmware update could very well bring that to the EG4 in the future.

 

[beastman]

I have the same situation; I would like to go w/ the lifepower eg4 LL ($1499/$1599) and want to get at least 30kw, possibly upwards of 40kw. These still qualify for 30% tax breaks in 2022 and my accountant says go for it since I live in California and we pay an ungodly amount of taxes here as a business owner. I may go w/ SOK but I can get more like 40kw of lifepower for the price of 30 kw of SOK. I do own an EV (Tesla MSP85) and want to charge off-peak w/ the bank 2-3 times per week.

I have an existing 8000 watt solar system w/ sunny boy SB7000US inverter grid-tied. Its been working perfectly for 7 years now - I installed it and permitted it myself - had $9k total invested into it. Even w/ all this power being generated, we still have a true-up of $3k per year on PG&E. I would like to reduce that down to less than $500.

We produce between 30kwh per day in the winter and 55-60kwh per day in the summer in Northern California. I bought a bundle of sunny islands (5048's) - two of them from 2011 for a little over $2k. They were functional from a working setup. I also got another SB6000US w/ the rs485 card the previous owner used in off-grid mode to work with the setup. The previous owners installers had problems w/ the settings on the sunny islands (they upgraded the software) but had issues w/ the generator start feature when using lithium on the homegrid stack'd batteries - this was probably installer error. The previous company got fired by the customer and the new company put in a sol-ark for the customer and that was the end of his problems.

So basically I want 30-40kw of storage using lithium (the agm's and lead acid batteries are crap for daily cycling) and I'm going to spend $10-15k to accomplish this and don't want to mess with them for a decade, similar to how my existing grid-tied solar has functioned so far.

Does anyone have the EG4 LL / Lifepower settings to work w/ sunny islands successfully? My goal is to also get another 2 sunny islands as slaves so I can run my 5hp well pump when power is out; the existing two units may do the job but it will be marginal.

I'm an EE/ Programmer and design automotive microcontroller add-ons for a living, so I am ok to do some mods w/ a raspberry pi or arduino w/ canbus. Not sure how installers these days are successfully doing these - on the forums there is a barrage of posts but not a lot of clear solutions. I have an arduino board now we use in automotive w/ canbus modules which might be able to talk to everything over can or rs485.

If lifepower cannot do closed loop but open loop can be done properly; i'll go that route. Otherwise i'll give up 5-10kw of storage to get SOK on the sunny islands and call it a day. I also have a 16kw generac that will be hooked up to the SI's in this setup and have a 100a transfer switch that can be used w/ the islands. The 16kw generac barely starts the 5hp well pump and I don't really want to use it any longer in that situation. I was told by a well service company I would really need more like 20kw for 3 seconds to start the 5hp well pump. It works but may cause premature failure of the generator, inverters and the 5hp motor without extra headroom.

Thanks

Brock

 

[BigPines]

I ended up going with the SOKs to get proper closed loop but I have heard several people have the EG4s working open loop.

 

[beastman]

I ended up going with the SOKs to get proper closed loop but I have heard several people have the EG4s working open loop.

SOK informed me today that SMA is not supported in closed loop according to this salesperson below. This is for a sk48v100; qty 8.

>>I also have a pair of sunny island 5048 and want to make sure there is documentation to support connectivity with these.

• Currently, communication with SMA is not supported. You would need to run your system in an open-loop or upgrade to an inverter such as the Sol-Ark where communications are supported. -- Jarrett -- info@currentconnected.com

 

[BigPines]

OK, that is really strange. When I purchased mine, they asked what inverter I was using and configured them for SMA before they shipped them. They even said SMA on the box! Now it looks like that option is no longer available on their web site. Not sure what is going on but it looks like they may have backed off supporting closed loop with SMA? There was a video showing the configuration and it looks like it may also been taken down. This sure doesn't give me the warm fuzzies. Feels shady. I wonder if they found a problem with this configuration? I would sure like to know what is going on since they sold me a product promised to work with the SMA. I guess I'll have to call them.

 

[BigPines]

Yep, looks like they took the video down. It was here:

 

[BigPines]

This page from the manual says SMA Communication is coming soon: https://www.currentconnected.com/sk48v100/manual-4-2/

SOK has some explaining to do.

I just sent them an email. I want to know what is going on. I'll report back what I find.