Tesla switched inverter to Delta (no optimizers) from promised solaredge (with optimizers)

[nwdiver]

NOTE the error log top page, I am not familiar with what "arc faults" mean in solar-speak, but sounds suspiciously like some wire, connector, or panel is loose and voltage is arcing across a gap.

Looks like the shade is capable of making the inverter think there's an arc fault. This is causing the inverter to shutdown.

That's unfortunate that you have 3 strings with shade on an inverter with 2 MPPT channels. But there's a good chance a firmware update should solve most of your problems.

 

[aesculus]

Tesla updated my FW to 3.18 back in early May as I was having some clipping. Note I have one set of solar shingles on PV1 and the other side of the building on PV2

Here is me before the FW update to 3.1.8 when I was on 3.1.2. I was originally on 2.2.13:


And here is me a few weeks after the FW upgrade to 3.1.8. It was a bit cloudy that day hence the

Now with some limb removal and cleaning the panels I have this:

 

[cali8484]

@dareed1 and @nwdiver

Here is my attempt to get yesterday’s (6/15) history of volt/amp graphs out of my delta M10-TL-US with one of versions of delta's mobile app (Mtool itself on the apple store or android was hardly helpful, i was able to find via a archive site mtool prof (M professional). and how to activate it (act code dareed gave actually). Firmware version seems to be 2.2.13.
The set of graphs are: volts/amps for one 24 hr window (06/15) PV1, PV2, (and PV3 :?)
PV1 (volt, amp, 24 hour window, dies off when the left 12 panels get shaded, so probaly tied to those)
View attachment 674265View attachment 674264
PV2 (maybe the MPPT with 2 strings tied to it? amperages are near double PV1, and its output drops at the time the right 24 panels in upper pisture start seeing shade. )
View attachment 674266View attachment 674267
PV3 ??? RE PV3: maybe just the firmware or software has the variables in it but no terminals to it. The app thinks there are 3x PV, or at least some layers of the software/firmware do. My "checkup" installer was pretty sure there was nothing in there to hook a third input up to, and it looked that way to me when we were poking around in there together ( i was an electronics tech for a number of years an age ago, so i am not totally new to this stuff, just new to solar.)
View attachment 674268View attachment 674269

NOTE the error log top page, I am not familiar with what "arc faults" mean in solar-speak, but sounds suspiciously like some wire, connector, or panel is loose and voltage is arcing across a gap. Please correct me if that is not the right idea. Looseness could be INSIDE the inverter too I guess, even bad solder joint, faulty circuitry. How would the firmware be trying to detect that? Maybe voltage drops faster than a threshold? Or maybe its just crappy firmware?
View attachment 674270

IN NEXT POST: AC active power, reactive pwr, current, freq, volts; plus the tesla apps Solar CT leg info for the same time window.

Both PV1 and PV2 DC voltages show large oscillations which seem consistent with MPPT not handling shading well at all. The inverter appears to always jump to 400V+ which suggests its algorithm is not finding the correct voltage point for shading.

 

[Heretic]

Tesla updated my FW to 3.18 back in early May as I was having some clipping. Note I have one set of solar shingles on PV1 and the other side of the building on PV2

Here is me before the FW update to 3.1.8 when I was on 3.1.2. I was originally on 2.2.13:

View attachment 674438
And here is me a few weeks after the FW upgrade to 3.1.8. It was a bit cloudy that day hence the
View attachment 674440
Now with some limb removal and cleaning the panels I have this:

View attachment 674441

hmm, it is a notable differnce from firmware. what had to happen to actually upgrade the FW? inperson (3 ft away to get blue tooth) or remote possible? if remote, do you have to bring local ethernet to be hooked up to it?

 

[Heretic]

This is great information guys, defiante Thank You's to @nwdiver @dareed1 @cali8484 @aesculus .

Should I send these screenshots from "M Professional" to Tesla people? Will they be telling me something like "you shouldnt be logging in with that app! only we can!" ?

I did manage to get my case responded to by a self-named "escalation team" with "Tesla’s Energy Resolutions Team" a couple days ago. I would really like to help them stay on track before they let me fall in the cracks again.

It sounds particularly like the arc fault triggers in log are meaning something needs to be tweaked, and maybe a firmware update could help. Unfortunately it seems the consensus is that combining 2 sets of 12 (=24 panels) panels onto one input in the inverter seems like a major source of the issues i am having with both shading and the inverter cutting out when a wispy cloud flys by (This is what i've suspected from the start, the behavior it just too weird to be normal), as well as an inverter not playing well with input changes from the array (clouds flying by repeatedly throwing the output to zero suddenly multiple times a day has really seemed like it should not be happening).

I now need to just figure out how to succinctly show tesla people the info that lets them see that too.

 

[dareed1]

Should I send these screenshots from "M Professional" to Tesla people? Will they be telling me something like "you shouldnt be logging in with that app! only we can!" ?
I did manage to get my case responded to by a self-named "escalation team" with "Tesla’s Energy Resolutions Team" a couple days ago. I would really like to help them stay on track before they let me fall in the cracks again.
It sounds particularly like the arc fault triggers in log are meaning something needs to be tweaked, and maybe a firmware update could help. Unfortunately it seems the consensus is that combining 2 sets of 12 (=24 panels) panels onto one input in the inverter seems like a major source of the issues i am having with both shading and the inverter cutting out when a wispy cloud flys by (This is what i've suspected from the start, the behavior it just too weird to be normal), as well as an inverter not playing well with input changes from the array (clouds flying by repeatedly throwing the output to zero suddenly multiple times a day has really seemed like it should not be happening).
I now need to just figure out how to succinctly show tesla people the info that lets them see that too.

I think you are mis-representing the concensus. Here's my understanding: 1) Connecting two strings in parallel can cause suboptimal power production during shading of one of the two strings. The voltage developed by the two strings must be equal and that can limit the total power developed by the two. For example, if one panel in string A is completely shaded and all others in strings A and B aren't shaded, then both strings can produce about 10A, but the MPPT will need to drop the string voltages to 11/12 of optimal. The power output of the two strings will be 22/24 of the output when no panel is shaded. If the two strings were on separate MPPT channels, the unshaded string would produce 12/12 (because it can operate at full voltage) and the shaded string would produce 11/12, for a combined power output of 23/24 compared to completely unshaded. If my arithmetic is correct, the worst case would be to have one half of the panels on string A shaded; then the power output would be 75% of the power yielded if the strings were wired to separate MPPT channels.

2) Strings in parallel shouldn't affect MPP tracking during cloudy days or poor sun angle (assuming the both strings are on the same azimuth). When you see power drops that last much longer than the transient insolation reduction caused by a passing cloud, or failure to produce power at low sun angles (but no shading) then that is very likely a MPPT algorithm (or hardware) failure. Similarly, if production near solar noon is clipped at less than inverter capacity the cause isn't parallel strings (unless MPPT channel current capability is exceeded, and your isn't). Taken together, that's why we have urged you to get the firmware upgraded, with two examples of actual improvement.

Assuming your inverter is connected correctly to the cloud, Tesla can upgrade the FW remotely. In my case, when we had some of the problems you're seeing, once I reported this to Tesla, and made contact with a technically savvy person, they upgraded the FW without me asking. I didn't know about M Professional at that time and had no idea that the FW was down revision.

I have sent screen shots of M Professional to Tesla, and so has a neighbor (the one whose issues were completley cured by a FW upgrade). It didn't cause any comment or issue that I could see. It's just data, but it allows them (if seen by a technical person) to quickly undertand that the system is not just producing less power than the homeowner expected, but is actually misbehaving. I sent emails summarizing in a short paragraph the behavior shown graphically on the sceenshots that I attached. I did this because in an early exchange I was told by the the Tier 1 technical support person that he could only see the text of my email and not the attachments. In that case, I asked if I could foward the email to him directly and he gave me his email address.

 

[sorka]

There is a TON of misinformation out there about optimizers. If you don't have a complicated array arrangement with minimal shading you will likely have better performance without optimizers. Even with shade there's not going to be a noticeable difference.

Optimizers what are they good for?

If the same thing had happened to me as the OP, I would have halted the installation the morning of. When they showed up, I made them show me the optimizer they were going to install.

 

[sorka]

No... it's the same. Exactly the same. A schottky diode is a schottky diode. And whether you have one panel or a string of 12 panels is also irrelevant since the current is the same. I guarantee 100% of UL rated solar panels have bypass diodes. They're not optional. They are required by code and for UL approval. Without bypass diodes if a leaf landed on a panel that section could easily catch fire.

But optimizers don't do what they do with large bypass diodes and the internal diodes still perform their normal function of shunting current around shaded areas which is why you lose just as much production when one cell group is shaded with an optimized system as you do with a non-optimized system.

There's nothing magic about optimizers. They're just performing the same MPPT function a string inverter does on the panel level instead of a string level. Which... if all strings in a panel face the same direction is ~100% unnecessary. Even if there's shade. If one cell group in a 300v string of 10 panels becomes shaded the MPP tracker will 'see' that on one of the sweeps it does several times per second and drop string voltage to 290v. The bypass diode shunts string current around the shaded section... exactly as would happen if there was an optimizer on the shaded panel.

Diodes can't give you panel level monitoring. I would have preferred micro inverters because I'm scared to death of SE dying in peak of summer and me losing 3 months of production in a part of the year where those three months is more than half a years entire production.

The optimizers were the compromise. Honestly, the only reason I went with Tesla is their really low price and the fact that I wanted powerwalls as well.

 

[Heretic]

I think you are mis-representing the concensus. Here's my understanding: 1) Connecting two strings in parallel can cause suboptimal power production during shading of one of the two strings. The voltage developed by the two strings must be equal and that can limit the total power developed by the two. For example, if one panel in string A is completely shaded and all others in strings A and B aren't shaded, then both strings can produce about 10A, but the MPPT will need to drop the string voltages to 11/12 of optimal. The power output of the two strings will be 22/24 of the output when no panel is shaded. If the two strings were on separate MPPT channels, the unshaded string would produce 12/12 (because it can operate at full voltage) and the shaded string would produce 11/12, for a combined power output of 23/24 compared to completely unshaded. If my arithmetic is correct, the worst case would be to have one half of the panels on string A shaded; then the power output would be 75% of the power yielded if the strings were wired to separate MPPT channels.

2) Strings in parallel shouldn't affect MPP tracking during cloudy days or poor sun angle (assuming the both strings are on the same azimuth). When you see power drops that last much longer than the transient insolation reduction caused by a passing cloud, or failure to produce power at low sun angles (but no shading) then that is very likely a MPPT algorithm (or hardware) failure. Similarly, if production near solar noon is clipped at less than inverter capacity the cause isn't parallel strings (unless MPPT channel current capability is exceeded, and your isn't). Taken together, that's why we have urged you to get the firmware upgraded, with two examples of actual improvement.

Assuming your inverter is connected correctly to the cloud, Tesla can upgrade the FW remotely. In my case, when we had some of the problems you're seeing, once I reported this to Tesla, and made contact with a technically savvy person, they upgraded the FW without me asking. I didn't know about M Professional at that time and had no idea that the FW was down revision.

I have sent screen shots of M Professional to Tesla, and so has a neighbor (the one whose issues were completley cured by a FW upgrade). It didn't cause any comment or issue that I could see. It's just data, but it allows them (if seen by a technical person) to quickly undertand that the system is not just producing less power than the homeowner expected, but is actually misbehaving. I sent emails summarizing in a short paragraph the behavior shown graphically on the sceenshots that I attached. I did this because in an early exchange I was told by the the Tier 1 technical support person that he could only see the text of my email and not the attachments. In that case, I asked if I could foward the email to him directly and he gave me his email address.

thank you, dareed. point taken: firmware could be a big issue, and the parallell strings alone would not cause as much problem as i am seeing with shading.... (i am seeing essentially complete drop out of the 24 panel doubled string set when only 1 panel of it is completely covered.)
I will try to get screen shots of e.g. arc faults, voltages, etc. to a responsive person at tesla before monday morning Thank you. Hopefully some one on the "escalation team" that knows the technicalities sees it. oh,... and assume they may not get to see the figures
My inverter is not cloud connected to my knowledge. Do you know if it pull an ether net cable to plug into its PCB if i can get m-prof to configure it? The inverter has no wifi board or cellular add on. Only BT and zigbee, but the manual says ethernet is std not optional. Is it dchp default or do i need to set it up statically inside the inverter, etc?

 

[sorka]

This is great information guys, defiante Thank You's to @nwdiver @dareed1 @cali8484 @aesculus .

Should I send these screenshots from "M Professional" to Tesla people? Will they be telling me something like "you shouldnt be logging in with that app! only we can!" ?

I did manage to get my case responded to by a self-named "escalation team" with "Tesla’s Energy Resolutions Team" a couple days ago. I would really like to help them stay on track before they let me fall in the cracks again.

It sounds particularly like the arc fault triggers in log are meaning something needs to be tweaked, and maybe a firmware update could help. Unfortunately it seems the consensus is that combining 2 sets of 12 (=24 panels) panels onto one input in the inverter seems like a major source of the issues i am having with both shading and the inverter cutting out when a wispy cloud flys by (This is what i've suspected from the start, the behavior it just too weird to be normal), as well as an inverter not playing well with input changes from the array (clouds flying by repeatedly throwing the output to zero suddenly multiple times a day has really seemed like it should not be happening).

I now need to just figure out how to succinctly show tesla people the info that lets them see that too.

You haven't actually paid Tesla yet, right?

 

[sorka]

My system went through 2 revisions that required submitting plan revisions to my county. My county wouldn't have accepted them had I not signed off on them. Tesla made me sign each one.

 

[nwdiver]

Diodes can't give you panel level monitoring. I would have preferred micro inverters ...... Honestly, the only reason I went with Tesla is their really low price.

As you say. Module level monitoring and increased redundancy are nice... just not for an additional ~$30 and ~$60 per panel respectively.

 

[sorka]

As you say. Module level monitoring and increased redundancy are nice... just not for an additional ~$30 and ~$60 per panel respectively.

Saving $30 / panel isn't even a consideration because Tesla only offers one configuration of different sizes and they were 1/3rd cheaper than the next lowest bid for a comparable system. If I wanted micro inverters, Tesla was half the cost the cheapest system I could pay an installer to install micro inverters.

In fact, Tesla was so cheap, the cost of the system was the same cost I would have paid for hardware alone had I bought the system and installed it myself.

 

[Heretic]

If the same thing had happened to me as the OP, I would have halted the installation the morning of. When they showed up, I made them show me the optimizer they were going to install.

No, I havent given them the check yet. I am surprised they arent quicker to try to get it from me. I told them we need a resolution first.
It's definitely been a learning experience over this. I should have been more concerned about checking if they were going to do what we had been writing/saying up to that point (before letting them unload in my driveway). 6pm at night it was "all done, See ya! "
From what i've seen SE does not have as many users complaining as this delta. seems SE better support, more information, clean power output curves from inverters ive seen posted, etc. The one thing i have seen that might have been an issue with solaredge is a couple articles saying they had too high a warrantee replacement rate. But that's been a couple years now. Really wish i'd been more concerned about it. Ive spent near $300,000 in the past 2.5 yrs on contractors. This one (tesla) has been more wasted time trying to get them to correct things than any other contractor i have dealt with. they are not all perfect by any means, but all of them that had some thing that needed to be corrected come up have been far better at handling things than these guys have been. E.g. Oh, right we talked about that panel being 200A, not 150A, no prob (done in less than a week, including re-inspection). ... or ... right, we had talked about 14 lights in that room, and somehow its 10. fixed in days. much more concerned about making it what we'd agreed on. Tesla's been in their own league on this one.

 

[Heretic]

As you say. Module level monitoring and increased redundancy are nice... just not for an additional ~$30 and ~$60 per panel respectively.

If that's all we'd have to pay on our end (end user) extra to get those features, I'd pay it. Small price (36x $30, or even 36x $60) to be able to pinpoint issues later down the road. Sort of like an insurance policy that only costs 2% up front for life. I've already wasted more time on this crazy back and forth than it takes me to make that much money. Maybe its not worth it for someone (like yourself it seems) that knows how to get up there and identify which panel is wonky.
Enphase microinverter system (closest bid price to tesla, same size array, but had a much smaller battery size) was still almost $60k, vs 48k. Another was going to do 2 lg batteries, and SE 10k/optimizers was over 60k, anotehr, electriq, was even more but was close to tesla batt array and inverter nominal capacity. 48 to telsa seemed right.
I expected the "big company lame CSR" type of issue. I didnt expect months of hangups just to get them to look deeply into it.

 

[nwdiver]

Maybe its not worth it for someone (like yourself it seems) that knows how to get up there and identify which panel is wonky.

Shouldn't really be your 'job' either. It sucks that Teslas service is so terrible.

For an installer it's easy to pin point a problem w/o module level monitoring (string level is sufficient) and the vast majority of customers stop checking their systems after ~4 weeks unless their electric bill is off. My Mom has a SE system and didn't notice a panel that had been failed for months. I happened to spot it on a visit.

 

[sorka]

No, I havent given them the check yet. I am surprised they arent quicker to try to get it from me. I told them we need a resolution first.
It's definitely been a learning experience over this. I should have been more concerned about checking if they were going to do what we had been writing/saying up to that point (before letting them unload in my driveway). 6pm at night it was "all done, See ya! "
From what i've seen SE does not have as many users complaining as this delta. seems SE better support, more information, clean power output curves from inverters ive seen posted, etc. The one thing i have seen that might have been an issue with solaredge is a couple articles saying they had too high a warrantee replacement rate. But that's been a couple years now. Really wish i'd been more concerned about it. Ive spent near $300,000 in the past 2.5 yrs on contractors. This one (tesla) has been more wasted time trying to get them to correct things than any other contractor i have dealt with. they are not all perfect by any means, but all of them that had some thing that needed to be corrected come up have been far better at handling things than these guys have been. E.g. Oh, right we talked about that panel being 200A, not 150A, no prob (done in less than a week, including re-inspection). ... or ... right, we had talked about 14 lights in that room, and somehow its 10. fixed in days. much more concerned about making it what we'd agreed on. Tesla's been in their own league on this one.

In your tesla solar account, you have a document titled "customer layout" that you e-signed that says exactly what you're getting. Did they update this document for the delta inverter and have you sign it? If not, then I wouldn't pay them until they install what they sold you.

 

[sorka]

Shouldn't really be your 'job' either. It sucks that Teslas service is so terrible.

For an installer it's easy to pin point a problem w/o module level monitoring (string level is sufficient) and the vast majority of customers stop checking their systems after ~4 weeks unless their electric bill is off. My Mom has a SE system and didn't notice a panel that had been failed for months. I happened to spot it on a visit.

How did you know which panel it was?

 

[nwdiver]

How did you know which panel it was?

I didn't. Saw that not all the panels were reporting. If it was one of my installs I would spot it by noticing that the string voltage wasn't correct. That would give me the string. Tracking down the exact panel could be as quick as ~5 minutes with a FLIR or ~20 if I need to take voltage measurements.

 

[sorka]

I didn't. Saw that not all the panels were reporting. If it was one of my installs I would spot it by noticing that the string voltage wasn't correct. That would give me the string. Tracking down the exact panel could be as quick as ~5 minutes with a FLIR or ~20 if I need to take voltage measurements.

Wait, I shut my inverter off and use my Raytek ST-80 on a hot sunny day, I don't see any difference in panel temperatures. I've tried. What do you do exactly?

I've noticed that things like even a mild breeze can make a huge difference in panel temps of the whole group and production but haven't seen differences with the inverter on or off.