Multiplus II 48/[5000|8000] Transformer Noise from Grid

Q&A and troubleshooting
, , , ,
1

Multiplus II 48/[5000|8000] Transformer Noise from Grid

Hello Victron community and developers,

I’m living in Ukraine, and as you know, we have a lack of electricity due to the war.
My two neighbors and I chose Victron Multiplus II 48/[5000GX|8000 Cerbo GX] inverters + DIY 280Ah batteries (16S) with JK BMS as ESS (without any PV systems yet). We live in private houses with a 3-phase (230V/400V) grid connection and use a single inverter (1 phase) only for critical loads. All three houses are connected to the same transformer substation, and most houses in that area use heaters, water pumps, water heaters, electric stoves, ovens, etc.

All systems were installed and configured similarly, and we faced the same Victron transformer noise issue. The noise is like pulsation, transformer vibration, or sinusoidal hum.

During a discussion about this issue in the Ukrainian Victron community Telegram channel, one of the installers suggested that the hum noise is generated by the Victron transformer due to high voltage/current harmonics in the grid.
We conducted a grid quality check using a Cesinel ReCon T electric power series logger device and obtained metrics/results. The auto-generated reports showed no critical grid issues, and all tests passed.

During troubleshooting, we found that:

  1. The hum noise occurs only when the grid is enabled, regardless of whether there are consumers or not.
  2. The inverter power consumption fluctuates (20–180 Watts) with the grid enabled and without consumers.
  3. There is no hum noise when we use a power generator as the “grid” input, with or without consumers.
  4. No hum noise occurs during inverter operation from the battery, with or without consumers.
  5. No hum noise was observed at another location. One of the Victron Multiplus II 48/5000 GX inverter was connected to the grid in an apartment (different transformer substation).
  6. Enabling, disabling, or changing grid codes (EN50549-1:2019; Other; None; with and without LOM), UPS function, Weak AC input, or ESS Assistant in the Victron inverter configuration had no effect.
  7. No installation or Victron hardware issues were found. These were checked multiple times by the community and professional installers.
  8. Installing a voltage stabilizer with transformer topology before the Victron inverter has no effect.
  9. The sine wave of each phase was checked with an oscilloscope and appears to have no distortions.

We still suspect this is a voltage/current harmonics issue. Is it possible to reduce or eliminate this hum noise by changing some grid parameters in the Victron configuration? Or by installing some device before the Victron inverter?

P.S. Four other neighbors have 3-phase Deye hybrid inverters (SUN-12K-SG04LP3-EU) with Non-Isolated/Transformerless topology, and they do not experience any grid hum noise issues.
P.s.s. Video files with hum noise and Cesinel ReCon T repors - Inverter_hum_noise - Google Drive

Thanks in advance

2

Perhaps you could share the results? The most likely cause of transformer noise would be a high harmonic content. The Inverter does check the quality of the Grid power to a certain extent, and will disconnect if this seems unacceptable.
I don’t think there are any settings within the inverter that can affect this noise. It might be worth using VE configure to tighten the mains acceptance voltage values, however, this might lead to nuisance disconnections of the grid.

3

Hi Mike,
The inverter doesn’t disconnect from the grid; it continues to operate.
From time to time, I get an overload error regardless of the type of load (low or high). As I understand it, the inverter (my model: Multiplus-II 48/8000) tries to “handle” this issue.

Cesinel ReCon T reports for 31-12-2024
EN_50160_compliance_report_31122024.pdf (2.8 МБ)
MEDCALScope_Power_Quality_Report.pdf (543.5 КБ)

P.S. As a new user, I am limited to uploading 2 files. All reports have been uploaded to my Google Drive.

4

next 2 reports Cesinel ReCon T for 31-12-2024
MEDCALScope_Power_Quality_Report_31122024_harmonics_histogram.pdf (360.4 КБ)
MEDCALScope_Power_Quality_Report_31122024_harmonics_time_evolution.pdf (1.3 МБ)

5

Latest Cesinel ReCon T report for 31-12-2024
NVE PQ Regulations FoL_230V_35KV_31122024.pdf (220.9 КБ)

6

Hi Just from a quick look at the first 2 reports:

  1. There is a possibility that the noise is caused by high order harmonics.
  2. There appears to be a supply impedance problem (Minor) - Power quality monitor report, comparing the voltage and current waveforms shows a ~2V dip when the current jumps to 15A.
  3. Do you know what is causing the steps in the current waveform?
    image
    image973×839 91 KB

    Does the noise from the transformer change in time to these current steps?
7

OK on the inverter overload - I sometimes see this on moderate to high loads if the battery is low, but NEVER with no load. Worth keeping an eye on.
Is your inverter connected on L1?
If so, can you try to connect to a different phase?

8

Do you know what is causing the steps in the current waveform?

It’s my 3-phase electric central heating flow boiler, Kospel EKCO.L2. However, I still hear the hum noise even when it is completely disabled. :pensive:

Does the noise from the transformer change in time to these current steps?

Yes, from time to time, it can change.

9

Yes, it is connected to phase L1.

If so, can you try connecting it to a different phase?

That was my first idea for fixing it. I checked all three phases, and the same hum noise occurs on all of them. :pensive:

10

@MikeD
What do you think? Could a filter with this schematic help with the issue?

filter
filter716×201 25.2 KB

compressed video with the sounds


11

Does the JK BMS communicate properly with the gx ?
What are your SOC settings? Is the Multi trying to charge the battery and just not getting anything in?
What is your battery monitor set to?
What AC and DC connection-cables do you use with an 8000 Multi?

12

Because of the 1) and 2) from your list, I believe it’s a “problem” with the inverter trying to keep in sync with the grid sinewave (which sometimes could be “influenced” by other possible injecting inverters) and the fact that your inverter tries to keep as close to the grid setpoint.

As you are using the systems as UPS, try to set the grid setpoint to a higher value, like 200W or more.

Also try to disable any type of feed-in. I know, you don’t do that, but keep the settings disabled.

13

Hi Steffen,

Does the JK BMS communicate properly with the gx ?

Yes.
The JK BMS communicates with Cerbo GX by RS485-> USB and charge/discharge controls by driver dbus-serialbattery v1.5.20241202.
I used the CAN protocol, but it had no effect.

What are your SOC settings? Is the Multi trying to charge the battery and just not getting anything in?

The inverter successfully charges and discharges the battery; there are no issues with that.

What AC and DC connection-cables do you use with an 8000 Multi?

The AC limit is 32A, so I use 6 mm² cable.
The DC cable is 50 mm². I know that Victron recommends 2 x 50 mm², but my daily power usage is about 1-2 kWH, with short spikes up to 6 kWH.

14

Hi Alex,
Thanks, I will test with those parameters.

image
image1521×383 32.4 KB

image
image1483×835 66.1 KB

15

Why is your battery not receiving any charging current?
it is at 55% and the charging current is 0 … is the Multi trying to charge the battery ? is the BMS closing ?
You didn’t take the battery with you to the other tests / connection points, did you ?

16

Hello Vladyslav,

i´m with Alex Pesarcu that the humming is caused by the inverter resynchronization. I think it has nothing to do with the dc-side or the bms. I´m not sure, but i also think it is no related to harmonics. That would have to be hell of a harmonic to cause a transformer to sound loud like that.

The aforementioned filter would have no effect in my opinion.

But i find it very strange that the humming occurs if you run the inverter in „dumb“ mode without grid-code and ESS.

I wonder because of the following – i have three quattros in a three-phase configruation and i know the “humming” phenomenon well :blush: – but it is far less pronounced than in your case. But when i for example shut down my cerbo gx, forcing the inverters in “dumb” pass-through mode, the humming is promptly gone and the transformers are very quiet – comparable as if in inverting mode without grid. But as soon as the cerbo is putting them back in active ESS mode, the slightly different and noticable varying humming reappears together with a slightly higher idle consumption than in passthrough or inverting mode.

Background:
Without grid-code, ESS and other fancy things like power assist etc. the MP works like a simple UPS. It passively takes what it gets and passes that trough and charges the batteries when needed.

With grid-code and ESS the MP behaves completely different and more „active“ – creating it´s own AC-waveform synchronous the the grid. And by varing phase and amplitude it controls the direction and magnitude of power-flow. That means the MP has to constantly regulate to stay in sync with the grid. In your report i saw lots of changes in voltage, espactially happening in short time frames.

For example – your MP is creating a grid-synchronous sine-wave with 220V rms – grid drops to 210V now we have 10V difference which causes huge currents – so as soon as a non-desired current flow is detected, the MP compensates for that by trying to re-match the grid amplitude. Freqency-wise the same. If you have a lot of fluctuations, you have a lot of re-match and re-sync events. Would be interesting if the voltage is more stable on the other substation where the other MP without the humming issue is located.

So in theory with only correct battery settings and the rest factory defaults and no grid-code, no ESS and no other fancy features active the excessive humming should not manifest.

As a quick-test you could also try powering off your cerbo to put the multi in passthrough mode and see if the humming disappears.

With the DEYE Inverters you mentioned the „humming“ problem physically cannot appear because they are using a completely different (transformerless) topology.

17

Since there is no interest in feeding into the grid here, I would only ever use inverters with such an unclean grid. High-power AC DC chargers on the grid side … such as Eltek, Huawai or meanwell
You could then use a multi if you want to add PV on the AC out side later. Otherwise you would really have to use an RS 6000 … because I think the topic of grid conformity is not an issue in the Ukraine … they are then also HF inverters and cannot hum.
But if the grid frequency and / or the voltage drops, they also disconnect … so again the chargers with a very wide voltage input …

take a look at these Voltage …

18

Both your battery cable and ac cable are too small for your set up. (Should be at least 2x50mm or 2x70mm for the battery side). (I would have had at least 6mm-01mm for the ac side.)
The core of the transformer is struggling to remain saturated the cabling can’t provide it.
Before you start adding other gadgets. Begin with that.

19

Are you sure the transformer needs to remain in the saturated area?
From what I remember, when the transformer is saturated, the output sine wave tends to become a square wave…

20

Not sure if you have used the MP2s yet Alex, you won’t have experienced this with the MulitRS.

Power assist comes from the way the transformer is used. As en example on grid connnect- Before connection the hum ramps/winds up then on connect power assists. (This would be fed from the battery side just before connect- it will fail or be unstable if the DC cable is undersized)
What i am hearing is an instability on that side and it is very visibly obvious the cable is too small in the video. Doesn’t look like it matches the recommendations for minimum sizing in the manual.

On the ac side now the cabling is still too small for the pass through and expected feed. On the input this is somewhat controllable through input current limit. But even then the output cabling is way smaller than i would expect to see for 8kVa.

I would start with getting the install at least to min spec before looking for elephants and chasing rabbits.