Hi

I have a question regarding PV inverter compatibility with victron inverters. 

Can you use any grid tie inverter on the AC-in side of a victron or only the listed ones (fronius, ABB, solar edge?)

My understanding is if you use "non-listed" pv inverters it wont/cannot be controlled via the GX device, but can still be installed and will be like two separate systems with separate control and monitoring?

My thinking is Three phase grid-tie system for daylight use and single phase victron with DC coupled PV to sustain backup loads. 

Most recently made grid-tie inverters can be put on the AC out of a Victron inverter.  Hoymiles micro inverters even have frequency shifting nowadays.

There are a number of standards but I can vouch that if it has the "VDE-4105" standard it seems to work, relatively stable.

The problem is if the Victron inverter is also grid-tied it has no mechanism to control the power output of the grid-tied inverter.

In other words it cannot perform frequency- shifting, so it will export to the grid, without some sort of throttling intervention.

There should also be 5kWh of LA battery for every 1kWh of PV inverter or alternatively  5kWH of Lithium battery for every 1.5kWh of PV inverter.

Also, there is a 1:1 max ratio of PV inverter to Victron inverter rule that should be observed.

What is lesser appreciated is that this downstream PV inverter can still be used to charge batteries as the Victron is a bi-directional device.  PV inverter use far higher dc voltages and so the panels can be a fair distance away without incurring the cost of preventing copper losses.

I use mine downstream of a Phoenix inverter, (the Phoenix has no AC-in), so it regulates by using frequency all day long. I have a small Solis 1.5 and an ABB uno 3kW downstream of it.

They were secondhand, I just plugged them, set up the charger on the Phoenix and they worked.

 

Let's just look at compatibility when you have the PV-inverter on the input side (cause then I don't have to discuss frequency shifting and all that stuff).

If you want to use the PV-inverter as a part of ESS, and you don't want to feed into the grid, then you must use a Fronius or ABB inverter. The reason it was limited in that manner is because there are inverters out there that appear to support the bits needed for grid limiting, but don't. So it will only explicitly support the ones it knows to work well.

If you are allowed to feed into the grid, then you can use whatever you want. ESS will see  the exported power and attempt to import it to charge batteries.

Then for the second half of the answer: If the PV inverter supports at least the basic mandatory bits of sunspec, the Venus device (aka CCGX or Venus-GX) can read power values directly from it. If the PV-inverter does not support it, then you can install a another carlo gavazzi energy meter and configure it to take the place of the PV inverter (so you can see that data).

So now I get to the brand names. SMA only recently fixed their sunspec implementation. Some missing bits in the older firmware means it simply could not be supported. So they are out. SolarEdge implemented only the most basic bits of sunspec (and rumour has it they are leaving the alliance), so I doubt SolarEdge will ever be supported. Both these brands show up on the venus device, the data is logged to VRM, and as long as you are allowed to feed into the grid it does work with ESS.

 

Thanks for the replies

I read the whole CCGX manual now and saw you can use a meter to get the pv inverter data onto the gx. 

Using a three phase Hoymiles system running with grid limiting on its own, (i can export to an extend because of old conventional eskom meters) and connected on one of the hoymiles phases is a 10kVA quattro on AC-in that only supports backup loads on AC-out.

Questions on this system:

1. When connecting a meter to the three phase supply, i gather i would be able to see all data as below:

AC loads box shows all three phases and power consumed from the hoymiles system. Critical loads shows that connected to the quattro

2. The Hoymiles system will switch off in the event of a power failure?, or will the phase on which the quattro is connected stay on as the quattro gives a "grid" from the  battery?

3. Should a single phase meter be connected to the quattro phase for ESS? 

4. Is it advisable to add DC coupled PV to the quattro to sustain those loads and prevent battery deadlock? Could the DC coupled pv replace 1 of the hoymiles phases?

3 minutes ago, MrVolt said:

1. When connecting a meter to the three phase supply, i gather i would be able to see all data as below:

If the system is aware of all three phases, it will display and log data for all of them even if you only have a Multi/Quattro on L1. You system seems to be working just fine.

4 minutes ago, MrVolt said:

2. The Hoymiles system will switch off in the event of a power failure?, or will the phase on which the quattro is connected stay on as the quattro gives a "grid" from the  battery?

You said it is on the input. Then it will go down when the grid goes down. The Multi does not power things on its input side (in fact, it purposefully avoids it).

5 minutes ago, MrVolt said:

3. Should a single phase meter be connected to the quattro phase for ESS? 

What for?

6 minutes ago, MrVolt said:

grid limiting on its own

I have to say something about this. Me and @Jaco de Jongh had some fun with such a system (someone spec'ed it without really looking at what is fully supported). It is possible to get yourself into a deadlock situation. If you have two systems both attempting to zero the grid supply, one of them might end up locking the other one out. For example, imagine you have only a few small loads running and then the battery suddenly fills up (as LFPs do, they stop charging like a plane stops against a mountain!). As the grid feed-in suddenly shoots down and negative, both the Quattro and the PV-inverter stops feeding energy into the grid. Then the Multi starts restarts first...

Now the PV inverter will never start back up again because it doesn't see any load: The Quattro is using battery power to hold it at zero.

You therefore have to configure the Victron system to import a bit of power, I would say at least 1% of what your PV inverter capacity is. The idea is that the PV-inverter tries to push it down to zero while the Victron system tries to pull it up to some low value, or in other words, the Victron systems always leaves a bit of load left for the PV-inverter to pick up. Without that, you could end up in a situation where the PV-inverter stops and doesn't restart.

The downside to that configuration is that once the sun is down, the Victron system will continue to have a higher grid setpoint and it will import a couple kWh from the grid at night. Which is why you want to use a supported PV inverter 🙂

 

9 minutes ago, plonkster said:

If the system is aware of all three phases, it will display and log data for all of them even if you only have a Multi/Quattro on L1. You system seems to be working just fine.

You said it is on the input. Then it will go down when the grid goes down. The Multi does not power things on its input side (in fact, it purposefully avoids it).

What for?

I have to say something about this. Me and @Jaco de Jongh had some fun with such a system (someone spec'ed it without really looking at what is fully supported). It is possible to get yourself into a deadlock situation. If you have two systems both attempting to zero the grid supply, one of them might end up locking the other one out. For example, imagine you have only a few small loads running and then the battery suddenly fills up (as LFPs do, they stop charging like a plane stops against a mountain!). As the grid feed-in suddenly shoots down and negative, both the Quattro and the PV-inverter stops feeding energy into the grid. Then the Multi starts restarts first...

Now the PV inverter will never start back up again because it doesn't see any load: The Quattro is using battery power to hold it at zero.

You therefore have to configure the Victron system to import a bit of power, I would say at least 1% of what your PV inverter capacity is. The idea is that the PV-inverter tries to push it down to zero while the Victron system tries to pull it up to some low value, or in other words, the Victron systems always leaves a bit of load left for the PV-inverter to pick up. Without that, you could end up in a situation where the PV-inverter stops and doesn't restart.

The downside to that configuration is that once the sun is down, the Victron system will continue to have a higher grid setpoint and it will import a couple kWh from the grid at night. Which is why you want to use a supported PV inverter 🙂

Could this deadlock/situation be prevented if going for 2 phases on hoymiles system with a Meter to log data to the GX and then the third phase on a DC coupled Quattro system?

 

8 minutes ago, MrVolt said:

Could this deadlock/situation be prevented if going for 2 phases on hoymiles system with a Meter to log data to the GX and then the third phase on a DC coupled Quattro system?

It's got nothing to do with meters or anything. The deadlock happens because two systems are doing the same job and one of them gets there first. To prevent it, make sure one of the systems doesn't hog all the work.

For example, if you set the Victron's grid setpoint to 200W, then it always tries to import 200W (for loads, of to charge the batteries). So the PV inverter's limiter sees 200W and ramps up power... the Victron system sees less than 200W and increases its import... and so it goes in a circle until the PV-inverter is maxed out.

4 minutes ago, plonkster said:

It's got nothing to do with meters or anything. The deadlock happens because two systems are doing the same job and one of them gets there first. To prevent it, make sure one of the systems doesn't hog all the work.

For example, if you set the Victron's grid setpoint to 200W, then it always tries to import 200W (for loads, of to charge the batteries). So the PV inverter's limiter sees 200W and ramps up power... the Victron system sees less than 200W and increases its import... and so it goes in a circle until the PV-inverter is maxed out.

Ok thanks. Best then to enable export or just let the hoymiles do limiting and the quattro to feed in?

On my question regarding 2 phases on hoymiles and one on the quattro, will this not be a better solution? If thats the case, still only one 3phase meter is used for data logging on the hoymiles side and logging on the quattro side? Not sure if you can set the GX to which phase is which system on the meter?

 

 

51 minutes ago, MrVolt said:

Ok thanks. Best then to enable export or just let the hoymiles do limiting and the quattro to feed in?

I'm still not quite sure what you want to do. An ESS system has a single meter on the input side (a carlo gavazzi) and the system tries to keep that meter at a specific value (default is 50W, but you can change it: It's the grid setpoint). If you have PV inverters in the system, even if they aren't supported and the Venus device doesn't know about them... it still works. The power flowing out into the grid pushes the import below the 50W setpoint and the system will counteract this by attempting to import that power.

It will import that power as long as there are loads or batteries to charge. Thereafter it just does the best that it can.

So if you have PV inverters on other phases, as long as you have a three-phase meter installed (an EM24 or ET340), it will just work. It will see the reverse power and attempt to import it. Even if that power is on another phase, it will still attempt to import it, balancing across all phases.

Remember that the Multi/Quattro always goes on L1. It's not a real limitation, since it doesn't matter which one you use as L1 as long as L2 and L3 follows in the correct order (so-called phase rotation).

10 hours ago, plonkster said:

I'm still not quite sure what you want to do.

I made a drawing of what id like to achieve. From the feedback you gave I know it will work, thanks. All that I'm not clear on is whether the GX device will show the PV inverter (Hoymiles) data as the screenshot in the post #4 above. 

 

28 minutes ago, MrVolt said:

All that I'm not clear on is whether the GX device will show the PV inverter (Hoymiles) data as the screenshot in the post #4 above.

If you want to see the power from the Hoymiles inverters (I assume they don't do modbus-tcp and sunspec?), then you must install a second Carlo Gavazzi meter in the yellow lines. Wire up only L2 and L3 and leave L1 unused (though you may have to apply power to the terminal). Configure that energy meter to be for a PV-inverter.

Unfortunately you cannot use two ET112 meters.

Other than that your picture is spot on. Just don't try to run the Hoymiles limiter in parallel with the Victron ESS one. If you do... make sure the Victron system leaves a healthy amount of power for the PV-inverter to pick up, in other words, jack up the grid setpoint.

Edit: If you wire it like this, then the Victron system will attempt to zero across all three phases. That means it will attempt to import power on L1 that is equivalent to what is being exported on L2 and L3. This assumes you are billed for the total across all three and not individually.

Edited September 9, 20196 yr by plonkster

I have just commissioned a system approved by COCT.

It consists of three 5KVA Multiplus inverters (the types that are not approved to NRS-097) married to a low cost three phase Solis converter on the A/C output side. The Solis is the 15KW G4 version which is COCT approved.

The Victrons are configured as UPS with the PV Inverter assistent using frequency shift from 50.5 to 52Hz to control the Solis output power in case of a grid failure. The Victons, now being a UPS fully compliant to SANS is no longer part of the SSEG application but is still shown on the diagrams submitted to COCT. The main point is that the SSEG is now NRS-097 compliant.

In addition to the above a Solis Export manager is used. This is used to regulate maximum allowed grid feed in in (or zero feed in as the case may be).

So from an operational point of view if grid is available the Victrons are in passthrough with inverters off.Only the battery charger is active. The Victrons are set to UPS mode so they can switch on in case of grid failure within 20mS. Any export (should that be desired) goes through the Victrons "backwards".

If Grid fails the Victrons open their passthrough relay (disconnecting the grid) - and in my case I have a further Ziehl anti-islanding solution downstream that also opens up (also to NRS-097). The Solis continues to produce and supplies local loads with the Victron regulating the output power via frequency shift. This works very nicely. The Victron uses the batteries as temporary load during load dumps to cater for the short periods where it needs to adjust frequency. For this reason there is a limit on the maximum power production your grid tie is allowed to do during this operating condition. Victron has the 1:1 rule for that - effectively your Victron inverter/charger must not be less than the grid tie in potential power production.

It's a bit more complex in detail as your battery capacity also comes into it but in general as long as the inverter is not over specified relative to the Victron there is no issue you will ever notice.

I used to use the Victron ESS before going this route. The new solution does have the added advantage that it is more power efficient as during normal (grid on) mode the Victrons use much less power than before as there is no need to run the inverters. This is particularly noticeable at night where the ESS base consumption accounts for 1/4 of my base consumption at night. Charging the batteries is now from grid (as required by COCT) so recharge after a power outage is less efficient than it was before - but it is a small price to pay overall.

Replacing the Victrons with the newer NRS approved inverters was not an option - firstly that is quite expensive but in my case I have the 100A transfer switch versions which I need and the newer types can only do 50A - that would force me to split by DB boards and in my case that is going to be a mess.

So - bottom line - for anybody stuck with the old Victrons a usable solution to get compliant is to bolt on a Solis at relatively low cost and "downgrade" the Victrons as UPS. Will work for single phase as well and you do not need the export manager as that is now built into the smaller Solis inverters (install the CT on the Victron A/C input). This whole setup is simple and quick to wire and get running. No complication anywhere and it does work very well.

 

Edited September 9, 20196 yr by The Bulldog

18 minutes ago, The Bulldog said:

In addition to the above a Solis Export manager is used

Is it possible to set a the "grid setpoint" with the Solis? In other words, can I tell it to hold the grid at 100W or 50W instead of zero?

Also, I'd love to poke in the guts of a Solis... if anyone has it wired up for modbus comms.

10 minutes ago, plonkster said:

Is it possible to set a the "grid setpoint" with the Solis? In other words, can I tell it to hold the grid at 100W or 50W instead of zero?

Also, I'd love to poke in the guts of a Solis... if anyone has it wired up for modbus comms.

With the Zero export setting it keeps the setpoint at 50W import very constant. I also noticed it reacts pretty fast to load dumps which should help those with overly suspicious pre-paid meters.

There are a lot of setup items in the Solis menu but the one thing that is lacking is decent documentation on them - I am in contact with Ginlong tech support though and they do seem quite responsive. I have a lot of questions.  🤨 I have the NRS-097 test report in detail - that helped a lot in answering some of the more intricate questions.

You are most welcome to poke into my Solis - it's got the RS485 feed and we have the protocol too. I have a prewired RS485 plug and cable so if you have something that can hook up - I'm usually around Saturday mornings. 

 

3 minutes ago, The Bulldog said:

we have the protocol too

Can you send that to me? I assume it's not stock standard sunspec... one can hope, right? 🙂

 

Just now, plonkster said:

Can you send that to me? I assume it's not stock standard sunspec... one can hope, right? 🙂

 

RS485_MODBUS-Hybrid-BACoghlan-201811228-1854.pdf

3 minutes ago, The Bulldog said:

RS485_MODBUS-Hybrid-BACoghlan-201811228-1854.pdf 147.68 kB · 0 downloads

Thanks for that. Now I know at least it's not sunspec. Which is a bit of a shame.

There are a number of projects on Github interfacing to Ginlong/Solis inverters.

Here is one: https://github.com/amfasis/ginlong-influx/blob/master/ginlong-listen.py

This one simply connects to the Wifi stick that seems to be supplied as standard with the Solis inverters. The Wifi stick in turn plugs into the inverters RS485 port.

There are a number of different protocols floating around - seems the current one applies for the G4 devices. I am not 100% sure if the protocol doc above would work for my Solis. Guess I need to hook up a serial port monitor and see what data appears on the bus.

11 hours ago, plonkster said:

Edit: If you wire it like this, then the Victron system will attempt to zero across all three phases. That means it will attempt to import power on L1 that is equivalent to what is being exported on L2 and L3. This assumes you are billed for the total across all three and not individually.

Not sure what you mean here, is it because there is a 3phase energy meter that measures the main consumption? This is the same meter that controls the victron, so export on the other phases and the victron will import to balance the phases? 

We are billed on the total consumption, but you cant go into negative readings on two of the phases. 

If the PV inverters are on separate phases than the victron as in the picture, wont they also operate separately in terms of grid limiting and import/export?

2 hours ago, MrVolt said:

is it because there is a 3phase energy meter that measures the main consumption?

Correct. And the Victron system won't balance the phases (very much the opposite actually!). It wants to save you the most money possible. If there is surplus PV (even if it is on another phase), it wants to use that for loads (on L1) or for charging the battery.

2 hours ago, MrVolt said:

If the PV inverters are on separate phases than the victron as in the picture, wont they also operate separately in terms of grid limiting and import/export?

If you use a three phase meter for ESS, then the Victron system will know about the power on the other two phases and will try to use it. So if you want them to be separate, you can't let the Hoymiles feed through that same meter.

I used to use the ESS on a three phase installation. It definitely has a  bug or two so be aware of that. I have tried my best to sort these out with Victrons tech support over the years and usually there where improvements but until last week it was not yet 100%. What typically happens is that under certain (unkown to me) conditions it will decide to feed large amounts of power into one phase only to import it back on another phase. This can extend to almost the maximum capacity of the inverter so is not trivial. I have the gut feel this has to do with just how the own consumption loads are like at the time. It is not related to any particular phase and can occur on any phase combination. When this happens resetting the CCGX does not fix the issue - once it restarts it merrily continues in the same way. This behavior is quite random it seems. I eventually gave up on this and just lived with it. It does not happen all the time. But if it does it is as persistent as a foot fungus.

I used to notice this typically because of the noise the Victron's fans make when they are running at high load - I would hear them and scratch my head as there where no big loads at the time and a quick punch of the buttons on the CCGX tells the story. Seems they are so eager to make power they will do anything... 🙂

1 hour ago, The Bulldog said:

What typically happens is that under certain (unkown to me) conditions it will decide to feed large amounts of power into one phase only to import it back on another phase

I remember that. It happened right when SVS was briefly pulled out, and improved when it was put back in. One phase thinks the battery is low and charges it, the other things it is too high and starts feeding power into the grid. You're quire right about the CCGX not being involved in that, it's something to do with voltage measurements and how the Multi/Quattro handles that. I think it was suggested that you swap two of the units (because voltage measurement is done by the master on L1, so swapping units changes the dynamics) but you decided against it.

Anyway, I'm glad you have a system that works now. This ESS stuff comes up with some pretty interesting corner cases.

27 minutes ago, plonkster said:

I remember that. It happened right when SVS was briefly pulled out, and improved when it was put back in. One phase thinks the battery is low and charges it, the other things it is too high and starts feeding power into the grid. You're quire right about the CCGX not being involved in that, it's something to do with voltage measurements and how the Multi/Quattro handles that. I think it was suggested that you swap two of the units (because voltage measurement is done by the master on L1, so swapping units changes the dynamics) but you decided against it.

Anyway, I'm glad you have a system that works now. This ESS stuff comes up with some pretty interesting corner cases.

I did swap out the units (i.e. reallocated the phases) in all combinations possible - but it made no difference at all. That was the point I gave up and just lived with it (since I had an old style mechanical meter at the time it only charges for the net power). 

It was just random - could not put my finger on the condition needed to get this going wrong. Yes when SVS was removed it went catastrophic - but I am not referring to that as this was quickly abandoned. I'm pretty sure it has something to do with how loads are distributed across the phases - in particular when one of the phases is less than 100W which sometimes happens. In all cases where it went wrong the remaining phase was on a very low load when I activate the screen on the CCGX. If I then go and switch on a high load on that phase the problem sorts itself out. The weird thing is that this is not consistent - at other times you can stand on your head with a similar load picture and it works just peachy. The other thing I noticed is that this seems to only happen if the batteries are fully charged. The  batteries have to be on float for this to happen it seems. Perhaps that is a hint ?

 

43 minutes ago, plonkster said:

Anyway, I'm glad you have a system that works now. This ESS stuff comes up with some pretty interesting corner cases.

I am not saying it was not working - it just still has room for improvement. ESS is a fantastic system - it really is and I am very sure, given some more time, it will be the gold standard if it is not that already. Three phase can get tricky - I'm sure ESS is bullet proof when run in a single phase system.

Had our first grid outage yesterday - new system handled it great. Just sends unnecessary over-frequency grid alerts when the Victron goes and shuts the Solis down completely (>52 Hz). Might have to tweak this a bit...

4 minutes ago, The Bulldog said:

m pretty sure it has something to do with how loads are distributed across the phases

On the surface one would think that, but I've studied that part many many times and there is simply no shuffling-across of loads from one phase to another (I'm the support guy who tried to figure it out!).

The grid setpoint sent to each phase accounts only for loads on that phase and the only time it will import extra power on a phase is if there is an export on another phase that needs to be cancelled out. Eg if you have a situation like this

    Export | Import
L1:        | ----->
L2:  <---- |
L3:        | ----->

Then it will adjust it to look like this:

    Export | Import
L1:        | ------->
L2:        |
L3:        | ------->

In other words, it will shuffle the export on L2 onto L1 and L3 so that all three phases are doing the same.

In this case the Quattro was acting on its own, it was seeing an overvoltage and feeding it into the grid, while another phase saw a low voltage and tried to charge. That is why removing SVS made it worse. I still don't know what caused it or how to solve it.