Hi, I have 3 Kings 5kW (RACK version) and each is handling one phase. They are connected to a pack of 6 pylontech US3000C batteries. I'm currently trying to figure out why they don't want to switch into bypass/eco mode when utility power is available.

I've noticed that when the battery is full, one inverter -- always the same one -- is nevertheless still trying to charge and reports a 68% battery capacity, the rest is reporting 100% and not charging. All of them report the battery voltage being at 52V. Does anybody have an idea what might be going on?

I've checked the cables as well as the voltages on many points along the connection between the inverters and the battery pack and everything seems to be the same and as expected. All cables have the same length. Batteries report 100% SOC via the console.

Btw. I use battery type USE with custom voltages set based on recommendations from this forum and the setting are the same on all inverters.

Thanks in advance for any hints on what to check or how fix this!

1 hour ago, solardan said:

I use battery type USE with custom voltages set based on recommendations from this forum and the setting are the same on all inverters.

Are you sure? The battery charge current settings can differ from inverter to inverter in a paralleled or 3-phase situation. It seems to me that perhaps one has a different valie for setting 16, Solar Energy Priority (which on other models determines charger source priority).

Though the different battery capacity suggests other settings may be different.

Can you check the settings via the front panel? Otherwise, you might have to move the RS-232 cable from one inverter to the next to read the settings via monitoring software.

1 hour ago, Coulomb said:

Are you sure? The battery charge current settings can differ from inverter to inverter in a paralleled or 3-phase situation. It seems to me that perhaps one has a different valie for setting 16, Solar Energy Priority (which on other models determines charger source priority).

Though the different battery capacity suggests other settings may be different.

Can you check the settings via the front panel? Otherwise, you might have to move the RS-232 cable from one inverter to the next to read the settings via monitoring software.

16 is set to SLb UCb on all of them. Otherwise the parameters are as follows, again on all of them:

battery_voltage                 48.0            V   
battery_recharge_voltage        46.0            V   
battery_under_voltage           45.0            V   
battery_bulk_charge_voltage     52.4            V   
battery_float_charge_voltage    52.0            V   
battery_type                    User                
max_ac_charging_current         60              A   
max_charging_current            60              A   
output_source_priority          Utility first       
charger_source_priority         Solar + Utility     
battery_redischarge_voltage     54.0            V

These were read via `mpp-solar -c QPIRI` for each inverter. I've checked these values via the front panel as well.

One more observation I have is that the reported battery capacity by the misbehaving inverter is 72% right before full charge and then drops to 68% when the battery gets to 100% SOC.

Edited December 16, 20222 yr by solardan

9 hours ago, solardan said:

One more observation I have is that the reported battery capacity by the misbehaving inverter is 72% right before full charge and then drops to 68% when the battery gets to 100% SOC.

My understanding is that the monitoring software (which I assume connects to only one inverter) is sending a PBMS or similar command, which contains the true SoC from the BMS. The master is supposed to gather that sort of information and distribute it to all the slaves. I haven't specifically checked that SoC is one of the pieces of data that gets passed around. I'm also hazy on whether the inverter that you send the PBMS command to has to be the master or not. I assume not, as masterhood is supposed to be random (even though as we all know, 99% of the time, the same machine becomes master).

Is the one that connects to the Pi the master most of the time now?

If you change the comms cable from the Pi to talk directly to the inverter with the bad SoC, does that fix that inverter's SoC? I imagine that then one of the others will have the bad SoC. Presumably the bad SoC results from guessing the SoC based on battery voltage. Strange that only one has the bad SoC.

 

5 hours ago, Coulomb said:

My understanding is that the monitoring software (which I assume connects to only one inverter) is sending a PBMS or similar command, which contains the true SoC from the BMS. The master is supposed to gather that sort of information and distribute it to all the slaves. I haven't specifically checked that SoC is one of the pieces of data that gets passed around. I'm also hazy on whether the inverter that you send the PBMS command to has to be the master or not. I assume not, as masterhood is supposed to be random (even though as we all know, 99% of the time, the same machine becomes master).

Is the one that connects to the Pi the master most of the time now?

I have assumed that in my setup, master is the one which handles Phase 1 and it is the one where I modify settings which subsequently get distributed to the other inverters -- well, only some get actually distributed, some do not like MAX utility charging current which I have to set manually on each of them.

Anyway, the inverter which reports "wrong" SoC is handling Phase 3. And it is always this one so far, after 2 restarts of the whole system. Phase 1 and Phase 2 always report 100% when the battery is really at 100%. Otherwise the values reported by all inverters are off but that is because I don't have the pylons connected to the inverter so that it can't override my inverter settings.

5 hours ago, Coulomb said:

If you change the comms cable from the Pi to talk directly to the inverter with the bad SoC, does that fix that inverter's SoC? I imagine that then one of the others will have the bad SoC. Presumably the bad SoC results from guessing the SoC based on battery voltage. Strange that only one has the bad SoC.

 

I have comms cable connected to all of them at the same time (via the microUSB ports on the inverter). So I have /dev/hidraw0 /dev/hidraw1 /dev/hidraw2 devices on my Pi and I can send commands to each inverter separately. I tried sending QPGS0 to each inverter and then QPGS0 QPGS1 QPGS2 via only one device. The reported capacity was the same or off by 1% from each other -- but I'm in charging mode now and the baterries are around 60%. I'm going to check again when SoC is 100%.

Looks like the readings are the same no matter which inverter I interrogate. The following is reading stats via one inverter:

$ for i in 0 1 2; do mpp-solar -p /dev/hidraw0 -c QPGS$i | egrep '(Phase|^battery)'; done
--
battery_voltage                 52.0            V   
battery_charging_current        18              A   
battery_capacity                68              %   
output_mode                     Phase 3 of 3 phase output           
battery_discharge_current       0               A   
--
battery_voltage                 52.0            V   
battery_charging_current        0               A   
battery_capacity                100             %   
output_mode                     Phase 1 of 3 phase output           
battery_discharge_current       8               A   
--
battery_voltage                 52.0            V   
battery_charging_current        0               A   
battery_capacity                100             %   
output_mode                     Phase 2 of 3 phase output           
battery_discharge_current       9               A

And this is the output when reading from each inverter separately:

$ for i in 0 1 2; do mpp-solar -p /dev/hidraw$i -c QPGS0 | egrep '(Phase|^battery)'; done
--
battery_voltage                 52.0            V   
battery_charging_current        17              A   
battery_capacity                68              %   
output_mode                     Phase 3 of 3 phase output           
battery_discharge_current       0               A   
--
battery_voltage                 52.0            V   
battery_charging_current        0               A   
battery_capacity                100             %   
output_mode                     Phase 1 of 3 phase output           
battery_discharge_current       8               A   
--
battery_voltage                 52.0            V   
battery_charging_current        0               A   
battery_capacity                100             %   
output_mode                     Phase 2 of 3 phase output           
battery_discharge_current       8               A

I've been reading firmware for the Axpert MKS IV, it happened to be a modern one I had open. The display knows the true SoC from the BMS, and it seems to communicate that to the DSP via the PBMS command. The second parameter of that command is the BMS SoC. That gets stored in a variable that is rarely used. But it's also put into a structure that is sent by CAN bus message to the master under certain conditions.

However, there is only one place that is sent, and that's when you're already a master. So this makes little sense to me, unless this message is intended to be sent only when an inverter changes from master to slave, e.g. when CAN bus comms are lost and some other inverter wins the race to masterhood when it comes back.

It seems buggy to me, but it's most likely because I don't fully understand the code. If the structure with the SoC in it was never set in slaves because of this theorised bug, then all slaves would show zero percent SoC, since all memory is cleared at the beginning of the firmware. Something randomly overwriting just the SoC in only one slave seems unlikely too.

Plus, the fact that you read the same value from all inverters indicates that the value is being passed around.

So something seems to be causing that one slave that the SoC is 68%. I don't know what that could be.

Okay, thanks for the responses anyway! I'm going to perform some random experiments like physically interchanging the inverters to handle a different phase, switch the parallel interconnect cards, etc. to see whether the problem is isolated to that one inverter. If I find out something new, I'll post it as an update...

I tried to change the setting 29 / Low DC cut-off voltage and noticed that when set 42V and lower, then the battery capacity reading (including the indicator on display) gets to 95%. Setting the lowest possible 40V value doesn't get me to 100% however and I probably would like to have such values set anyway. The capacity reading on other inverters are still 100% (when battery says 100% SoC).

Haven't tried any other experiments yet but it seems to be a software bug to me. All inverters read the same voltages, yet only one is reporting wrong capacity. But it is weird that only one is buggy as it has the same firmware as one of the other two I have (U1 75.05). I have also an older one with an older U1 version which I bought first but it is working / reporting correctly. Maybe I should try to sync to one version on all of them. Would anyone have a recommendation on which firmware to try?

Edited December 19, 20222 yr by solardan

I've managed to get the P3 inverter to report 100% capacity now. Changed setting of output source priority to SUB and it reports 100% when the FV panels are connected and providing some charging current. When I disconnect them, it reports below 100% again. Only the P3 inverter is behaving that way. Now I at least know that the bypass / eco mode not working is not due to the capacity under-reporting because it does not switch to bypass even if all inverters report 100%. I guess it has something to do with the 3-phase setup as other users are reporting it not working for them with such setup.

23 hours ago, solardan said:

then the battery capacity reading (including the indicator on display) gets to 95%. Setting the lowest possible 40V value doesn't get me to 100%...

After 2am here, so I don't have time to check, but there is definitely a case where the guessed SoC maxes out at 95%, and in fact is set to 95% in some modes. It depends on whether it's in battery or line mode, from memory.

I'm not familiar with three phase operation, but I assume that they should all be in battery mode together, or all in line mode together. And of course Kings are a little weird with their two line modes (bypass and double conversion).

I have 75.08, if it's not already in the files section. So that might be worth a try.

46 minutes ago, Coulomb said:

After 2am here, so I don't have time to check, but there is definitely a case where the guessed SoC maxes out at 95%, and in fact is set to 95% in some modes. It depends on whether it's in battery or line mode, from memory.

I'm not familiar with three phase operation, but I assume that they should all be in battery mode together, or all in line mode together. And of course Kings are a little weird with their two line modes (bypass and double conversion).

Thanks! I'm going to try how it behaves in SBU mode which is the last one I didn't try and then the FW upgrade 🤞

46 minutes ago, Coulomb said:

I have 75.08, if it's not already in the files section. So that might be worth a try.

Found it, thanks again!

Hello! I have got a similar problem.

I finally got to the point where I connected the batteries (3 pcs Pylontech US5000, later i will double the pack) and set the 3 inverters (Axpert King II all the same 50.01 firmware) to P1 P2 P3…
So far, there have been 2 inverters up in parallel (one Master, the other Slave) and both communicated the Pylontech pack, the battery icon was flashing, without an error code.
Now it only sees what is connected to the battery. P1 is connected to the battery with a factory cable. It's flawless. The other two fail with error code 61.
The communication cables between the inverters are connected according to the description. The current sharing cables (black with red screw ends with green clips) are not connected.
Neither solar panel nor AC is connected yet.
Do you have an idea why only one inverter sees the battery and the other two don't? 🤔 (Unfortunately, plugging the cable into the A/CAN or B/RS485 port of the battery does not change anything.) Or does it need 3 phases AC? 🤔

 

On 2022/12/16 at 3:41 PM, solardan said:

I'm currently trying to figure out why they don't want to switch into bypass/eco mode when utility power is available.

The bypass/ECO mode in the King is not available when connected in parallel.

This is not well documented.  You will find it at the bottom of page 29 of the manual on the AEVA site (USER’S MANUAL (mppsolar.com))

The inverters not connected to the battery BMS should be configured with battery type USE, not PYL as you would expect. Warning code 61 only arises if the battery type is one of the BMS types (PYL, WEC, LIB, etc). I assume that the master will control the charging for the slaves. The master is not necessarily the one connected to the battery BMS. In theory, which one becomes master is random, though it can be forced by staggered turn-on. The first one on will remain the  master.

Thank you! It works! 🙏