The battery cannot vary its parameters that quickly. The cycles for reducing charge or ccl=0 for balancing or cells going high, is much slower. You can see this in the widget chart you provided.
The battery power chart you showed in the recent post seemed reasonably smooth.
Nick, many thanks for the pointer. It was a good, long read, my take from it is the following:
-
In case i would have Enphase solar panels they could cause the same fluctuations that i have, specially in the morning with sunrise.
-
In case i would use a Shelly meter instead of a Victron, the lower sampling rate (not sure if i understood correctly but the Shelly seems to average the values) would mean no feedback hence no osciallations.
-
In case i use Hoymiles inverters (although it is not clear which type), i can expect oscillations as well.
-
Victron will distribute in the foreseeable future a firmware version with the VM-3P75CT that would make i possible to use an average of the readings, hence reduce the feedback rate.
-
This could also yield for the solaredge, however, i do not see strong fluctuations in the morning when the sun comes up, only later on when the MP’s are a few hours busy filling the battery. What still might happen is that fluctuations from the PV inverters are visible in my EM. This morning with a PV wattage of about 3000 watt and a battery filling wattage of more than 2000 watt i noticed bigger fluctuations then without a solar wattage, yet when i switched on the water cooker (about 2000 watt) the fluctuations went back to as if there was no solar wattage ….
-
What i can do is make a Node Red virtual meter that averages the values and that once a second delivers information, like the Shelly. What i so far do not know is how to switch the two EM’s the victron one with the EM one …
-
As noted before, even with switching off the Hoymiles the oscillations occor. If i switch off the SolarEdge they are also there but not as big, because the Hoymiles only delivers 1200 Wattp and the solaredge can go up to 10000 Wattp.
-
Not sure if that would do the trick, i will proceed with (2) and produce a Virtual EM, although victron tells us not to use it, testing should be fine.
Once again, many thanks for you pointers, i have not cracked the nuts yet, but still think i have excluded a few things as well.
It could well be that the total effect is a combination of two or more pieces, like CT Clamp interference with PV inverter fluctuations … I hope I get there in the end ;-).
I would be careful with virtual EM’s, Victron don’t want these used to run ESS’s as they can be unreliable or experience issues due to the nodered layer. No need to create more problems or instability.
They are fine for measuring loads and other third party devices.
All, i have been changing some settings on my system and investigating if temperature has an influence on the large deviations. It certainly does not help if the system gets heated up, but the influence is not that strong. I have not been able to reproduce the large swings lately, hence me not being able to reproduce the large swings might be because of lower temperatures (at least where i live).
I have also been interested in the basic noise level of my system and have conducted a test without solar influence (so night time recording). The first set of figures shows the system between midnight and roughly 6 am in the morning (no solar influence). You can see that one phase (phase 1) has the majority of the (changing) loads on it, that is the freezer in the kitchen, the refridgerator in the kitchen, the (dormant heat pump) and the wine fridge. That latter is the cause of the peaks you see on phase 1 every hour (a bit faster ..). I have recorded the PEN voltage out of curiosity.
The second set of graphs shows when i switch the system to grid power instead of battery power. Unfortunately the wine fridge still interferes with the battery, will do a run without the wine fridge. But what also can be seen is that there is some “residual noice” in the system, either caused by the loads on the AC OUT or by the Victron system (the EM ???). I will do some investigations next month on the size of that noice (fourrier transformations etc). The data is available.
My question to you, is this normal (so should i not look for the cause of this noice) or are your systems behaving a bit more elegant ???
@MarcKeulers I also had fluctuations when commissioning my Multiplus-2 GX.
I’ve started with current trafo and then moved to EM540 energy meter. But I’ve forgot to place the short at current trafo location (as per manual).
After short wire was placed, all went nice and calm. (No RedNode, no external calculations. All Victron stand alone.)
Would that help you? Just pay attention.
@Viorel_C , many thanks for the pointer. Is it possible you show by photo where you have attached the short at current trafo location? I can not find that in the VM-3P75CT manual (but as so often i might not look in the right spot).
In the manuaI I read about lenght of the wires and the induced noise …
What i got from Schneider Electric that they bind all the negatives to each other and to the ground, but i am not sure if that is doable or advisable for the VM-3P75CT …
@MarcKeulers It was not on energy meter side, but on Multiplus-2 GX side.
In my case, if that short was not placed, and the system defined to take info from energy meter, the Multiplus gets confused.
Would it be same for your system? No clue. I just share my experience.
@Viorel_C , super, many thanks for this, i will check what i have done. See below one of my Multiplusses with the Current Sensor shortcutted as per manual (we did the right thing in the end, bit long ago).
Many thanks for your pointer as this one makes sense, unfortunately for me, i need to keep on searching …
What else I think of, is what I’ve experienced: Do another round check of all settings / commissioning. I’ve experienced a loss of ESS assistant. No clue how. In there also check that current transformer is not ticked. To me, what you have is a miss understanding of Victron to what to look for. Trafo or energy meter.
Make sure that if you use modbus, IP’s and their ranges are important. I’ve end up with IP in Victron same as my main IP of one of my house routers. And of course… troubles …
What is your network? mono, tri (4 wires, 3 wires)?? If you have 3x230 without neutral, check if your energy meter is suitable for 3P network.
@Viorel_C thanks for the pointers again and good you share your experience … i learn from it. We have checked the settings / commissioning three weeks ago and could not find anything out of the ordinary. I will make a download of all the settings (have to find out how) so i can refer to them more easily.
On the IP settings site, these are fine and checked. No strange double IP addresses found or anything that is not working. The VM-3P75CT is connected through VE-BUS so that should not give me any trouble. The only item directly connected to the internet is the Cerbo GX.
I am running a normal 4-wire 3 phase Dutch system, nothing fancy and the meter has been set up correctly as such (double checked).
The one strange thing i find is that my fluctuations increase in magnitude when i deliver back say 3000 Watt compared to 280 Watt, but they do not increase any further when i deliver back 5000 watt or even 10.000 watt.
One more time, many thanks for your pointers, i will publish a comparison between ESS mode 1 and 2 in the next posting …
Dear @nickdb you pointed out phase control, i never did a whole day, that is what happened now last monday (normal mode 1 ESS) and tuesday (mode 2 ESS). I share the picture (it is big), the biggest difference i see is that the temperature of phase 2 increases more than the others in mode 2 as there is more PV power coming through that phase due to the additional PV inverter on that phase. Otherwise i can not see much difference in the pictures or the amplitude of the fluctuations. Note the situations are not 100% alike, that is quite difficult to create. Good thing is that in all cases i no longer have 2000 watt fluctuations … i have removed the top cap of the cabinet the units are in so the air can flow better. One of my next projects is to include 2 Artic P8 Max with temperature control (as they make a lot of noice) to reduce the temperature of the unit in summer. You have to click on the picture (if it is too small, let me know, if you want details, let me know)
@Viorel_C this morning i have done another check on the configuration and all three Multiplusses are identically configured. The ESS assistant is there and the external AC Current is not checked. Thanks again for the pointer, i think it is good to double check ! Takes time but at least we can ommit that kind of error …
When I got stuck, I’ve asked a friend with experience to look remotely in the system. He has done that using my VRM code, so without onsite visit.
@Viorel_C you got great ideas, thanks. Tomorrow this will be done 
… let”s see what is found !