FoxESS Community - Owners & Installers Forum

Hi everyone,


Let’s say I’m using a 3-phase FOXESS hybrid inverter and an EP11I/EP12 battery.

I have a question about the reaction speed of the system when handling sudden power demands, especially from PWM-driven loads like induction cookers (or other appliances with rapidly fluctuating power usage).

Specifically, I’m wondering:

How quickly does the system switch between charging and discharging modes (battery-wise) when there’s a spike or surge in power demand?

For example, let’s say I have 4kW solar producing power, an average 2kW load, and thus the battery charging at 2kW. If a short load spike suddenly happens and surges by another 5kW [ 15sec on 50 sec off ] (total load now becomes 7kW), with only 4kW solar available, the system would need an additional 3kW for about 15 seconds. In this case, would the battery handle this surge, or would the system pull from the grid instead? ( yes took easy numbers to explain the case real loads will be different)

Has anyone had experience with this, or does anyone have any technical details on how the system manages these types of peak loads?

Additionally, technically, would this damage the battery if the system needs to handle such surges for hours and hours ?


Any insights would be greatly appreciated!

Thanks in advance!
Noel

I don't have a 3-phase inverter, but KH7 with EP11.
They should act similar really.

There will always be a lag with any inverter before it reacts, though I find it quick to respond. I have an electric shower and the demand hits 10kW, the battery will quickly supply upto 7kW (depending on how much solar generation) and the grid fills in the remaining 3kW.

A cooker going on and off reacts the same.


I always think of the electric as waves on a beach, with a sort of "ebb and flow" so if you are in your example producing excess solar, then you are "pushing" against the ocean to export, once you demand more local power, the waves come rushing in to fill the gap.

I am sure the battery will survive this abuse, EV cars can pull several hundred kWs when sending it, and then regen back all day long.

Yes, that’s the question: How long (or how big, time-measured = reaction speed ) is this lag, when switching from charging to discharging and vice versa?

I would like to have some specific numbers or timeframes here.

This pattern has been the most resource-consuming for me over the past three years of monitoring various types of data.

The main concern is whether I can overcome these 'short' surge peaks and avoid paying grid prices, which are more than 20 times (in price value) the value ( for what i get back for ) what I inject.


I want to repeat once more: a surge peak lasts about 15 seconds, followed by 50 seconds of normal operation. This pattern continues for hours.
If the battery were charged in an ideal case [ fast switching, / allmost no lag ( < 1sec, ideal would it be 0 ( what is not possible i agree)] this pattern would be covered by the battery, and not by the grid.



If the INVERTOR-LAG ( charge vs discharge) is 15 seconds or more (in this specific case), adding batteries won’t solve anything.





Noel

Short: I don't know the exact rate of the CT Clamp reaction times.
ModBus that is connected reports the Inverter power back every 2s, this has to be controlled so as not to overload the data in/out.

So it reacts at least every 2s, but when you look at the Inverter screen, the power shown fluctuates faster.

Also if you have not already done so, you can ask to have your Inverter set to "push" back on the grid say a permanent 50W to reduce your import costs.


Maybe someone knows the actual reaction times, we do have users with Shelly EM monitoring on their electrical circuits to give a better refresh rate, maybe their results show the reaction of the Inverter is quite quick.

Noel_V70 wrote: ↑Wed Mar 18, 2026 11:54 am Yes, that’s the question: How long (or how big, time-measured = reaction speed ) is this lag, when switching from charging to discharging and vice versa?

While I can't give you hard numbers on the reaction time, it must be very quick and pretty accurate because even without grid compensation set ours was keeping imports to ~40Wh/h, with grid compensation of -75W imports were ~5Wh/h, but the exports went up to 30+ Wh/h, still trying to find a balance.

What you may not have considered is the inverter constantly draws from the batteries, I've seen posts claiming 200W for 3 phase inverters, which it needs to able to react quickly.

As for solar hot water systems, we had one but the tubes had all cracked so was useless in the end and the company that made them stopped doing solar so no replacement tubes either, we were quoted AU$5,000 to replace + labour, which it turns out was better spent on a battery.

Our meter data is in 5 minute blocks to 10,000ths of a kWh or 10ths of a Wh.

From yesterday's data when the setting was -30W we exported between 2 to 3Wh/h and imported between 14 and 23Wh/h.

When it changed to -50W we exported between 2 and 15Wh/h, and we imported between 8 and 16Wh/h.

I’ve been digging a bit deeper into various FoxESS documents (datasheets and such), but I still haven’t found the information I need to make the right decision.

It makes me wonder—how difficult would it be for the FoxESS to measure this and include it in an “advanced” section of the inverter datasheets?

Regards
Noel


PS: When looking at the specs of some very cheap batteries, I came across this…

So for a sudden load spike:
The battery does NOT respond in milliseconds
Instead, it ramps output over hundreds of ms to a few seconds
Practical interpretation
Sudden load increase ~0.5–2 s typical

Updates are discrete power setpoints, not continuous analog tracking
Typical control loop cadence: ~0.5–2 seconds (inferred from smart meter + latency)
Output is often stepped (e.g., 100–800–2500 W) rather than fully smooth

Noel_V70 wrote: ↑Thu Mar 19, 2026 8:05 am I’ve been digging a bit deeper into various FoxESS documents (datasheets and such), but I still haven’t found the information I need to make the right decision.

It makes me wonder—how difficult would it be for the FoxESS to measure this and include it in an “advanced” section of the inverter datasheets?

I've found other documentation sorely lacking so not surprised to be honest.

What I've also come to realise with the app/website and the display of instantaneous data is it can show lots of importing and exporting imbalances with our new air fryer, but the meter data doesn't show any and it's wired and I would have thought a lot more sensitive than CT clamps.

The only ramp up I've noticed is after a battery firmware upgrade, it took up to 5 minutes before it would charge at 5kW again.