Yup, a Kill-A-Watt would work too. And like you, I can see the Fac on the Sunny Data Control screen from all the inverters.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Powerwall 2 + UPS Connundrum - and solution
- Thread starter NinjaVece
- Start date
-
- Tags
- ups 65 hertz
Power Saver, last Sunday I actually did TWO tests. The first was EXACTLY like yours, only I started at 100% SoC. I tracked the freq and found that the PW2 initially jumped to 63 Hz and held that level until the PW2 got down to 96% SoC. Then the PW2 lowered the freq to 60 Hz and within 5 minutes I was off to the races with 3.5-4 kW solar input. When the PW2 recharged to 99%, it raised the freq to 60.5 Hz and a majority of my M215s cut out which lowered my solar out to 0.8-1 kW - a little less than my house was using. This SLOWLY dropped the PW2 back to 96% and the PW2 lowered the freq to 60 Hz and I was fully producing again.
This cycled once more: up to 99%, freq to 60.5 Hz, down to 96%, freq to 60.0 Hz. It never did again go to the full 63 Hz while cycling this way, which is good. That means that only the initial solar inverter smackdown goes to the 63 Hz level when starting at 100% SoC, the rest of the outage, the freq adjust is much milder. I ran this initial test for 45 minutes and concluded that I'm OK at this point.
The second 2 minute test at 100% SoC and solar production was just to actually video the output of an APC UPS as it went to battery mode for a second and then reacted to the PW2 spooling up the Hz rating. The LCD on the APC is quite good at showing this. I wish I could provide the video I took of this sequence but TMC doesn't allow for video formats (at least I couldn't get it to work) . Here is a quick pic of the PW2 at the 60.5 Hz level from the first test.
If there is a way to show video in a post, let me know.
Hmm, interesting. It is my experience that you have to keep on Tesla support to get issues solved. I have been relentless with Tesla and my third-party installer (and even their general contractor I work SGIP through) to get this solved. I've talked to Enphase and the original folks that put my solar panels in for good measure.
Suffice it to say, you should test this scenario in a controlled setting. Some weekend, pull the Service Disconnect under the same conditions you experienced your outage with and track what happens. You will find that if you want testing done right, you're going to have to do the testing yourself.
I have done quite a bit of testing of my new setup lately with generally good results, but also with some worries. I recently saw the dance of the line frequency during a test power failure. I started at 100% SOC on the PW2 and the line frequency went to 63 Hz to shut down the solar inverter, but the solar came on quickly as needed to deal with power needs in the house that exceeded the PW2 capabilities during this test (I have only 1 PW2 at this point). The PW2 turned the solar inverters on and off to keep the SOC above 96%. So far so good.
In another test I did a few days ago however, when I purposely overloaded the PW2 with no grid power, the PW2 shut down and the generator came on as I expected. I also expected that the PW2 would use the solar during that downtime to up its SOC from ~50%. However in the 20 or so minutes I waited, the solar and the PW2, which was in standby, never made the connection and neither came online until after I restored grid power. During that time, the solar inverters reported no voltage, so the PW2 was not trying to make contact with the solar. It was not a frequency issue. That was a disappointment. Perhaps without a load, the PW2 does not feel the need? I had a different result when I manually threw the generator transfer switch without overloading the PW2. In that case, the PW2 and solar played together nicely while the house was being powered by the generator. Obviously more testing is necessary, but getting the PW2 out of standby without the grid up may be an issue.
In another test I did a few days ago however, when I purposely overloaded the PW2 with no grid power, the PW2 shut down and the generator came on as I expected. I also expected that the PW2 would use the solar during that downtime to up its SOC from ~50%. However in the 20 or so minutes I waited, the solar and the PW2, which was in standby, never made the connection and neither came online until after I restored grid power. During that time, the solar inverters reported no voltage, so the PW2 was not trying to make contact with the solar. It was not a frequency issue. That was a disappointment. Perhaps without a load, the PW2 does not feel the need? I had a different result when I manually threw the generator transfer switch without overloading the PW2. In that case, the PW2 and solar played together nicely while the house was being powered by the generator. Obviously more testing is necessary, but getting the PW2 out of standby without the grid up may be an issue.
Well it seems when you are at 100% and solar is producing more than house load, opening the main breaker set the PW to raise the frequency so high without a ramp up. Maybe that is to protect the PW when it has no more capacity and to forcibly shut down the inverters. Seems a bit extreme. I'm at 100% now and could do this test... but opening the main breaker usually knocks my inverters off-line during the transfer, especially when the PW is not discharging.
Did you try toggling the PW switch after the overload and shutdown? According to this information on the FAQ site, that may be necessary to restart it without grid power.
To resume Powerwall operation, turn off all lights and appliances in your home, then toggle the on/off switch located on the side of Powerwall. If, after an extended utility outage, Powerwall is depleted, then wait for the next sunny day before toggling the on/off switch.
Yes I did try flipping the switch on the PW2 to get it to wake up and feel the solar, but in the time I waited, likely about 10 minutes, it never put out any voltage to which the inverters could sync. Perhaps if I waited longer? I do plan to do that test again.
I think that I have also learned that one can't think of the PW2 as a monolith when it comes to syncing. If it is powering the house and the power fails, then the PW2 inverters are running and synced to the grid and solar and things go smoothly with a minimal disruption. If however the PW2 is charging and the inverter is not running then it takes a while for the solar to kick in after a power failure. As discussed above, the third case is that the PW2 is in standby and then there may or may not be a problem getting it going and synced with the solar.
I think that I have also learned that one can't think of the PW2 as a monolith when it comes to syncing. If it is powering the house and the power fails, then the PW2 inverters are running and synced to the grid and solar and things go smoothly with a minimal disruption. If however the PW2 is charging and the inverter is not running then it takes a while for the solar to kick in after a power failure. As discussed above, the third case is that the PW2 is in standby and then there may or may not be a problem getting it going and synced with the solar.
I just opened the main breaker while the PW was at 100% charge and solar was less than 0.1kW production... and the frequency immediately jumped to 62.97Hz. With the house load at 1.5kW, it dropped to 99% charge but the frequency is still way up at 62.83Hz. So it would seem that at high % charge, the PW will immediately raise the frequency to about 63Hz and hold it there until the % charge drops. This is different then when the % charge is approaching 100% and it can gradually increase the frequency to curtail production.
After 30 minutes now, charge is at 98% and the frequency is back to 60.00Hz. It dropped linearly as the charge dropped from 99% to 98%.
With clear weather in the forecast, this weekend is going to be an off-grid test for the next two days...
After 30 minutes now, charge is at 98% and the frequency is back to 60.00Hz. It dropped linearly as the charge dropped from 99% to 98%.
With clear weather in the forecast, this weekend is going to be an off-grid test for the next two days...
Over the last few days I have conducted several forced power failures to see how the system responds to various situations. I had two perplexing problems that seemed to limit the use of the PW2 with solar in a power failure, but I think I may understand both issues now.
As to getting the PW2 out of standby after a discharge overload, I was recently successful on that issue. I shut down everything including the generator, with the house going dark for a few minutes, and then power cycled the PW2. That brought it back to life and it started powering the house and the PW2 synced with the solar after the typical 5-minute wait. Whether I waited longer in the power off position this time, or the PW2 seeing the load from the house was what made the difference, I don’t yet know, but at least there seems to be a solution to this issue.
The second problem was the ability of my solar to charge the PW2 during a power outage, either with the generator powering the house or the PW2 powering the house. Up until today I assumed this sync problem was likely to be a PW2 line frequency stability related issue, but now I think otherwise.
I should add at this point that I have 2 different SMA inverters a 7 kW and a 4 kW and I have been largely testing around noon. What I have seen quite a few times now is that after a power failure, one solar inverter would come on line, outputting e.g. 2-3 kW, and then the second one comes on may 10 or 15 seconds later, I see a total of 7 kW or more from the solar, and then within seconds, they both go down. I have watched the solar go on and off many times the last few days, trying to power the house and charge the PW2 with the grid off. After every failure it would take 5 minutes for the solar to try again and then it would turn off within seconds. PW2 would gradually lose SOC, even though the sun was shining, and solar output was enough for all of the above.
Perhaps too much solar output, is my current theory. It seems that if the solar is putting out more than the sum of the house needs and the single PW2’s max charging rate, 5 kW, it shuts down the solar completely with a rapid frequency change. To test this theory, I tried again late in the day when the solar was only putting out 3.5-4 kW and in this case, once synced the solar output from both inverters was stable and fell gradually as the sun went down.
I thought that the PW2 could up its frequency to curtail the solar output. If that is the case, then perhaps the PW2 is over reacting to the sudden potential charging overload when the second inverter come on and ends up shutting the solar down completely due to a dynamics issue as the inverters can’t respond quickly enough. If that is the case, then a slow ramp of solar power, e.g. in the morning, may produce a better result with the PW2 able to throttle the solar output. I have to find a morning when I don’t have meetings to test this case. My SMA inverters are ~ 10 and 4 years old.
What I really need is a second PW2 as 2 would be able to take all the excess power that my solar panels could deliver and is really the ultimate solution to this problem. A 2ndPW2 would obviously significantly lessen the chances of accidently overloading the PW2 in the discharge mode. I have had one on order for more than a year, but my local installer tells me that he thinks that my 2ndPW2 is still 6 months off.
Is my understanding of the above correct?
As to getting the PW2 out of standby after a discharge overload, I was recently successful on that issue. I shut down everything including the generator, with the house going dark for a few minutes, and then power cycled the PW2. That brought it back to life and it started powering the house and the PW2 synced with the solar after the typical 5-minute wait. Whether I waited longer in the power off position this time, or the PW2 seeing the load from the house was what made the difference, I don’t yet know, but at least there seems to be a solution to this issue.
The second problem was the ability of my solar to charge the PW2 during a power outage, either with the generator powering the house or the PW2 powering the house. Up until today I assumed this sync problem was likely to be a PW2 line frequency stability related issue, but now I think otherwise.
I should add at this point that I have 2 different SMA inverters a 7 kW and a 4 kW and I have been largely testing around noon. What I have seen quite a few times now is that after a power failure, one solar inverter would come on line, outputting e.g. 2-3 kW, and then the second one comes on may 10 or 15 seconds later, I see a total of 7 kW or more from the solar, and then within seconds, they both go down. I have watched the solar go on and off many times the last few days, trying to power the house and charge the PW2 with the grid off. After every failure it would take 5 minutes for the solar to try again and then it would turn off within seconds. PW2 would gradually lose SOC, even though the sun was shining, and solar output was enough for all of the above.
Perhaps too much solar output, is my current theory. It seems that if the solar is putting out more than the sum of the house needs and the single PW2’s max charging rate, 5 kW, it shuts down the solar completely with a rapid frequency change. To test this theory, I tried again late in the day when the solar was only putting out 3.5-4 kW and in this case, once synced the solar output from both inverters was stable and fell gradually as the sun went down.
I thought that the PW2 could up its frequency to curtail the solar output. If that is the case, then perhaps the PW2 is over reacting to the sudden potential charging overload when the second inverter come on and ends up shutting the solar down completely due to a dynamics issue as the inverters can’t respond quickly enough. If that is the case, then a slow ramp of solar power, e.g. in the morning, may produce a better result with the PW2 able to throttle the solar output. I have to find a morning when I don’t have meetings to test this case. My SMA inverters are ~ 10 and 4 years old.
What I really need is a second PW2 as 2 would be able to take all the excess power that my solar panels could deliver and is really the ultimate solution to this problem. A 2ndPW2 would obviously significantly lessen the chances of accidently overloading the PW2 in the discharge mode. I have had one on order for more than a year, but my local installer tells me that he thinks that my 2ndPW2 is still 6 months off.
Is my understanding of the above correct?
I think you need more data, like the line frequency when the solar is going in and out, in order to complete your understanding of what is happening. In any case, having a second Powerwall would definitely help the situation. In the meantime, if you are around when this happens, you can just turn on the breaker to one inverter so that it stays under the power limit. The Powerwall software may have an elegant way of handling the curtailment when the SOC is getting high, but it may not for the condition where the solar is starting to output more power than the Powerwall can absorb. Clearly, it's possible to implement an elegant curtailment for this case, but it may simply not be implemented yet. It is also possible that the current ramp rate on the solar inverters is too high for the Powerwalls to signal the solar inverters to curtail proportionally.
I'll certainly do more experiments. I really hope that a slow ramp of solar output in the AM won't cause this same issue. If so, it greatly reduces the magnitude of the problem. As to watching the line frequency, I may have to move my Kill A Watt from the garage to an outlet nearer where I am playing with the breakers, watching the inverters, etc. My power central, where PG&E, solar, the generator and now the PW2 all get together, is about 200' from the house, and I currently running back and forth to monitor it all! I am also concerned that the frequency ramp is so fast I may not catch it, but it is worth a try.
I thought of switching off the breaker to one of my solar panels and I was going to try that before I realized that yesterday afternoon they were both producing less the charge max. But frankly I would hate to have to do that other than as an experiment as I really want to use all the solar I can get during a power failure! That was one of the main reasons to get a PW2.
One temporary solution might be to increase the load from the house, during the middle of the day (during a power failure) to essentially dump the excess solar power into a resistive load (water heater) or charge one of our Teslas. However, that would be far from automatic and a bit tricky to find the right load that would not overpower the PW2 before the solar came on line but would take enough current to prevent the PW2 from objecting to too much solar when it did come on. I do like the idea of the car somehow sensing the PW2 gateway on the same network and automatically charging at the right level when there was excess solar capacity, including when the PW2 was at 100% SOC, but that is another story.
I thought of switching off the breaker to one of my solar panels and I was going to try that before I realized that yesterday afternoon they were both producing less the charge max. But frankly I would hate to have to do that other than as an experiment as I really want to use all the solar I can get during a power failure! That was one of the main reasons to get a PW2.
One temporary solution might be to increase the load from the house, during the middle of the day (during a power failure) to essentially dump the excess solar power into a resistive load (water heater) or charge one of our Teslas. However, that would be far from automatic and a bit tricky to find the right load that would not overpower the PW2 before the solar came on line but would take enough current to prevent the PW2 from objecting to too much solar when it did come on. I do like the idea of the car somehow sensing the PW2 gateway on the same network and automatically charging at the right level when there was excess solar capacity, including when the PW2 was at 100% SOC, but that is another story.
Last edited:
Car charging is a good way to dump the excess solar. You can sit in the car and adjust the amperage while watching the Power Flow screen. When it reaches a value that would give you enough headroom to turn on the second solar inverter, you could walk over and do that. The adjustable charging current is really handy on Teslas. You're right though, it would be nice if it could be remotely and automatically adjusted to consume excess solar from the cloud.
I had my installer move my Grid CTs from the grid input on the Tesla Gateway to the actual grid main circuit. When he was done, I showed him the app and the charging current adjustment in the car. He was really impressed because he didn't know the Tesla cars could do that and that it would reflect in the app in near real time.
I had my installer move my Grid CTs from the grid input on the Tesla Gateway to the actual grid main circuit. When he was done, I showed him the app and the charging current adjustment in the car. He was really impressed because he didn't know the Tesla cars could do that and that it would reflect in the app in near real time.
When I got home from the morning’s meeting, the solar panels were putting out 4.5 kW with the PW2 charging at 3.4 kW and a 68% SOC. I figured that this was perfect time to see if a slow ramp in solar power, as the sun continued to rise, would achieve stable operation with the PW2 able to control the solar output and keep it in an acceptable range for charging with a minimal load from the house.
I killed the breaker to disconnect from the grid. Soon thereafter the solar was putting out 5.6 kW (both inverters active) with the PW2 charging at 4.5 kW. However, 20 minutes later both inverters were off line for whatever reason with the PW2 at 69% SOC. Over the next half hour, I watched both inverters come online for a short period of time and them dump, same as before. At this point, while briefly online, the larger inverter was putting out 5 kW and the smaller, 2.5 kW. So the slow ramp in solar power, really didn’t help. Once one inverter went down for whatever reason, it had trouble coming back online without taking both down. The sudden rise in power for charging to above its max charging rate seemingly caused the PW2 to shut down all solar with a rapid frequency rise.
The next step was to kill the breaker for the smaller solar panel. This allowed the larger panel to come on line and stay there without issue charging the PW2 at 4 kW. After 15 minutes of continuous operation this way, I started my Model S charging at 15 amps, which more or less zeroed out the 4 kW PW2 charging as now all of the solar was going to the house and car. Time to turn back on the smaller inverter. Now I was able to see a reliable operation with both inverters putting out a total of up to 8 kW. I ran this way till the PW2 SOC got to 98% at which point first one inverter then the other shut down. At that point I turned the grid back on as it was almost 2 PM and time to start sending solar to the grid!
This would seem to be one way of operating during a power failure while using both solar panels and getting some charge for the car. That is the good news. What seems more difficult is having the PW2 control the output of the inverters to keep the charging rate at an acceptable limit with a minimal load from the house.
According the SMA document on Frequency Shift Power Control (SFPC), both my inverters are listed as compliant with this standard and therefore I would expect that the PW2 should be able to control the power output of the two inverters . See http://files.sma.de/dl/7910/SB-OffGrid-TI-US-en-19.pdf
As I read this document it suggests that power control happens proportionally starting at 61 Hz and the output goes to 0 at 62 Hz. I never really saw much if any curtailment of power or frequency rise before the inverters shut down in first test of the morning the morning, and I only saw the frequency rise to 60.3 before the inverters shut down on the later test, and only when the PW2 achieved a 98% SOC. The last 2 Hz must have been so fast I missed it. Obviously FSPC works to protect the PW2 from overcharging in terms of SOC, but it is not obviously whether it will use FSPC to limit the charging rate at a lower SOC.
I will try again to see if I can see stable power control via FSPC with both inverters running but so far, I have not been successful in seeing that work.
I killed the breaker to disconnect from the grid. Soon thereafter the solar was putting out 5.6 kW (both inverters active) with the PW2 charging at 4.5 kW. However, 20 minutes later both inverters were off line for whatever reason with the PW2 at 69% SOC. Over the next half hour, I watched both inverters come online for a short period of time and them dump, same as before. At this point, while briefly online, the larger inverter was putting out 5 kW and the smaller, 2.5 kW. So the slow ramp in solar power, really didn’t help. Once one inverter went down for whatever reason, it had trouble coming back online without taking both down. The sudden rise in power for charging to above its max charging rate seemingly caused the PW2 to shut down all solar with a rapid frequency rise.
The next step was to kill the breaker for the smaller solar panel. This allowed the larger panel to come on line and stay there without issue charging the PW2 at 4 kW. After 15 minutes of continuous operation this way, I started my Model S charging at 15 amps, which more or less zeroed out the 4 kW PW2 charging as now all of the solar was going to the house and car. Time to turn back on the smaller inverter. Now I was able to see a reliable operation with both inverters putting out a total of up to 8 kW. I ran this way till the PW2 SOC got to 98% at which point first one inverter then the other shut down. At that point I turned the grid back on as it was almost 2 PM and time to start sending solar to the grid!
This would seem to be one way of operating during a power failure while using both solar panels and getting some charge for the car. That is the good news. What seems more difficult is having the PW2 control the output of the inverters to keep the charging rate at an acceptable limit with a minimal load from the house.
According the SMA document on Frequency Shift Power Control (SFPC), both my inverters are listed as compliant with this standard and therefore I would expect that the PW2 should be able to control the power output of the two inverters . See http://files.sma.de/dl/7910/SB-OffGrid-TI-US-en-19.pdf
As I read this document it suggests that power control happens proportionally starting at 61 Hz and the output goes to 0 at 62 Hz. I never really saw much if any curtailment of power or frequency rise before the inverters shut down in first test of the morning the morning, and I only saw the frequency rise to 60.3 before the inverters shut down on the later test, and only when the PW2 achieved a 98% SOC. The last 2 Hz must have been so fast I missed it. Obviously FSPC works to protect the PW2 from overcharging in terms of SOC, but it is not obviously whether it will use FSPC to limit the charging rate at a lower SOC.
I will try again to see if I can see stable power control via FSPC with both inverters running but so far, I have not been successful in seeing that work.
@powersaver,
Thanks for that input. Obviously FSPC works now for shutting down the solar near 100% SOC, but not, it seems, to prevent charging at over the max rate at lower SOCs. Will FSPC work in that latter case if the inverters are reconfigured for off-grid?
Thanks for that input. Obviously FSPC works now for shutting down the solar near 100% SOC, but not, it seems, to prevent charging at over the max rate at lower SOCs. Will FSPC work in that latter case if the inverters are reconfigured for off-grid?
I believe that is a function of the PW and not the inverter. Similar to your test, if the SOC is high (>97%) and solar is on when the grid fails, the PW will send the frequency to 63Hz to shut everything down so it doesn't overload. I think that is a protective measure since it doesn't know how fast the inverters will respond. 63Hz will shut every inverter down immediately.
As for your charging situation, if the solar is more than 5kW per PW, and the grid fails, even if SOC is < 97% it will do the same, as the solar will overload the PW. As you demonstrated, bringing up house load will prevent this. I presume you have only 1 PW and your total solar is more than 5kW so this will happen to you. But under normal charging, with solar below 5kW per PW, as the SOC gets to 98%, the PW will begin a ramp up from 60Hz at 98% to 63Hz at 99% to curtail production. I tested that, monitoring the frequency ramp versus SOC, and it works.
You would have to monitor the frequency to confirm if over production during charging is curtailed or just shutdown. I suspect it will just shutdown, since it is an overload. I'm actually surprised Tesla let you connect that much solar to 1 PW since overloading is clearly an issue for you.
As for your charging situation, if the solar is more than 5kW per PW, and the grid fails, even if SOC is < 97% it will do the same, as the solar will overload the PW. As you demonstrated, bringing up house load will prevent this. I presume you have only 1 PW and your total solar is more than 5kW so this will happen to you. But under normal charging, with solar below 5kW per PW, as the SOC gets to 98%, the PW will begin a ramp up from 60Hz at 98% to 63Hz at 99% to curtail production. I tested that, monitoring the frequency ramp versus SOC, and it works.
You would have to monitor the frequency to confirm if over production during charging is curtailed or just shutdown. I suspect it will just shutdown, since it is an overload. I'm actually surprised Tesla let you connect that much solar to 1 PW since overloading is clearly an issue for you.
wbhokie
Member
Any updates on this?
If so :
1) Will the new gateway release do anything to change it?
2) Has anyone found any additional UPS systems that might work?
3) Has the Powerwall 2 been updated to resolve this frequency issue that normal UPS systems will work?
If so :
1) Will the new gateway release do anything to change it?
2) Has anyone found any additional UPS systems that might work?
3) Has the Powerwall 2 been updated to resolve this frequency issue that normal UPS systems will work?
wbhokie, as of the last couple of firmware updates to the PW2, the Hz increase when the battery is fully charged and the grid power goes out so the Gateway 'islands' the house current has decreased from 65 Hz to 63 Hz (62.9).
This level is acceptable to nearly ALL UPS systems that can get one over the short one second 'hump' when the relay snaps shut to 'island' the house. Thus, you don't need a special UPS any longer if using the latest PW2 firmware releases.
Hope this helps you.
wbhokie
Member
Similar threads
