Loud humming noise and voltage fluctuations

DIY
, , , ,
1

Hi

Installation with Victron Multiplus II GX 5000kva and FelicitySolar 25kwh battery.
Initial boot looks good, works for a whole day and then randomly goes into an
Overload L1 alarm, it then enters a vicious loop where it tries to charge the battery, you can see it go up to 500, 700watts, then it drops back immediately.
After rebooting the unit it works again, but some day later it will again go into this behaviour.
There are no alarms, no events, nothing to be found on the vrm website or inside the gx menus.
Connection is pretty simple :
Only AC-IN is connected, grid tied and there is a BMS connection to the battery.

It is set to external control, a piece of software tells it when to charge and discharge.
Checked that software logs where I was able to see the software asking it to charge, but the victron does not want to.

Does anyone have any ideas on what this could be ?
I have 4 other installations with same battery and same victron, software running stable for months now.

Videos :
Victron loud humming : https://photos.app.goo.gl/BcF8Aaqe37QtCuiA6
Voltage swings during this time : https://photos.app.goo.gl/HYwF6Q5EXnLBzrnR8

2

External software commands are only overridden by information from the BMS. It also depends on whether you are using mode 1 or mode 3 in the ESS.
Even in mode 1, I’ve seen some unexpected variation in power levels where the setpoint is supposed to be fixed (Either manually or by external software). I think that this may be caused by variations in the CVC Voltage level sent by the BMS.

3

Their is no sound on the video ?

4

I had checked if the BMS was limiting power but it was at 200A max discharge and charge the entire time, the victron is in some kindoff deadloop until rebooted, bms parameters are good.

image
image780×307 54.3 KB

It has now been running flawlessly since 19.30

5

I just checked the link at work and I had sound, its a cyclic big hum you can hear on the video where it goes to 700 watts then drops back down to 200 or even lower sometimes, this repeats itself, while the setpoint asked is for example 1000.

6

Here is some more info from the situation when it occurs.

System Grid

  • Power draw 387watts

VE.Bus State

  • Here we can see the dotting between passthrough and bulk, so it tries to bulk (load battery) but fails and goes back to passthrough.

AC Input

  • Within boundaries 234.7v

AC Input Frequency

  • A little polluted but still well within boundaries.

BMS Limits

  • No changes, stays at 200A for charge and discharge.

The power draw on L1 is 387watts, which does not sound to me like an L1 Overload.
I do not know what causes the L1 Overload :frowning:

image
image1618×918 88.8 KB

I also noticed that on 25/09/2024 this alarm was also raised, our installation date is 20/06/2025
The Sales Invoice date on the official victron request states it was sold on 07-08-2024.
This would mean the unit was powered on at 25/09/2024 and had the error we are seeing today already active.

image
image1026×255 11.4 KB

7

Yea video has sound i just forgot to klick

Your ac in voltage seems to dip bellow 220v ?

Also whats your battery voltage ? And ve.bus voltage? If you have a bad conection then its possible your mp2 hits its charge voltage limit ?

8

It is hard to say without more detailed info but the drop in AC Input voltage, as L1 loads up looks like a smoking gun to me. Especially if you have weak AC or Dynamic current limiting enabled.

You absolutely shouldn’t be seeing voltage swings like that on AC In, it suggests a poor connection or badly undersized cables.

I suspect your issue is the charger backing out to protect the grid but right now that is just a hunch…

Note that… Almost all measurements, reported by the system, are somewhat delayed and or smoothed. I cant comment on actual latency just that attaching external instrumentation will, in my experience, show differences during transient events.
I am not saying this is a problem just that the peek wattage you are seeing reported, may not actually be the electrical peek.

Are you seeing the voltage fluctuations everywhere, just on the input or just on the output?
Can you check where the Victron is connected to the grid, in the distribution unit?
There is no way the Victron should be able to significantly affect the voltage there, unless there is an underlying issue with the grid.
I should qualify that…
A massive overload could potentially cause your grid connection to read a low voltage for a very short period but that doesn’t seem very likely.

Is the unit working when it is set to Invert only? This will tell you if there is a load issue not associated with the charger, and or grid. Or an inverter fault I guess, it will depend what you find.

9

BMS CVL remained unchanged the entire day.

image
image563×341 11.1 KB

VE.Bus voltage

image
image550×927 38.1 KB

10

weak AC or Dynamic current limiting enabled

  • Not enabled

Are you seeing the voltage fluctuations everywhere, just on the input or just on the output?

  • Input (output is not connected, only AC-IN)

Can you check where the Victron is connected to the grid, in the distribution unit?

  • Its a newly built electrical cabinet, its on a 20A with 4mm wires (we do not exceed 3800watt charge/discharge)

I could try rewiring this part to check for cable fault?

Is the unit working when it is set to Invert only

  • The unit is working for a long time without any issues, like it will work for an entire night and morning but then during mid-day we had this issue.
    Powering the unit off and on solves the issue.

I can try the suggestion Invert only when the battery reaches a high enough state to power the house for an entire day, this would be the real test, as it can run without issues for hours.

11

You are going to have to explain that… I realise you could have a system that didn’t have a load yet but you say the house is connected..?

Is this setup trying to use ESS export capability to offset your local grid loads, on the input?
If so you are going to have to say where the metering points are and how they are set up.

That also makes L1 overload even more interesting because it would have to be during export.

Do you have a drawing of your installation, showing all connections, metering points, cable sizes and lengths.?

It also means that the voltage swings should literally not be possible unless there is a huge issue somewhere. New may not be good in this instance, and simply because it is stable for a while doesn’t mean it wasn’t stable and compromised!

I am not saying this couldn’t be an inverter issue, but you are a long way from concluding that right now!

How long is this 4mm cable?
4mm, depending on cable type and installation method would be good for anything between 25A and almost 40A (Verry ballpark) and you would typically expect a voltage drop of not less than 11mV / A / m.
Unless your 4mm is the best part of 100m long, which I do not expect to be the case, it is likely fine.
Heck, 5% volt drop, so about 11.5 V, would be circa 48m at 22A

Something isn’t adding up here… Which probably means we are lacking information.

EDIT… If you don’t have anything on AC Out, invert only will not help you, diagnostically or otherwise.

12

You are going to have to explain that… I realise you could have a system that didn’t have a load yet but you say the house is connected..?
Is this setup trying to use ESS export capability to offset your local grid loads, on the input?

  • AC-IN is the only line connected, we will offset usage from the house
  • If the house has 500watts surplus the charger will take it
  • If the house is using 500watts the inverter will provide it from the battery

If so you are going to have to say where the metering points are and how they are set up.

  • The metering point is at the digital meter from the grid operator using the P1 port.

That also makes L1 overload even more interesting because it would have to be during export.
Do you have a drawing of your installation, showing all connections, metering points, cable sizes and lengths.?

  • I can provide pictures and a drawing.
  • I only installed the victron the rest is done by the electrician and I am unaware.
  • 20A breaker comes from 40A 10mm
  • 20A breakes goes to victron on a 4mm for 1.5m on the AC-IN

How long is this 4mm cable?

  • 1.5m

Picture
40A > 20A > 1Fase Meter (to monitor battery power in/out, not the grid meter) > victron AC-IN.

image
image768×804 76.5 KB

13

OK well that isn’t a layout you would see in the UK but I am also not seeing anything obviously wrong and different to my expectations doesn’t mean bad in any way.
In fact it looks well put together, from what I can see.

Clearly that is a 3 phase board and it appears that you are only using a single phase and have single phase meters, at least associated with the Victron.

Cable sizes and lengths ‘look’ OK, again so far as I can tell, I am not suggesting I have done any calculations, have an knowledge of your country’s regulations or even this specific systems design details, but they look sensible…

That brings me back to voltage swings…

Those meters will likely have a demand register/display that shows the maximum demand in the last 15 or 30 mins, depending on the meter and where you are.

If that ‘peek’ draw is only a few kW you should not be seeing a significant voltage swing with those cable lengths. And it looks like the charger is backing out as the voltage falls, as I said earlier.

If that is the case then there is a voltage drop, that there shouldn’t be, somewhere on your system.

I don’t know enough about the internals of the Victron to comment an a mechanism that could load the grid, even briefly, to the point where the voltage falls be 25V, although that seems unlikely, even as a fault, to be occurring without tripping the 20A MCB, which, depending on its type, would probably trip between 60A and 100A within 0.5S

I do not know if the voltage drop is a symptom or a cause, with respect to the inverter, but I do know you shouldn’t be seeing one that big.

Please check for the voltage swing at several points, starting at the inverter and moving towards the main incoming connection.

Also make sure you have the positions of AC loads and the meter/s properly defined in the console setup.

If you are able, please draw a single line diagram showing what is connected, where and with what size wire. This is not a circuit, just a drawing showing the relationship of things and their connections. Like this, but obviously not this arrangement.
Including breaker sizes and types along with cable length for anything outside of an enclosure would also be helpful. You may find it easier to give things numbers on the diagram and then provide a detail schedule, separately, to stop the diagram becoming too messy.

lastly…
Please turn off the inverter entirely and then add a significant load to the house, or a spare way in that board, a couple of 2kW electric heaters would do nicely, and observe the voltage drop.
Preferably at whatever points you measured before.

Its seems like an odd problem, I grant you.

Disclaimer… Please take my comments in context…

I am NOT a Victron expert and have no formal Victron training.

I do have two Victron systems in my house and have helped friends in the past with various setups, so I am reasonably familiar with setup and potential issues in ‘some’, but absolutely not all, circumstances.

I design and install micro hydro installations for a living and qualified as an electrician in the UK several decades ago.

14

The problem just started again now, it flips from charging to discharging too.

image
image678×427 18.1 KB

image
image621×422 16.9 KB

220v - 237v

image
image1071×772 43.6 KB

We rebooted the unit and it was back OK, I am voting on defective unit, everytime we reboot it its fine again, voltage was also within range.

15

I agree it looks very odd, but I am not saying the voltage isn’t in rage, just that the voltage swings are ‘probably’ influencing the system and that you shouldn’t be seeing any noticeable voltage swing at those power levels, let alone 20 or so volts.

I could well be incorrect about the reason for the cycling between charge and discharge but I am absolutely sure that the voltage swings are not normal.

Your inverter is a tiny fraction of the grids’ potential to source or sink energy, there is no way it should be able to drive voltage swings like that on your loads. Even if it is faulty and assuming it is somehow drawing several times the rated current of the breaker.
Something that is, as yet, unobserved and unlikely to happen, repeatedly, without the thing failing completely!

I don’t know what the problem is, I don’t even know where the problem is, but without additional measurements it isn’t likely that you will pin it down.

Sorry buddy, there are few shortcuts when diagnosing issues in systems, any systems, you just have to gather the data and see where it takes you.
Starting with a conclusion and looking for a cause is generally a poor plan and seldom gats you to where you need to be.

If you were describing this as an inverter fault, to a dealer for example, what would you say?
My inverter is inexplicably drawing sufficient current to influence the voltage of the local grid?

I am pretty sure they would then ask how much current is it drawing and where is this voltage drop!

Even if you don’t have a clamp meter you could look up the volt drop, per A per meter, for the 4mm wire you are using and measure that, with your meter set to Min/Max and then calculate the current. It wouldn’t super accurate but it would tell you if you are looking at 400W or 8000W

I don’t know what else to tell you, if you want an answer you have to test.
Even though I have been at this, electrical systems in general, not Victron, for the best part of 35 years I would be methodically documenting tests before jumping to conclusions…
Because you have to, to get an answer you are confident about!

You ‘could’ be right, it ‘could’ be the inverter, but if it is, you can’t demonstrate that right now!
the fact that restarting it stops the behaviour just tells you whatever it is is either being triggered by something external or only happens when the unit is attempting something it doesn’t do immediately after start-up.
You can probably also rule out thermal issues unless you are letting it cool when rebooting.

EDIT…
Have you tried just rebooting the GX device, rather than power cycling the entire thing, is the outcome different?
To be fair for that integrated model, I don’t know if that is possible. You absolutely can reboot from the console but I do not know if that reboots just the GX device or the inverter too.

16

Dont use the p1 port for the victron system its iligal in belgium and definitely not recommended everywhere else

17

Also where are you from ?

Because you speak Dutch? (TKruidenhof)
Im from the region of antwerpen and love a good electrical problem

18

Its the victron causing the high voltage swings.

Voltage while the victron is in faulted state

Voltage after shutting down the victron

19

Yes dutch, Gent region, feel free to share any ideas :slight_smile:

20

Mesure voltage drop over the wire and cince you are from belgium you have to install an “enfluri” meter

So buy yourself an Vm-3p75ct and install that
That will also tell you if the voltage swings are from the grid or from the mp2