Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Powerwall 2 + UPS Connundrum - and solution

This site may earn commission on affiliate links.
So I did a test tonight after the sun was certainly down and turned off the grid power to the house. UPS did not like the power and my LED lights were flickering like crazy. I don't have a way to measure the power frequency but all signs point to the Powerwall providing 66hz even with no input from the PV system.

I have an open ticket with Tesla but has anyone else tested this?

I'm not trying to sound like a snarky member here, but have you read this thread? There is a lot of examples of people here doing exactly what you are describing, testing it, and what their current work-arounds and discussions with Tesla have been.
 
So I did a test tonight after the sun was certainly down and turned off the grid power to the house. UPS did not like the power and my LED lights were flickering like crazy. I don't have a way to measure the power frequency but all signs point to the Powerwall providing 66hz even with no input from the PV system.

I have an open ticket with Tesla but has anyone else tested this?

The PWs don't know if the sun is up or not. The decision to jump to greater than 60Hz is solely based upon the battery state of charge. So whether it's day or night, if your battery levels are north of 92% or so, your frequency is going to be raised in order to prevent any inverters form coming online.

The E-mail I received from Tesla regarding this phenomenon indicated that there is no hard and fixed battery level that will trigger the increased frequency. But it is generally somewhere between 85 and 95%.
 
I'm not trying to sound like a snarky member here, but have you read this thread? There is a lot of examples of people here doing exactly what you are describing, testing it, and what their current work-arounds and discussions with Tesla have been.

Appreciate the help, and yes of course I have read the thread. Other than specific UPS suggestions there aren't any actual workarounds to the issue and I haven't heard back from Tesla yet myself. I could have been more clear that my specific question was about the night time situation where the PV is not producing. vlad_kalish had a night outage back in January, but it wasn't a test and he also did not have a way to check the frequency. NinjaVece 's reply shortly after makes complete sense but didn't address the night question directly. NinjaVece also said he started this thread with a much earlier version of the firmware (1.17) and Tesla has changed this behavior multiple times since so even fairly recent testing could already be out of date as 1.34.3 came out here sometime in late March or early April.

The PWs don't know if the sun is up or not. The decision to jump to greater than 60Hz is solely based upon the battery state of charge. So whether it's day or night, if your battery levels are north of 92% or so, your frequency is going to be raised in order to prevent any inverters form coming online.

The E-mail I received from Tesla regarding this phenomenon indicated that there is no hard and fixed battery level that will trigger the increased frequency. But it is generally somewhere between 85 and 95%.

So I mostly agree however hasn't it been shown by other folks on this thread that the Powerwall will ramp up the frequency over time if the PV is still producing and then ramp it down to let the inverters come back once the battery has enough room? That would imply that it has the capability of monitoring the input and be smarter about how to adjust the frequency so that the panels don't come straight on as the sun rises or whatever. Naively it could be that it simply does a time based (X hz for N seconds) or Powerwall SOC based frequency adjustments. Regardless, since my inverters are UL1741 certified 60.5hz should be just enough to keep the inverters offline.

@NinjaVece was the 60hz +/- 3hz UPS consideration the only reason you were thinking of 62.9hz as your "steady state" for turning off the PV? Seems to me that for UL1741 certified inverters 60.5hz should be just fine.
 
Appreciate the help, and yes of course I have read the thread. Other than specific UPS suggestions there aren't any actual workarounds to the issue and I haven't heard back from Tesla yet myself. I could have been more clear that my specific question was about the night time situation where the PV is not producing. vlad_kalish had a night outage back in January, but it wasn't a test and he also did not have a way to check the frequency. NinjaVece 's reply shortly after makes complete sense but didn't address the night question directly. NinjaVece also said he started this thread with a much earlier version of the firmware (1.17) and Tesla has changed this behavior multiple times since so even fairly recent testing could already be out of date as 1.34.3 came out here sometime in late March or early April.



So I mostly agree however hasn't it been shown by other folks on this thread that the Powerwall will ramp up the frequency over time if the PV is still producing and then ramp it down to let the inverters come back once the battery has enough room? That would imply that it has the capability of monitoring the input and be smarter about how to adjust the frequency so that the panels don't come straight on as the sun rises or whatever. Naively it could be that it simply does a time based (X hz for N seconds) or Powerwall SOC based frequency adjustments. Regardless, since my inverters are UL1741 certified 60.5hz should be just enough to keep the inverters offline.

@NinjaVece was the 60hz +/- 3hz UPS consideration the only reason you were thinking of 62.9hz as your "steady state" for turning off the PV? Seems to me that for UL1741 certified inverters 60.5hz should be just fine.

gpez, the PW2 does NOT increase the Hertz level at night if the grid goes off-line. The ONLY reason it would increase the Hertz level is if the house circuit is islanded and the solar panels are still attempting to inject power into the system. To prevent over-voltage when the SoC of the battery is at 100%, the firmware will raise the Hertz level to deactivate the solar inverters. At night there is no reason for the PW2 to raise the Hertz level if the house circuit is islanded as there are no solar panels looking to input voltage into the system.

SSonentag is incorrect. The Hertz raise is in response to solar inverters attempting to inject voltage into the system. No voltage, no reason for the PW2 to raise the Hertz level.

To answer your question, most UPSs only work with power +/- 3 Hertz from 60 Hertz. APC UPSs are an example and I have a few of them. When the Hertz level goes above 62.9 Hertz, the APC UPSs will not see the PW2 power as 'good' so will run off their batteries until the Hertz level drops. Other UPSs like Eaton and MinuteMan can handle 70 and 66 Hz respectively as 'good' power and will pass it through to devices behind them. I have an Eaton for my critical loads as it will work with PW2 power all the way up to 70 Hz. Whether all the devices like the 70 Hz power is another issue...
 
The PWs don't know if the sun is up or not. The decision to jump to greater than 60Hz is solely based upon the battery state of charge. So whether it's day or night, if your battery levels are north of 92% or so, your frequency is going to be raised in order to prevent any inverters form coming online.

The E-mail I received from Tesla regarding this phenomenon indicated that there is no hard and fixed battery level that will trigger the increased frequency. But it is generally somewhere between 85 and 95%.

SSonentag, the PW2 firmware only raises the Hertz level if the circuit is islanded because the grid circuit goes dead, the battery is at 100% SoC AND the solar inverters are attempting to bring in power from the solar panels that can over voltage the PW2 due to the high SoC and no outlet to the grid. You are correct in that it doesn't care if it is day or night but what it DOES care is whether the inverters are attempting to inject power into the system. No attempt at power injection, no reason to raise the Hertz level.

So it isn't about whether your PW2 is 92% or greater, it's about whether your SoC is so high (generally 97-100% SoC) AND the inverters are pushing power into the system. If Tesla support is telling you it is between 85-95%, the person that sent the email doesn't know what they are talking about. The cycle from 66 Hz to 60 Hz within an islanded circuit is between 100% to 97% and back again - not the percentage they are stating. I've tested this MANY times. Again, the rise to 66 Hz is based on a confluence of factors: 1) islanded circuit where solar power has no outlet to the grid, 2) a battery SoC of 97-100% charge and 3) solar inverters looking to continue to inject voltage to the house circuit.

You can even add 4) where the solar input is GREATER than what the house is actually using at the moment. If all THREE were true but the solar input is LESS than what the house is actually using, there is no reason for the PW2 to raise the Hertz level. The power has an outlet and thus can be used.
 
You can even add 4) where the solar input is GREATER than what the house is actually using at the moment. If all THREE were true but the solar input is LESS than what the house is actually using, there is no reason for the PW2 to raise the Hertz level. The power has an outlet and thus can be used.

So...if I find myself in this situation, one workaround could be to plug in my Model S and start charging it to soak up the power? (Let’s assume the car can draw more than the solar panels put out.)

Bruce.
 
So...if I find myself in this situation, one workaround could be to plug in my Model S and start charging it to soak up the power? (Let’s assume the car can draw more than the solar panels put out.)

Bruce.

Bruce, works for me. As long as your charging breaker is under PW2 control, it should work. But remember to limit the charge to 20-30 minutes as it will suck down the wattage in the PW2 unless you have some intense solar output.
 
So...if I find myself in this situation, one workaround could be to plug in my Model S and start charging it to soak up the power? (Let’s assume the car can draw more than the solar panels put out.)

Bruce.

Back when I was first testing the PW2 and my solar output was more than it could accept, I got in my Model S and was able to vary the charging amperage to both charge the car and the PW2 at an acceptable rate, using a combination of the app to monitor the PW2 charging and the screen in the car. It certainly occurred to me then that this could all be automated either via the app or the car. I would love to try this again, but my PW2 has not been working for a couple of months now. Sigh.
 
gpez, the PW2 does NOT increase the Hertz level at night if the grid goes off-line. The ONLY reason it would increase the Hertz level is if the house circuit is islanded and the solar panels are still attempting to inject power into the system. To prevent over-voltage when the SoC of the battery is at 100%, the firmware will raise the Hertz level to deactivate the solar inverters. At night there is no reason for the PW2 to raise the Hertz level if the house circuit is islanded as there are no solar panels looking to input voltage into the system..

So this was NOT my experience during my test this week on 1.34.3. I turned off my grid connection at 1am and immediately after the change over to the Powerwall the frequency spiked. My UPS showed 63hz though that was "output" so I'm not sure I can trust it. Assuredly the UPS was not happy with the wall power and my LED lights flickered until the SoC dropped enough. I am going to test again once I hunt down an oscilloscope or at least something that can measure frequency.

As an aside when I did that 1am test a bunch of my AFCI breakers and GFCI outlets tripped...
 
  • Like
Reactions: abasile
Your are correct, it appears to be solely based on SOC and not whether solar is on or not. I did a test at 9pm with 100% SOC and the same thing happened until the SOC dropped to ~97%. I don't think the PW looks to see if solar is generating or not, because even if it weren't at the moment the grid failed, it could come on-line and they want to keep it off-line.

Just wish they would limit it to 63Hz or let us set a value that works for our inverters instead of this one size fits all approach.
 
I am going to test again once I hunt down an oscilloscope or at least something that can measure frequency.
You can see the PW output frequency from the JSON output of your PW. Add "/api/meters/aggregates" to your web UI URL and look for battery.frequency.

Here is the output from my PW. I have a JSON parser extension installed on my Chrome browser.
  • battery:
    {
    • last_communication_time: "2019-05-02T18:01:36.690200438-04:00",
    • instant_power: 1890.0000000000002,
    • instant_reactive_power: -20,
    • instant_apparent_power: 1890.1058171435802,
    • frequency: 60.007999999999996,
    • energy_exported: 4312920,
    • energy_imported: 4867430,
    • instant_average_voltage: 246.45,
    • instant_total_current: -48.5,
    • i_a_current: 0,
    • i_b_current: 0,
    • i_c_current: 0,
    • timeout: 1500000000
    },
 
For anyone wanting an easy way to show their line frequency, a Kill-A-Watt device is a super simple and inexpensive solution.

https://smile.amazon.com/P3-P4400-Electricity-Usage-Monitor/dp/B00009MDBU/ref=sr_1_5

61FOAk4UA5L._SL1500_.jpg
 
  • Helpful
Reactions: gpez and bmah
So this was NOT my experience during my test this week on 1.34.3. I turned off my grid connection at 1am and immediately after the change over to the Powerwall the frequency spiked. My UPS showed 63hz though that was "output" so I'm not sure I can trust it. Assuredly the UPS was not happy with the wall power and my LED lights flickered until the SoC dropped enough. I am going to test again once I hunt down an oscilloscope or at least something that can measure frequency.

As an aside when I did that 1am test a bunch of my AFCI breakers and GFCI outlets tripped...

I believe he meant "should not" and not "does not" but in any case.... all night time outages I've had say you are correct. It operates the same way whether or not your solar inverter is receiving power from the panels and instead it "should only" be raising the Hz if the solar inverter is actually producing power. My experiences have been the same as yours. At night when the power went out frequency still went to 66Hz until the PW dropped into the mid 90%s and then it went back to 60hz and the solar panels weren't producing any power at all.

If the PW isn't supposed to work that way then it isn't working properly. I am using Eaton UPS units though so everything seems to stay online currently.
 
You can see the PW output frequency from the JSON output of your PW. Add "/api/meters/aggregates" to your web UI URL and look for battery.frequency.

Here is the output from my PW. I have a JSON parser extension installed on my Chrome browser.
  • battery:
    {
    • last_communication_time: "2019-05-02T18:01:36.690200438-04:00",
    • instant_power: 1890.0000000000002,
    • instant_reactive_power: -20,
    • instant_apparent_power: 1890.1058171435802,
    • frequency: 60.007999999999996,
    • energy_exported: 4312920,
    • energy_imported: 4867430,
    • instant_average_voltage: 246.45,
    • instant_total_current: -48.5,
    • i_a_current: 0,
    • i_b_current: 0,
    • i_c_current: 0,
    • timeout: 1500000000
    },

+1, this is super helpful.

I'll be able to do this once Tesla tells me how to fix the wifi antenna so that my TEG can connect on something other than 3G...