And you have no loads between the Carlo and the inverter?
Carlo is just after the main breaker.
Let me explain like this what I see.
When LS is over:
1 - The Carlo shows the full draw, 100% as expected.
At the same time, the inverter is not synced yet, 2) the AC Loads are exactly the same as the Carlo’s draw from Eskom.
Once the inverter has synced, only then I will see the picture above, the 1) Eskom draw drops as the MPPT/battery takes over again.
So your AC loads are connected to the AC in.
Which would fit the behaviour you described.
I connect things differently.
The Carlo stays where it is.
Critical loads on AC1 out.
AC loads on AC2 out.
During LS you will lose AC2 out and keep your critical loads as you do.
AC1 remains on.
When the power returns AC2 will only get power after the delay.
AHA, that makes sense.
Did not want to run the non-critical loads via the inverters AC_Out2.
But still, no matter which way it is connected, Eskom stays on the “other side” of the main breaker. Want to stop it there. Protect the entire system.
I did actually specify:
Can I ask why you chose not to use AC2 out?
It may cause E/L issues as well.
Is your inverter on an E/L supply or just your AC loads?
Do you have a second E/L on your critical loads?
I think so too.
When the grid comes back, the relay test will be done immediately. So you will immediately hear the clickety-clack. You will also hear the hum of the Multi change just a bit for about a second (that’s synchronising the AC waveform). And then it will sit there for another 60 seconds, making sure the voltage and frequency is good, and THEN it will close the transfer switch.
After that, there is a 10 minute ramp required by NRS097, so you may well find that the Multi stops feeding in energy and your grid consumption is initially all of your load, and then things will slowly return to normal.
As TTT observed, the grid meter returns and shows on the overview before the Multi reconnects, so for a short period of time the numbers may not completely line up. Ironically this is a recent change (in 2.80 I think), because in the past the grid meter value was not shown unless the Multi’s input was connected (because we needed to know WHICH input is active, in order to know if it is a generator or grid). Later versions now make an educated guess… and as a result, you get a grid meter reading while the Multi is still disconnected.
If TTT’s inverter is being fed by an E/L supply are there any components of the test that will trip an E/L?
Not anymore. Multiplus-II uses a different way to test TN bond. The old Multigrid in the Aluminium case passed a small DC current which sometimes tripped RCDs.
The relay test can however cause transients that can nuisance-trip an RCD, usually if it already has a high standing leakage.
The context here, however, is that for loads behind a Multi, you already have a device that makes sure everything is in good shape before connecting your loads, which is what you said. So you only need this additional device for your non-essentials.
Let me ask the reverse:
What are the limitations of running an entire house via an MPII 5kva, using only AC_Out 1 for Critical and AC_Out2 for all Non-Critical loads?
As per Plonk, the MPII is different now, been running like this for years, nary a problem with EL.
The inverter is connected to the main DB with a 50a 2P breaker, so yes, on the main DB E/L.
Critical loads DB has its own EL yes.
I thought you dedicated an entire thread to the frustration caused by your E/L tripping when the power returned after LS?
The one has nothing to do with the other. Never had EL problems.
Had an error, a very well-documented error at that, that worked in the reverse i.e got the error when the grid is off, not when syncing with the grid when Eskom returns.
It is sorted now … as I posted there, the combination of a spare panel wire measuring a voltage due to radio noise, highest at 8am, with the MPPT connected directly onto the inverters battery poles. The error is gone since I pulled all that apart.
So to confirm, the Multi is feeding nothing from Eskom to the Critical Loads DB until it is synced, the ±60seconds?
Put it another way … for those 60 seconds, the Critical Loads DB is still far far away from Eskom?
I don’t want to lose any loads when the grid fails so my whole house goes to AC1 (Critical) nothing connected to AC2
The only limitation is the size of your transfer switch.
Which would limit you to the sum of 50A from ESKOM and 5kVA.
But using AC2 out would allow you:
To dispense with the CG - you’d see your load anyway.
To avoid the problems of appliances destroyed by an unstable supply when the grid returns.
Precisely.
When the grid returns, it does the relay test, and it syncs its own frequency to the grid’s, but it doesn’t connect to the grid. The regulations says you can only do that after 60 seconds. For that 60 seconds, your critical loads continue to run from the battery and you can still overload the inverter. In fact, that is a danger. Your family goes: Oh! The power is back on! And immediately overloads the inverter 
If you plug in your mk3 and run veconfigure, it will even show you the countdown until it connects.
1 Like
That’s an option as well; you are not forced to take advantage of the flexibility.
I also think with assistants, you can choose to maintain AC2 and keep all loads on until a certain SOC, and then dump AC2.
It is also possible to have your PV inverters dedicated to a certain AC out and dump everything on the other AC out at a certain SOC to maintain the ability to charge batteries. Thereby avoiding the need for an MPPT for black starting.
I stop them at the Critical Loads DB before they overload the inverter, breakers trip first. 20a main breaker, and plugs on a 10a breaker. Caught the buggers already plugging in 4kw dryer whilst there was LS.
So, if I use AC_out1 and 2, then the house has a max of 11.5kw (50a x 230v) in total, right?
And that is why I chose the Carlo route. Critical loads are 5kva, and the rest of the loads in the house are limited by the 63a main breaker.
And it is a good thing I did, as there are 7.2 people using power in this house at times. Critical Loads are the sacred circuits.
Yes, Victron downgraded model MP ii to 50A from the 100A of the previous 5kVA models.
Do you ever use 15.5kW?
We have touched a time or more on >14kw yes with family visiting, in winter. You live in Cpt, you are every family member’s favorite stopover.
I kinda “gooi a vloermoer” (tantrum) when I see that … one of notable proportions.
EDIT: For the price back then for a Carlo, it is worth it, less strain on the MPII.
Also, re the 60s delay, I want ±5 minutes, we have had twice now where the transformer has tripped after the Victron synced, street overloading it, or whatnot.
And I have no worries about overloading either.
NRS 097 only stipulates a minimum time; Victron has imposed >60s which complies.
Victron has not added the facility for a user to choose a higher custom time. I think they should add this ability before their next bout of certifications.
Although it is easily accomplished with an external timer.