question

mysik avatar image
mysik asked

Victron ESS + ET112 = slow response

Hi,

Recently I've installed ET112 grid meter on my system (which was using CT sensor previously) and I've noticed that the Multi is very slow on response to load. Currently my system looks like this:

The grid set point is set to -10W. The multi needs a good 15-18 seconds to stabilize after large load kicks in and another approx 15s to stabilize after load drops off. Check the video from VenusGX (the load is switched on at about 16s of the clip and switched of at 43s of the clip):

Is that how it should be? The difference in the load is just about 1.4kW, why is it taking so long to handle that?

Previously I was using CT Sensor extended with Cat5e cable to over 25m long, the system looked like this:

The response to load was pretty much instant. I've recorder video using the same load as on the previous one (load switched on at about 10s and switched off at 23s of the video):

The system need seconds to stabilize

Also you can see a clear difference on the Grid use graph from VRM portal:

As you can see, with CT sensor the system is very close to grid set point, with ET112 grid meter it is way off and hardly ever stay close to set point. The situation is even worst during cloudy day when PV power is fluctuating (you cen see it on my other youtube video).

Is it normal behavior? It looks like Multi can handle this load easily, the problem seems to be in the feedback from the grid meter (or algorithm when ET112 is in use). Is there anything I can do to improve this reaction time??? Please advice.

Kind Regards

Multiplus-IIESSgridct
ess-diagram1.jpg (20.0 KiB)
ess-diagram2.jpg (23.6 KiB)
capture.png (29.4 KiB)
capture1.png (32.1 KiB)
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

16 Answers
Meine_Energiewende avatar image
Meine_Energiewende answered ·

This looks exactly like my system with an ET340 System. I even did not know that i can respond fast with a CT Sensor. Do you know if 3 Phase setups are possbile with the CT Sensor ? It would be a very nice option for me then.

Thx

Jens

1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Maciek avatar image Maciek commented ·

Yes. You can use 3 external CT or internal in MPII.

0 Likes 0 ·
Rob Duthie avatar image
Rob Duthie answered ·

Hi

Remember a CT is a analog device and will be much faster in doing calculations than a digital meter like the ET112 which does have built in delays in the baud rate and in its sending of data out

which can be set in the ET112 configuration software you can set baud rate and DMD calculation time etc to upload the data.

Note: the ET112 is way more accurate (Meter Grade Quality) in measuring values than the CT is, the CT is only average readings, as advised by Victron. And they advise the setting of the grid set points should be greater than -10 around about -50 or +50 as the inverters are not revenue grade type readings as a Fronius inverters etc which are. Standard setting is 50. If read the ESS manual guide it tell you this.

I have the same setup running all the latest firmware updates (need to check yours?) which has a lot of improvements regards to readings in power etc. And i have not noticed any lack of speed of the readings which appears to be around a second. I have mine set to zero grid set point, which works OK in my application.

And i have zero feed in and all export when i have surplus PV power. Also mine is running over a wireless RS485 link as well.

Regards

Rob D

NZ

3 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mysik avatar image mysik commented ·

If I can see a load measurements that I have on the GX screen are from ET112, so the system knows that there is a change in the load, it is just not reacting fast enough.

It pretty much looks like ET112 is sending data to VenuxGX, but Multi is lagging behind. So from what I can see the problem is more in VenuxGX or/and Multiplus settings or comms between them. If ET112 would be a problem then you would not see the load on the VenuxGX screen and that is not the case.

0 Likes 0 ·
rainerb avatar image rainerb commented ·

I can confirm the behaviour. When I switch on the 2000W load, the correct measurements of the load and the grid usage (from ET340) are shown on the Remote Console after about 2-3 seconds. BUT it takes another 10 seconds until the ESS regulation has compensated this load to the grid setpoint.

Maybe the quite slow regulation speed is intended this way?

Maybe @mvader from Victron can bring some light into this?


Regards

Rainer

0 Likes 0 ·
n-dee avatar image n-dee rainerb commented ·

Hi, contacted mvader already with this some time ago. (but not as prezise as you do)
Answer was that it is not (and willl not be) worked on this.
I hope that with more people reporting he will change his mind.

0 Likes 0 ·
rainerb avatar image
rainerb answered ·

Hi,

I have a Multiplus with ET340 meter and my ESS system needs about 15 seconds to compensate a 2000W load. Is this a normal time span or can I increase the speed?

What is the recommended baud rate for a ET340 with about 5 meter cable length?

Thank you.


Regards

Rainer


1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

pau1phi11ips avatar image pau1phi11ips commented ·
I use 9600 with the ET112. This seems to work fine for me.
0 Likes 0 ·
n-dee avatar image
n-dee answered ·

Same Problem here (with EM24) - System needs approx 10 seconds to compensate

Additionally:
1) Grid setpoint is never reached (regardless of setting) but fluctuates around setting.
2) Fluctuation of setpoint is as high as +-1500watts (if load kicks in) so that even water warming (set to -500w) kicks in (after load ends and ESS goes to e.g -1000w) and drains battery.
3) In Austria feed in from Battery in the night is actually forbidden but due to the points above this is not configurable.

Seems as ESS function is technically not mature. I'm hoping for improvements in future firmware releases.


2 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

laurentiupitran avatar image laurentiupitran commented ·

hello,


please see attached picture with reaction time in a ESS system.


laurentiu

0 Likes 0 ·
answering-time.png (107.6 KiB)
mysik avatar image mysik commented ·

But ESS function with CT clip looks muuuuch better than when it using grid meter. That what I still don't understand, especially that using grid meter you have more accurate information and this data is available on GX device because it is displayed, it just look like the system is not using it or is poorly tuned to use it.

0 Likes 0 ·
Daniël Boekel (Victron Energy Staff) avatar image
Daniël Boekel (Victron Energy Staff) answered ·

Hi @Mysik and others,

I've asked around, and this is a known issue that is being worked on.

Thanks for finding and reporting the issue!

43 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

n-dee avatar image n-dee commented ·

Hi,

Additional to the occuring slow response I'm adding a Video of (relative) stable Loads but somewhat "confusing" fluctuation of grid setpoint:
My settings: Grid setpoint -20 -> Please note that I dont have AC Loads connected. The whole house is on Critical loads. therefore AC Loads should stay on zero.

regards
Andy

0 Likes 0 ·
grid.mp4 (1.6 MiB)
gyrovague avatar image gyrovague commented ·

Any progress on this? I'm facing a similar issue, which might be related to ESS + ET112, where it is back-feeding into my grid and causing the prepaid electricity meter to reset. @Daniël Boekel (Victron Energy Staff)

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

Are you on one of the tshwane newer smart meters? They are very sensitive and have been causing issues.

Double check what your grid setpoint is in ESS, raising this can help.

Make sure grid feed-in is disabled.

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

No, I have a 12+ yr old Conlog BEC23 prepaid meter, in George. It allows 40W for 15s or 600W over 1s of back-feed before it will cut-off the grid, which sounds reasonable to me, not very sensitive. The Victron is responding very slowly to load changes, and often feeds in over 500W (I've seen up to 1800W) over multiple seconds (as many as 15) after a load such as a kettle is switched off.

My default grid set point was 50W, but I've tried raising it to as high as 300W, and it still doesn't solve the problem, meanwhile causing loss of efficiency of the ESS. Grid feed-in is disabled.

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

copy of answer in other topic:


for faster response: have the loads behind the inverters, this gives a faster response than with the energy meter.

having the energy meter as source works for 99% of customers, but in this mode all kinds of grid rules have to be followed also (that's one of the reasons it's difficult to improve the reaction time), when loads are behind inverter ('critical loads'), the inverter can provide power almost instantly.

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

I'm not sure I follow 100% what you're suggesting. Are you saying that the non-essential loads should all be moved to essential? Problem with that is (I think) then if there is grid failure, those loads could un-desirably use up battery energy unexpectedly (oven, stove, geyser etc.).

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

From the other thread, I am not sure why you have a grid meter. Your setup does not need one, I asked if this could be affecting response times. Perhaps @Daniël Boekel (Victron Energy Staff) can answer if it is an idea to remove/bypass it and if that could help?

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

Maybe, I'm not sure. The grid meter has been disabled now in the ESS configuration, but it made absolutely no difference, the inverter response to loads changing (in both directions) is still very slow.


Edit: see here my other post (which was diverted to this thread), where there is some more info about my setup and the issues: https://community.victronenergy.com/questions/61374/victron-ess-slow-response-to-load-changes.html

0 Likes 0 ·
daryl avatar image daryl gyrovague commented ·

You could move them behind AC-output 2 and cause that to fail during a grid failure

0 Likes 0 ·
gyrovague avatar image gyrovague daryl commented ·

Oh, I was actually under the impression that is how it already is... but, looking at the installation line-diagram now, I see that it's not, the non-essential seems to be connected to the municipal side after the energy meter but before the inverter.

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

It's not unusual to do that, in a solar setup you'd really want to attach appliances that you want powered by solar or battery. The rest leave off the inverter.

0 Likes 0 ·
n-dee avatar image n-dee Daniël Boekel (Victron Energy Staff) ♦♦ commented ·

Don't know if I understand you - but I have everything behind the inverters (only critical loads) and it is not faster than before. (well smartmeter is there aswell)

regards,

Andreas

0 Likes 0 ·
gyrovague avatar image gyrovague n-dee commented ·

My solar installer is saying the same or similar -- he says they actually had worse overshoots when using AC2 on previous installations with multiplus II and quattro, and that having non-essential before the inverter with grid meter worked far better :shrug:


Unfortunately in my case it will require significant and costly rewiring to move the non-essential to AC2, so it's not a quick and easy experiment at all. I would need to jump and commit to it, and if it doesn't work out (as installer predicts) then I've wasted even more money. It's 50m of extra wiring (thick expensive cables too) and multiple hours of electrician labor, so not cheap.


I'm at somewhat of an impasse it seems. The municipality blames the victron ESS system, so they won't replace my prepaid meter (and they have an exclusivity deal with Conlog anyway). The installer says the Conlog prepaid meter is too sensitive (though the Conlog spec allows 40W/s over 15s or 600W in one second, which supposedly conforms to some standard and sounds quite reasonable to me), and I need to replace it with L&G, but I can only replace it privately with a 4-quadrant L&G, but at great cost (around 400euro, imported from Europe, because the locally South African manufactured ones apparently require some power-line comms grid infrastructure which my town does not have). If I have the system rewired so that non-essential is on AC2, it will cost less than a new prepaid meter (but far from nothing), but it very well might not solve the problem (dead loss if so). If I don't do anything, I cannot use my oven at all (as has been the case the past few weeks since the system was installed). From my point of view, I agree with the municipality, that the Victron ESS overshooting hundreds of Watts for 10-20 seconds is unacceptable and causing the prepaid meter to disconnect the grid pro-actively. If Victron @Daniël Boekel (Victron Energy Staff) can give me more assurance that rewiring to AC2 is the correct solution, I can go that way.

0 Likes 0 ·
solarhack avatar image solarhack gyrovague commented ·

HI

I have two observations/suggestions may somebody can comment on or weigh in on. Would either of these be viable. The first is something for Victron and the second a wiring question.

1) According to the technical spec the ET112 has a maximum response time of 500ms. If Victron adopted another meter with a faster response time they could change this. The Fronius grid meter (Victron have a relationship with Fronius) samples at less than 200ms. My logic is a faster sampling rate would slightly reduce the peak and significantly reduce the duration of the feedback. I was wondering if Victron could not change their software to support the Fronius 63A-1 meter (It also has a Modbus/RS-458 interface. I know that not all meters return the same measurement data and that this isn't as simple as wiring up a new meter) @Daniël Boekel (Victron Energy Staff) would this be a possibility?


2) Could you not rewire your setup as follows (Is this legal and or possible) and would it work

Present wiring:


Proposed wiring


I realize this will proably mean you don't get any benefit of solar for your oven, and its usage wont be part of you reporting. Would it work?


Any thoughts?



1 Like 1 ·
grid-par.png (13.3 KiB)
Phil Gavin avatar image Phil Gavin gyrovague commented ·

Just put a temporary equivalent load anywhere convenient on the AC2 output and test the response before committing to the expense of permanent measures.

0 Likes 0 ·
michelg avatar image michelg gyrovague commented ·

Hi, the easiest and cheapest way to test the AC2 will be using 1 or 2 grill / micro-wave owens, so you can install them easily near your AC2, and having them sucking around 2kW ...

You can borrow them from your neighbors, just for a few hours, during your tests. No costs, no cabling, and a result.

Michel

0 Likes 0 ·
gyrovague avatar image gyrovague michelg commented ·

Great ideas @Phil Gavin and @MichelG, thanks! I have some electric oil fin heaters than can draw 2kW on demand, pretty similar to what the oven on non-essentials currently uses.

1 Like 1 ·
n-dee avatar image n-dee n-dee commented ·

Yesterday I removed the smartmeter and configured cerbo in -> ESS -> grid metering to Inverter/Charger.

The speed on load-changes is MUCH (!) better. Only two or three seconds.

regards, Andy

0 Likes 0 ·
n-dee avatar image n-dee n-dee commented ·
Any news on this?

regards, Andy


0 Likes 0 ·
peternielsen avatar image peternielsen commented ·

@Daniël Boekel (Victron Energy Staff) is there an update on this ? Why is there a difference between the reaction of a CT vs Energy Meter ? It seems it's not the communication and it's ESS how acts differently, in my head it's just a different way to give the same information to the system.

0 Likes 0 ·
Noisy Boysie avatar image Noisy Boysie commented ·
0 Likes 0 ·
elel avatar image elel commented ·

Hi @Daniël Boekel (Victron Energy Staff) ,

I encountered the same problem in which the response of my ESS+MP-II+ET112 system to a sudden change of load is very slow.

(For those said that you don’t have this problem with their ET112, try something with high load and turn it on/off repeatedly. You will find that when you turn it on, the grid at once takes up the load, then the battery gradually takes up the load, and then the grid resumes to the setpoint. The whole adjustment would last for 10 to 20s depending on how large the load is. The big problem is when you turn off the load. In such case, once you turn it off, the battery does not decrease its output at once but does it gradually. It causes the extra power (from the battery) being fed into the grid. It takes 10 to 20s for the grid to go back to the setpoint. It is really annoying to see how my hob uses the grid power (i.e. not being able to use the battery) when it turns on and then causes the battery to push extra power to the grid when it turns on/off for 10 to 20 seconds, repeatedly, at its temperature control mode.)

I have now changed to use CT instead of ET112 as CT gives a very good response to the change of load. However, with such change, I lose the accuracy of ET112 and the Venus screen does not show the Critical Loads as a separate box

I agree with @peternielsen that it is not a communication problem between ET112 and Venus/OS but how MP-II/ESS reacts to sudden change of load after getting the same information from different sources (i.e. from CT or grid meter). Why can ESS+CT response to sudden change of load quickly but ESS+grid meter can't ? What is the difference?

Can you give us an update the status of how Victron is now solving this problem after your update nearly 2 years ago?



0 Likes 0 ·
shaneyake avatar image shaneyake elel commented ·
The difference between the CT and the meter is where the power regulation happens.

With a CT the power regulation is done inside the inverter and is only limited to the 400W per second ramp rate limit.

When using the meter the data goes from the Meter to GX to Inverter, this can be anywhere from 1-5s

The truth is that this doesn't really matter, most utilities are using a 30 min window with some using a 1 min window. They take total power in-out = reading and report that to their billing systems. So if you look at those readings you will see the reported value is 0 or close to 0 because the grid usage was cancelled out by the export at the end. You can easy confirm if this is the case with your utility by looking at your bill. In 99% of places the little export and import you are seeing doesn't matter so I don't expect Victron to dedicate lots of time on this, however the new meters released do look like they will have better performance in the future and that Victron is still aware of this problem. This is fundamental limitation to the way the GX works that will take a lot of work to fix.

0 Likes 0 ·
elel avatar image elel shaneyake commented ·

Hi @shaneyake , thanks for the reply.

I have another view regarding the information you provided.

I don’t have too much idea what “ramp rate limit” is but what I see in my current CT+MP-II+ESS setup is that the inverter responds instantly to the switch-on of, say, my 2.2KW kettle and provides the full power while the grid is still maintained at the setpoint. When the kettle is switched off, the inverter drops the power quickly within 1 sec and only pushes the grid power down a little bit and the grid can resume to the setpoint within 2 to 3 seconds.

Transferring a few digits of meter reading from human to human to human can be completed well within 1 to 2 seconds. We are not living in stone age. Why does the transfer of signal from meter to GX to inverter needs 1 to 5 seconds? If this is really the case, such delay should no way be caused by technology but deliberately created by the designer of the system (and I have no idea why).

On the other hand, we are now talking about 10 to 20 seconds of delay of the system with grid meter to respond to the sudden change of power demand. Even with the 1-5s delay in the communication you mentioned, the additional delay is still much more than that we see in the case of CT.

Another view to whether it does matter or not is that in some common household situations, the ESS+grid meter system creates more trouble than expected. For example, all hob cookers are designed to turn off when the set temperature is reached and turn on again when the temperature falls below a threshold. Such off/on cycle can repeat within 20 to 30 seconds to keep the hob temperature close to the set point. This is very bad to ESS+grid meter system. What I saw was that when the hob turned on, the grid took up the power. While the inverter was trying to take up the power, the hob turned off. The inverter then pushed the power to the grid until it resumed normal. In other words, during the cooking period, power to the hob mostly came from the grid (not the battery) while the battery was used to push its power to the grid!

Regarding how the utility companies record the in/out power, my observation is different. In my case (in UK), the smart meter has a reading for incoming power and another reading for outgoing power (so no averaging). I don’t have feed-in arrangement with my utility company so all feed-in are lost. Anyway, feed-in is not welcome unless one has arrangement with their utility company.

From my point of view, this is a design fault in which the ESS+grid meter system cannot utilize the instant response ability of the inverter (as in the ESS+CT system). I am thinking…..why can’t the inverter get the grid meter reading from the Venus/OS and do the same as if it has a CT?


0 Likes 0 ·
shaneyake avatar image shaneyake elel commented ·

The Ramp rate limit is 400W per second on all Victron inverters. The inverter output won't change more than 400W per second. So if the inverter is commanded output 2000W it will take 5 seconds to get there. This is hard coded and is worse with some grid codes. If you have inverters in parallel you get faster ramp rates.

You can read more here
https://www.victronenergy.com/live/ess:ess_mode_2_and_3#response_times_and_ramp_speed


Meter -> GX Modbus Client -> DBus -> ESS control -> Dbus -> VE_BUS -> Inverter

The reason it takes 1-5second for the meter valves to get to the Inverter is because it is a multi step process.

The connection between meter and GX is a poll type, the GX continuously polls the meter at about oncer per second for new valves. The meters are also doing some averaging and there are limitations on how fast your can actually poll the meters.

Dbus can be fast less than 1ms but reading can take time depending on implication.

Then you have ESS loop on some set time, like every 1second, I can't find actual ESS script, it might be event driven but only outputs every few seconds.

Then Dbus again

Then finally it gets to a script that sends it to the inverter. This also only happens every second or so.

Once the setpoint is in the inverter the inverter ramps up or down to hold AC_Input at that setpoint. Which is limited to 400W per second per inverter.

The delay more comes from the fact that at the core of a GX is the Dbus, this makes things extremely flexible and is what has allowed Victron to become the most innovative company out there but doing things this way does have a speed limitation. Victron has investigated switching to MQTT instead which would increase response time but also means a rewrite of all current code.
https://github.com/victronenergy/venus/issues/168


When using a CT the inverter control loop will hold the CT at the set point, the only limit is 400W per second but there is no lag in the response as the control is all done by the inverter.

Example of how it works:
Lets say there is a 1000W load directly connected to grid and 200W on inverter output. Grid setpoint of 0W.

Meter reading will be 1000W
Inverter input will be 0W.
Inverter output will be 200W (This is calculated not measured just FYI)

You turn ESS on, GX calculates Setpoint - Meter = Error (-1000W)
GX commands inverter setpoint to -1000, inverter ramps up to hold -1000W on it's input, so total power it is making now is 1200W and grid is 0W.

If you add or remove load from the output of the inverter it doesn't need the GX to change inverter setpoint but if loads before the inverter change it has to go through whole GX chain. By using a CT you are effectively moving all loads to the output side of the inverter in terms of control.

0 Likes 0 ·
elel avatar image elel shaneyake commented ·

Hi @shaneyake , thanks a lot for the detail explanation. It let all who suffer have a better undersanding what the current design is and why there is a delay in response when grid meter is used. For those who is thinking of using a grid meter, they now have more information to make an informed choice. It would be even better if such information can be shown in the grid meter web pages.

I appreciate very much the flexibility and versatility of GX/VenusOS and this is also one of the main reasons I chose Victron products. It is a pity that I have to give up the grid meter and move to CT. CT is fine in my case, but I have to give up the high accuracy of the grid meter and look at an incorrect device connection topology of my system in the VenusOS screen (in which all the loads are incorrectly shown in a box connected to my MP-II).

Hope that I can one day use my grid meter again.

0 Likes 0 ·
shaneyake avatar image shaneyake elel commented ·

There is actually nothing stopping you from

GRID -> Grid Meter -> CT -> Loads/Inverter

You should be able to set system back to grid meter and still have fast response as CT - Meter = 0W and so the slow loop will always be 0 and inverter will use fast loop.

You non critical loads reading will also always be 0W as all loads will be in loads box but that is maybe worth it for you.

0 Likes 0 ·
pedaaa avatar image pedaaa shaneyake commented ·

but there is a showstopper:

the max. length for CT is limited. And i think its not allowed to extend it?!

In my case i would need 90-100m CT cable length, which is not allowed?! So i will use EM24 RS485.


BTW:

Do you know:

- is EM24 Ethernet or RS485 version equal in regards of load response time, or does it make no difference?

- are loads in "AC-Out" affected of the grid-meter->GX->multi control loop delays? Or are they handled "as quick as always"

0 Likes 0 ·
shaneyake avatar image shaneyake pedaaa commented ·

Yeah, I think the max for CT is 20m now.


I think it is the same but the EM24 ethernet version has constant problems with locking up, I would not use it.
Victron also just announced a bunch of new meters have look, polling rate is much faster on the new ones so response will be somewhat better.

AC-Out is handled by the inverter and will give you fast response. The limit is 400W per second, inverter can't go faster.

0 Likes 0 ·
elel avatar image elel shaneyake commented ·

Hi @shaneyake,

Hi thanks a lot for giving me this CT+grid meter option.

I connected both CT and grid meter to the system and tried the effect of choosing the option “External Meter” (i.e. using the grid meter) and “Inverter/Charger” (i.e. using CT) under ESS/Grid Metering menu.

I timed both options and found that the response time (to get back to the grid setpoint) for the option “Inverter/Charger” is around 5 to 10 seconds better than that of the “External Meter” option when a large load is turned off. The response to turning on a large load is similar.

I will appreciate it very much if you can further explain, when both CT and grid meter are connected to the system,

  1. what are the differences in choosing the 2 options I mentioned above?
  2. Would the system calculate the grid consumption (and report the data in VRM) based on the grid meter or CT? Would it depend on which of the above 2 options I chose?
0 Likes 0 ·
jeanphi500 avatar image jeanphi500 elel commented ·
I'm facing the exact same issue. Just ordered the CT clamp to try the same as you.

When you have the grid meter set as the ET112, do you get consistent readings in VRM? I got readings that were really not consistent when the load / feed in would vary. And the ET112 would regularly show as not connected in the remote console
thx!


0 Likes 0 ·
elel avatar image elel jeanphi500 commented ·
I didn't encounter what you described. I have the grid setpoint at 50w and therefore basically have no feed-in (other than those small amount of feed-in cased by the slow response of the grid meter). For the load, I did not notice any inconsistent from what I expected. My ET112 never showed as not connected. My grid meter is working quite as expected other than the slow response.
0 Likes 0 ·
shaneyake avatar image shaneyake jeanphi500 commented ·
We use the ET112 at all our single phase installs. Never really had a problem, all readings are the same as utility meter +-1%.

Meters always show as connected unless grid outage.

What RS485 to USB cable did you use? Victron one works best. Did you have to extend the cable? If so what cable did you use? Cat5e/Cat6 works well. Shielded is better for long runs like 50m+.

0 Likes 0 ·
shaneyake avatar image shaneyake elel commented ·

It is the selecting which control loop should be used, GX/Grid Meter or Inverter with CT. By selecting Inverter you are removing all of the lag/delay that is cause by reading a meter, processing the data and setting the inverter setpoint. As it just does the last step, sets the inverter setpoint to what is the grid setpoint in the menu.

2. Yes, it depends on what you have selected as your meter.

Strange that you actually saw a difference, I wonder if it is actually just a display delay as the meter data is delayed when grid meter is selected. Can you look at the inverter screen on the GX when timing or better yet a clamp meter on your in coming connection.

0 Likes 0 ·
elel avatar image elel shaneyake commented ·

My observation is somehow different from what you described. Just connecting both CT and grid meter already made a good improvement in response time (no matter whether I select "External Meter" or "Inverter/Charger" when both are connected.) It seems that whenever CT is connecrted, the long signal transmission is skipped no matter whether the grid meter is connected or not. I will try to take some video of the screen to show what it actually happens.


0 Likes 0 ·
shaneyake avatar image shaneyake elel commented ·

When you connect a CT to an inverter the CT replaces the input current sensor in the inverter.

If wired Grid -> Grid Meter - > CT -> Loads
The ESS value should alway be 0. CT - Meter = 0W
So the long control loop should always just be sending the grid setpoint value to the inverter same as when set to Inverter/Charge. There shouldn't be a difference when changing that setting other than where the kwh counters come from.

You said "I timed both options and found that the response time (to get back to the grid setpoint) for the option “Inverter/Charger” is around 5 to 10 seconds better than that of the “External Meter” option"

I said I think this is just because the GRID block on the GX is the data from the meter which can be delayed when in "External Meter” option and probably not what is actually what is happening.

I think we are in agreement, but a Video would be helpful to confirm.

1 Like 1 ·
elel avatar image elel shaneyake commented ·

Hi @shaneyake , I took a record of both cases. In both videos, the load used is a hair-dryer of around 1.9kw. When "1" was entered in the notepad, the hair-dryer was turned on. When "0" was entered, the hair-dryer was turned off. Grid setpoint was 20w.


With grid meter and CT connected, and selected Inverter/Charger (i.e. CT) in ESS:

The grid reading took 10s. 11s and 12s to return to above zero when the hair dryer was turned off


With grid meter and CT connected, and selected grid meter in ESS:

The grid reading took 14s. 17s and 20s to return to above zero when the hair dryer was turned off


It seemed that with grid meter selected (when both CT and grid meter were connected), the grid reading took longer to get back to the setpoint.


0 Likes 0 ·
shaneyake avatar image shaneyake elel commented ·
Thanks for the video but can you re-run the tests looking at inverter screen on the GX -> menu -> Multiplus-ii


AC input Power

The problem with looking at the main screen is the grid block gets switched from CT to meter when you change the meter setting. The meter will always be delayed somewhat, potentially skewing your results.

0 Likes 0 ·
Show more comments
Maciek avatar image Maciek shaneyake commented ·

Hi @shaneyake. Thank you for suggestion. I tried this “Combo” ET meter plus CT sensor and it gives the best results!!!

From all different configurations I tried this is still the best choice but not as good as it should be when inverter reading values from meter directly. Inverter is still working far way from grid setpoint and readings are from different place sometimes going slightly below 0 in stable load …. But that’s what we have,

Thanks again !!!

0 Likes 0 ·
christian-dorst avatar image
christian-dorst answered ·

I have the same problem with EM24 in 3-phase-setup with Multiplus II here.

https://community.victronenergy.com/questions/54038/multiplus-ii-ess-mode-1-latency-and-limits.html

Sunny greetings

Christian

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

jacola avatar image
jacola answered ·

Same problem here but with a "twist". I have an ET340 and poll it from an ethernet based linux controller approx. 5 times per second. ESS IN "EXTERNAL" mode, i want to control it via setpoints.
Ialso read the relevant modbus registers showing the actual production values for the multiplusses from the venus GX approx 5 times per second.
I calculate the needed setpoints with those 200ms intervals, so i could update the setpoints via modbus/tcp 5x per second, but the system then behaves very erratic and sometimes even the AC load side is not showing on the display any more.

So now i have reduced polling and setpoint calculation and setting (via mb/tcp) to about once every 1.2 seconds. i can see that despite writing back those setpoints every 1.2 seconds via mb/tcp, they freqently are ignored/dropped (but modbus write is acknowleddged) ! maybe that gives a hint.


ESS 1: -44/ -160/ -230 2: -24/ -804/ -801 3: -33/ -684/ -641 BAT: 69.0,49.8new setpoints should be -146,-811,-682

ESS 1: -62/ -160/ -146 2: -31/ -804/ -801 3: -28/ -684/ -641 BAT: 69.0,49.8new setpoints should be -128,-804,-686

ESS 1: -54/ -160/ -128 2: -27/ -804/ -804 3: -28/ -684/ -686 BAT: 69.0,49.8new setpoints should be -136,-807,-687

ESS 1: -48/ -160/ -136 2: -32/ -804/ -807 3: -36/ -684/ -687 BAT: 69.0,49.8new setpoints should be -142,-802,-679

ESS 1: -28/ -160/ -136 2: -24/ -804/ -802 3: -42/ -684/ -687 BAT: 69.0,49.8new setpoints should be -162,-810,-672

ESS 1: -21/ -177/ -162 2: -25/ -812/ -810 3: -49/ -663/ -672 BAT: 69.0,49.8new setpoints should be -187,-818,-645

ESS 1: -5/ -177/ -187 2: 2/ -812/ -810 3: -25/ -663/ -672 BAT: 69.0,49.8new setpoints should be -203,-844,-668

ESS 1: -1/ -177/ -203 2: -13/ -812/ -844 3: -23/ -663/ -668 BAT: 69.0,49.8new setpoints should be -206,-829,-670

ESS 1: -10/ -177/ -203 2: -29/ -812/ -844 3: -28/ -663/ -670 BAT: 69.0,49.8new setpoints should be -197,-813,-665

I have high power pumps (11kwh 3P), when they come up they run for several seconds and then switch off. This causes complete instability of the victron stuff, said simple - it cannot follow at all and usually i have to feed the pumps for many seconds from the grid, then when multiplusses finally are up to produce, the pumps switch off and i feed the grid.

I would be more than happy to see this getting fixed. Maybe those observations help - i guess it's not the ESS regulation or meter speed (because both is not having an influence here in my system), it seems to be loss of set setpoint values, eventually, and very slow responses from the multiplusses to GX

Johannes


2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

jacola avatar image
jacola answered ·

is there maybe a way to simulate the CT Sensor, to take advantage of the much faster response handling ? I guess the "correct" way would be that victron addresses this in the software, but i am willing to read my meter, do my own control loop and provide split second setpoints, even by simulating a CT sensor, as long as the multiplus follows my wishes quickly ..

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

matthias-roetzer avatar image
matthias-roetzer answered ·

Hi all,

I've seen this slow response to load and PV inverter output jumps as well in cases where the multiplus II does the grid measurement. I was then curious how it would behave in off grid mode and good thanks in this case the response is ultra fast.

@Daniël Boekel (Victron Energy Staff), nevertheless this is a nasty issue, glad to hear that Victron is working on it!

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Rob Duthie avatar image
Rob Duthie answered ·

Hi All

I am running a ET112 with Mulitplus2 and Fronius Grid tied inverter and 10 kw Pylon tech battery pack, and my ET is wireless linked to the Meter board to my raspi controller, and i have no issues of slow response at all. I have set my feed to zero feed for the Victron but the Fronius injects what it likes etc. But the meter response is very quick in flipping from supply load and battery and grid within a couple of seconds, no delays anywhere at all.

But being in New Zealand we don't have silly lines, or power retailers company's restricting or penalizing you for injecting power like other countries do. You need to protest your rights to the solar movement. They are are being a bully boy.

Rob D

NZ

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

peternielsen avatar image
peternielsen answered ·

@Mysik @RainerB did you find a solution to this, what is the reason one cannot use CT sensor on each grid phase ? Is there a benefit using the Energy Meter, i can see you @Mysik changed from CT-> Energy meter, why ? Can ESS not work with 3 x CT sensors and do the same as an Energy Meter ?

2 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

rainerb avatar image rainerb commented ·

Hi, unfortunately I did not find a solution for this. I am still using my ET340 energy meter and it is still slow.


Regards

Rainer


0 Likes 0 ·
Maciek avatar image Maciek commented ·

Let me answer your questions since I’m in exactly the same situation and tried different solution, hardware and did all measurements.

You can use 3 CT sensors for 3P configuration. Reaction is quite fast but readings from CT while Multiplus are under load are completely inaccurate So inverter is compensate wrong load and generates Big feed to grid (while showing in the system that all is OK). It looks Victron can not implement proper algorithm for reading signal from CT sensor in 3P configuration for many years.

that’s why people (including me) try external energy meter, which is very accurate but as we discovered inverters due to communication issues can not adjust quickly and it takes more than than 25 seconds for higher load.

for example I have 3 pumps and 3AC and my System is constantly getting crazy feed to grid about 2000w !

I don’t understand how it is possible that company like Vivtron can not solve such a big problem. And what is even worse it’s not described and communicated to clients properly.

0 Likes 0 ·
Rob Duthie avatar image
Rob Duthie answered ·

Hi

Never had a problem with slow responce from my ET112 meter with Mulitplus2 i have installed to date.

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Maciek avatar image
Maciek answered ·

I have same problem here in three phase configuration. 3x MultiPlus II 5000/48V.

I tried all configurations:

1. 3x external CT sensors directly on grid lines.

2. Energy meter ET340

3. All loads directed through MultiPlus (IN and OUT) using internal sensor for measurements.

Non of these works correctly.
1. Gives feed-in (verified on pro external meter) due to wrong readings of power from CT . Probably wrong implementation by Victron while calculating power in 3 phase configuration (active reactive) because it require more data… but system react fast as expected when increasing and throttling!

2. Gives accurate readings from meter but inverters reacts with such big delay (up to 25s) that system is useless … feeding more power to grid than producing for consumption ( this situation repeats every 2-3 minutes when I have 2 pumps and 3 AC. Perfect example of totally wrong design of the system ….

3. Same as 1 with little less error. In fact readings go off in different way than on external sensor so this solution is the best for now.

After huge investment from my side (time and money) I see i on this forum that its old problem which Victron is not able to solve for years.

And dont accept explanation found here that they can not send reading from energy meter to multiplus in less than 1 second and replace it with reading from CT - then after 3 steps it should go down. 3-5 seconds to throttle down is acceptable but 25 sec definitely not with 5 consumer turning on and off with 1-3kw system się getting crazy on both sides inverter and grid.

Clearly implementation of external energy meter is not done correctly. I can see from my external meter connected for checking that Victron system throttle down quite fast for 75% but then slow down and for next 15 seconds struggle doing strange things. It looks like a joke not professional solution. All these energy meters can be used only to display values in Venus system but not to control it due to delays introduced by Victron on the way to inverter. We have 2022 and how long does it take to read data, send , do simple math and send further … Or let them clearly say they use wrong protocols and algorithms which blocks proper implementation.

In my opinion for this momemnt ESS is not mature and not ready for sale and installation. They should stop advertise it.

UPDATE: after suggestions from one of active user here I’ve tried combined solution Energy meter and CT sensors in the same time. And got more satisfying results. Still not as expected but much more precise readings on one device with quick reaction as using CT sensor or ET meter only. Unfortunately these readings are not used by ESS. Thanks @shaneyake !!!

Maybe Victron @Daniël Boekel (Victron Energy Staff) can use data from both devices to develop algorithm which will react quick using signals from CT first but then adjust precise readings using data from meter when the load is cons?? For example calculate delta between CT and ET and then every 3 seconds apply to CT readings

12 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Hi @Mysik , what grid-code are you using?
0 Likes 0 ·
Maciek avatar image Maciek Daniël Boekel (Victron Energy Staff) ♦♦ commented ·

@Daniël Boekel (Victron Energy Staff)

Hi. I'm using "no grid code" (first or second option on the list) because country where I live (UAE) is not on the list. Would appreciate if you recommend country from the list which is close to local requirements.

Today I spend another few hours on measurements and can only confirm what written above. Test done with quite stable 3p load 1.2kW on each phase turning on and off every 60s.

- external CT completely not usable. I used short cables 1m and tried to isolate them from any interferences. While ESS and inverters believe to keep Grid Point 150 (readings around 50+50+50) in reality on my calibrated Iskra meter and also on ET340 its back feeding -120 -30 -40 jumping like a lottery (but still for ESS nice and constant grid input around 150 :)))

Build in CT sensors, much more stable readings but still far away from reality. Instead +50+50+50 showing in the system it's working on +30+0+40 (fluctuating of course under 0 quite often).

Two interesting observations in this case. first: under small load around 100-200W per phase readings are more accurate but under load 1200W per phase difference in readings from reality goes down at least 30-40W which looks like some correction factor is missing for this setup of CT in 3P config under load (VAR??).

And of course when ESS controlled by ET340 readings are pretty accurate (comparing with calibrated Iskra meter is +-1W) but system is working in completely unacceptable way taking ages to stabilize under load and then the same to throttle down.

In general it's very sad situation. People like me spending a lot of money on Victron MPII and CT or energy meters like et340 which are not working correctly in the system as intended.

It should be clearly informed to clients that in this case Victron ESS is not working correctly.

Now temporary solution is to put Grid Point extremely high but then during 24h operation we will have quite high energy consumption.

Or to go Off Grid by switching off AC IN for as long as possible but as I see from some answers on this forum Victron is not supporting this solution while not providing correct one.

0 Likes 0 ·
Hi @Mysik

the ET340 is very slow, and also has an internal filter making the effective readings even slower.
The EM24 is already much faster.

the new EM540 will be much faster than the EM24, so improvement is coming.

And as you've seen, the CT's are much faster due to the direct communication with the ESS assistant and fastest readout.

Note that a system with the loads behind the inverter is as fast as a system with external CT's installed.

0 Likes 0 ·
samuel-brucksch avatar image samuel-brucksch Daniël Boekel (Victron Energy Staff) ♦♦ commented ·

@Daniël Boekel (Victron Energy Staff) could you please provide more information on the improvements related to Em540? From my understanding so far, the restriction is not from the update rate of the meter, but from that 400W/s limit. So with a faster update rate of the meter will that limit be higher? Because if not we will see the same problem with EM540, as people already confirmed, that the meter showed updated values quite early, but the Multiplus was still slow in adopting the power.


Are there also any Venus OS or MP-II improvements (software side) coming with the EM540, that faster readings from the meter will also allow a higher limit for the power calculation? Or is that 400W limit totally decoupled from the meter updates? Because then we would not see any improvement with the new meters.

1 Like 1 ·
Daniël Boekel (Victron Energy Staff) avatar image Daniël Boekel (Victron Energy Staff) ♦♦ samuel-brucksch commented ·
Hi @Samuel Brucksch


Unfortunately I don't know all the answers, I know that the meter is faster.


The 400W/s limit is probably a grid-code limit?

0 Likes 0 ·
jetstram61 avatar image jetstram61 Daniël Boekel (Victron Energy Staff) ♦♦ commented ·
Hi Daniel, I am using an external CT with my Multiplus 2 and ESS system. This results as you know in a slightly incorrect VRM display because the multi believes all the loads are flowing through it. Because of this I am getting some L1 overload messages however can I assume these are false as the loads are not actually passing through the Multi ?

Thanks

0 Likes 0 ·
mysik avatar image mysik Daniël Boekel (Victron Energy Staff) ♦♦ commented ·
@Daniël Boekel (Victron Energy Staff) , you responding to wrong person. @Maciek was asking the question, not me!
0 Likes 0 ·
My apologies!
0 Likes 0 ·
Maciek avatar image Maciek Daniël Boekel (Victron Energy Staff) ♦♦ commented ·

Hi @Daniël Boekel (Victron Energy Staff),

I trust what you say But it's difficult to accept this excuse when I see readings from ET340 quite accurate (very accurate comparing to CT sensor build in MultiplusII) and updated morer than once per second on my GX screen!

For me it looks like, the way how you designed system, to connect, read, process and forward these data through different devices and protocols makes it not usable or simply not implemented in optimal way.

Few years ago I was using some super cheap 3x Chinese inverters with external power meters communicating through RS485 cable and it was adjusting to load without any issues, like in your solution.

In all your materials Victron says 'Energy meter....used as control input for an ESS systems" What we see its not true! It's good only to display correct readings on GX device. But when you switch control to "External meter" to control input it behaves like blind person moving in crowded metro station.

Now you mentioned some new product coming (EM350). Looks like new expenses. Can you share more info please? When we can expect it to come and what improvement it will bring?

If I can suggest, you should work more on CT sensors implementation in MPII. I'm sure it's either something wrong in algorithms or some parameter is missing. Its suspicious especially when ESS shows positive readings from CTs nicely adjusted to Grid setpoint and quite stable when in fact they are negative and fluctuating a lot on 2 independent meters! Definitely must be something wrong with data processing under higher load in my 3P configuration. Under 1KW load per phase CT readings are lower for about 40-50W than correct readings from meter .... so its matter of correction table based on load or similar solution.

Another improvement would be for the system is controlled by "individual phases" in ESS to setup and control Per phase grid setpoint not Total L1+L2+L3.

If you have some solutions suggestion how I can improve or test this configuration I would be happy to try.

Thx and Best regards.

0 Likes 0 ·
Hi @Maciek

CT's or the loads behind the inverter is the quickest way possible.

did you try selecting ESS 'per phase' regulation?

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

Of course working only with this setting. But what I mean is when you select Per phase regulation to set grid point and control it per phase not as summary of 3phases. I observe that’s MP even on stable loads like to work for example (60,5,50) with 2 phases quite above from 0 and one just around which cause backfeeding during fluctuations. If it will keep all 3phases far away from zero it will not happen so often.

All these happens under higher but stable 3p load.

0 Likes 0 ·

@Maciek disconnecting grid is also an option.

basic function purely in assistants is written here:

https://www.victronenergy.com/live/assistants:ignore_ac_input_using_the_generator_assistant

of course you can connect the Cerbo's relay with generator start/stop function to an input of the L1 inverter and use that to ignore input when wanted.

0 Likes 0 ·
h-copter avatar image
h-copter answered ·

@Mysik long and interesting thread. As the originator, what solution have you now implemented?

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

gazza avatar image
gazza answered ·

I agree... a long and interesting thread several years!! My take is that a faster meter will make (almost)zero difference here. If the ET112 is 500mS response time and we replace with a 200mS or less (obviously) more expensive meter, the problem is clearly elsewhere in the system to produce the 10-15second response time we are discussing here? It is an order of magnitude that is unacceptable in the whole philosophy of the ESS that is advertised by Victron.? Sure I get that the analog feedback of the CT clamp is almost realtime and the energy meter requires processing but please do not insult our intelligence... that is not the problem, there is a design issue, whether hardware or software. The CT clamp is just not a viable option for many as the solar hardware is not always near to the grid input point.
For those that publish that they have NO problem, which are in the minority on this thread, please show us the light? Show some videos, what are you doing that we are not? I just don't buy it when there is so much evidence of failed ESS grid set-point following. Indeed it is so critical for certain innocent customers who could actually be penalised for not complying with grid feed in restrictions?

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mrrikp avatar image
mrrikp answered ·

Same issue here. With CT, response almost instant, but as I have non-critical loads on the input side of the inverter, those aren’t accounted for. Hence, I use the ET112 on grid side and observe very slow and ‘wobbly’ results using it. Turning in a kettle results in wild fluctuations!

Given this is an important part of ESS, Victron really ought to officially comment - even if to give a reasonable explanation as to why they won’t fix it, or can’t. It’s clearly a problem for many users spanning several years, and since update time of ET112 is extremely fast, the issue appears to be between Cerbo and Inverter.

2 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

daryl avatar image daryl commented ·

Isn't that exactly what they are doing in the v3.0 Venus?

https://community.victronenergy.com/questions/164781/venus-os-v3004-available-for-testing.html

ESS: increase control speed of systems using todays supported meters (ET112, ET330, EM24, and also the ABB ones) a little. But, for faster meters, such as the EM540 which is not available yet, increase it very significantly. Requires updating the built-in MK3-chip, see next bullet.

0 Likes 0 ·
1234enough avatar image 1234enough daryl commented ·
I had been using ET112(grid meter) / multiplus ii and was disappointed in the slow response times.

So tried a CT clamp and wow, response is how I would expect- very fast like a second or less.

Great the new update will improve things with ET112, but I doute it will match a CT clamp for speed, especially when the official response is a new energy meter is to improve things more(yet to be released and no idea on cost).

Think I'll stick with CT clamp, as I doute even the new energy meter will match the response time, but we'll wait and see.

0 Likes 0 ·