Offgrid Powerwall at 59.6 Hz (no solar curtailment)

[oholladay]

Is anyone familiar with why Powerwalls would not operate at ~60Hz when off grid absent the need to curtail solar production?

Now that we have a "Go Off Grid" button, I am able to more easily observe how my Powerwalls behaves in "Off Grid" mode. One thing I have noticed is that, when I am "off grid," the Powerwalls consistently operate at 59.6Hz (bouncing rather erratically between 59.55 and 59.65). I get this data from the Powerwall API and have confirmed it with a multimeter. Obviously, when I am connected to the grid, they operate very close to 60Hz. I am able to compare this to the real time frequency data posted by ERCOT and it is very close.

I know that when the Powerwalls get close to full, they will raise the frequency to curtail solar production. However, my Powerwalls are doing this at modest levels of charge (60% to 80%) and at all levels of solar production.

 

[gpez]

Is anyone familiar with why Powerwalls would not operate at ~60Hz when off grid absent the need to curtail solar production?

Now that we have a "Go Off Grid" button, I am able to more easily observe how my Powerwalls behaves in "Off Grid" mode. One thing I have noticed is that, when I am "off grid," the Powerwalls consistently operate at 59.6Hz (bouncing rather erratically between 59.55 and 59.65). I get this data from the Powerwall API and have confirmed it with a multimeter. Obviously, when I am connected to the grid, they operate very close to 60Hz. I am able to compare this to the real time frequency data posted by ERCOT and it is very close.

I know that when the Powerwalls get close to full, they will raise the frequency to curtail solar production. However, my Powerwalls are doing this at modest levels of charge (60% to 80%) and at all levels of solar production.

59.6hz is pretty close to standard grid operating range of 59.7-60.3hz and still within the IEEE 1547 normal operating range of 59.5-60.5hz. Even though it's low it shouldn't be causing you any problems (or curtailment).

Can you log in to your gateway and see what https://teg/api/site_info returns? (where "teg" is the IP of your gateway). That may shed some light on how your Powerwall is configured.

 

[slcasner]

59.6hz is pretty close to standard grid operating range of 59.7-60.3hz and still within the IEEE 1547 normal operating range of 59.5-60.5hz. Even though it's low it shouldn't be causing you any problems (or curtailment).

Except for any AC-synchronized clocks running slowly (losing 9.6 minutes per day if at 59.6Hz all 24 hours).

 

[oholladay]

59.6hz is pretty close to standard grid operating range of 59.7-60.3hz and still within the IEEE 1547 normal operating range of 59.5-60.5hz. Even though it's low it shouldn't be causing you any problems (or curtailment).

Can you log in to your gateway and see what https://teg/api/site_info returns? (where "teg" is the IP of your gateway). That may shed some light on how your Powerwall is configured.

I agree - I don't think it's causing me any issues (I don't have any AC-synchronized clocks, but the point from slcasner is noted). I checked the site you mentioned and it indicates the following, which looks right to me:

"grid_code":"60Hz_240V_s_IEEE1547a_2014","grid_voltage_setting":240,"grid_freq_setting":60

I guess I am just curious as to why the Powerwalls seem to be targeting 59.6Hz rather than 60Hz and was wondering if other people's Powerwalls exhibit similar behavior in Off Grid mode.

 

[wjgjr]

I agree - I don't think it's causing me any issues (I don't have any AC-synchronized clocks, but the point from slcasner is noted). I checked the site you mentioned and it indicates the following, which looks right to me:

"grid_code":"60Hz_240V_s_IEEE1547a_2014","grid_voltage_setting":240,"grid_freq_setting":60

I guess I am just curious as to why the Powerwalls seem to be targeting 59.6Hz rather than 60Hz and was wondering if other people's Powerwalls exhibit similar behavior in Off Grid mode.

Yes - I have seen mine sitting in the 59.7 Hz range off-grid. How full are your PWs when you are seeing this?

 

[mountfield]

Powerwall+ here - when running battery only and no solar production, I see 59.6. As solar ramps up this creeps up to and above 60hz when solar needs to be curtailed.

from some investigation, it appears that being slightly offset from 60hz helps with detecting and resolving islanding conditions, and is a way to meet the requirements of UL 1741 for anti - islanding protection

 

[gpez]

Powerwall+ here - when running battery only and no solar production, I see 59.6. As solar ramps up this creeps up to and above 60hz when solar needs to be curtailed.

from some investigation, it appears that being slightly offset from 60hz helps with detecting and resolving islanding conditions, and is a way to meet the requirements of UL 1741 for anti - islanding protection

I don't follow. IEEE 1547 (and by proxy UL 1741) have the ride-through range as 59.5hz - 60.5hz so 59.6hz is no different than 60.0hz or 60.4hz from a compliance/standards perspective. Curtailment should only happen above 60.0hz regardless of configuration, and I'd guess most would start at 60.3hz or similar depending if you're on grid or off grid (off grid could be much more lenient).

 

[sorka]

PG&E around where is live is 60.0Hz all the time except for rare events when they might raise it to kick solar off the grid area wide but I've never experienced that since I've had solar.

My Powerwalls also run about 59.6 when off grid. Its not a huge issue but does increase the retries for my devices that communicate via powerline line...which numbers over 200 devices(insteon). It would be worse if it wasn't for the fact that most of my insteon is now dual band (powerline and wireless).

 

[oholladay]

Yes - I have seen mine sitting in the 59.7 Hz range off-grid. How full are your PWs when you are seeing this?

The times I've looked at it, they have ranged from ~50 to the low 90s. I reconnect to the grid when they get to around 90% to avoid curtailment (I'd rather push to the grid than have curtailment).

 

[oholladay]

PG&E around where is live is 60.0Hz all the time except for rare events when they might raise it to kick solar off the grid area wide but I've never experienced that since I've had solar.

My Powerwalls also run about 59.6 when off grid. Its not a huge issue but does increase the retries for my devices that communicate via powerline line...which numbers over 200 devices(insteon). It would be worse if it wasn't for the fact that most of my insteon is now dual band (powerline and wireless).

Glad to see that my Powerwalls are not out of line with what others (you, mountfield, and wjgjr) have observed. I suppose that means this behavior is intentional - I just don't understand why. Interesting point on the interaction with your Insteon devices. I wonder if there could be some negative long-term effects on AC motors (like air conditioning compressors).

 

[mountfield]

I don't follow. IEEE 1547 (and by proxy UL 1741) have the ride-through range as 59.5hz - 60.5hz so 59.6hz is no different than 60.0hz or 60.4hz from a compliance/standards perspective. Curtailment should only happen above 60.0hz regardless of configuration, and I'd guess most would start at 60.3hz or similar depending if you're on grid or off grid (off grid could be much more lenient).

Not an expert by any means, so I may have this wrong. But what triggered my line of thought was a webinar that highlighted frequency variation to ‘push’ the frequency. When grid attached this would not succeed as the grid is too strong. When islanded this will modify the observed frequency so can be used to detect this condition.
Now as for why this would be required when manually off grid - I have no idea, but it could simply be simplification of logic.

 

[mountfield]

Just to add some more data to the picture. Here’s a graph of frequency - where it was at 60 I was on grid - and when off grid - above 60 is where solar was being curtailed due to full powerwall. As you can see it ranged from about 59.6 to 60.5

 

[oholladay]

Just to add some more data to the picture. Here’s a graph of frequency - where it was at 60 I was on grid - and when off grid - above 60 is where solar was being curtailed due to full powerwall. As you can see it ranged from about 59.6 to 60.5 View attachment 706520

I like this graph. What software did you use to record and graph this data?

 

[mountfield]

I like this graph. What software did you use to record and graph this data?

This is a grafana dashboard I setup. For data collection from powerwall, I started with liveaverage ( GitHub - liveaverage/docker-powerwall-dashboard: Grafana dashboard showing trend & historical data polled from Tesla Powerwall 2 API served up in an easy to use docker image )

And then made a number of modifications to integrate with my existing influx / Grafana setup and to add additonal data points (including the frequency data)

 

[oholladay]

This is a grafana dashboard I setup. For data collection from powerwall, I started with liveaverage ( GitHub - liveaverage/docker-powerwall-dashboard: Grafana dashboard showing trend & historical data polled from Tesla Powerwall 2 API served up in an easy to use docker image )

And then made a number of modifications to integrate with my existing influx / Grafana setup and to add additonal data points (including the frequency data)

Sounds like a lot of work, especially for someone like me who doesn't have experience working with APIs (other than just using a browser to look at the long string of text it spits out) or any of those tools.

Any idea why the numbers fluctuate so much when it is off grid? It seems that, at least when the Powerwall is synchronized with the grid, it is perfectly capable of putting out a fairly clean 60Hz.

 

[mountfield]

Sounds like a lot of work, especially for someone like me who doesn't have experience working with APIs (other than just using a browser to look at the long string of text it spits out) or any of those tools.

Any idea why the numbers fluctuate so much when it is off grid? It seems that, at least when the Powerwall is synchronized with the grid, it is perfectly capable of putting out a fairly clean 60Hz.

If I have time I may do a write up - it’s not as hard as it seems but does require you have some experience with the tools currently.
Regarding the frequency - it’s correlated with solar production. When it’s calling for more solar it’s below 60hz and when it’s calling for less production it’s above 60hz with a gradient between the extremes

 

[oholladay]

Except for any AC-synchronized clocks running slowly (losing 9.6 minutes per day if at 59.6Hz all 24 hours).

I didn't realize how relevant this point was to me until I noticed that my coffee maker clock had lost nearly 10 minutes over the past week and was way behind.

Initially, I thought I was immune to this problem because I was under the impression that this was only an issue affecting older motorized clocks that run on AC power, of which I do not have any.

When I saw the digital clock on the coffee maker lose so much time, I did a little research and found this article, which indicates that a common chip used in digital clocks is actually tied to the mains frequency. I don't know what is in my coffee maker, but the correlation between it losing time (which I've never noticed before) and my testing of "off grid" mode suggest that it is indeed tied to the mains frequency.

 

[h2ofun]

I didn't realize how relevant this point was to me until I noticed that my coffee maker clock had lost nearly 10 minutes over the past week and was way behind.

Initially, I thought I was immune to this problem because I was under the impression that this was only an issue affecting older motorized clocks that run on AC power, of which I do not have any.

When I saw the digital clock on the coffee maker lose so much time, I did a little research and found this article, which indicates that a common chip used in digital clocks is actually tied to the mains frequency. I don't know what is in my coffee maker, but the correlation between it losing time (which I've never noticed before) and my testing of "off grid" mode suggest that it is indeed tied to the mains frequency.

My wind up clock losing time quickly

 

[oholladay]

If I have time I may do a write up - it’s not as hard as it seems but does require you have some experience with the tools currently.
Regarding the frequency - it’s correlated with solar production. When it’s calling for more solar it’s below 60hz and when it’s calling for less production it’s above 60hz with a gradient between the extremes

From a more "macro" perspective, that makes sense on the frequency. It's just that the lines are so much more "jagged" when off grid vs when on grid, even when the "trend" is relatively flat.

 

[slcasner]

I didn't realize how relevant this point was to me until I noticed that my coffee maker clock had lost nearly 10 minutes over the past week and was way behind.

Initially, I thought I was immune to this problem because I was under the impression that this was only an issue affecting older motorized clocks that run on AC power, of which I do not have any.

When I saw the digital clock on the coffee maker lose so much time, I did a little research and found this article, which indicates that a common chip used in digital clocks is actually tied to the mains frequency. I don't know what is in my coffee maker, but the correlation between it losing time (which I've never noticed before) and my testing of "off grid" mode suggest that it is indeed tied to the mains frequency.

Over the long term, the 60Hz line frequency is more stable than a crystal oscillator (which is on the order of a second or a few per day). But the grid operators have proposed being released from the requirement to maintain a stable 60Hz because that would ease their job of managing grid load.