Deye inverter harmonics/UPS buzzing/humming

[madmes]

Hello,
I have 3 deye SUN-12KW-SG04-LP3-EU inverters running in parallel. During night time, the inverters only power a few things (mostly non linear loads) that never exceed 2kw. i have been noticing that at night time, the computers' UPSs are humming/vibrating a lot.

I do not have an oscilloscope at hands, so i used my phone to record the noise using Spectroid. It seems the inverters are pushing all fundamental harmonics.
(50Hz, 100Hz, 150Hz, 200Hz, 250Hz, etc...) so they are outputting everything but a sinusoidal wave. might as well be a square wave considering all the harmonics.
What could be the reason for this? is there a way to mitigate it?

I will try to isolate the inverters tomorrow and run each load separately. hoping its something stupid and not a fried/damaged inverter

Image #1: offgrid mode during night time, ~1.2kw load consisting mainly of ac-dc items (non linear).

 

[RCinFLA]

The PWM H-bridge L-C low pass filter is a compromise between low AC load current and high AC load current demand. The high frequency EMI output filter will vary based on inverter AC load.

When you have a sudden AC surge current on AC output, 50/60Hz low frequency current can drive filter inductor toroid close to saturation of core causing a drop in its effective inductance reducing the HF filtering.

Some get 'ratty' looking at high AC loading, some have a lot of output overshoot/undershoot ringing at low AC loading.

A poor AC sinewave PWM filtering design can have a significant effect on inverter's idle current due to L-C output filter's reactive load on output PWM sinewave H-bridge.

 

[madmes]

@RCinFLA can this kind of output overshoot/undershoot damage the loads(LEDs, computers) ?
I will also test different levels of loads tomorrow and verify that the harmonics disappear as the load increases up to a certain point.

 

[timselectric]

The first thing I would do is make sure everything is properly grounded. Ground loops can cause noise.
Half way grounded, can cause noise.
Your plan to isolating the loads and inverters will help narrow it down.

 

[Nobodybusiness]

The first thing I would do is make sure everything is properly grounded. Ground loops can cause noise.
Half way grounded, can cause noise.
Your plan to isolating the loads and inverters will help narrow it down.

Speaking of isolation.
I’m going to ask a stupid question.

Do they make Isolation transformer big enough for say 15kw continuous usage?

Is it practical or even worthwhile to have one in between your inverters and house loads??

 

[timselectric]

Do they make Isolation transformer big enough for say 15kw continuous usage?

You can get any size transformer. I'm going to have a 37.5kva isolation transformer for my main system.

Is it practical or even worthwhile to have one in between your inverters and house loads??

It's the safe way to get split-phase from a 240v single phase inverter. (If you have grid input to your inverter)

 

[Nobodybusiness]

You can get any size transformer. I'm going to have a 37.5kva isolation transformer for my main system.


It's the safe way to get split-phase from a 240v single phase inverter. (If you have grid input to your inverter)

Already have split phase inverters.

Was just thinking of eliminating noise and potential sags and what not.

If you already have split phase inverters would it help or be beneficial in any way to have an isolation transformer between them and house loads?

 

[timselectric]

Already have split phase inverters.

Was just thinking of eliminating noise and potential sags and what not.

If you already have split phase inverters would it help or be beneficial in any way to have an isolation transformer between them and house loads?

Doubtful it would help.

 

[Nobodybusiness]

Doubtful it would help.

That’s what I needed to know.
Thanks!

 

[madmes]

Update: grid power came(slang used in my country as blackouts are standard, with the rare 1hour/day of grid power).
The majority of the noise created by the UPSs hooked onto the inverters is gone in ongrid mode.

Image #2: offgrid at later night. All harmonics appear to be lower. Confirming suspicions that i need to check all loads individually.


Image #3: ongrid at early morning (electricity came). Low frequency harmonics non existant compared to image1 and 2, higher order harmonics still present (but lower) as the whole village is charging hybrid inverter batteries.

 

[madmes]

The first thing I would do is make sure everything is properly grounded. Ground loops can cause noise.
Half way grounded, can cause noise.
Your plan to isolating the loads and inverters will help narrow it down.

I might have that. Will check all wirings.

Meanwhile i had 10kw linear load(heaters, induction motors) and the issue remained so i restarted the system and the problem is fixed for now.
Seems like either the firmware has bugs, or switching a specific motor on or off is creating this mess.

Image #3: after system restart. Issue fixed temporarily, it will happen again soon.


Note: i mixed up between 2 UPSs, hence why one vibrates at 50hz and the other at 100hz, however 150hz and above are now negligeable.

 

[madmes]

Update: the issue happens whenever the inverters switch from ongrid to offgrid mode.

 

[timselectric]

That leads me back to a grounding issue.
Or lack of N/G bonding, in battery mode.

 

[madmes]

That leads me back to a grounding issue.
Or lack of N/G bonding, in battery mode.

The grid neutral, load neutral and gen neutral are connected together at all times.
The grid N and gen N have always been connected.
The difference now is we have installed grounding(didn't have one before) but it doesn't reach the generator yet. There is usually 5-10V difference between N and PE at all times.

 

[madmes]

Update: tried all sorts of combinations amongst the bonds, even connected grid N to my PE, nothing changed.
Then thought one of the phases is damaged, so i tried the ups on all 3 phases. Whichever had some load led to lower noise up to a point where the noise goes away entirely.
I think its a regulation issue in the inverters whenever they go from high load(like 20kw) to low load(2kw), such as the increment size they use in voltage or current output isn't dropping when the load drops.
I will try to contact deye and inform their engineers of this issue. Hopefully it can be fixed in an update. Meanwhile i am looking for an oscilloscope i can borrow to check the sin wave when the noise happens.

 

[timselectric]

Update: tried all sorts of combinations amongst the bonds, even connected grid N to my PE, nothing changed.
Then thought one of the phases is damaged, so i tried the ups on all 3 phases. Whichever had some load led to lower noise up to a point where the noise goes away entirely.
I think its a regulation issue in the inverters whenever they go from high load(like 20kw) to low load(2kw), such as the increment size they use in voltage or current output isn't dropping when the load drops.
I will try to contact deye and inform their engineers of this issue. Hopefully it can be fixed in an update. Meanwhile i am looking for an oscilloscope i can borrow to check the sin wave when the noise happens.

If all neutrals are connected together. And everything (Inverter , enclosures, solar panels, any metal raceways or boxes) is bonded correctly. And you have only one N/G bond. There wouldn't be any ground loops. And there shouldn't be any voltage between neutral and ground anywhere at any time.
Can't be sure if this is the cause of the noise. But can't rule it out, unless it's correct.

 

[madmes]

If all neutrals are connected together. And everything (Inverter , enclosures, solar panels, any metal raceways or boxes) is bonded correctly. And you have only one N/G bond. There wouldn't be any ground loops. And there shouldn't be any voltage between neutral and ground anywhere at any time.
Can't be sure if this is the cause of the noise. But can't rule it out, unless it's correct.

My grid operator doesn't supply a ground wire(they don't even supply power beyond an hour per day for the past 2 years) . I have TT grounding for that and its not bonded to earth, although i tried bonding them today as well after disconnecting the grid and the issue persisted.
The issue gets better or worst depending on the load size and type of load regardless of the phase. I think its a software issue but i will keep investigating the wiring.

 

[timselectric]

My grid operator doesn't supply a ground wire(they don't even supply power beyond an hour per day for the past 2 years) . I have TT grounding for that and its not bonded to earth, although i tried bonding them today as well after disconnecting the grid and the issue persisted.
The issue gets better or worst depending on the load size and type of load regardless of the phase. I think its a software issue but i will keep investigating the wiring.

When I refer to grounding and bonding. It's mainly about the N/G bond. And bonding all conductive equipment to it. While connecting earth to your system is also important. Without a N/G bond, there is no grounding system.

 

[madmes]

From what i know about TT systems is that i do not bond N and PE, it is done at the transformer.
I will consult with an expert on my country's electric regulations to see what can i do

 

[madmes]

I've linked the whole issue to 1 of the inverters. It was running hotter than the rest during the cold night, taking 200w more than it is feeding, and producing a weird a high pitched noise.
I think it is damaged due to how bad the parallel system starts/stop (one inverter at a time, not all 3 simultaneously) so it ran all the load for a second before the other 2 kicked in to carry their parts.