Hey guys,

 

My setup: Axpert 5kVA, 1 Pylontech 48V US3000B, 9 x 330W PV.

 

I noticed that during the day (when the PV is supplying a decent amount of power, say 1.5kW) I can switch on the kettle (2.5kW) and the combined power of the PV and the battery seems enough to power it, since the inverter stays on battery/PV. But at night (I've got it configured through ICC to run off the battery until 35% SoC), the minute I switch on the kettle, the inverter switches to grid.

 

The charging amps for the Pylontechs is 30A - would the discharge amps be the same? So that it can deliver max 30A * 48V = 1440W? I guess that would explain why the battery alone cannot power the 2.5kW kettle?

 

1. Is this correct?
2. How does the Axpert "know" to switch to grid? Is it aware of how much power the battery can supply, or does it somehow detect that it's not supplying what it needs to power the load, and then switch to grid?
3. Would adding a second Pylontech (in parallel) mean the combined batteries can supply 60A, so 2.8kW, meaning I would be able to run the kettle on the batteries? (I understand buying a R14000 battery just to run a kettle at night makes no sense - the question is more theoretical; and to know what the next step might need to be if I were to want to go fully off-grid at some point.)

 

As an aside question, I saw in some posts from last year that @plonkster was recommending people using Pylontechs (or LFPs in general, I suppose) to have at least 2 batteries anyway. I didn't quite understand the reasoning, but is that still a recommendation, and why exactly?

 

Thanks!

Frans

4 minutes ago, SolarFrans said:

The charging amps for the Pylontechs is 30A - would the discharge amps be the same? So that it can deliver max 30A * 48V = 1440W? I guess that would explain why the battery alone cannot power the 2.5kW kettle?

Those Axperts go back to grid if the voltage sags too low. I don't think there is any other reason behind this. The Axpert can't communicate with the battery either, and as far as I know ICC doesn't pull data for anything other than basic voltage/current/soc. No charge current control is done as with other more capable setups (usually those with canbus capability).

During the day the voltage sags less because you have incoming PV.

About the discharge rates. The battery requests a max discharge rate of 30A per module, which is roughly C/2 (ie at that rate it will be empty in 2 hours). Why? Cause if you go faster then the cells heat up and things are less efficient. But the inverter has no way of enforcing the discharge rate, that is dictated solely by your loads. What is more, the discharge rate is more like a recommendation from the BMS. You may go higher or lower, you can go as high as 2C (empty in 30 minutes) on LFP batteries, though usually only for a minute or so. The BMS will disconnect you if you overdo it.

But of course the voltage will sag if you discharge it that hard. So that is why one will generally recommend at least two racks, cause with just one rack you have 2Kwh storage on a 5KVA inverter... that's more than 2C at the peak.

13 minutes ago, plonkster said:

Those Axperts go back to grid if the voltage sags too low. I don't think there is any other reason behind this. The Axpert can't communicate with the battery either, and as far as I know ICC doesn't pull data for anything other than basic voltage/current/soc. No charge current control is done as with other more capable setups (usually those with canbus capability).

During the day the voltage sags less because you have incoming PV.

OK, so no magic there.

Back-to-grid-voltage on the Axpert is set to 48V (since I have ICC handling the Battery/Grid switching based on actual SoC I intentionally set it to a low value) and I can't see the voltage dropping below 48V from the metrics that ICC reports. But then I guess ICC only samples once every 10 seconds or so, so I might still be missing a quick dip below 48V?

I'll play around a bit more with the back to grid voltage then. I'm hesitant to put it much lower than 48V since, if ICC would fail (if somehow the Pi loses power or crashes), the Axpert might very well drain the batteries completely before it even gets to that voltage -- the Pylontech seems to have very steady voltage and doesn't seem to go outside the 49V - 51V at all, even at low SoC. But let me try and lower it temporarily for an experiment.

 

Quote

But of course the voltage will sag if you discharge it that hard. So that is why one will generally recommend at least two racks, cause with just one rack you have 2Kwh storage on a 5KVA inverter... that's more than 2C at the peak.

OK, that makes sense. We're usually not pulling more than 1kW so I guess 1 x 3kWh is fine.

 

Thanks for the help!

Frans

3 hours ago, SolarFrans said:

Back-to-grid-voltage on the Axpert is set to 48V (since I have ICC handling the Battery/Grid switching based on actual SoC I intentionally set it to a low value) and I can't see the voltage dropping below 48V from the metrics that ICC reports. But then I guess ICC only samples once every 10 seconds or so, so I might still be missing a quick dip below 48V?

Yes, I believe that's what is happening. In unpatched firmware (and patched firmware before the latest revision "e"), the battery voltage used was not load compensated. In other words, it just uses the raw battery voltage, regardless of whether that's because of low SOC or because of a load.

Patched firmware 73.00e and 72.20e use a rough estimation of the battery's internal resistance (it can be adjusted in 5 steps) to compensate for the internal resistance. If this worked perfectly, there would be no change in compensated battery voltage for a sudden load like a kettle, so there would be zero reason to switch to bypass mode. In practice it's not perfect, but it doesn't have to be; it just needs to "blunten" the sag enough so it doesn't go to bypass.

I've been running it for a while, and it does help a lot, especially at night as you've noted. We call this feature KettleKomp™ (of course, the ™ is tongue in cheek, we don't trademark it at all). If your machine is eligible for patched firmware, perhaps give it a go. It's a lot cheaper than another battery.

I've dropped my Back to Grid voltage way down to 45V and it's still switching to bypass when I switch on the kettle early morning or late at night. It's not instant, though, only after a few seconds.

I'm wondering about what, e.g., this page says regarding discharge power:

Quote

• Deliver up to 5KW* with single module (2.4KWh) – * please note this value is inverter/charger dependant, typically the maximum achieved value is 1.2kW per battery

Wondering what "typically the maximum achieved value is 1.2kW per battery" means and if that is perhaps my issue.

Is anyone aware of limits to discharge power/amps in the Axpert 5KVA ?

 

Thanks,

Frans

The firmware upgrade won't work for me - I've got a 5kVA model with a power factor of 1, 48V max charge voltage, and firmware 74.10.

57 minutes ago, SolarFrans said:

I've dropped my Back to Grid voltage way down to 45V and it's still switching to bypass when I switch on the kettle early morning or late at night.

What about the battery low cutoff voltage, setting 29? If the battery voltage sags below that value plus 4.0 V, then the battery is considered "weak", so it will switch to line mode. If you were able to use the LFP flavour of patched firmware, then the 4.0 V would be come a much more reasonable 1.0 V, but sadly, as you say, you don't have that option.

Low battery cut-off is at 45.5V right now. You're saying it would switch at 49.5V already? OK, let me try and lower that significantly and see what happens... Thanks!

 

What is it about this model that there is no patched firmware available? Is it just that no one has made the effort to write it; or is it something about the hardware that doesn't allow custom firmware to be uploaded?

I'm a programmer myself; if there's some way to port the changes in the current patched firmware to the 5kVA 48V model, I might want to give it a try. Any information available on this?

 

Thanks!

Frans

Hi @SolarFrans

I just watched the following youtube video from CentrurionSolar this morning.  It cover/solves just that topic.

https://www.youtube.com/watch?v=M9PLPzRQjdo

 Using ICC and a cable to connect the batteries to ICC the real battery level is monitored.

 

6 hours ago, SolarFrans said:

What is it about this model that there is no patched firmware available? Is it just that no one has made the effort to write it; or is it something about the hardware that doesn't allow custom firmware to be uploaded?

It's mainly that no-one has access to a 74.XX firmware to patch from. If we had that, it's about 4 person days of work, plus testing. Weber and I don't have such a machine for testing. There is also the potential for bricking the machine if we make certain types of error.

6 hours ago, SolarFrans said:

I'm a programmer myself; if there's some way to port the changes in the current patched firmware to the 5kVA 48V model, I might want to give it a try. Any information available on this?

It takes a bit of a special programmer to do this; you need to be familiar with DSP assembly language. The only information on doing this is in the .txt file inside each patched firmware's .zip file. There are brief explanations for what each of the up to sixty something patches do, and assembly language source code for the larger blocks of patched code.

1 hour ago, 1Hallux said:

I just watched the following youtube video from CentrurionSolar this morning.  It cover/solves just that topic.

Neat. So you can switch from mains to battery on a proper PylonTech SOC percentage.

But I note that you still have to set the battery cutoff voltage quite low, so the machine doesn't go switching due to low battery voltage.

So the battery cut-off voltage was the culprit. Even though it was set at 45.5V, I guess the magical +4V that @Coulomb mentioned meant that it was still switching to grid when it dropped below 49.5V. I've lowered this to 42.0V and just tested now with a kettle, and it stayed on battery, with the battery supplying 3kW. (I can probably increase it to ~44V now that I know this is the culprit.)

 

I was already running with ICC and the Pylontech cable to get real SOC readings, and had already adjusted the "back to grid" voltage to a low value so the inverter's normal logic wouldn't override ICC's decisions. But the inverter was still switching to grid when I turned a kettle on because of the cut-off voltage setting and the +4V thing.

 

Thanks @Coulomb!

 

Frans