question

Klaus Naumann avatar image
Klaus Naumann asked

MultiPlus II ESS is feeding into the grid while showing "bulk"

Hi,


I have a newly installed ESS system with 6x MP2 5k on three phases (two MP2 feeding one phase each).

The system acts weird in that it drains the battery and feeds about 1000W into the grid. It shows "bluk" while doing so, which as I understand is a charging mode instead of a mode where it should discharge the batteries.

When I put the mode from "On" to "Inverter only", this behavior stops and it will only produce as much power as is needed.

Firmware is 506 on all MP2s.


Any help is appreciated!


Klaus

ess discharging
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

8 Answers
Alexandra avatar image
Alexandra answered ·

@Klaus Naumann

The Bulk state will be triggered when you reach re bulk offset on the inverters. So likely the voltage on the battery drops to that threshold.

2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Klaus Naumann avatar image
Klaus Naumann answered ·

Hi Alexandra, Thanks a lot for your reply!

So this went on for hours, the battery dropping to SOC of 50% (according to the JKBMS) or 77% (according to the RaspberryPi GX device). This was yesterday evening, with no sunshine (I have a PV system behind the AC out, or in front of it - depending how my override switch is positioned).

Today the situation is as follows: There's about 6000W of PV power. The MP2s ignore any setting in the GX ESS settings regarding limiting the charging or inverting power but instead again reach roughly 800W grid feed. This means that 5200W are pushed into the battery at the moment.

It somehow seems like the grid set point is around -800W and that's the only thing that counts... in the GX ESS settings, grid set point is set to 0W.


2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Klaus Naumann avatar image
Klaus Naumann answered ·

More input... I have an EM24 as grid meter. Currently I have the system switched in such a way that the PV system is not passing through the MP2s. If I switch "Grid Meter" from "External" to "Inverter/Charger", the system suddenly takes about 1100W again from the battery and feeds it into the grid, always staying in state "Bulk", no matter if it's charging the batteries (as described in the post above) or drawing from them.

2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Klaus Naumann avatar image
Klaus Naumann answered ·

Lots of things were tried today, including a Firmware downgrade to 3.01, which didn't change anything. Setting the "Grid setpoint" to 1000W makes the system run roughly around the 0W mark on the EM24, which seems really obscure.

2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Klaus Naumann avatar image
Klaus Naumann answered ·

Actually that was a coincidence of sorts. It's now back to using weird values.

2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Klaus Naumann avatar image
Klaus Naumann answered ·

I've done quite some more debugging:

The system runs smooth when I just use 3 MP2s (either three work, there's no difference). Using 6 again creates the same weird result. I've further tried just L1a and L1b (so both devices on the first phase) and the same with the second phase. Both create weird results, in that they drain 1000W or 700W from the battery and push the excess into the grid (e.g. when the house uses 300W, then 700W or 400W will end up being pushed into the grid).

I've tried to measure around a bit with a clamp multimeter and it looks really weird. On the AC side, the readings are consistent. On the DC side, the readings are all over the place, with devices on the same phase having one device that reads e.g. 14A and the partner -2A. This also stays the same when only using two devices on one phase.


Any ideas?

2 comments
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Alexandra avatar image Alexandra ♦ commented ·
@Klaus Naumann

A couple of thoughts. You have spent this much on a three phase parallel but not bought an official GX device. Did you know a Cerbo S is as about cheap as a PI and an MK3?

I am assuming all your cable set up are as required by the manuals and have followed the rules of three phase parallel set ups here. Meaning it is balanced.

Is feeding in from AC and DC switched off?

Is ESS programmed on all the inverters?

What is the grid set point?

0 Likes 0 ·
Klaus Naumann avatar image Klaus Naumann Alexandra ♦ commented ·
@Alexandra


I think the main reasons were:

- I'm an avid user of RaspberryPi's and I had a RPi4 that wasn't used

- I would have needed the MK3 anyway to program everything

- I didn't (and don't) see any difference in the Cerbo devices and the RPi


Yes, all cabling is done according to the wiring guide. 70mm2 is used between the batteries and the MP2s, 6mm2 for connecting them to the wiring box and 16mm2 going away from there. All AC and DC cables from and to the MP2s have the same length.


Feeding from DC is off, feeding from AC was on in all scenarios as I intend to have a PV system behind this, once it runs properly. I've tried to switch it off as well now, it doesn't change anything in the behavior.


Yes, the ESS Assistent is loaded (and sent) to all inverters.


Grid setpoint is -30W at the moment (because I'd rather like to have something pushed into the grid, rather than using it) but the setting doesn't affect the behavior in a great way it seems.


0 Likes 0 ·
Klaus Naumann avatar image
Klaus Naumann answered ·

To close this: many (read: 40...50) hours of debugging were spent on this error on my/our end. A support case with the distributor and Victron in the background concluded that this must be an issue with the cabling, even though it was pointed out multiple times that the system is following the cabling guide, including descriptions and photos.

I've then - by chance - updated the firmware from 506 to 508 and now the system is running without any issues. No cabling was touched. I'd wish that Victron would be more willing to help customers and less assuming that the issue couldn't possible be with the product.

2 comments
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Alexandra avatar image Alexandra ♦ commented ·
@Klaus Naumann

Glad you have it sorted now. Some installs are time consuming.

I hear what you are saying about support. However a few notes on that. The product assumes a person is professional and is capable of taking responsibility for the install. So installer puts the time. Sometimes it bites, but that is life.

Debugging over the internet not in person is almost impossible. Yes data helps but boots on the ground is it at the end of the day is the way. It also does not help if there are a few people out of thousands with a weird problem that not everyone is experiencing, so it is not an unreasonable conclusion to think it is more likely install related when the product is fine elsewhere.

Again due diligence of the installer is required to put in the time to make sure it is not something overlooked in installation.

In Troubleshooting one of the prescribed steps is update software. The firmware is usually tested before release. But with the flexibility of the product things are very much bound to be different in the field.

So the advice is there. And the manuals are there as a min guide. Sometimes the min is not enough and an adjustment is needed. Or at least I have seen that in my experience.

0 Likes 0 ·
Klaus Naumann avatar image Klaus Naumann Alexandra ♦ commented ·

@Alexandra


I agree that debugging over the Internet is not of great value, that's why the installer (a certified electrician and proficient with Victron products) and myself (engineering background) decided at some point to ask the distributor for additional support, even if paid. The answer was that Victron doesn't do in-field support for end customers and that the distributor doesn't have a partner that can do support.


The due diligence was done. Multiple scenarios were tested (3 units alone on 3 phases, two units on one phase, all of this in several combinations, cables' resistance measured, individual currents on cables, and a lot more things, too many to list here).


I agree that firmware updates are standard in trouble shooting. When the units arrived, they were on 4xx (I believe 497, but not 100% sure anymore) and all of them were updated to 506 which was the newest firmware at that point. In the weeks after, which were used to run the above mentioned tests and checks, eventually new firmware versions came out. One mentioning fan noise (507) and the other an update for Quattro-IIs (508). Neither have any indication that they could fix an issue with parallel systems. Had Victron said "please update the firmware" instead of "it must be the cabling", it would have happened a bit earlier. But this was not the case either.


Don't get me wrong: I still like the product, I still like its flexibility and everything. But the assistance in case of issues could really be better - that's all I'm saying.


0 Likes 0 ·
manta avatar image
manta answered ·

I still think your cable resistance is too low, for the multiplus to balance out the phases.


Maybe make something, in node-red to switch off the parallel inverters when on low load. I use 2.5mm2 cables on AC side to deliberately create some resistance on low load to keep things balanced.

On the flip side : On continuous load of 4000W they tend to get a little warm over time, and so its costing a bit of efficiency, nothing critical. Still it's annoying to waste power.

2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Related Resources

Additional resources still need to be added for this topic