SMA Sunny Island SI6048 SOC

Trukinbear

I've been fussing with this SMA Sunny Island SI6048 for a year now, and I'm still stumped as to how to 'fix' it's SOC calculations.

My System: SMA Sunny Island SI6049, 2-Midnite Classic 250 charge controllers controlled by the SI via MODBUS/Canbus Comm adaptor, 18-330W JA Solar panels, 24-125-11 986Ah 48V industrial battery. Off the grid.

The Sunny Island's SOC is all over the place - take yesterday for example. Yesterday my solar panels brought my SOC (according to the Classic Charge controllers) to 100% - the SI read a SOC of 91% (pretty close for a change). Last night I used 122Ah, and the charge controllers were calling this 89% SOC, but the SI6048 said it was 20% and cranked the generator. Some days it will let me pull 400Ah or so and still give me a SOC of 50%, and other times I use near nothing and it says my SOC is 20% or less, and wants to boil the piss out of the battery. The SI is reading the energy coming in from the charge controllers, and the shunt has been calibrated (plus I have a WhizBang Jr). What gives? I'm about ready to sell this system and get something that operates as I expect. 

Can anybody help? SMA service is no use, they are German they just say everything works like it is supposed to - we're German.

Trukinbear

Now after putting 135Ah to the battery, the Classic charge controller says its 100% SOC, and Sunny Island says its 87% SOC. That seems to be a really big jump in SOC for so few Ah, it is a 986Ah-C20/628Ah-C6 battery. I can accept that maybe the battery is actually 87% vs 100%, but the 20% it said it was this morning vs the 89% as indicated by my WhizBang Jr equipped MidNite Classic 250 charge controller has got to be fixed. Ideas? Why would my Sunny Island think that 122Ah was 71% of the capacity of my battery? I can assure you it WASN'T at only 20% SOC.

Trukinbear

After pulling out 90Ah now my Classic charge controller reads 84% and the SI's reading dropped from 87% to 76% - now the readings are close together again. 

BB.

Do you have the external battery shunt option installed?

https://files.sma.de/downloads/SI4548-6048-US-BE-en-21W.pdf (page 54, section 6.4.4)

And has it been configured per page 73?

Another thing to try... Reset to factory defaults and reprogram all of the settings (electronic devices sometimes get the configuration memory corrupted, and do flaky things until the memory is cleared and reconfigured with the correct values).

-Bill

Trukinbear

BB. said:

Do you have the external battery shunt option installed?

https://files.sma.de/downloads/SI4548-6048-US-BE-en-21W.pdf (page 54, section 6.4.4)

And has it been configured per page 73?

Another thing to try... Reset to factory defaults and reprogram all of the settings (electronic devices sometimes get the configuration memory corrupted, and do flaky things until the memory is cleared and reconfigured with the correct values).

-Bill

Yes to all of the above, I have a shunt and it has been calibrated - last month at the one-year anniversary I reset the system and loaded all my parameters back in. My battery parameters are dictated by the owner of the battery company who personally extended my battery warranty to 10 years if I used his numbers. The Sunny Island sees the MidNite Classics as though they are SMA charge controllers (not available in the USA) and is accurately tracking and controlling what they do. Shunt values are mirrored by my WhizBang Jr so I assume them to be correct. 

Trukinbear

Sometimes they are really close - today for the first time I've ever seen them agree on SOC (other than at 100%) they agree at 89%. 

BB.

One other question, the two small sense wires from the shunt to the SMA... Are they twisted (say one turn every 2-3 inches)? Are those wires separately routed from AC and DC power cables?

Just making sure that you are not getting crosstalk between those wires that measure 1-50 mVolts and the other cabling that has very large switching currents.

Try to avoid parallel wire runs with the sense wires, and they cross high current cables at 90 degrees.

Also double check that the sense lead connections are clean and tight.

And to be paranoid, you have copper cable for the sense lead (different metals can generate voltage at junction points).

Bill

mcgivor

What may be occurring is the difference between how the information from the shunt is processed by each device, the sampling rate would likely differ. The current flow across the shunt is bidirectional, which in ideal conditions, such as a standby charging /dischargeing, would likely be relitivly accurate. When used in an offgrid application the current is dynamic, constantly changing direction, depending on loads and PV input, a momentary change from a passing cloud or heavy load for example may be captured at different moments in time causing drift in the respective devices.

This is something I've observed on my BMS particularly in the early morning or late afternoon, the display is rapidly changing between charge, storage and discharge which would indicate the PV input and loads are pretty close to one another. Whilst this is occurring a clamp on ammeter will either display a positive or negative reading but at a much slower rate than the BMS is indicating, due to its lower sampling rate. The Watt hour meter on the BMS drifts over time indicating more energy put into the battery than taken out, this is undestandable because the LFP bank is ~98% efficient, but is relitivly close to the charge controller reading on a daily basis.

There are other factors which may come into play such as battery inefficiencies, wiring losses etcetera, FLA are ~80% efficient meaning 20% more energy is needed to replace the deficit, I'm not sure how that compensated for by the equipment used but food for thought none the less. Perhaps using a separate watt hour meter with its own shunt would provide verification......then again it may add to the confusion, personally I wouldn't worry about it too much and certainly wouldn't rip everything out to replace it with something else that will likely have it own set of idiosyncrasies.

One possibility? Has the battery Ah capacity been programed into the SMA, I'm not familiar with the device but would assume it needs the information to calculate the SOC.
 

Dave Angelini

A no brainer here would be to ask boB at Midnite what he thinks of this combination of 2 completely different brands. This is often a path to this kind of poor behavior. Nothing worse than non repeatable instrumentation. Right up there with C19.  Maybe you have done that?

Trukinbear

Dave Angelini said:

A no brainer here would be to ask boB at Midnite what he thinks of this combination of 2 completely different brands. This is often a path to this kind of poor behavior. Nothing worse than non repeatable instrumentation. Right up there with C19.  Maybe you have done that?

I think boB designed this system for the Classic charge controllers to be controlled by the SMA inverter via the MODBUS/Canbus SMA communication adaptor they sold me.

The Whizbang Jr is a super accurate current sensor that the Classics use, it is attached to the same MidNite shunt that the inverter is, but is supposed to be more accurate. These two readings are never off by more than a few hundredths of an amp.

and yes I have...

Trukinbear

Sadly recently when one of my Classics went belly up and left me at 50% power and I contacted them about it, they suggested (at the close of business on a Friday) that my unit which was purchased the year before was like 9 years old, and out of warranty - that I should send it in with a blank check - I melted down over the weekend and shot my mouth off. They did cover it under warranty, and I did apologize but they hardly respond to me anymore. This is a shame as I've got a year's worth of my income invested in their products. I LOVE their products, and others swear they are super helpful... that just isn't my experience. I've got an open service ticket right now that nobody has so much as replied to... I even want to buy more but they will not answer specific questions about their products, at least not when I ask them. :-(

Trukinbear

I found this in the manual - I've seen the SOC drop from 50%-60% to 20% in the blink of an eye and start the generator.
This might be what is happening to me - now I just have to determine why.

Recalibration of the State of Charge: The Sunny Island recalibrates the display value for the state of charge of the battery to 20% as soon as the battery voltage per cell has reached a limiting value that is dependent on the battery type and the battery capacity. In addition, recalibration of the state of charge always takes place if one of the following conditions is met:
• The nominal capacity is entered incorrectly.
• The battery cable resistance is too high.
• The battery is not connected correctly.
• Differing concentrations of acid in FLA batteries (if necessary, reset the parameters in the menu "222# Battery Charge Mode).
• The battery has aged so much that the SOH is below 80%.

Trukinbear

Ok, after bringing the battery to a full charge, and letting it stabilize with no load for an hour, making sure the wires to the shunt were installed correctly we did 'new battery' Changed the battery rating from the C-6 rating to SMA's suggested calc to C-20 which gives us a number about midway between the two at 775Ah, calibrated the shunt, reentered my charge rates, charger output and all the other little things that change to default when you do 'new battery.' Powered her back up, verified the shunt readings via the readings from the Whizbang Jr which were spot on. However, the Sunny Island having been told it has a NEW battery and is hooked to a battery sitting at 49.9V (under load) thinks it's SOC is 46%. :-(
Maybe it will recalculate after a cycle.

Trukinbear

That should have been 49.9V as corrected above - fat fingers and poor eyesight :-)

Trukinbear

Fun fact for anyone looking for ways to rapidly shut down this side of their system - if you break the circuit to the shunt the inverter, and the generator if it is running, both abruptly shut off. 

Trukinbear

Now it is starting to look right!!

Trukinbear

SMA support says the Sunny Island should not be hooked to the DC shunt when the MODBUS/Canbus communications adaptor is used. While not in the documentation from MidNite that came with this adaptor, nor mentioned anywhere in SMA's difficult to read manual to get this to work right during initial setup I should have:
a) #225.01 BatCurSnsTyp - set to NONE
b) #250.28 ChrgCtlOp - set to SMA

BB.

I know pretty much zip about what you are talking about...

The only reason (hardware wise) I can think that they want the DC shunt removed is because it can put a 50 or 100 mVolt offset in the System's DC ground... Depending on what type of signalling they are using (R422 or other)--That could be an issue (but normally it should not be--The signalling levels should not be affected by that low of offset voltage (and you can still have similar offset voltages just from the DC wiring and heavy current flow (voltage drop, xx AWG, yyy current, etc.).

I am sorry I cannot help you more about that.

We have had a few members that have done MODBus and CANBus projects--Hopefully one of them can help.

-Bill

Trukinbear

Hopefully, this is the solution to the MidNite Solar/SMA networking problem. Would have been nice if they'd told me this last year, in the interim I've pointed the finger at everyone only to circle back. Not in the manual, not in the documentation but necessary settings for it to work - it's almost like they were trying to sabotage each other. Sadly I'm the one that had the year of 'fun.' HOWEVER, with these changes having been made I'm not certain they did anything... they were just supposed to be that way and weren't. I'll follow up with the results. 

Edit: they didn't want the shunt removed, it is still necessary for the whizbang jr - they just don't want the sensor wires hooked to it, and the unit programmed to have no shunt installed. 

daklein

Trukinbear said:

 they just don't want the sensor wires hooked to it, and the unit programmed to have no shunt installed. 

That sounds a little nuts, how would the SunnyIsland know the battery current?  Does the SunnyIsland receive battery current back from the modbus/canbus adapter from the Classic?   I thought it was a one way device, with the SMA directing the Classic what charging mode and voltage to choose.  If it does return some data to the SMAs, I would __Love__ to know what are the CAN messages to send to set any parameters in the SunnyIslands over CAN

I have a pair of the SunnyIslands, basically a DC solar trailer installed in my garage.  I don't have anything yet to log and plot the data from the SMAs yet, but I have a CAN adapter that is able to dump out the messages and now I need to get it logging.

Today I think I noticed a drop in the SMAs soc estimate, after a cloudy day and starting from grid connected overnight and low 90s SOC and staying mostly in silent mode.  I the late afternoon, I thought I saw it around low 80s, then within an hour or so it had connected to the grid to charge because it hit 70. Ooooh Noooo, charging on peak rates.  The battery is getting colder now, only about 10dC or under 50dF, which shouldn't be too low yet, but maybe that's related, decreasing the actual capacity.

Trukinbear

My SI6048 gets current information via the MODBUS/Canbus adaptor from the WhizBang Jr equipped Classic 250. Since SMA replaced my SI, and I've gotten it dialed in this way it is operating flawlessly with the SOC within a few points of that calculated by my Classic 250, and verified with my HydroVolt. 

cwr

I also have a DC Solar trailer. My issue is that even when I am producing more than I am using my soc continues to drop and my soh indicates 139%

I have replaced the ten 250 watt panels that came with and put on eight 390 watt Canadian solar bi-facial panels which puts me up to around 3000 watts input at peak and the most i ever use is 1.2.

I expect discharge but it seems I get that continually even when I am producing more than I am using. It is quite frustrating.