diy solar

diy solar

AC Coupling Design

We have. There are settings inside California 21 that are adjustable within the Mode Q(U)+Q(P) settings
My assumption was that the SolarEdge was able to modulate fairly well and that if you see cycling it is a result of the AC coupling algorithm in the Sunny Islands. I think Hedges offered some insight but since his Grid Dependant Inverters were older he just saw the same on off cycles that you are seeing. Since your SolarEdge inverters are Rule 21 and UL1741SA I still believe the issue is the Sunny Island algorithm, but putting an EV charging load should alleviate that.
 
Sunny Island will run frequency all the way up to 65 Hz if necessary to curtail output.

Rule 21 or SA seem clear regarding ride-through indefinitely at full power, and stay on-line (I think at full power) for 299 seconds mandatory, disconnect by 300 seconds mandatory.

Near Nominal is now 58.5 to 60.5 Hz. I was able to fix my problem by setting low frequency of Sunny Boy to 58.9 Hz instead of 59.3 Hz.


What I haven't seen clearly defined is frequency-watts.
Could be your inverter ramps down to zero watts right at 60.5 Hz, and it implements that mandatory cutoff above 60.5 Hz after 300 seconds.
I think, if frequency-watts has ramped down to zero output, it shouldn't be required to disconnect ever, for the moderately off-nominal frequency; just wait for frequency to come back within near-nominal.

I suggest reducing the frequency where your inverter reaches zero output, so there is some margin between that and 60.5 Hz, to allow for error/disagreement, also response time.

Here's something I just found, haven't studied:

 
Sunny Island will run frequency all the way up to 65 Hz if necessary to curtail output.
I understand that. What my theory is all about is the algorithm in the Sunny Island that decided when to start the ramp up in frequency and how fast that ramp was based on some measurement of battery capacity. Furthermore why does the Sunny Island tell the SolarEdge grid dependant inverter to ramp up to full bore then shut it down soon. If the batteries are near full why does it not start the SolarEdge at a modulated value? If we had frequency logging on this issue it would answer a lot of questions. If the OP could find a consistent load to put on the system as the batteries reach capacity that would give us some data as well..

Somewhere I recall a statement that someone said the SMA inverters did that to keep the 60 Hz average so the clocks would be on time. Is that true and is there a way to disable that?
 
Last edited:
Yes, average of 60 Hz is by default. ”250.11 AfraEna” parameter.
Disabling that is recommended by SMA if you are using standard UL-1471 inverters (without frequency-watts or Rule-21), so they don't get knocked off by low frequency.

The frequency ramp rate of Sunny Island is slow, maybe to allow motors to keep up and inverters to not get knocked offline.

Supposedly Sunny Boy for offgrid has parameters set including:

"dFac‑Max max. rate of change 4 Hz/s"

Sunny Island gradually ramps frequency up and down, not as fast as loads that switch on/off of course. It makes up the difference by charging/discharging battery. It would track gradually changing loads, like a compressor working harder as pressure builds.

I wouldn't be surprised if lithium batteries suddenly shooting up in voltage could be a problem. But that would show up as battery disconnecting.
If Sunny Island asks Solar Edge for full output, that should be because either loads and/or battery charging need the current.

My Sunny Boys will display why they disconnected, e.g. frequency error at 59 Hz.
If Solar Edge doesn't display, just watching with a meter could show.
 
Back
Top