Powerwall/Solar Edge behavior in power outage

[blulegend]

Hi,

I simulated a power outage today by shutting the main breaker. My powerwall is set up in a partial backup setup with a subpanel that is backfed by both the powerwall and solaredge inverter. The powerwall was about 90% charged at the time per the app. Everything in the subpanel stayed on as expected but the solar inverter was shut off. Is this correct behavior? Does the powerwall need to be drained a little further before the inverter is allowed back on? I will need to do some further testing but wanted to get an idea of expected behavior first.

Thanks!

Jack

 

[cwied]

Yes, this is expected. The Powerwall will raise the frequency it generates to shut down the solar inverter if it is close to full. If you search the forum, you can probably find a post with the exact percentage where it will turn on again.

 

[power.saver]

The PW starts curtailment by raising the frequency at 97%, so that should not have been a factor in your test. However, flipping the breakers could cause a momentary outage as the PW kicks in, and the inverter would shutdown due to the disturbance. After 5 minutes, it should have restarted. Did you leave the breakers off for at least 5 minutes?

 

[blulegend]

It would’ve been close to 5 minutes. Pretty sure it was longer but can’t remember for sure. I will have to run a longer test next time.

I do know that it was after 2 pm which is when the powerwall starts discharging based on my rate plan and letting all the solar back to the grid. So would a main breaker shut off let the inverter experience this brief outage if the powerwall was already backfeeding into the same sub panel as the inverter?

 

[power.saver]

It has been speculated that when the PW is in discharge mode there is less chance of a disturbance because it is already pushing out power. When it is charging and has to switch to discharge, a disturbance seems more likely. If you can do the test again, and if the inverter shuts down, try to wait more than 5 minutes to see if it recovers. You might also look for an indication on the inverter as to why it shut down.

 

[blulegend]

Thanks. When I looked at the inverter, it behaved like it was turned off with no lights. It’s a solar edge HD 7600 inverter. I was able to check the Tesla app backup history and it reported a 7 minute outage today.

I’ll check again tomorrow if I can for longer.

 

[cr0ntab]

I too have a SolarEdge HD 7600 inverter.

Mine will shut down when the power gets flipped but the readout on it says it will resume in 5m.

After the 5m is up the inverter fires up and resumes. When the batteries get full the powerwalls bump the frequency and the inverter slows down.

It all works great.

In fact, I can only fire my 5 ton AC off grid if the inverter is up and both it and the batteries are putting out power.

 

[blulegend]

Thanks guys! Today I tried it again and because it’s the weekend the powerwall was at the reserve of 25%. This time when I shut off the main breaker, the inverter shut down but remained powered. After a countdown, it restarted and is now charging up the powerwall again. So it’s working great. I’m guessing the powerwall was too charged last time so the system kept the inverter off. Loving it!

 

[abasile]

We have two Powerwalls and a solar system consisting of SunPower PV panels with micro inverters. Everything seems to work fine when the grid is up. However, when there's no grid power (either due to an outage or me flipping the main breaker for a test), no charging has ever occurred from solar as far as I've been able to tell.

Given an ambient temperature in the range of 5 C, I wonder how low the Powerwalls' state of charge needs to go before they'll lower the frequency to turn on the solar inverters. I tried a test earlier today, with the Powerwalls at 63% charge. I shut down the main breaker and waited over half an hour, and no solar production occurred. This doesn't seem right. You can see the gap in solar production in the below screenshot. (I unplugged an EV from its charger a minute or so before turning off the main breaker. And today's solar production is lower than average due to clouds.)

The Powerwall Support team previously told me that everything is working "normally" when I tried a similar test (to no avail) at 89% --> 84% charge. We're expecting some storms, but after the weather calms down, I'll probably try testing again with the Powerwalls at only 20 or 30% charge.

Being able to charge from solar during an extended outage was a key selling point, and the Powerwall Support folks have assured me that this should work. Particularly in cooler/colder temperatures, have others experienced an upper limit on SOC (state of charge) above which their Powerwalls won't allow solar production during outages? It is my expectation that the Powerwalls should be able to recharge from solar, during outages, even if the ambient temperature is significantly below freezing!

 

[miimura]

You need to monitor the line frequency after the outage (simulated or otherwise). A simple device like the Kill-A-Watt will display it for you. I have had this problem during a real outage. I noticed that the solar was generating for a short time, then went offline. I called Powerwall support and they found that the Powerwall was driving a frequency of 63Hz even when the SOC was below 80%. Normally, solar inverters will sleep for 5 minutes after any power interruption. If the frequency is normal, they should start up again. If the frequency is outside a narrow band around 60Hz, the solar inverters will remain offline. This was acknowledged as a "that shouldn't happen" issue by Powerwall support and my case was supposedly escalated. However, I have not received any feedback that my case has been resolved.

 

[abasile]

Thanks, I'll have to follow your advice and monitor the frequency.

I'm sorry to hear that your case hasn't been resolved yet. As we live high in the mountains, I really want to make sure we can recharge from solar during an extended outage, so I've followed up with Powerwall Support a number of times. The people who answer the phone have been kind and sympathetic, but the "second tier" support folks have so far insisted (behind the scenes from my perspective) that they see nothing wrong. Well, I think *something* is amiss if our Powerwalls are at 63% and still not charging from solar when the grid is down! (They have no problem charging all the way up to 100%, from solar, when the grid is up.)

You need to monitor the line frequency after the outage (simulated or otherwise). A simple device like the Kill-A-Watt will display it for you. I have had this problem during a real outage. I noticed that the solar was generating for a short time, then went offline. I called Powerwall support and they found that the Powerwall was driving a frequency of 63Hz even when the SOC was below 80%. Normally, solar inverters will sleep for 5 minutes after any power interruption. If the frequency is normal, they should start up again. If the frequency is outside a narrow band around 60Hz, the solar inverters will remain offline. This was acknowledged as a "that shouldn't happen" issue by Powerwall support and my case was supposedly escalated. However, I have not received any feedback that my case has been resolved.

 

[NuShrike]

We have two Powerwalls and a solar system consisting of SunPower PV panels with micro inverters. Everything seems to work fine when the grid is up. However, when there's no grid power (either due to an outage or me flipping the main breaker for a test), no charging has ever occurred from solar as far as I've been able to tell.

Given an ambient temperature in the range of 5 C, I wonder how low the Powerwalls' state of charge needs to go before they'll lower the frequency to turn on the solar inverters. I tried a test earlier today, with the Powerwalls at 63% charge. I shut down the main breaker and waited over half an hour, and no solar production occurred. This doesn't seem right. You can see the gap in solar production in the below screenshot. (I unplugged an EV from its charger a minute or so before turning off the main breaker. And today's solar production is lower than average due to clouds.)

View attachment 373160

Have you tried going into your SunPower Monitor, and changing the panels' grid-profile from IEEE-1547a to one of the CA CPUC Rule21 profiles? That may make the micro-inverters more compatible.

It looks like there's a new profile called CPUC R21 Reactive Power Priority. Don't set that yet.

You should only do this when the panels are online (sun is up and powering panels).

 

[abasile]

Now that's interesting! Thank you for sharing this. Normally, I access the old, "advanced" SunPower portal at SunPower Monitoring System but that doesn't seem to give me any way of changing settings on the micro-inverters. The newer "dumbed down" SunPower monitoring site certainly doesn't have anything as useful as that! I've tried directly accessing the SunPower box that's plugged into my router (direct Ethernet cable), but it doesn't respond on http port 80. Do you have SunPower micro-inverters? How do you access their settings?

I will say, though, that I wouldn't expect any of this to be necessary. My micro-inverters are reporting an "average operating frequency" of 59.99 Hz. I would think that if the Powerwalls can supply 60.0 Hz during an outage (simulated or actual), the micro-inverters should be happy and turn on. At no point during the Powerwall order and installation process did Tesla tell me that I'd have to do anything special to get my particular micro-inverters to work with the Powerwalls, and they did ask multiple times what model of panels/inverters I have.

Have you tried going into your SunPower Monitor, and changing the panels' grid-profile from IEEE-1547a to one of the CA CPUC Rule21 profiles? That may make the micro-inverters more compatible.

It looks like there's a new profile called CPUC R21 Reactive Power Priority. Don't set that yet.

You should only do this when the panels are online (sun is up and powering panels).

 

[miimura]

I think the inappropriate frequency shifting is/was a Powerwall bug. I was on 1.29 when it happened to me. I should really test it again now that I'm on 1.31.

 

[abasile]

I think the inappropriate frequency shifting is/was a Powerwall bug. I was on 1.29 when it happened to me. I should really test it again now that I'm on 1.31.

For what it's worth, I'm on 1.32. The Powerwall Support folks "pushed" the update to my Powerwalls in response to my prior calls, but something still seems amiss.

Anyway, I just went on Amazon and ordered a new multimeter (https://www.amazon.com/gp/product/B00KXX2OYY) that should enable me to monitor the line frequency.

 

[NuShrike]

Now that's interesting! Thank you for sharing this. Normally, I access the old, "advanced" SunPower portal at SunPower Monitoring System but that doesn't seem to give me any way of changing settings on the micro-inverters. The newer "dumbed down" SunPower monitoring site certainly doesn't have anything as useful as that! I've tried directly accessing the SunPower box that's plugged into my router (direct Ethernet cable), but it doesn't respond on http port 80. Do you have SunPower micro-inverters? How do you access their settings?

I will say, though, that I wouldn't expect any of this to be necessary. My micro-inverters are reporting an "average operating frequency" of 59.99 Hz. I would think that if the Powerwalls can supply 60.0 Hz during an outage (simulated or actual), the micro-inverters should be happy and turn on. At no point during the Powerwall order and installation process did Tesla tell me that I'd have to do anything special to get my particular micro-inverters to work with the Powerwalls, and they did ask multiple times what model of panels/inverters I have.

It's accessed directly through the "debug" Ethernet port inside the PVS5 box -- which is behind its firewall.

I have SunPower 360w AC "-D" panels (the last with the SolarBridge inverters before the switch to Enphase aka "-E"), and I look at their settings via that internal "installer" web portal.

There's ways to make that debug-port generally accessible via WiFi without having to open up that box again. Otherwise, it's a black-box for any sort of "setting".

I would think your Neurio readings should be accurate enough to get realtime idea of what frequency the waveform is. Whenever my PWs get installed, I'll see if I have the same issues.

 

[iSolar]

We lost power today at around 7:20am. When the some came up around 830, the solar didn't come on. I have SolarEdge HD - it was displaying Error 18x7B, which is: MainError - Grid measurements are out of range. Contact the grid operator per SolarEdge manual. Also, I could tell the frequency was off as my UPS devices were showing grid outage and died.

When solar was still not on by 10am I called Tesla. The person said the battery was too cold to take a charge and would turn on the solar once it warmed up. I'm in Virginia -- it's cold here -- with the battery in my garage, the Tesla agent said it was showing a temp of 2.6 degrees celsius. She said the battery has a heater to warm itself up. 4 hours into the outage, the battery discharged to 55% and never brought the frequency in line to turn on the solar. At this point the power came back on.

When the power came back on, 5 minutes later the solar turned on (this was to be expected) and the battery immediately began taking a charge at over 4kW, peaking at 5kW until full a bit over an hour later. This seems to be a clear bug or at least a problem. If the battery was too cold to heat itself after 4 hours, that's a problem -- the heater is insufficient. However, this seems bogus because as soon as the grid returned, the battery began charging off the solar.

When I first got the system in June, I did some self tests by flipping the grid off. The system worked as expected during those tests -- the solar came on after 5 minutes (tripped off by the initial switch over) and charged the power till full at which point the solar turned off as the powerwall brought the frequency out of spec. I don't know what firmware that was, but it seemed to perform better. If the reason the battery wouldn't take a charge is because it was too cold, as Tesla says, that is unacceptable as it needs to be able to heat itself in a reasonable amount of time. And that doesn't explain why it charged itself as soon as the grid came online.

I will be following up with Tesla. I'm thrilled the powerwall kept my house running -- it was 10 degrees out when we lost power and we kept the lights on, the TV/internet running, and furnace going throughout the outage Neighbors came over as our schools were on a 2 hour delay and the kids watched movies here till it was time for school. However, I'm very concerned that the system didn't ever bring the solar online. If this outage has been more prolonged, we would have eventually run out of power despite the fact that it was a bright sunny day. Fortunately, our outages tend to be short. I'm curious if eventually, as the Powerwall got low enough, it would have switched on the solar eventually, but that would still be very flawed software as peak sun may have been gone by then.

 

[abasile]

Tesla’s “second level” support gave me the same excuse about the temperature being low, at 12C in my case. That’s totally nonsensical. I wonder if there is indeed a software defect that could be affecting a large number of Powerwall users. Most people won’t notice because most outages are short. But it’s an absolutely fatal problem in a true disaster scenario when the Powerwall is most needed. I do expect that we’ll get to the bottom of this.

 

[Tbryant87]

Today, my installer contacted me to test out a simulated grid outage. My installer has another client who is experiencing the same issue mentioned here. They wanted to test with me to see if they get the same results. Tesla is claiming the low temp issue too. In my testing, I too was not able to charge via solar during an outage. My SolarEdge inverter was saying the AC frequency was too high. This was even after my battery level was below 80%. My PW is located in my garage which sits around 45f. My installer is going to ask Tesla an update to figure out what's going on. I'll share once I find out. Does anyone here have an update?

 

[dortor4ev]

my powerwall keep my solar online through an outage and the solar powers the house - charging is a secondary consideration - i’ve even seen when the powerwall is full that my solar output is curtailed to match house demand.

i have enphase micro inverters and live in CA where temps are more mild.