Victron ESS Response Time Measurement

[ricardocello]

Below is a chart of the Victron ESS (Grid Code Other) responding to a 6 kW water heater turnoff event.
The general consensus is that Victron has a hardcoded 400W/second limitation.
With split-phase Quattros, this should be 800W/second.
This chart shows a faster response time than that (about 3000W/second). @1201
Note the oscillation at the end as the control loop tries to stabilize at 500W Grid Point.

 

[ricardocello]

Here is another example with an electric kitchen range heating element shutting off.

 

[ricardocello]

From Victron Documentation:

There are multiple factors that determine the response time to a (digital-) command to feed in:

1. Latency and communication speed of all components in the communication chain: ModbusTCP or MQTT, GX Device, MK3 microprocessor, ESS Assistant, internal communication in the Multi itself.
2. Rate limiting imposed by the used Country Grid code. Code “Other” has no rate limiting, Code “Europe” allows installer configurable rate limiting, many other codes have fixed a powerup ramp up.
3. Hard coded rate limiting in the inverter/charger firmware: as per ESS version 162 it is set to 400W per second. The reason for this rate limiter is that without that there are regulation problems (overloads and such) when the mains is weak (long cables and such resulting in a relatively high impedance). More information on that here, as well as in the main ESS manual.

 

[1201]

Below is a chart of the Victron ESS (Grid Code Other) responding to a 6 kW water heater turnoff event.
The general consensus is that Victron has a hardcoded 400W/second limitation.
With split-phase Quattros, this should be 800W/second.
This chart shows a faster response time than that (about 3000W/second). @1201
Note the oscillation at the end as the control loop tries to stabilize at 500W Grid Point.

View attachment 203969

Hey thanks for posting this!

I also found a vid that showed pretty fast response in some cases so I also wondered and the 400w/sec rate

Also I have a question- can grid setpoint be set to a negative value?

If so then it should be able to sell from battery to grid which could be profitable in the future

 

[1201]

From Victron Documentation:

There are multiple factors that determine the response time to a (digital-) command to feed in:

1. Latency and communication speed of all components in the communication chain: ModbusTCP or MQTT, GX Device, MK3 microprocessor, ESS Assistant, internal communication in the Multi itself.
2. Rate limiting imposed by the used Country Grid code. Code “Other” has no rate limiting, Code “Europe” allows installer configurable rate limiting, many other codes have fixed a powerup ramp up.
3. Hard coded rate limiting in the inverter/charger firmware: as per ESS version 162 it is set to 400W per second. The reason for this rate limiter is that without that there are regulation problems (overloads and such) when the mains is weak (long cables and such resulting in a relatively high impedance). More information on that here, as well as in the main ESS manual.

Is the 400w/s limit for both feed in and feed out? Or only feed in?

 

[ricardocello]

Yes, you can set the grid set point to be negative. I do not do this.

When you have a net metering agreement, that says to the ESS: Try to sell back as much as possible.
Then you are more concerned about the rising edge response (big device turning on) to ensure that you maximize sell back.

 

[ricardocello]

Is the 400w/s limit for both feed in and feed out? Or only feed in?

I assume so, but I don't know for sure.
I can try asking on Victron community, but Victron never seems to really talk about this topic.
Neither of my charts above show 400W/second (or 800 for split phase), so I don't know what to make of it.

 

[ricardocello]

The other observation I have is that conventional advice says to raise the Grid Setpoint high enough so these events don't occur.

That doesn't really address the issue at all, it just makes the bottom of the valley a little less negative.
Instantaneously, the power has to go somewhere. Not sure why it takes almost a second to respond either.

 

[1201]

The other observation I have is that conventional advice says to raise the Grid Setpoint high enough so these events don't occur.

That doesn't really address the issue at all, it just makes the bottom of the valley a little less negative.
Instantaneously, the power has to go somewhere. Not sure why it takes almost a second to respond either.

You're right. It's definitely faster than 400w/s but still too slow for an unpermitted connection. It looks like it's a software thing though so it may be adjusted in the future

Yes, you can set the grid set point to be negative. I do not do this.

When you have a net metering agreement, that says to the ESS: Try to sell back as much as possible.
Then you are more concerned about the rising edge response (big device turning on) to ensure that you maximize sell back.

So I'm assuming let's say it's set to -800w then the victron will try to sell back 800w from battery to grid continuously

 

[ricardocello]

You're right. It's definitely faster than 400w/s but still too slow for an unpermitted connection.
It looks like it's a software thing though so it may be adjusted in the future

Depends how ancient your meter is!

They say they are working to improve it, but it will likely be an ESS Assistant update in the inverters,
which I hate because I'll have to break out the Windows laptop and MK3-USB.

So I'm assuming let's say it's set to -800w then the victron will try to sell back 800w from battery to grid continuously

That is correct, but you need to make sure you have a minimum SoC set so you don't totally deplete the battery.
Of course, solar will be prioritized if it available.

 

[1201]

@CuriousJohn

 

[ricardocello]

Here is a well-pump turn on transient.
Note: The well pump is on the output of the inverters (critical load). Because water is critical!

 

[Brucey]

So it's trying to use 500W minimum from the grid at all times to avoid grid feedback, but in these cases the firmware doesn't slew fast enough to avoid some rare conditions where the impedance to grid is high?

 

[ricardocello]

Well Pump Off Transient for Completeness

 

[ricardocello]

So it's trying to use 500W minimum from the grid at all times to avoid grid feedback, but in these cases the firmware doesn't slew fast enough to avoid some rare conditions where the impedance to grid is high?

That's right, but these aren't rare, they happen here all the time when big loads switch off that are parallel to the inverter.
It's a physics/math problem really, nothing can respond instantaneously. Control loops take time.

I'd be interested in how other manufacturers do (SolArk zero-export, eg4 zero-export, etc.) But this is challenging to instrument to get the data.
The EM530 gives me almost 10 Hz data which is nice.

 

[1201]

That's right, but these aren't rare, they happen here all the time when big loads switch off that are parallel to the inverter.
It's a physics/math problem really, nothing can respond instantaneously. Control loops take time.

I'd be interested in how other manufacturers do (SolArk zero-export, eg4 zero-export, etc.) But this is challenging to instrument to get the data.
The EM530 gives me almost 10 Hz data which is nice.

.

 

[Q-Dog]

Does this same feedback happen when these appliances are operated from the grid? Faster? Slower? Or does the feedback thing not happen on grid power?

 

[ricardocello]

Does this same feedback happen when these appliances are operated from the grid? Faster? Slower? Or does the feedback thing not happen on grid power?

The ESS is grid-interactive. The input of the inverter is bi-directional.
If the ESS assistant was not installed, there would be no feedbck to the grid.

When inverting from battery/solar, the inverter input is disconnected from the grid for safety.
Only the critical loads are powered in that case, and nothing above is applicable.

These plots show that the ESS control loop is not fast enough to prevent undesired feedback to the grid.

 

[Q-Dog]

The ESS is grid-interactive. The input of the inverter is bi-directional.
If the ESS assistant was not installed, there would be no feedbck to the grid.

When inverting from battery/solar, the inverter input is disconnected from the grid for safety.
Only the critical loads are powered in that case, and nothing above is applicable.

These plots show that the ESS control loop is not fast enough to prevent undesired feedback to the grid.

I get that. My question is, if the appliances are powered directly from the grid, does this feedback occur when they shut down? Does the meter "turn backwards" for a split second? Or is this only a thing that happens with inverters?

 

[ricardocello]

Or is this only a thing that happens with inverters?

Sorry, I was being dense.
I’ll get the inverters out of the circuit, and retest with the grid meter running.

 

[ricardocello]

@Q-Dog Pure resistive loads do not show any issue (as you would expect).

 

[ricardocello]

No ESS Connected, HVAC Startup, with EasyStart 368

 

[ricardocello]

No ESS connected, HVAC shutdown

 

[ricardocello]

Clarifications on the charts above:

Water Heater is on input of inverters (L1 and L2), in parallel to the grid.
Kitchen range is on input of inverters, in parallel to the grid.
Well Pump is on output of inverters (critical loads panel).
HVAC is on input of inverters, in parallel to the grid.

 

[Hedges]

Also I have a question- can grid setpoint be set to a negative value?

If so then it should be able to sell from battery to grid which could be profitable in the future

That is exactly what I did with Sunny Boy Storage.
Allowed value for maximum import is shown as 0W ... 500,000W.
I entered -2000W. The box was highlighted in red. And it worked to force export.

In my test setup it thinks the CT is at grid input. It is actually right at the battery inverter AC connection. For a while I had an electric heater wired through same CT, to simulate house load. Reversed heater wire to simulate PV, testing shaving of PV to battery.

There are two ports for WattNode energy (power?) meters. You would think we could put one on grid connection, one on PV connection, but so far SMA denies that is supported (or rather said I don't need two meters.) Given comms to GT PV, the data could be obtained. Ought to be able to shave imports from grid, and charge only to the extent of PV, so allowed to export to grid.

Eventually after I work out the configuration bugs I plan to get utility PTO and deploy this to property with new NEM 2.0 agreement. We can add battery after that PV system is approved. Then I should be able to game the $0.30 summer and $0.20 winter spread, get larger credits.