18kpv loud bang on day 3 of operation

[zanydroid]

Isn't that what this thread is?

Oh it was crowdsourced and the technicians were contacted and responding. TBH the techs sounded about reasonable capability for mass market residential equipment (sometimes people on here expect the service level for a $25K gross margin small business account. Go figure), so for me I would have tried to progress on my own still. I would have told them it broke with a config that passes their calculator and specs, and insisted on a timely replacement instead of trying to debug after letting out the smoke.

Or. Go through their script for how to manage RMAs for an infant mortality unit

 

[Deere103]

Yes, I skimmed the thread too quickly, quite right.

I would continue dialog with SigSolar. See if they can tell you what the maximum observed voltage was.
Your goal should probably be an RMA at this point, and they need to stand behind their absolute maximum ratings.
Probably marginal components.

Personally, I would still remove one or two panels from the string.
I would treat 500V as my Voc limit rather than 600V.

Yes depending on what the resolution is I likely will reduce string size to 11 and parallel 2 11 panel strings inside the unit on mppt 1, but the recommendation to aim for sub 400 by SS is crazy, I would kill my output, I'd rather just turn my pv off anytime it gets sub freezing than loose that much power generation in summer.

 

[Deere103]

Oh it was crowdsourced and the technicians were contacted and responding. TBH the techs sounded about reasonable capability for mass market residential equipment (sometimes people on here expect the service level for a $25K gross margin small business account. Go figure), so for me I would have tried to progress on my own still. I would have told them it broke with a config that passes their calculator and specs, and insisted on a timely replacement instead of trying to debug after letting out the smoke.

Or. Go through their script for how to manage RMAs for an infant mortality unit

Thanks, I was out of town and they wouldn't discus ideas or look at monitor history without being in front of the unit, so I gave a bit of a pass, but being there to diagnose real time wasn't so far more helpful. It feels like home insurance where they are going to keep requesting more pics and data on my wiring until they find a gotcha or free pass out of standing by the product.... But hey, maybe they'll surprise me... I'm not to the point of bitching yet! Just wish I had a video of the banging.....i clicked it off fast as I was worried about a fire

 

[AntronX]

video of PV behavior following the "pops/bangs" ; sorry no way to upload video bc of size

Is this how slow screen refresh rate is on these inverters? That's really bad.

 

[fnnwizard]

Thanks, I was out of town and they wouldn't discus ideas or look at monitor history without being in front of the unit, so I gave a bit of a pass, but being there to diagnose real time wasn't so far more helpful. It feels like home insurance where they are going to keep requesting more pics and data on my wiring until they find a gotcha or free pass out of standing by the product.... But hey, maybe they'll surprise me... I'm not to the point of bitching yet! Just wish I had a video of the banging.....i clicked it off fast as I was worried about a fire

I only have mobile access now so very hard to have close look at the log. But a quick viewing shows mppt to be fine, inverter not tho. And as wpns mentioned, vbus V is all over the place at some points.
there ar a couple warnings triggered for around that time.

Can u go into data and pull event log?

 

[Deere103]

Is this how slow screen refresh rate is on these inverters? That's really bad.

Seems kinda slow but could be from my phone shutter too, it's not too bad in person, the touch screen well that's another story... Haha

 

[Deere103]

I only have mobile access now so very hard to have close look at the log. But a quick viewing shows mppt to be fine, inverter not tho. And as wpns mentioned, vbus V is all over the place at some points.
there ar a couple warnings triggered for around that time.

Can u go into data and pull event log?

These are the only 2 events I have seen ever, and happened right after I turned off pv due to the banging noise. I killed the grid power too and the batteries would not discharge at this time, actually I still don't know if the batteries will power the load, they power the unit fine and charge for a few seconds until pv voltage bleeds off, not sure this gives you any clue, but maybe. Also should be noted no faults on the battery side that I have ever seen either.

 

[fnnwizard]

These are the only 2 events I have seen ever, and happened right after I turned off pv due to the banging noise. I killed the grid power too and the batteries would not discharge at this time, actually I still don't know if the batteries will power the load, they power the unit fine and charge for a few seconds until pv voltage bleeds off, not sure this gives you any clue, but maybe. Also should be noted no faults on the battery side that I have ever seen either.

I got a chance to review your data and can't really make sense of some things.

There was:
* 1 point of 3min segment where only 1 battery pack was reporting.
* Inconsistent battery temp logs, it's usually reporting a fictitious number of 2 or 3 (°C) but yours went to 38,38,45,45. for 4, 3-minute segments

* and big ? issue was, for at least 8 minutes, the packs were charging at about 2.7kW level and 18kpv not reporting any of it.
This is right after you rebooted system at 10AM from previous shutdown at 7:47AM

Do you have 2 packs of the EG4 wall mounts?
Do you have another charger hooked to the batts?

You might as well post pictures of install so more eyes can vet for you.
We'd need to see settings of 18kpv to further analyze. ---> Maintenance tab, it should default to Remote Set ---> Read ---> Export PDF

Also logs of all settings changes --->Maintenance tab ---> Set Record ---> Export Data

Lastly it sounded as if you had 18 back online when speaking with EG4/SS? Any additional data logs of that?

 

[Deere103]

I got a chance to review your data and can't really make sense of some things.

There was:
* 1 point of 3min segment where only 1 battery pack was reporting.
* Inconsistent battery temp logs, it's usually reporting a fictitious number of 2 or 3 (°C) but yours went to 38,38,45,45. for 4, 3-minute segments

* and big ? issue was, for at least 8 minutes, the packs were charging at about 2.7kW level and 18kpv not reporting any of it.
This is right after you rebooted system at 10AM from previous shutdown at 7:47AM

Do you have 2 packs of the EG4 wall mounts?
Do you have another charger hooked to the batts?

You might as well post pictures of install so more eyes can vet for you.
We'd need to see settings of 18kpv to further analyze. ---> Maintenance tab, it should default to Remote Set ---> Read ---> Export PDF

Also logs of all settings changes --->Maintenance tab ---> Set Record ---> Export Data

Lastly it sounded as if you had 18 back online when speaking with EG4/SS? Any additional data logs of that?

Thanks for the attention and input. Attached is the PDF of the settings and the log from EG4/SS "troubleshooting", it was simply turning on the system with PV2, and then switching to PV3 input. No load or discharge from batteries other than to power the inverter. I could not locate the setting change log you mentioned, still learning the monitor center nuances. where is the "set record" tab?

I do have 2 eg4 indoor wall mount batteries, the 14.3 kWh packs.

I have an eg4 chargeverter attached, but it is completely isolated and is only able to energize when a generator is connected to the outside generator inlet port. I essentially added this as my grid down, no solar backup solution.



The spike you are talking about around 10 I think was just when I turned the PV input back on, it charged at 3 kw for a few seconds but then the voltage quickly bled down as shown in the video, this is the same thing that happened with PV input 3 when troubleshooting with SS, it was a much warmer day and the string size is 11 currently, so voltage was even further from Voc limits. Not sure about the inverter not reporting or where you're seeing that?

Do you know if the PV inputs are completely isolated? ie if there was damage to PV input 2, would that bleed over and damage the others?

You seem to have integrated circuit component knowledge, any ideas what component failure would lead to the PV voltage bleed like this?



I see what you're talking about, I don't know for sure, but would suspect the erratic battery readings could be related to me shutting down the system quickly as things were popping and smelling like electrical burning.

 

[rogerheflin]

Seems kinda slow but could be from my phone shutter too, it's not too bad in person, the touch screen well that's another story... Haha

The rolling shutters (that are in just about every camera--except some really expensive ones) produce some odd looking refresh patterns in both videos and photos.

 

[Deere103]

No sound or narration, so I can't really tell what's going on, but your MPPT definitely isn't tracking right. Maybe your HV bus is hitting some maximum, can you see if that's showing up in your logs? It's be VBus1, VBus2, and VBusP (which is about half of VBus1).

Mine tend to run around 400, 325, and 200, respectively, if that helps...
View attachment 261315View attachment 261316

Posting my Vbus info, did you plot multiple days using the monitor program, or did you send to excel or something, I can only do a chart compare within the monitor for 1 day that I've found?

here is 11/30 when things were functioning properly, FYI, the huge spike in charging power at beginning of day was me testing the AC grid charge capability.



Data from 12/1 when things went south:



and the troubleshooting session with SS, spike 1 PV2, spike 2 PV3

 

[Quattrohead]

At the end of the day whatever we think about your PV strings, the inverter should not have gone bang and in theory is supposed to protect itself. So I'm pretty sure you have a valid warranty claim and it will get taken care of. Pity it's taken so long.

 

[Brucey]

At the end of the day whatever we think about your PV strings, the inverter should not have gone bang and in theory is supposed to protect itself. So I'm pretty sure you have a valid warranty claim and it will get taken care of. Pity it's taken so long.

Not sure if @SignatureSolarJess was looped in apologies if already done.

 

[Quattrohead]

She's on holiday

 

[wpns]

Posting my Vbus info, did you plot multiple days using the monitor program, or did you send to excel or something, I can only do a chart compare within the monitor for 1 day that I've found?

I’m polling modbus registers to my own databases and plotting the data of interest to me. Kind if a kludge, but that’s the DIYer in me.

 

[ricardocello]

I’m polling modbus registers to my own databases and plotting the data of interest to me. Kind if a kludge, but that’s the DIYer in me.

I'm a big fan of writing your own code to poll MODbus registers (RTU and TCP).
It gives you total control over what you collect and see. I've done it for the Cerbo and energy meters.
It's old school, but that's ok I'm old.

 

[wpns]

I'm a big fan of writing your own code to poll MODbus registers (RTU and TCP).
It gives you total control over what you collect and see. I've done it for the Cerbo and energy meters.
It's old school, but that's ok I'm old.

Probably overkill but:

Poll both inverters, all six batteries, Acrel grid monitor, ambient temperature and humidity, and (soon, maybe tomorrow?) two DC ammeters.

Write immediate data to Redis and update the LAN MQTT broker.

Once a minute, scrape the latest data from Redis and write to SQL.

On the machine that’s running the MQTT broker, write everything published to MQTT into Redis, then once a minute into SQL as above.

Once a minute, create all the graphs and update the appropriate web pages. Make timelapse movies of battery balancing at 8PM (well after sunset).

 

[ricardocello]

Probably overkill but:

Poll both inverters, all six batteries, Acrel grid monitor, ambient temperature and humidity, and (soon, maybe tomorrow?) two DC ammeters.

Write immediate data to Redis and update the LAN MQTT broker.

Once a minute, scrape the latest data from Redis and write to SQL.

On the machine that’s running the MQTT broker, write everything published to MQTT into Redis, then once a minute into SQL as above.

Once a minute, create all the graphs and update the appropriate web pages. Make timelapse movies of battery balancing at 8PM (well after sunset).

Nice job, That’s a comprehensive setup, and you know everything that’s happening.
If you like detail (and it appears you do), you’ve got it.
I definitely like the web page updates, it hides all of the machinery you’ve built.

I’m not a database guy other than occasional mySQL, but Redis looks particularly useful.

Victron does lots of this with VRM, but having your own setup locally is more rewarding.

 

[fnnwizard]

Sorry so late to reply, just got back from dropping my daughter off at school.... finals week for her... and vball club tourney for my other all weekend and I;'m beat so need to make this quick.

where is the "set record" tab?

Right under Remote Set.

Was chargeverter ever charging the bats?

Do you know if the PV inputs are completely isolated? ie if there was damage to PV input 2, would that bleed over and damage the others?

You seem to have integrated circuit component knowledge, any ideas what component failure would lead to the PV voltage bleed like this?

Separate MPPT's so yes is isolated.

Moving from 2 to 3 would act same since logs hint that nothing is wrong with your #2 mppt.

There's no bleed as you call it. That DC Bus Voltage is normal. There's a dc booster in there that will take PV that is <400V to ~400V so inverter can convert to AC more efficiently. Peak AC is actually around 339V, (240*√2) since we measure at RMS of 240V.

If PV is over 400V, it passes all of it up to the 600V max.

I'lll take a look at your settings and other log when I have more time.
But its all for academics as I always keep an open mind for the what and whys so this thread is of interest to me... unit needs to be replaced from EG4/SS

 

[SignatureSolarJess]

Obviously some of your debugging will involve measuring Voc unloaded on a cold morning plus trying a different MPPT input. Let us know how you resolve this. @SignatureSolarJess might be able to help

Thank you for tagging me! Yes, you're welcome to reach out and we'll do all we can to help

 

[Deere103]

Sorry so late to reply, just got back from dropping my daughter off at school.... finals week for her... and vball club tourney for my other all weekend and I;'m beat so need to make this quick.

Right under Remote Set.

Was chargeverter ever charging the bats?



Separate MPPT's so yes is isolated.

Moving from 2 to 3 would act same since logs hint that nothing is wrong with your #2 mppt.

There's no bleed as you call it. That DC Bus Voltage is normal. There's a dc booster in there that will take PV that is <400V to ~400V so inverter can convert to AC more efficiently. Peak AC is actually around 339V, (240*√2) since we measure at RMS of 240V.

If PV is over 400V, it passes all of it up to the 600V max.

I'lll take a look at your settings and other log when I have more time.
But its all for academics as I always keep an open mind for the what and whys so this thread is of interest to me... unit needs to be replaced from EG4/SS

Thanks, yeah I'm curious too, but wish I wasn't having to troubleshoot a brand new unit and potentially rewiring another one if am exchange happens...haha hoping it's just a lemon, but I still want to understand what failed and why no faults so I don't have to worry about my house burning down in future.

and yes, i tested charging the batteries using the chargeverter when I installed it, it was powered by the generator, the inverter was not wired yet when I did this so I don't have the data for this, I ran it for about an hour to test and see SOC change. here's a short video clip of that I sent to a friend of mine, no narration. FYI the chargeverter wiring comes from the gen 50 Amp inlet, and shares house entry conduit with inverter grid wiring, but is isolated from inverter otherwise.

WeTransfer | Send Large Files Fast - Up To 2GB Free

The simple, quick and secure way to send your files around the world without an account. Share your files, photos, and videos today for free.

we.tl

 

[Deere103]

Thank you for tagging me! Yes, you're welcome to reach out and we'll do all we can to help

Thanks, I have reached out to tech support twice, last monday and friday. I sent a bunch of requested info to tier 2 support, but I have not heard back yet

 

[SignatureSolarJess]

Thanks, I have reached out to tech support twice, last monday and friday. I sent a bunch of requested info to tier 2 support, but I have not heard back yet

If you'd like to shoot me a DM with your contact info, I can look into the status for you!

 

[Deere103]

I was contacted by SS tech and given some info on the system. They are being helpful and starting the RMA process with EG4. It looks like they will ship me a new unit and have me return my damaged 18kpv for investigation. She said it may take over a month, but that SS/EG4 would inform me as to what the damage was and components affected. I will report back what they find and give update when new unit is installed.

I appreciate the input from everyone here. Also, FYI as a result of this thread there has been some clarification from EG4 on voltage limits and behavior when exceeded. There was a lot of confusion between us, and with SS techs but the official word from EG4 is after exceeding 500 vdc on a MPPT, the inverter will switch off PV input until a sub 500 vdc range is reached. It was my impression that there were efficiency losses out of the 230-500 vdc range, but it appears it simply bypasses the PV power when >500 vdc. in any case, it's not supposed to go boom and let the smoke out. Hopefully EG4 sends a tech bulletin or clarifies in the manual. it's a bit strange to me to rate the DC input up to 600 vdc but then limit the MPPT to 500 vdc usable, it makes little sense to me, but maybe this is just their way of forcing in some headroom before damaging the unit.

We'll see how it goes, but I will likely operate 2 strings of 11 panels instead of 12 and input into MPPT to parallel inside the inverter. I will still have 6 panels left over from my pallet, I may add them to another MPPT input, but for me that's another 600 ft of wire run, so If i do it, I'll probably do 3 strings of 10 panels.

thanks again everyone, and thanks SS for helping to make things right!

 

[ricardocello]

... the official word from EG4 is after exceeding 500 vdc on a MPPT, the inverter will switch off PV input until a sub 500 vdc range is reached. ...

Now your data on 1 Dec makes sense.