question

dylan avatar image
dylan asked

Victron Multiplus II not switching correctly to inverter when Grid fails?

1663648814330.png

As you can see in the graph above sometimes when the grid is switched off during loadshedding ac input will drop, as above but ac output of inverter will follow. This is causing the system to not seamlessly switch over. I did change the low cutoff for ac in to 195v, but it still does this why?

There are 2 X multiplus II in parallel. 2 X e-towers. All ac and DC cables are the same length. I have measured that 1 inverter would push less power to grid side loads than the other.

Do I have a inverter that is not working correctly?

The alarms I get is overload, high DC ripple and low battery, but that clears and all works fine. But the switchover is about 5 seconds.


Multiplus-IIESS
1663648814330.png (68.5 KiB)
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.

13 Answers
Warwick Bruce Chapman avatar image
Warwick Bruce Chapman answered ·
1 comment
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.

dylan avatar image dylan commented ·
Yes it is 100%
0 Likes 0 ·
nickdb avatar image
nickdb answered ·

Grid failures are not always clean, they sometimes brown-out.

The multi tracks the input voltage to it's lower and upper cutoff limits, it does not smooth the voltage drop, so sensitive gear will restart in this scenario.

This is not uncommon with eskom.

I would ask your installer to get you access to the BMS (freedom won do permit this) to check what the battery is reporting.

You don't say what the state of charge is at the time and ripple can have many causes, from cabling, excessive load to an iffy inverter. You can check the battery limits in VRM to see if anything is being throttled, as well as min/max cell voltages to rule out imbalances as contributing.

Being a parallel system also makes it more complex.



3 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.

dylan avatar image dylan commented ·
The load is around 400w ay the time. SOC was around 70%. Cable is 50mm2 from inverter to Lynx distributor and from battery, all same length.


I believe its a grid brownout.

0 Likes 0 ·
Munnik avatar image Munnik dylan commented ·
I have the exact same problem on 2 x Multiplus 2 sites. I have tried multiple settings and cable/terminal checks. The one client is very frustrating and does not understand. I am at the point to remove the system and give him a full refund. I sold Victron to him as one of the best inverters on the market, but because off al the dipping when loadshedding kicks in he does not agree with me. I also find this problem extremely irritating because it is taking up alot of my time. I have installed multiple inferior cheaper brand inverters and have no dipping issues at all. I have searched, Google and spoken to other Victron installers abot this problem, but cant get a answer that works. Hopefully someone on the forum can help me.


Thanks Munnik



0 Likes 0 ·
Munnik avatar image Munnik Munnik commented ·
Just some feedback on my situation. I properly went out to site about 10 times. I tried every possible test that i could think off, I even installed an AVS and contactor, nothing worked. I was literally on my way to remove the inverter and battery and sent it back to the supplier. I friend of mine phoned and said he had the same issue with the dipping on a Multiplus 2 and BSL battery site and he installed the latest V500 firmware. The latest was the V497. This came out this week or last week. I upgraded the inverter with the firmware...and i couldn't believe it.. problem solved. Me and the client are over the moon because this was an extremely frustrating and time consuming exercise...
0 Likes 0 ·
Alexandra avatar image
Alexandra answered ·

@Dylan

Change the LOM detection and see if that helps.

But it does seem like the system is not symmetrical. Which as others have mentioned it needs to be.

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.

nickdb avatar image nickdb ♦♦ commented ·

A clamp meter should easily show if things aren't balanced.

0 Likes 0 ·
dylan avatar image dylan nickdb ♦♦ commented ·
All cables are symetrical, but you are correct 1 inverter was pulling 6 amps and the other 3 amps.


The cables are 100% the same, so could it be the inverter?

0 Likes 0 ·
marekp avatar image
marekp answered ·

@Dylan

If one MP-II is puling 3A and the other 6A you have a BIG imbalance.

I had similar problem with two MP-II in parallel and the problem was with the hardware revision difference between two MP-IIs in parallel.

All cables were equal and AC cables were as small as possible and no go.

Changing MP-II for a proper hardware revision, matching the other in a pair, fixed the problem.

Here is the link to my problem and solution.

https://community.victronenergy.com/questions/91403/how-to-balance-2-mp-iis-in-parallel.html

1 comment
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.

dylan avatar image dylan commented ·
Thanks will check the hardware revisions.


0 Likes 0 ·
nickdb avatar image
nickdb answered ·

@Dylan try update the GX to the latest RC 2.91. In DVCC, you can now enable SVS for Freedom won. See if that makes any difference to system behaviour.

8 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.

dylan avatar image dylan commented ·
Should this not be off for these batteries as it is forced off. Is 2.91 a beta version
0 Likes 0 ·
nickdb avatar image nickdb ♦♦ dylan commented ·
Yes, currently in beta. Have a read of the release notes in the modifications section.

There is a FW specific change.

0 Likes 0 ·
Warwick Bruce Chapman avatar image Warwick Bruce Chapman dylan commented ·
This change has been implemented at our request with the concurrence of FreedomWon. We have been finding in some systems there is a different of opinion between the FreedomWon Orion BMS and the Victron as to the battery voltage. SVS means the BMS voltage will be preferred. We have only been testing 2.91~1 for a day now but the issues we have had are resolved since the update and enabling of SVS.
0 Likes 0 ·
nickdb avatar image nickdb ♦♦ Warwick Bruce Chapman commented ·
And there is no indication as to where this loss is occuring when tested independantly? There is always some loss, but larger disagreements tend to have a cause.

Saying that, had large differences of opinion with BSL which caused grief.

0 Likes 0 ·
Warwick Bruce Chapman avatar image Warwick Bruce Chapman nickdb ♦♦ commented ·

I don't think it is a loss per se, more of a measurement error / difference. But, it is being investigated.

0 Likes 0 ·
nickdb avatar image nickdb ♦♦ Warwick Bruce Chapman commented ·
If I can ask, what magnitude of a difference are you experiencing?

With a different vendor I had on average 0.5V difference that under load exceeded 1V. Couldn't get to the bottom of it, recabling included. It eventually disappeared when batteries were changed.

That delta was enough to throw ESS/MPPT charging off but never caused instability.

(thread hijack off)

0 Likes 0 ·
Warwick Bruce Chapman avatar image Warwick Bruce Chapman nickdb ♦♦ commented ·
0.5-1.0V is enough to stop it ever getting to 100% SoC and cause SoC drift over time. One site went down to 38% (but above 54V) and is now corrected after SVS was enabled.
0 Likes 0 ·
nickdb avatar image nickdb ♦♦ Warwick Bruce Chapman commented ·
After the most recent FW update, I have inherited daily drift, far more often than before (have a case open with them) so am also testing the SVS setting, will see if it makes a difference.


0 Likes 0 ·
dylan avatar image
dylan answered ·

@Daniël Boekel (Victron Energy Staff) I have a client that has 2 sites each with multiplus II's also doing exactly the same as above. This rules out that its from inverters in parallel. The grid is during loadshedding not being switched immediately but has sort of a brownout and the Multiplus does not cut out quick enough. The output of the inverter follows the grid voltage down to 83v in this case causing a overload and low batttery alarms. Sometimes it even switches the system off. Can you please advise.

18 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.

nickdb avatar image nickdb ♦♦ commented ·
Have you tried with higher lower cutoff values above the 195V you used before?

You may need to follow the formal support process via your distributor to escalate for assistance.

0 Likes 0 ·
dylan avatar image dylan nickdb ♦♦ commented ·
Yes, I have tried setting as high as 201V. I work for the distributor. So Am looking for a solution before these are returned for Warranty.
0 Likes 0 ·
Daniël Boekel (Victron Energy Staff) avatar image Daniël Boekel (Victron Energy Staff) ♦♦ commented ·

Hi @Dylan

FreedomWon?

I've seen some other sites where the battery dipped, resulting in this behavior.
But generally, if your grid voltage is normally a stable 230V, set the disconnect value even higher, like 215 or 220 V to try to prevent.

The problem is that the inverter will follow grid voltage down to the set level, but when doing this it pulls from the battery, if the battery drops, it cannot keep up.

0 Likes 0 ·
dylan avatar image dylan Daniël Boekel (Victron Energy Staff) ♦♦ commented ·

Thanks that makes sense. Yes grid voltage is normally around 230. Will try this.

Yes it is a freedomwon on 2 sites and a Hubble on the other.

0 Likes 0 ·
martin-delp avatar image martin-delp dylan commented ·
I am having the same issue. Tried this. Multi disconnects and connects correctly (at say 215v and reconnects at 225v) multiple times a day, but as soon as Eskom loadsheds, the inverter still shut down (low battery voltage) and restarts after about 30s. This happened for the first time around September 2021 (when I was running 1x MP2). Currently running 2x MP2 in parallel and same issue still occurs. Someone mentioned in another post that a contactor on AC1-IN could solve the problem. Not ideal, but could this be a reasonable work-around?
0 Likes 0 ·
marekp avatar image marekp martin-delp commented ·

@martin_delp

If system tells you "low battery voltage", maybe you should check your battery if it is still up to the task.

0 Likes 0 ·
martin-delp avatar image martin-delp marekp commented ·
Thanks. I have tested this by tripping the main breaker on the DB under load, say at 4.5kw (microwave and kettle going), and the inverters switch over correctly. I have 2x 10kwh lithium batteries with 100 DCL each. It would therefore appear grid-related, so trying to understand if there is anything we can do externally to disconnect the MP2's quicker
0 Likes 0 ·
marekp avatar image marekp martin-delp commented ·
@martin_delp

It could be related to grid, but I would first measure voltage drop in various places of your DC line when problem happens.

0 Likes 0 ·
nickdb avatar image nickdb ♦♦ martin-delp commented ·
Ripple and low battery are consistent with a sudden surge that the battery can't deliver.

I have heard of number of sites with issues on Freedom Won and Hubble also has a bit of a rep (lower voltage, current limits).

The challenge is that it happens a bit too quickly for vrm to capture.

At least on the Freedom won they should be able to pull data from the orion's for some insight.

0 Likes 0 ·
dylan avatar image dylan Daniël Boekel (Victron Energy Staff) ♦♦ commented ·
@Daniël Boekel (Victron Energy Staff) Is Victron working on a solution? The inverter does not cutoff at any voltage you set the lower voltage cut off too. It does not happen quick enough. Is their firmware to address this? This is leading clients to want a full refund.
0 Likes 0 ·
Daniël Boekel (Victron Energy Staff) avatar image Daniël Boekel (Victron Energy Staff) ♦♦ dylan commented ·
Hi @Dylan

The sites I know about solved it by installing a different or an extra FreedomWon battery.

So unfortunately it looks like a battery problem, it drops out to quickly. Please contact FreedomWon if they have a solution in firmware or hardware.

0 Likes 0 ·
dylan avatar image dylan Daniël Boekel (Victron Energy Staff) ♦♦ commented ·
@Daniël Boekel (Victron Energy Staff) It is definitely not the batteries, the 2 multiplus inverters follow the input voltage down to 52v , then only switch over, causing issues. Can it be that I have 2 different versions on inverters as suggested before. The inverters don't have Hw rev stickers on. Can this be checked by their serial numbers? Can I give you access to my site?


1 Like 1 ·
nickdb avatar image nickdb ♦♦ dylan commented ·

What are the first 4 digits of each? That does highlight the manufactured year/week.

1 Like 1 ·
dylan avatar image dylan nickdb ♦♦ commented ·
Is this the first 4 numbers after HQ. They are 2204 2203.
-1 Like -1 ·
nickdb avatar image nickdb ♦♦ dylan commented ·
Yes, correct. Ok, they are all close enough so that eliminates a known cause of asymmetry in parallel systems. Have you put them under load and measured how even the split is? Would be good to rule this out as a contributing factor. Another suggestion is to try get the system charted on grafana, VRM just averages out the data too much so you may be missing detail at the time of failure.

What size inverters are these? iirc 2 e-towers are around 160-180A max, FW support restricts this to 8kVA of inverter so exceeding this can cause issues.

If you have a loaner battery you could try add one to see if it does resolve the issue.

0 Likes 0 ·
dylan avatar image dylan nickdb ♦♦ commented ·
That is the other problem they are not pulling symmetrically. Even with all cables same length. I will next be taking a day to just check all wiring is 100% and maybe adding a automatic voltage switch. I have 2 multiplus 2 5 kva with to etowers. Bms allows 210 amp discharge. So doubt that is the problem.
0 Likes 0 ·
Daniël Boekel (Victron Energy Staff) avatar image Daniël Boekel (Victron Energy Staff) ♦♦ dylan commented ·
What grid disconnect voltages have you set?
0 Likes 0 ·
dylan avatar image dylan Daniël Boekel (Victron Energy Staff) ♦♦ commented ·
Set to 215 low disconnect
0 Likes 0 ·
dylan avatar image
dylan answered ·

@Daniël Boekel (Victron Energy Staff) any answers on this. Another client has put in a contactor and the system still does this.


6 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.

nickdb avatar image nickdb ♦♦ commented ·
I know of another system with same specs doing the same thing. It really appears to be the battery dipping, as it shows tons of DC ripple at lower SOC. The system works fine with solar running to buffer. If your issue is reproducible, the best thing you can do is connect an RS232 cable and log from PBMS tools while creating the fault.

Willing to bet that some cell voltages get out of whack and the BMS dies.

Don't bother asking FW for help, they will just tell you its a victron problem or bad wiring.

Best you can do is gather the data to show this isn't the case.

0 Likes 0 ·
dylan avatar image dylan nickdb ♦♦ commented ·
The data on VRM clearly shows the AC output voltage follows the ac input voltage, this is the issue, but will check the battery as well. Had FW check and they said it is fine.
0 Likes 0 ·
nickdb avatar image nickdb ♦♦ dylan commented ·

That is their default reply and the quality of the onsite/frontline support tech leaves a bit to be desired. I wouldn't put any value in their VRM-based assurances. VRM would also not catch the problem, the sample rate is too low and the BMS detail is hidden. PBMS is much more granular and it will capture every cell. It exports to CSV so easy to chart from there - it's easy to do and should give you a bit more ammo to get someone to help.


0 Likes 0 ·
dieter-lubbe avatar image dieter-lubbe commented ·
Hi Dylan. I have 3 x Multiplus ii units in parralel. The only thing that worked for me was change the grid code setting to "Germany, internal NS protection". I get almost zero disconnects now when load shedding hits.



0 Likes 0 ·
dylan avatar image dylan dieter-lubbe commented ·

Did you have to change any wiring? Is this compliant? What does the internal NS protection mean?

0 Likes 0 ·
dieter-lubbe avatar image dieter-lubbe dylan commented ·
Yes it is complaint and my wiring stayed the same.
0 Likes 0 ·
Andrew Poluta avatar image
Andrew Poluta answered ·

Hi Guys


I am having the exact same issue.

Fortunately I've recently installed three identical systems in one complex in Linden JHB.

The systems comprise of the following:

1x Multiplus II 5kVA

1x E Tower

1x Smart Solar 250/60 and 6x 450w panels.


When the brownout occurs, two of the systems change from pass through to inverting seemlessly but one follows the grid right down to 90v AC and VE Bus state indicates 'off' for a few seconds.


There is clearly an issue here that we all need to urgently solve before our brand gets associated with this issue. End users are being badly affected by this 'non UPS' issue and word spreads fast.


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.

driaan-de-clercq avatar image
driaan-de-clercq answered ·

Hi all,

We got a similar problem with a system consisting of:

1x Victron Multiplus

1x SMA PV inverter

3 x SS202 Solar MD batteries.

With the troubleshooting we found that the BMS's of the three Solar MD batteries were faulty and did not balance the cells properly. We replaced the BMS's of all three batteries, with assistance from the Solar MD team.

Next we disconnected the PV inverter to see if it influences the 'voltage dips' when loadshedding occurs, but it did not make any difference.

The DC cables from / to batteries are the same length, adequate thickness and there are proper connections. All the wiring on the AC side is also correct.

The voltage dips can not be seen on VRM or the CSV file as the logging interval of VRM is to slow.


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.

leetpro avatar image
leetpro answered ·

Also having the same issue on a Quattro 15kva, only difference is that it wont turn back on until the grid is restored (I'm testing a short on the battery terminals at the inverter)

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.

Matthew Walsh avatar image
Matthew Walsh answered ·

For those who are still experiencing this issue what are your grid settings?

1 comment
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.

dylan avatar image dylan commented ·
I have set mine to other and upped the cutoff volatge to 210V, do you have a solution? It still happens
0 Likes 0 ·
dylan avatar image
dylan answered ·

@Daniël Boekel (Victron Energy Staff) @Guy Stewart (Victron Community Manager) Is there any solution, it seems to happen less often, but I now have 2 sites. One a site that is almost 4 years old and it comes up with phase master missing and then resets. The other has Dc ripple fault and battery low and then resets.

All wiring is fine, its been double checked and it only happens when grid goes down in loadshedding and only some times.

3 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.

Daniël Boekel (Victron Energy Staff) avatar image Daniël Boekel (Victron Energy Staff) ♦♦ commented ·
that points to the battery cutting out / dipping.


try a bigger or different battery.

1 Like 1 ·
dylan avatar image dylan Daniël Boekel (Victron Energy Staff) ♦♦ commented ·
Ok will try get another battery and add it., thanks
0 Likes 0 ·
dylan avatar image dylan Daniël Boekel (Victron Energy Staff) ♦♦ commented ·
One site have 2 X solarmd 7.4kwh batteries though
0 Likes 0 ·
martin-delp avatar image
martin-delp answered ·

Just checking if this issue has been resolved? I installed a voltage monitor/relay with a contactor between meter on AC1 In but still have the same issue of low voltage disconnect upon loadshedding - therefore cannot be a brownout issue. Some people of other threads have had success disconnecting the ve.bus cable between the ccgx and multi before loadshedding hits which seems to indicate a comms issue - however to disconnect each time is not going to be sustainable. Has anyone else had any luck?

3 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.

nickdb avatar image nickdb ♦♦ commented ·

That’s most likely annectodal. Almost always it has been found to be a battery issue, those that upgraded or changed products no longer have experienced the issue. Most common with pace based BMS’s so in this region, Hubble, FW etower, BSL.

0 Likes 0 ·
jakloosterziel avatar image jakloosterziel nickdb ♦♦ commented ·

Any feedback on solving any of the above, i have the same issue. I have changed Ac in wiring and DC cables several times but no luck.

0 Likes 0 ·
martin-delp avatar image martin-delp jakloosterziel commented ·
Someone solved it through changing certain BMS settings. Another person did a "work-around" it by using some home automation and ESP loadshedding app's API to automate the MP2's switch to "inverter only" prior to loadshedding and back "on" shortly thereafter. I have not been successful, but have not yet been able to access the BMS to see what is going on.
0 Likes 0 ·