What I want to “sort”:
2 x 2kw Geyser elements in 2 geysers.
Both are on timers to schedule them to run consecutively.
Heatpump/EV/Flat panel options are not viable for me ito their costs and lifetimes.
Victron grid-tied system currently doing the job with 4.2kw array, and I’m running a wee bit short to recharge the battery bank.
So my thinking:
Remove the geyser heating off the 4.2kw array by adding 2kw paneling and get a grid-tied Solis, connected to the house, to handle the heating of the geysers, being grid-tied to the house.
The 4.2kw on the Victron is then more than enough to then recharge the 560ah bank, load-shedding.
Or is it better to get an MPPT that can handle 2kw on a separate array, keeping in mind that the Solis is probably more efficient?
Put the solis on the input, use a second Carlo just for the solis. In times of no heating the Victron will use the power to assist in charging batteries or feeding loads.
I have done a similar system for @Ironman an he was quit surprised with the end result…
I was thinking to connect the Solis direct to the main DB, the sensor before the Victron’s Carlo? Simply to keep it as simple an install as possible.
I reasoned that if the Solis lowers the Eskom draw of the main DB, that will result in the existing panels on the Victron system being more “dedicated” to battery charging?
I’m really scared about that, cause of my bad luck all the time … and stupidity at other times. But I’m not averse to learning.
You will configure the second carlo as a pv inverter, that way the victron system knows that it’s there and will use the available power as much as possible.
Your panels might go into rest after you are done heating the geysers… If the victron knows of its existence, it will try and use the power… Big benefit.
I put a 1500W on the output of a 5kVA Phoenix. Works well, supplies the loads, and excess power charges the batteries through the Phoenix.
It responds to frequency shifting when the batteries are full.
The Phoenix doesn’t have an AC in, so there is no chance of ever exporting AC, but I believe there is a fairly cheap Solis CT option to control export independently when paired with a Multi.
It’s been rock solid, quiet and it is deceivingly lightweight. Its operating voltage range also starts lower than most PV inverters.
This Solis size has only one MPPT. I have noticed at least on two other brand PV inverters there is an advantage to having two MPPTs.
The overall rating of the PV inverter may be 3.6kW, but from what I have seen the individual rating of each MPPT may be 2kW.
So it is possible to get a 2kW output from a 3.6kW just using one MPPT.
Anyway, I am still playing with this to see if I can exploit it further with my East & West arrays.
Hence connecting it on the input/output of the Victron, like:
Output would be connecting the Solis on the Critical Loads DB?
Input like after the main breaker to the inverter in the main DB?
Just for some more info. If you already have a grid meter installed (An energy meter configured to measure the grid), and you have a PV-inverter on the inside, the system will see that there is power that wants to flow outwards, and it will tell the Multi to import that. In other words, even without the extra energy meter (to proxy for the unsupported PV-inverter), the system already works correctly.
Without the extra meter, the numbers does not make sense. The system cannot accurately estimate the loads, and especially if the PV-inverter is on the output, when the reverse power (into the Multi) goes negative (which it can), the overview will show zero (it is capped at zero to prevent people from asking dumb questions). So that is why you want the meter, so it makes sense, and so you can see the data on VRM.
Since you cannot limit the power, when the batteries get full, it goes into the grid. Unless you ALSO fit the Solis’s own CT into your installation, and then there is some additional caveats I am not familiar with.
It would technically not be hard to integrate support for this into dbus-modbus-client. Just need to look at ev_charger.py and do something similar. All the building blocks for mapping registers onto dbus is there. Even making it limit power is possible according to the modbus documentation I have for it. Who’s game?
Correct, but you will loose the display and logging off the data. Small price to pay to have it available on the system. And we both know how difficult it is for Victron to display and calculate things if it doesnt know where power is coming from.
I have a colleague up North who runs a SolarEdge system like that. Too far away from the main system to tap into Sunspec, and he does not have the additional grid meter. The system works just fine, in the sense that the power is used for his loads, it charges the batteries if there is any left over. Literally the only issue he has is that the numbers don’t line up. And being in Europe, of course it does not matter if the batteries are full and the surplus starts to flow into the grid
I have an CG EM24, because I have 3 phase.
I was messing with its placement in VRM, maybe I am wrong, but it looked like it had the capability of being a 1Ph grid and also a PV inverter meter simultaneosly. @plonkster ?
Yes. You can configure it so that L1 is used as the grid meter, and L2 is used as the PV-inverter. Since Venus 2.80, you can also use both L1 and L2 as “acload” meters, to measure AC loads and log it to VRM. The feature was never extended to L3, because that’s basically a non-existent scenario (two phases and a single-phase PV-inverter), and when adding acload capability, it was easy to add this to L1 and L2, but also adding L3 would take WAY more work.
When L2 is used as a pv-inverter, we call it “piggybacking”. You can also do this with the ET340.
Right, we seem to say the same thing, yes?
I was thinking to connect the Solis direct to the main DD i.e after the main breaker on the main DB.
Then add the 2nd Carlo for the Solis, to also register it on the Victron system as a PV Inverter, as suggested.
How to connect that Carlo, it is not clear, but I’ll pay someone to do this in any event if I go this route.
That means one does not have to go and dabble with the dbus-modbus-client part, yes?
Still figuring out how to stop the Solis from feeding back if say the Victron is off, or there are insufficient loads if there is no current sensor for the Solis?
Unless the Carlo takes that role over for the Solis?