Sunnyboy DC Overvoltage

[cliveb]

My SB4000TL has been running without incident since it was commissioned on 31st August last year. Then today, I noticed the red error LED was on. The display showed "DC OVERVOLTAGE", then "DISCONNECT GENERATOR" and "RESTART IN 25 S". I didn't know whether "DISCONNECT GENERATOR" is supposed to be an instruction to do something or a message of what the inverter is doing by itself. So I didn't throw any of the isolation switches and just left it alone. After a while it recovered and seemed fine - the red LED went out and the display returned to normal. Unfortunately I failed to take note from the disply of the reported DC voltage for the two strings. (I have 16 Suntech 245W panels, two strings of 8 each)

I fired up Sunny Explorer and checked the events, to discover that there were several DC Overvoltages yesterday morning and there have been 2 so far today. I went back a few more days and there were no events prior to yesterday.

What might this error condition indicate? I can't see how the panels could possibly be producing too much voltage - it is sunny, but it's January! Might the inverter have developed some sort of software malfunction that is throwing up these errors as phantoms?

I'm asking here before I contact the installer because I trust the people here to know what this might mean. Once I know what it means, then I'll have the necessary ammunition when I call the installer.

[eabadger]

hello, read my post on tristar mppt, in very cold weather your o/c voltage goes up, morningstar are saying my unit may have expired due to this.
what is o/c voltage of pv and was it cold to day when bright sun hit them?

steve

[GavinA]

are you sure it's run in 2 strings of 8?

The cold weather will drive up the string voltage significantly, but I'd be surprised if 8 of those suntech panels would get anywhere close to being over voltage. Sunny design does recommend strings of 10 and 6 for a 16 panel system, and I could more believe that a 10 panel string might temporarily go overvoltage due to the cold panel temperatures than I could an 8 panel string.

[JohnS]

I am surprised that 8 panels is giving an overvoltage situation.

The SMA 4000TL has a max DC voltage of 550V.  The Suntech 245W panels have a Voc of 37.1 at STC of 25oc and a temerature coefficient of -0.34%.

Assuming a temperature of -5oc, this would raise the voltage by 0.34*30/100 = 10.2%.  Ie from 37.1 to 41.1V.

And 8 * 41.1 = 329V which is well inside the 550V limit.

What sort of DC volts do you normally get?

[cliveb]

Thanks for the responses. First of all, I have to say I do not know whether the 16 panels are indeed configured in 2 strings of 8. I had just assumed that. Checking the handover pack I got from the installer, it doesn't state how they are configured.

When I look at the SB4000TL's display, it alternates between two outputs (and highlights two different sides of the icon of the panels). I assumed these two displays were the two strings, and they always report very similar DC voltages, so that made me assume the two strings had the same number of panels. For example, right now it's quite sunny with a bit of hazy cloud. The inverter is reporting a total power of 950W, and the alternating displays both show 254V and 1.9A.

Question 1: does the Sunnyboy's two alternating displays indicate the two strings? Or is it showing something else?

Secondly, regarding the higher voltage at low temperatures. Well yes, it's pretty damn cold today (as it was yesterday). But I reckon it was equally cold on Saturday and Sunday, and there were no DC overvoltages on those days. And the sun is low in the sky in January - surely the peak voltages during September would have been much higher? I cannot believe that either string could deliver over 550V in January. What on earth would they be delivering in June?

Question 2: if these overvoltages are caused by the cold weather, can I safely ignore them? Is the inverter in danger of being damaged in any way?

[Ben55]

Have a look at your DC cables coming into the Sunnyboy.

I have just two cables,  therefore one string.

If you have two strings, there will be four cables from the panels to the sunnyboy.

Hope this helps

[cliveb]

Have a look at your DC cables coming into the Sunnyboy.

I have just two cables,  therefore one string.

If you have two strings, there will be four cables from the panels to the sunnyboy.

Hope this helps

The DC cables are hidden inside trunking, so I can't tell how many there are. What I can say is that the cables enter a DC isolation box, then there's another DC isolation box right next to that and the cable(s) then go to the inverter (inside yet more trunking). Presumably there is a DC connection between the two isolation boxes. Does the fact there are two DC isolators have any significance?

Let's suppose for a moment that I have 16 panels in one string. What's going to happen in June, when they are all maxed out? 16 panels at 37V is nearly 600V - above the 550V limit mentioned by JohnS. That said, I have seen the inverter display an indicated power over 4kW during September last year, and there was no overvoltage error then. But I don't remember what the indicated DC voltage was.

One final question before I call the installer: is it remotely reasonable to install 16 x 245W panels in one string on an SB4000TL?

[Richard Owen]

What I can say is that the cables enter a DC isolation box, then there's another DC isolation box right next to that and the cable(s) then go to the inverter (inside yet more trunking).

Two DC isolators suggests two strings.

[cliveb]

Two DC isolators suggests two strings.

Does this still hold when all of the DC cables from the roof go into one isolator, and all of the DC cables to the inverter emerge from the other isolator? (Clearly there is some non-visible connection between the two isolators).

I've called the installer and they immediately said they would come out on Thursday to take a look. Sod's Law means there will be no more overvoltages between now and then! At least I have the Sunny Explorer exported events data to show them.

[ovonrein]

At least I have the Sunny Explorer exported events data to show them.

If you have SE, then it's easy enough to see the config.  Look under "DC Side" and it will probably show you two string voltages.  You also report ~254V above - that's consistent with a string of ~8x30V - the most likely config for 16 panels.  And I agree with every other post on here so far - pretty darned unlikely that cold temps should send such a (short) string into over-voltage.

[Richard Owen]

Does this still hold when all of the DC cables from the roof go into one isolator, and all of the DC cables to the inverter emerge from the other isolator? (Clearly there is some non-visible connection between the two isolators).

I've called the installer and they immediately said they would come out on Thursday to take a look. Sod's Law means there will be no more overvoltages between now and then! At least I have the Sunny Explorer exported events data to show them.

Don't see why not. We've routed all the cables through a pair of isolators to make everything look neat. The first pair of cables terminate in the first isolator and the second pair just pass through and vice versa in the second isolator.

Anyway, sounds like you'll find out soon enough.

[Laurence]

Clive, check out the performance curve graph in the spec sheet for the solar panels you have.

You will find out that PV panels generate a remarkably constant voltage, but that the amount of current you can potentially draw from them goes up, with increased solar input. The MPPT tracker(s) in your inverter will "surf" up and down along the V/I curve, looking for peak power. That will generally be at around the same voltage (plus or minus a bit), whatever the light level.  Think big 12V battery and little 12V battery - both produce 12V but all other things being equal the larger battery can deliver more current, hence more power.

Yes, the O/C (Open circuit) voltage that your array can produce will go up a little in cold weather, but as previous posters point out, it's only a few percent. The system should have been designed to cope with likely voltages on cold, sunny days.

You need to study either the panel on your inverter, or fire up Sunny Explorer on your PC. If you see TWO pairs of numbers (i.e. 300V, 2A and 310V 1.9A), then you know you've got two strings, each pushing out about 600W, give or take.  If it says one pair of numbers (i.e 600V at 2A), then THERE is your problem!  Two DC isolators suggests that your panels could have been set up as two arrays, but they might have been wired on the roof in error as one array. Or even 12+4 panels.  the Volts and Amps figures would tell us that.

Your SB4000 is a clever box. It will be logging and storing any error messages, regardless of the weather, when your installer shows up!  I like the fact that your installer can get to you so quickly - presumably he's not as busy as he was, late last year...

[cliveb]

You need to study either the panel on your inverter, or fire up Sunny Explorer on your PC. If you see TWO pairs of numbers (i.e. 300V, 2A and 310V 1.9A), then you know you've got two strings, each pushing out about 600W, give or take.  If it says one pair of numbers (i.e 600V at 2A), then THERE is your problem!

OK, I fired up Sunny Explorer and under Inverter > Spot Values > DC Side it shows two sets of values (2 each for Current, Voltage and Power). The figures in the two sets are remarkably similar, so I guess my 16 panels must be in two strings of 8 each.

We'll see what the installer has to say when he comes tomorrow.

[cliveb]

We'll see what the installer has to say when he comes tomorrow.

A quick update. A chap from the installer (Hatch Sustain) turned up this morning and after I explained what had happened, he said that it does indeed sound like a software glitch. He checked the wiring to the inverter and said everything looked fine (and in passing mentioned there are indeed two strings).

He then took a complete export of all data from the inverter and said he was going to call SMA to find out exactly what the error codes mean. We left it that I would keep an eye on things and if these errors continue it may require an inverter replacement. He seemed unfazed at the thought of having to replace the inverter - presumably they would pass on the cost to SMA through warranty.

All in all I've been pretty pleased with the service from Hatch Sustain.

[GavinA]

when you say he checked the cables, did he actually use a proper tester, or just check them visually?

I ask because I've got a sneaking suspicion that this could have been caused by leakage current flowing between the 2 strings to create a path for the 2 strings to effectively run in series.

This could possibly either be due to cables getting nipped in the frame, or leakage from the panels to frame, or badly fitted connectors between the panels, probably made worse by the dew getting into whatever the issue is (leakage from the panels is definitely worse when wet, as is insulation resistance in the connectors), or just a loose connection / cable that had been nicked in the isolators. If it was this, then it's possible that it would be at such relatively low ampages that the inverter would only pick up on it at start up when the available power was very low anyway, then at higher outputs it'd just ignore that tiny spike on the power curve.

I can't entirely work out how this would work, but the 2 strings somehow temporarily ending up being in series is about the only way I can see that the voltage could possibly go above the inverter maximum voltage with this set up, so this is what I'd be checking for to rule it out, and this can only be done via proper insulation resistance tests, and continuity tests on the circuit, not a visual appraisal (assuming the installer hasn't somehow managed to reduce the peak allowable input voltage on the inverter).

FWIW, in cold dewy conditions, I have measured a leak on a 4000TL from an individual panel of 30V, but only 8 nano amps, so there definitely can be some very low level leakage even from a properly installed system. I've never known of the above scenario actually occurring, but suspect it could be a possibility if a combination of factors were involved.

I'd be interested to know how this pans out anyway.