Solar Panel system making high pitched noise

[sidraja999]

Try reducing panel voltage by removing few panels from the circuit. If that does not help then bypass all tigos and connect panels direct to inverter pv input.

Growatt tech support confirmed that I am under the operating voltage max of 550v. With 9 panels, it is about 400v. The inverter throws an error code if the voltage is too high.

 

[sidraja999]

I would still double and triple check all connections.
I can't count how many times I found a wrong connection on the third round of checks.

I will do that tomorrow.

 

[zanydroid]

The sound appears near the panels and in the DC disconnect box. What could this mean?

Disconnect

• Bad connection (b/c the 1kHz or 20kHz, whatever it is, could well be going down the whole wire, it will not be limited by distance)
• Some component is resonating (wires can't do that but maybe there are things in the SPDs that can do it)
• (unlikely) sound conduction from the distant array, this will fall off by distance

Panels:

• plenty of active components and things that can resonate, inside the optimizers.
• the solar panels are just wires and cells, so probably less likely than the optimizers

All that said, I think divide and conquer , check all junctions, is way more productive than trying to think of root causes (root cause = shot in the dark given the limited amount of data available, divide and conquer / other diagnosis steps will get more data). I mean I know there's not much that can be done at this hour, sitting in front of a keyboard. But even still writing down the plan for how to divide and conquer diagnosis this is more productive than randomly throwing out root causes.

 

[timselectric]

Since it's a ground mount array. It should be easy to access the wiring. I would start with bypassing each panel/optimizer, one at a time. Then, the SPD's.

 

[zanydroid]

Growatt tech support confirmed that I am under the operating voltage max of 550v. With 9 panels, it is about 400v. The inverter throws an error code if the voltage is too high.

I think the idea here is, "Trust absolutely nothing, something wack be happening".

In my engineering operations day job I remind folks to do this all the time.

 

[sidraja999]

Disconnect

• Bad connection (b/c the 1kHz or 20kHz, whatever it is, could well be going down the whole wire, it will not be limited by distance)
• Some component is resonating (wires can't do that but maybe there are things in the SPDs that can do it)
• (unlikely) sound conduction from the distant array, this will fall off by distance

Panels:

• plenty of active components and things that can resonate, inside the optimizers.
• the solar panels are just wires and cells, so probably less likely than the optimizers

All that said, I think divide and conquer , check all junctions, is way more productive than trying to think of root causes (root cause = shot in the dark given the limited amount of data available, divide and conquer / other diagnosis steps will get more data). I mean I know there's not much that can be done at this hour, sitting in front of a keyboard. But even still writing down the plan for how to divide and conquer diagnosis this is more productive than randomly throwing out root causes.

Thanks for the details! I will check the wiring tomorrow.

Could something as simple as a zip tie that is too tight be causing the issue?

 

[timselectric]

Could something as simple as a zip tie that is too tight be causing the issue?

Not unless it's pulling on a connection.

 

[sidraja999]

FYI....attached is a screen shot from the Tigo app at 1:15 PM today when panel production was high.
I thought I would see some issues here if the optimizers were failing.

Ignore the black and gray icons. These are Tigos on panels that are not in the same string. Still setting those up.

 

[Hedges]

That's high panel production? One 129W, one 214W? Are they different orientations, or shaded?

 

[zanydroid]

That's high panel production? One 129W, one 214W? Are they different orientations, or shaded?

Based on how the string was described earlier, I wouldn't expect such a big difference at mid-day.

(Also, that I think brings up the idea that, if we think the panels should be outputing uniform amounts of power, outliers in this wattage or voltage monitoring at panel level might point at which optimizer(s) are being weird. Is there voltage monitoring per panel?)

 

[sidraja999]

That's high panel production? One 129W, one 214W? Are they different orientations, or shaded?

The 129W panel is shaded by a tree. The others are not at the time of the day. These are 405W panels. I am assuming due to it being January the sun is not right above. Also, the panels are ground mount with a 10 degree tilt.

 

[Hedges]

Is 131W shaded? And there is 236W.

10 degree tilt in winter would explain reduced output. All same orientation, so only shade is expected to make them differ by more than tolerance, about 5%.

 

[sidraja999]

Is 131W shaded? And there is 236W.

10 degree tilt in winter would explain reduced output. All same orientation, so only shade is expected to make them differ by more than tolerance, about 5%.

131W might have been. A cloud at that time. Not sure. All facing south.

Attached is another screen shot at 2:15 pm.

 

[zanydroid]

Assuming full cut panels.

If the panel wattage are close or at 2/3, 1/3 ratios the optimizers don’t have to do that much.

In between those is when they do current matching

 

[zanydroid]

What is the runbook to try for tomorrow?

 

[sidraja999]

Solar system is generating power today. Sunny day.
No error codes.
Noise is still there.
Before I check the wiring, I am in the inverter app to see if there are tools to monitor the system.

If there was a loose wire connection, the AFCI should have triggered.

 

[timselectric]

If there was a loose wire connection, the AFCI should have triggered.

It should, on the PV circuit.
Maybe not, on the panel to optimizer connections. (Depending on how isolated they are)

 

[sidraja999]

The noise resonates even through the EMT conduit which is 50 feet away from the panels. The noise is not present now since it is almost sunset and energy production has reduced. Only during peak sunlight does it occur.

I think this noise is coming from the inverter and being sent through the wiring. I have reached out to the Growatt tech support and sent them the recordings. They should respond tomorrow.

 

[AntronX]

You need to isolate this problem to either Growatt or Tigo. Try connecting another array to that MPPT input.

 

[JRH]

I can't access your recordings; you need to make them public.

Anyway, it's possible you are hearing the switching frequency of the optimisers, either the cores or even the semiconductors. This would tie in with the noise being more prevalent when the system is working hard.

Many young(er) people can hear these supposedly "out of hearing range" frequencies, it's how the ultrasonic "teenager repellers" work.

Do you have a dog, he may be going crazy!

Teenage repellers….. haaaaaaaaaaaaaa ?