I had to implement the Fronius Zero feed-in on my CCGX/Multiplus because the Grid meter was changed to an electronic meter. i have all my load on the Multi put 1 and 2. Fronius is on Output 1.

Zero Feed-in works but is limiting the PV while there is plenty of charging to be done. setting ESS to without Battery life increases charge but still have PV limiting especially with fluctuating load. The charge limit is not holding it. if i switch to keep batteries charged then i get better charging but then grid draw goes up... so this is not working for me... I notice also that CCGX will drastically reduce PV output, very quickly, and then it takes its time to bring it back, bit by bit...

I have now switched Zero Feed-in on CCGX off. Re-installed my smart meter onto the Primo. My logic is that the Multi must get on with it function of charging etc. When the multiplus starts exporting the smart meter will pick it up and pull back PV. At least i am hoping that this direct control will be quicker than the sluggish Multiplus/CCGX.

Should this set up work? Anything else i could have done to the "official" way of linking Victron to Fronius to make it work better? Any settings that i need to pay attention to for it to work well, either with the smart meter, or CCGX way? i know they cant work together and i gather that Victron say it cant work any other way?

On 2019/08/10 at 5:22 PM, Keith said:

Zero Feed-in works but is limiting the PV while there is plenty of charging to be done.

I have seen that, when I don't run Optimized with Battery life. My CCGX/VenusGX is also set to Zero feed back and my batts get charged and loads powered to the full extent of what the array can produce at all times.

ESS ideally should be set to Optimized with Battery Life. Using Keep Batteries charge will use Eskom to charge as it is meant to be keep batteries charged AT ALL COSTS for like rolling blackouts which as we very well know, can happen day or night and every few hours.

 

Edited August 12, 20196 yr by Guest

18 minutes ago, The Terrible Triplett said:

I have seen that, when I don't run Optimized with Battery life. My CCGX/VenusGX is also set to Zero feed back and my batts get charged and loads powered to the full extent of what the array can produce at all times.

ESS ideally should be set to Optimized with Battery Life. Using Keep Batteries charge will use Eskom to charge as it is meant to be keep batteries charged AT ALL COSTS for like rolling blackouts which as we very well know, can happen day or night and every few hours.

 

Both with and without battery life unnecessarily limits my pv. Without life has less limiting. So i am not sure why in my case it is not operating right.

assume you have fronius zero feed-in on ccgx or venus?

yesterday i had zero feed-in with the smart meter control and good charging. with ccgx in control up to 3% feed-in and poor charging.

The ESS mode should have no effect on how the PV limiter works. Let me explain roughly what it does.

As soon as the sun is up and the Fronius turns on, the CCGX can see that it has turned on (sunspec model 101 or 103, there is an operating status field). When this happens, it increases the grid setpoint. There is some logic to how much, but generally it increases the grid setpoint by about 70 watts. In other words, it instructs the Multi and/or your Carlo Gavazzi to attempt to hold the grid import to whatever you set your grid setpoint to (on the ESS menu) PLUS 70 watts. Then it adjusts the Fronius to attempt to eliminate this 70 watts, in other words, it attempts to put the Fronius and the Multi into a fight where one attempts to pull it up by 70 watts and the other attempts to push it down.

This is implemented as a loop around the other control stuff and is almost entirely independent, hence should not be affected by any of the other settings. Any differences you see are most likely purely coincidental.

Also, for @The Terrible Triplett's benefit, this is AC-tied PV and works completely different to feeding excess DC-side PV into the grid. Completely different mechanisms.

On 2019/08/10 at 5:22 PM, Keith said:

i know they cant work together and i gather that Victron say it cant work any other way?

It can work... Victron just won't help you if it doesn't. Got to draw the line somewhere 🙂

I've advised several people with non-supported inverters that have their own smart meter to set it up this way. Simply set the grid setpoints in such a way that the PV inverter tries to push to the grid a bit harder than the Multi does, in other words, make them fight each other over 50 watts in the middle, and it should work just fine. Downside: higher AC setpoint at night.

19 minutes ago, plonkster said:

The ESS mode should have no effect on how the PV limiter works. Let me explain roughly what it does.

As soon as the sun is up and the Fronius turns on, the CCGX can see that it has turned on (sunspec model 101 or 103, there is an operating status field). When this happens, it increases the grid setpoint. There is some logic to how much, but generally it increases the grid setpoint by about 70 watts. In other words, it instructs the Multi and/or your Carlo Gavazzi to attempt to hold the grid import to whatever you set your grid setpoint to (on the ESS menu) PLUS 70 watts. Then it adjusts the Fronius to attempt to eliminate this 70 watts, in other words, it attempts to put the Fronius and the Multi into a fight where one attempts to pull it up by 70 watts and the other attempts to push it down.

This is implemented as a loop around the other control stuff and is almost entirely independent, hence should not be affected by any of the other settings. Any differences you see are most likely purely coincidental.

Also, for @The Terrible Triplett's benefit, this is AC-tied PV and works completely different to feeding excess DC-side PV into the grid. Completely different mechanisms.

It can work... Victron just won't help you if it doesn't. Got to draw the line somewhere 🙂

I've advised several people with non-supported inverters that have their own smart meter to set it up this way. Simply set the grid setpoints in such a way that the PV inverter tries to push to the grid a bit harder than the Multi does, in other words, make them fight each other over 50 watts in the middle, and it should work just fine. Downside: higher AC setpoint at night.

Thanks. makes some sense. however, if the multi is in a battery mode "with life" there is only a few hundred W charge. As soon as i change that to "Without life" then the charge goes up, in an observed instance from 900 W to 1.3 kW. it might have been better to have access to that 70 W (although it could be some proportional value? who knows?) to eliminate effects of fluctuating load.

the Multi is particularly fast in cutting the Fronius output if a load is switched off. for example an Iron switching on/off often as the thermostat operates. then the PV limit shown on the Fronius device page starts stepping up little by little.

In the control mode i have now i have a grid setpoint quite high 500 W. i had tried higher values when the ccgx was in control but that did not seem to help. if the grid is below the 500 at night then there is nothing that now tries to meet the 500, it just draws whatever the load is.

Ill watch it a few days and see what the graphs say. previously the PV on VRM looked like a rugged tooth shark got hold of it....:)

4 minutes ago, Keith said:

Thanks. makes some sense. however, if the multi is in a battery mode "with life" there is only a few hundred W charge. As soon as i change that to "Without life" then the charge goes up, in an observed instance from 900 W to 1.3 kW. it might have been better to have access to that 70 W (although it could be some proportional value? who knows?) to eliminate effects of fluctuating load.

In optimised mode there is one particular value on the Multi that runs the show, it's what we call the "hub4 setpoint", and it determines whether power is to be imported from the grid or exported to it. What it tries to do is to keep your "point of control" at zero (or close to it, whatever you set the AC setpoint to on the ESS menu, this is different to the hub4 setpoint I'm also talking about in this paragraph). This works in both directions, if there is power flowing out into the grid, then the setpoint is increased to attempt to import that power (and charge batteries with it). If power is flowing in from the grid, then the sets the setpoint lower (and even negative).

In all cases, if you set a setpoint that the Multi cannot comply with (eg you're telling it to import 1000W but it can't because it has less loads than that and the batteries are full), then the Multi simply does the best that it can.

As I explained already, when there is a PV inverter in the system, the hub4 setpoint is always offset by some 70W so that there is always a slight import: If the Multi wants to charge the batteries, it will always have at least 70W to do it with (and as it starts charging, the 70W margin is eliminated and the limit will be moved up until it charges at full power).

If the Multi is not charging the batteries... it is because for some reason it thinks that it can't. Hence there is nowhere to go with the power, and the limiter simply squeezes it away.

You have to figure out why the Multi doesn't want to charge.

You can DM me your email address and I'll get someone from support to contact you if you want someone to look at this in more detail.

3 hours ago, plonkster said:

The ESS mode should have no effect on how the PV limiter works. Let me explain roughly what it does.

As soon as the sun is up and the Fronius turns on, the CCGX can see that it has turned on (sunspec model 101 or 103, there is an operating status field). When this happens, it increases the grid setpoint. There is some logic to how much, but generally it increases the grid setpoint by about 70 watts. In other words, it instructs the Multi and/or your Carlo Gavazzi to attempt to hold the grid import to whatever you set your grid setpoint to (on the ESS menu) PLUS 70 watts. Then it adjusts the Fronius to attempt to eliminate this 70 watts, in other words, it attempts to put the Fronius and the Multi into a fight where one attempts to pull it up by 70 watts and the other attempts to push it down.

This is implemented as a loop around the other control stuff and is almost entirely independent, hence should not be affected by any of the other settings. Any differences you see are most likely purely coincidental.

Also, for @The Terrible Triplett's benefit, this is AC-tied PV and works completely different to feeding excess DC-side PV into the grid. Completely different mechanisms.

It can work... Victron just won't help you if it doesn't. Got to draw the line somewhere 🙂

I've advised several people with non-supported inverters that have their own smart meter to set it up this way. Simply set the grid setpoints in such a way that the PV inverter tries to push to the grid a bit harder than the Multi does, in other words, make them fight each other over 50 watts in the middle, and it should work just fine. Downside: higher AC setpoint at night.

This seems to be profound! currently i have the Multi grid set at say 250 W. On my Primo i have it set at 0 W. So my logic is saying that the PV is trying to neutralize the 250W by pushing harder. correct?

3 minutes ago, Keith said:

This seems to be profound! currently i have the Multi grid set at say 250 W. On my Primo i have it set at 0 W. So my logic is saying that the PV is trying to neutralize the 250W by pushing harder. correct?

Yup, that is exactly what I have in mind. The Primo tries to push it down to zero, and the multi tries to pull it up to 250W. The effect should be that while there is enough loads and/or space in the batteries for charging, you will import 250W and the Fronius will run flat-out. Then as the batteries starts to fill up, the Multi won't be able to import all of the power the Fronius is making and the Fronius will start to win the fight and push it down to zero. And then its limiter will activate.

20 hours ago, plonkster said:

Yup, that is exactly what I have in mind. The Primo tries to push it down to zero, and the multi tries to pull it up to 250W. The effect should be that while there is enough loads and/or space in the batteries for charging, you will import 250W and the Fronius will run flat-out. Then as the batteries starts to fill up, the Multi won't be able to import all of the power the Fronius is making and the Fronius will start to win the fight and push it down to zero. And then its limiter will activate.

So i set grid down yesterday to 250 W. Primo still on 0 W. last night no grid charging. and earlier on the system started charging. had a look at the PV graph on VRM portal. di do a little pull back at about 1148, that's ok. tried a little lower than 250W but then it is a bit close so i think for now i will leave it at 250W and watch it.

Much better results.

i am going to add a MPPT 250/60 with 8x330 Wp panels on the DC side and doubling the battery capacity, now 200 Ah. plus at some later stage i can add PV on MPPT2 of the Primo, in a different direction to extend the noon harvest into late afternoon. getting there, slowly, bit by bit.

first screen shot is today smart meter/Primo control. second screen shot is ccgx/multiplus effort.

Edited August 13, 20196 yr by Keith

Hi I pulled in an ethernet cable for the comms between Fronius and my router. Also assigned to ccgx/multi to do the control/charge.

a comment i got from one installer where i got my Multi from said the control is bad because 1. that comms to the battery must be good else ess dont work well. (My system has Li-ion and the bms not supported. So the Multi is the  charge controller. Is this a problem? and would a BMS 700 make matters better? Victron ESS says if the multi is the only charger i dont need a BMS) 2. I dont have DC MPPT on my system. (Well that would add an additional pv source and add another thing to control... So i am not entirely convinced that it is not handling things well because i lack a MPPT on DC side.)

All loads are on AC out 1 and 2 so i dont need a victron power meter. the Multi does this measurement. however, would a meter improve the control?

On 2019/08/17 at 6:05 PM, Keith said:

Hi I pulled in an ethernet cable for the comms between Fronius and my router. Also assigned to ccgx/multi to do the control/charge.

Wait? You didn't have that? I was pretty sure I saw the PV-inverter on the VRM device list?

The ESS limiter uses modbus-TCP... which means it needs TCP, aka internet. It needs to be able to talk to the PV-inverter. There is a part of sunspec specifically implemented for limiting the power, though only a few PV inverters fully support it (to my knowledge only Fronius and ABB does it properly, and SMA has caught up in their latest firmware, or so I am told).

Just follow the instructions for setting it up, especially the part about the DNO editor here. It's a common mistake to set the Limit option, but it should be set to unlimited (non-intuitive, but otherwise the inverter ignores the sunspec instructions and simply runs flat-out). Especially when the inverter used to run with its own smart meter and you convert it to Victron ESS, it is easy to overlook this setting.

On 2019/08/17 at 6:05 PM, Keith said:

comms to the battery must be good else ess dont work well

Nonsense. ESS can work with "dumb" batteries just fine, as long as you have some way to estimate the SOC of the battery. And even without that, it will still work fine, but it might not be so good for the batteries.

On 2019/08/17 at 6:05 PM, Keith said:

So the Multi is the  charge controller. Is this a problem?

No. You don't have DC loads or DC chargers (aka MPPTs). The Multi is the only way into and out of the battery, so it is perfectly capable of tracking the SOC. Just make sure the charging efficiency configured in VE.Configure is accurate. I think the default is 85% and this is too low for LiFePO4. When this setting is too low, your SOC estimate will tend to drift downwards (ie it will underestimate the SOC of the batteries).

A BMV is slightly more accurate, but in your setup I don't think it's warranted.

On 2019/08/17 at 6:05 PM, Keith said:

All loads are on AC out 1 and 2 so i dont need a victron power meter. the Multi does this measurement. however, would a meter improve the control?

No need to add a meter. That just makes the load tracking even slower (because of the overhead in communicating with the meter).

Thanks. No problem setting up the fronius. Just below the fronius zero feed switch is a no if things are not right in the settings but mine is yes. So ill check on how well it is controlling.

i was also concerned with the frequency in ess start at 50.2 hz and i have been seeing eskom drifting up to 50.36  hz on vrm. But in the documents it has start at 51 hz. So i went in earlier to change that. Confusing documents. 1. Says no need to change the frequency. Next point 2. Says change it and gives a table for 50 and 60 hz 🤷‍♂️.

1. Note that it is not necessary to change the default settings in the Assistant. 
    
2. Complete the rest of the Assistant and write the new settings to the MultiPlus or Quattro.

 

Table for settings for 50Hz and 60Hz frequency

State	50Hz	60Hz	.
Start	51.0	60.2	Hz
Minimum	52.7	62.7	Hz
Disconnect	53.0	63.0	Hz

More information on adding Assistants is here.

 

Thanks. No problem setting up the fronius. Just below the fronius zero feed switch is a no if things are not right in the settings but mine is yes. So ill check on how well it is controlling.

i was also concerned with the frequency in ess start at 50.2 hz and i have been seeing eskom drifting up to 50.36  hz on vrm. But in the documents it has start at 51 hz. So i went in earlier to change that. Confusing documents. 1. Says no need to change the frequency. Next point 2. Says change it and gives a table for 50 and 60 hz .

1. Note that it is not necessary to change the default settings in the Assistant. 
    
2. Complete the rest of the Assistant and write the new settings to the MultiPlus or Quattro.

 

Table for settings for 50Hz and 60Hz frequency

State	50Hz	60Hz	.
Start	51.0	60.2	Hz
Minimum	52.7	62.7	Hz
Disconnect	53.0	63.0	Hz

More information on adding Assistants is here.

49 minutes ago, Keith said:

no need to change the frequency

The defaults are usually good enough. But sometimes you may have to change it. The Multi does a closed control loop anyway, it adjusts it upwards until it reduces enough, so by modifying these settings you may just cause the Multi to react a bit faster. For example, if you tell the Multi that power reduction only starts at 51Hz, but the Fronius actually starts at 50.2Hz, then the outcome will (likely, I have not tested it) be a much faster initial reduction in power.

The closed control loop actually works even without a PV-inverter. I learned this completely by accident. Some manufacturer in Europe sells a device that diverts excess PV into your geyser (hot water cylinder for the non-saffers). It works by looking at the frequency. Normally it is installed with a PV-inverter, so that the element is turned on as the PV-inverter is limited by the Multi. The manual said however that it works even without a PV-inverter... and upon testing I found this is 100% correct. If you install the PV-inverter assistant (which is also included in the ESS assistant) on an off-grid system, it raises the frequency as the battery voltage goes up, so even if there is no PV-inverter and the PV is DC-tied, you can still make it work, the MPPTs just has to charge a little higher than the Multi so that the "overvoltage" causes the frequency to be raised.

However, in the case where these are sitting on the grid, the frequency can not be changed by the multi. Only if it disconnects off the grid and runs on inverter can it dictate the frequency. Thus they are not using frequency in the case of fronius to do pv reduction.

if one selects the fronius inverter on the ccgx devices. In the table below L1, 2, 3. There is a value of inverter limit. In my case when it is not limited it is 4000w it seems this is sent to the fronius as a set point. Ccgx then changes this number down to whatever new set point below 4000 is needed. Then as matters settle the number gradually climbs back up, slowly. This must be sent to fronius by modbus tcp. It is also recorded in the last column of the vrm data download.

I find the multi very sluggish to increase charge power, it seems the first priority is to reduce the fronius power. I have asked victron but that community site is most unresponsive. And i am not the only one who sees the pv limit operating in stead of increasing charge power.

 

5 minutes ago, Keith said:

However, in the case where these are sitting on the grid, the frequency can not be changed by the multi. Only if it disconnects off the grid and runs on inverter can it dictate the frequency. Thus they are not using frequency in the case of fronius to do pv reduction.

Right. You can't shift the grid. The stuff I was talking about is for people using PV-inverters in off-grid systems. It's more common than you'd expect 🙂

6 minutes ago, Keith said:

I find the multi very sluggish to increase charge power, it seems the first priority is to reduce the fronius power. I have asked victron but that community site is most unresponsive. And i am not the only one who sees the pv limit operating in stead of increasing charge power.

As I said earlier, the two are independent. The Multi has its own brain (a little 8-bit microcontroller) and it does its own thing. The limiter module for the PV-inverters runs on the CCGX (or Venus-GX, Venus device for short) and doesn't care about what the Multi or the solar chargers are doing, it only looks at the power level and makes adjustments as necessary. That is to say, the limiter follows what the Multi is doing rather than the reverse.

If the Pv-inverter didn't limit the power it would not make the Multi any faster. It would just end up pushing into the grid.

I still don't understand why it does that though. Normally the Multi will spool up to full charge power in a matter of seconds.

1 hour ago, plonkster said:

Right. You can't shift the grid. The stuff I was talking about is for people using PV-inverters in off-grid systems. It's more common than you'd expect 🙂

As I said earlier, the two are independent. The Multi has its own brain (a little 8-bit microcontroller) and it does its own thing. The limiter module for the PV-inverters runs on the CCGX (or Venus-GX, Venus device for short) and doesn't care about what the Multi or the solar chargers are doing, it only looks at the power level and makes adjustments as necessary. That is to say, the limiter follows what the Multi is doing rather than the reverse.

If the Pv-inverter didn't limit the power it would not make the Multi any faster. It would just end up pushing into the grid.

I still don't understand why it does that though. Normally the Multi will spool up to full charge power in a matter of seconds.

maybe it only looks bad on the VRM trends. I took a look earlier. Monday is wash day so not much spare PV. Still there was a few control points. I exported some out of VRM. I note that every so often there is an error, without a reason code. normally at night i would get the code in that field. Should not be as i have a cable in for the comms.

Wondering how i can track this. PV pumps all the time it is not disconnecting, and as far as i know the Fronius is always present on the ccgx, don't see it disappearing as was the case when i was on wifi and fronius missed with the wifi power.

Attached PDF output of VRM

17Perth_20190819-1000_to_20190819-1507 1.pdf

Good day Plonkster.

So yesterday pm i disconnected the smart meter modbus. Got the Fronius comms cable and ccgx direct onto the router. i have had a look at the last 12 hours vrm output. Still a few PV errors but they don't seem to affect the charging. They are momentary, maybe a timeout?

The reason why the load seems to fall away when the error is there is because the output power is the sum of the number ccgx gets from the fronius plus what the multi is measuring on the output. so if the ccgx get nothing from the fronius (error probably i asked but got no response?) then it logs an error and the measured load. Charging continues, fronius is happily generating, but for some strange reason did not respond, time out maybe?

So i think possibly a few things. initially the fronius firmware update i installed when i did zero feed in set up messed with the wifi strength, so i got bad comms. Fronius is on the edge of my router range. Then i reinstalled the smart meter to try something different. This worked ok. But i had already ordered an ethernet cable and at the same time Fronius had a new firmware to fix the wifi strength. Seems if one leaves the Smart meter in and modbus connected to the Fronius (monitoring only) it might clash with the CCGX tcp modbus communication (well, i have a LOT less errors and they don't seem to be affecting the operation).

So my understanding on VRM logging is that irrespective of the log interval it will still log every change of state. Thus in the log you had yesterday the log lines are not 30 minutes apart.

Thanks for all your responses. Think a whole lot of issues combined, but seems to be running nicely now.

 

On 2019/08/19 at 2:08 PM, plonkster said:

It works by looking at the frequency

would this diverter work with a Goodwe hybrid GTI set at zero grid feed in? I am still trying to find an automated solution to divert excess PV power. 

3 minutes ago, Fuenkli said:

would this diverter work with a Goodwe hybrid GTI set at zero grid feed in? I am still trying to find an automated solution to divert excess PV power. 

Best ask Plonkster. i have only been wrestling with Fronius and a Multi.

3 hours ago, Fuenkli said:

would this diverter work with a Goodwe hybrid GTI set at zero grid feed in? I am still trying to find an automated solution to divert excess PV power. 

I don't know. The product is the Elwa-F, which it seems might be discontinued. Here is a link to  the Elwa-E. It's a European product, no idea if it is available locally.

3 hours ago, Keith said:

So my understanding on VRM logging is that irrespective of the log interval it will still log every change of state. Thus in the log you had yesterday the log lines are not 30 minutes apart.

Some things are logged immediately (like alarms, certain state changes, etc). Others are logged only on interval, which for you was 30 minutes. So the effect is unevenly spaced records, sometimes two or three within a minute, and then nothing again for 20 minutes or more. By lowering the log interval this still happens, but the gaps are smaller.