6 minutes ago, The Terrible Triplett said:

Assuming the fans are clear, nothing really changed, Eskom did not cause any damage, another point to ponder is that I have come across electrical devices that just become erratic overnight or in higher temps. When you take it back to the supplier, you may be told that the electrical parts inside have reached their intended design life read, time to replace. Very very annoying.

The typical scenario is with laptops. After some time, they start running hot and become a complete pain to deal with. With desktops I've managed to get a bit more time out of them by replacing the heatsink compound. I've had an ADSL modem where I saw the caps are bulging, and even replaced the largest one. Bought me about another month before I finally gave up. It would be fine as long as the connection didn't fail, but would not reconnect after a power failure. By the final week, I had to put it in the freezer for a few minutes to get it to reconnect.

So inverters aren't ADSL modems, but in some ways they are not that different either :-)

4 minutes ago, plonkster said:

The typical scenario is with laptops.

Jip, and cell phones ... and a Victron 375va Phoenix inverter that had a disagreement about 12/24v DC connection.

Caps where replaced and it worked perfectly afterwards for a long while till the ambient temp increased. Then it started to behave erratically.

Upon deeper inspection by a Victron dealer, whom could not pinpoint the problem either, the conclusion was reached to replace the circuit board. Thing is, at that level, it is a better to buy a new one.

6 hours ago, The Terrible Triplett said:

YES!!! It was Eskom. Apparently it was on TV tonight, that there are "issues", apologies from Eskom.

Man I'm smiling, MG preformed above my expectations.

@plonkster, I think there are now two of us wanting to sort the MG tripping the MG DB, as a matter of importance. 

 

This may be your problem

2 hours ago, Chris Hobson said:

This may be your problem

Jip, after your suggestions the other day of inverters being affected by Eskom, that was my first thought. That it scared the dinges out of me ... I won't lie. :-)

When I got confirmation, what really impressed me was the sensitivity of the MG. If it was not for the quirk of it tripping the DB, it could have gone unnoticed.

Unless you sit next to a UPS and hear it clicking as Eskom dips or overshoots, one seldom knows it is happening, unless it is really bad one and lights dip, stuff reboots - that is seriously bad. The unseen ones are not as bad, but still not good in the long term for sensitive electrical parts.

Therein my like of online UPS'es. Only way to really sort it.

Ok,

Will put the inverter back into service today and see whether the issue then persists or not.  That being said... Whatever Eskom did, an error 51 is very, very bad MOJO.  As I said, you even need to disconnect batteries to get the Inverter to restart.

How / What can one do to prevent these surges entering the Inverter and thus causing the error condition?  I do already have surge protection devices, but they obviously don't detect this as none of them blew.  Knowing what causes the problem is half of the problem solved, the other half, is keeping my Inverter powered, as restarting is a huge PITA.

1 minute ago, Chris Knipe said:

How / What can one do to prevent these surges entering the Inverter and thus causing the error condition?  I do already have surge protection devices, but they obviously don't detect this as none of them blew.  Knowing what causes the problem is half of the problem solved, the other half, is keeping my Inverter powered, as restarting is a huge PITA.

I not saying it won't blow but the Axpert is fairly tough kit. Last summer I have had a surge not blow the SPDs  (still not sure why not) spider- arc up the wall - cause the Axpert to go into Error 9  (Bus soft start failed). I switched off and switched on again and have had no hassles since.

1 hour ago, Chris Knipe said:

How / What can one do to prevent these surges entering the Inverter and thus causing the error condition?

From what I have gleaned over time, it is a tough one. 

Worst time is when the power comes back after a Eskom failure. That is when you keep things off, let the grid stabilize first. Friends of mine lost PC's, fridges, sound systems, you name it, even UPS'es during the Eskom woes. Me, nothing.

So the question is: Is it a real problem where one is situated, it happens all the time or regularly, only then take action. 

THE most secure way I found is to keep Eskom away by using a online UPS, which is NOT a cheap solution. How they work:
- Batteries are charged via Eskom
- The batteries in turn power the inverter
This way NOTHING Eskom does, gets out the other side, as the battery charger can take any volts ito dips and overshoots. Ideal for like generator startups, for the batts are the small 7-9ah ones, so it is designed for only a few minutes at full power, till gennie has settled.

Other side of the coin, if Eskom was / is the cause of equipment failure, like when we had power failures, it may also be an insurance claim. Check that with your insurer / broker.

My woes last night with the Multigrid, 4 times it kicked over into UPS mode, after picking up disturbances. I was seriously impressed as I did not think it would notice. A cheap UPS on my son's PC also kicked in at the time.

UPS'es are designed for this, solar inverters generally not, therein me being impressed with the Multigrid.

Edited October 9, 20187 yr by Guest

Well for me on good days, the Inverter will trip with a Error 51, 2 to 3 times a day - so yes, it's a big one for me.  I most certainly don't want to loose power that amount of times per day, and then having to go through the manual and exhaustive process of disconnecting *all* power (incl. batteries) from the Inverter to start it up again, only to have it trip again after a few hours is very, very troublesome.  This is just as damaging to equipment IMHO.

So IF the supply is the cause then there's three things that can be done:

1) The Axpert's firmware needs to be fixed.  A surge on input power should result in a TEMPORARY error condition, not a permanent one (i.e. in an ideal world, the Inverter should recover automatically).  Frankly speaking, the Inverter should disconnect the mains supply and run off batteries / PV, not go ape sh!t (Error 51) and power off everything.  That is the real problem, and this is where it should be fixed IMHO.

2) A small on-line UPS in front of the Inverter to continue supplying power to the Inverter in the event that Eskom goes crazy (6KVA or so, with maybe 5 minutes run time?).  This should provide a stable supply to the Inverter and thus prevent the Error 51 from occurring (again assuming the surge is on the supply side and therefore the cause of the Error 51).

3) Just move off the grid, and get it over and done with already (in planning TBH).

Given that I want to drastically increase my system and move off grid in any case, I doubt the costs for a UPS will be justified at this stage.  IMHO this is a issue in firmware, and should be fixed in firmware.  I do know and realize however, good luck with that one.  It's not going to happen.

You also may have a faulty unit.

And I agree the inverter should handle it, within T&C's obviously.
And I do not think a online UPS before the Axpert is by any means a good solution ito costs as you need to match the UPS with the Axpert's size.

I hazard a guess and say you are based in Cpt?

Off-grid is an option in Cpt yes, in JHB most definitely yes.

Me, I did the numbers and came to the conclusion that a Multigrid hybrid is the better option. The new Multiplus II is even better, IF you need batteries as backup.

And there are Goodwe's too.

37 minutes ago, Chris Knipe said:

manual and exhaustive process of disconnecting *all* power (incl. batteries) from the Inverter to start it up again, only to have it trip again after a few hours is very, very troublesome. 

You should have a battery disconnect.

I do.  That's not the point.

On 2018/10/08 at 10:54 AM, Coulomb said:

There is a restart on overload option, which defaults to off.

Ah, my bad. I've just been reading some more firmware, as you do, and have found that it's "ordinary" overloads (fault code 7) that are restartable; fault code 51 ("over current or surge") is not restartable. (Several other fault codes are also restartable, but are not relevant here.) It seems to be triggered by an instantaneous current measurement of 40 A or 80 A (I think it starts with 40, then goes to 80 if all is well). 80 A is the peak of a 56.6 A RMS sine wave, which is some 260% of rated current (5000 VA / 230 V = 21.7 A). If it's still working on the 40 A limit, that's only 130% of rated current, and it can take that for 5 seconds (in fact, 5 seconds of up to but excluding 150% of rated power). The 260% figure is 250% if the output happens to be set to 220 V instead of the default 230 V. So that's a "never exceed" value, hence it's not restartable. Something is seriously wrong.

So I'm leaning towards a hardware fault, likely caused by overheating when the SCC works hard. You might try reducing the maximum total charge current temporarily to see if it makes things better. It might also be good to find out what the inverter temperature is reporting just before the error 51s. It may be possible to extract that information from ICC logs or similar.

It would also be nice to see what the current readings are; there may be a command I can dig up that will tell you what the current reading is. You'd presumably have to read dozens of values, to build up an idea of the average and peak values. That could tell you if the readings are wildly too high. It looks like the information could be available in the "snatch graph"; I've always wanted to poke my nose into one of them 

One possibility is that the inverter IGBTs are shooting through at high temperatures. I have no idea if that's likely or not. If it is happening, then these errors aren't just a nuisance, they are saving your inverter from a catastrophic failure event. Perhaps some transient from Eskom weakened them, and they are now prone to shoot through at high temperatures. If so, a new set of 4 IGBTs should fix the problem, at least until the next surge.

Now 40A to 80A sounds reasonable to me in terms of a surge and/or over current.  Any idea of where we are talking here in terms of the event, or could this be on the supply, load, battery and/or PV? I will double check this, but I think I was on 80A for a total current, and I brought this down to 40A after my firmware updates - will double check it.

I don't really monitor the inverter - it's something that always got a "back seat" and needless to say it's never been done.  The only thing I have is WatchPower (I can prob. get something organized for ICC too).  Let me get everything back up and running, and I'll see about leaving my laptop connected to it during the day, with the hopes that we can get a trip.  Then at least we have something to look at in terms of data values up to the event occurring.  I just don't think PowerWatch monitor the temps.

From what I remember, touching the heatsinks at the top it wasn't untouchable hot, but it was definitely > 60C - that's the actual outside part of the heatsinks.  Loads of the inverter is in the 700VA region, so not a lot.

It's just strange as my input & output on the inverter is protected by 20A breakers.  I'm really, really surprised that I am not getting a trip before the inverter goes to throw it's little fit.  Again, why I am still leaning PV.

Let me get some data, let me get a re-occurrence and let's see if there's anything that can be picked up in the data I have / can get.  If we can narrow it down to a couple of components, one can always as you suggested replace those components to just be on the safe side, and see how things go from there. What you are saying seems plausible enough, even so that it can verify my story that it only happens on warm, hot days...

Just an FYI as well... Generally I wouldn't be worried but it may be applicable here now.

Inverter's batteries and PV connected.  No supply, no load.  Almost 700W coming in over the PV.

Batteries are fully charged at 58.2V (the clampon says 1.1A current, not sure how accurate that is), but are cooking - quite a lot.  Never used to cook so badly. 

Either way, not sure where that energy from the PV is going to right now... 

Program 02 (Max charging current), is and was at 20A.

Edited October 9, 20187 yr by Chris Knipe

3 minutes ago, Chris Knipe said:

(the clampon says 1.1A current, not sure how accurate that is),

Does the clampon have a dc selection, if it has and its a reputable make then they are pretty accurate. Not all clampons can be used for reading Dc current

1 minute ago, Jaco de Jongh said:

Does the clampon have a dc selection, if it has and its a reputable make then they are pretty accurate. Not all clampons can be used for reading Dc current

That's more than likely the issue then.  I was also wondering how a clamp-on would read DC current.  

Thanks for clarifying Jaco.

39 minutes ago, Chris Knipe said:

I was also wondering how a clamp-on would read DC current.  

Hall sensor. Usually accurate to around 100mA.

AC clamps are usually current transformer, and those can only see the AC component.

Edited October 9, 20187 yr by plonkster

Had a look now yes. I only have transformers here (other than my clamp on), nothing for DC

Anyways, will undo the bypass in an hour or so, and then I guess we wait for the next warm day....

8 hours ago, Chris Knipe said:

Now 40A to 80A sounds reasonable to me in terms of a surge and/or over current.  Any idea of where we are talking here in terms of the event, or could this be on the supply, load, battery and/or PV?

Oh, sorry, I neglected to say. This is all instantaneous current at the inverter's output, which is also its input when AC charging. But this isn't during AC charging, so it must be inverter output current.

I've also just realised that my shoot-through idea is shot down; shoot-through would cause a surge of current through the IGBTs and from the bus capacitors, but not at the inverter output. Perhaps the output MOV (Metal Oxide Varistor) is breaking down at too low a voltage, as a result of damage from an Eskom surge. Again, I don't know how plausible this is. 

A quick google search indicates that every "hit" the MOV takes degrades it, and some say that this results in the "knee" voltage lowering. Perhaps they are temperature sensitive as well, though I did not find that in my quick search. That makes the MOV a good candidate in my opinion, especially in areas (South Africa?) where mains surges are more common. A visual inspection of the MOVs near the AC input and output may show discolouration, which would be a good indication that they should be replaced.

I discuss this more on AEVA here.

Hate to say this guys... MOVs look brand new.  There's hardly even any dust in the unit.  Even the fans are clean.

2 MOVs on the main board close to the power out, and 2 MOVs on the AC Input directly, all 4 looks brand new.  I don't see any visible burning on any components for that matter.  The entire internals of the unit looks brand new.

Main board MOVs

A shot in the dark.

On the forum there is also a Infini 3kw no batteries tripping on startup, no errors and the fact that the other night I had my inverter and a UPS switching 4 times due to Eskom.

Assuming the inverter is not faulty and settings are right, what are the chances that Eskom is causing the inverter to trip - and not the heat?

Any chance of putting test equipment on the inverter input / Eskom, to constantly read and record the incoming current, to see what came in when the inverter flips again?

Still busy with disassembly currently and working my way through the service manual to test things.  The amount of test equipment I have is very limited unfortunately.  Thus far, I do not see anything spectacular and/or obviously wrong.  Its most certainly not heat/dust, the inverter's cleaner than my kettle TBH.

Only once the main board is out, will you get to the nice stuff (drivers, mosfets, etc).  

Well everything is now taken apart and laying all over the place... No visible damage anywhere... Not a MOV, not a mosfet, not a cap, nudda.  No smell of burning.  Hardly even any dust.  Everything for a lack of a better word, looks brand new...