I am new here, and I hope this is on topic for this channel.

I have a bit of a strange situation with my off-grid solar setup... it works better on cloudy days!  

So I replaced a Axpert VM III 3kW system at the start of the year (because I had a battery warranty refused (2x eleksol Gel 250Ah)) a failed cell after less than 8 months use...

Anyway it is a learning process, and I upped my budget and went for 2x pylontech US3000C with a new 48V inverter, an 'Axpert 11K E'.  (There was not much choice since the 5kW inverters had all gone app only.)

The system has been running good up until last month, and then summer arrived...

I dont have many panels, only 6 for a total of ~2.7kW, but I adjust my usage to the weather, and making sure I am close to fully charged at the end of each day.

So I decided to add a 3rd US3000C to the system, since I had spare generation just not being used, or so I thought...

The inverter had been peaking at ~37A battery charge, and I made the assumption that it was a limit of only having two batteries, or the BMS playing it safe since I read that LiPo can be a bit slower to react?

However, what I found is that on 'full sun' days the system is getting stuck ~113V.  I'd previously noticed a ~90V sticky point, which it eventually passes.

I tried all kinds of things to resolve it, including splitting the single string into 2 strings (the inverter has 2 string inputs), trying one of the original batteries as master, going back to 2 batteries, trying a USR battery setting, cooling the panels (it gets hot here) and even giving them a good clean at last

So after 3 weeks of brainstorming, bam! a cloudy day, and instead of 1.9kW generation, it was suddenly hitting the magical 2.7kW again, between clouds of course

btw this thing sounds like it is taking off when it goes into high gear, I cant imagine what it would sound like with 11kW of generation!

So the next week, back to full sun and same problem.  Yesterday, clouds and working again!  So yes, I have a setup that prefers cloudy days, bizarre no?

Sorry I dont have graphs to share, it turns out the 'BT and WiFi' is actually WiFi only, with a .cn cloud app I wont install.  Although I have started a project to reverse engineer the WiFi protocol, but first my A/C is hungry for more power!!  37C forecast on Tuesday any ideas or tips please?

Cheers, Peter

PS Firmware U1 57.04, U2 12.21, U3 00.00

19 minutes ago, whitehexagon said:

I am new here, and I hope this is on topic for this channel.

I have a bit of a strange situation with my off-grid solar setup... it works better on cloudy days!  

So I replaced a Axpert VM III 3kW system at the start of the year (because I had a battery warranty refused (2x eleksol Gel 250Ah)) a failed cell after less than 8 months use...

Anyway it is a learning process, and I upped my budget and went for 2x pylontech US3000C with a new 48V inverter, an 'Axpert 11K E'.  (There was not much choice since the 5kW inverters had all gone app only.)

The system has been running good up until last month, and then summer arrived...

I dont have many panels, only 6 for a total of ~2.7kW, but I adjust my usage to the weather, and making sure I am close to fully charged at the end of each day.

So I decided to add a 3rd US3000C to the system, since I had spare generation just not being used, or so I thought...

The inverter had been peaking at ~37A battery charge, and I made the assumption that it was a limit of only having two batteries, or the BMS playing it safe since I read that LiPo can be a bit slower to react?

However, what I found is that on 'full sun' days the system is getting stuck ~113V.  I'd previously noticed a ~90V sticky point, which it eventually passes.

I tried all kinds of things to resolve it, including splitting the single string into 2 strings (the inverter has 2 string inputs), trying one of the original batteries as master, going back to 2 batteries, trying a USR battery setting, cooling the panels (it gets hot here) and even giving them a good clean at last

So after 3 weeks of brainstorming, bam! a cloudy day, and instead of 1.9kW generation, it was suddenly hitting the magical 2.7kW again, between clouds of course

btw this thing sounds like it is taking off when it goes into high gear, I cant imagine what it would sound like with 11kW of generation!

So the next week, back to full sun and same problem.  Yesterday, clouds and working again!  So yes, I have a setup that prefers cloudy days, bizarre no?

Sorry I dont have graphs to share, it turns out the 'BT and WiFi' is actually WiFi only, with a .cn cloud app I wont install.  Although I have started a project to reverse engineer the WiFi protocol, but first my A/C is hungry for more power!!  37C forecast on Tuesday any ideas or tips please?

Cheers, Peter

PS Firmware U1 57.04, U2 12.21, U3 00.00

Attach some pictures of consumption, production, battery power etc so we can see whats up with this puppy

Edited July 8, 20233 yr by HendrikBigChief

I dont have anything currently set up on the monitoring side, due to it wanting to use a .cn cloud server/app combi.  Is there another way to get the data locally? I dont have windows for the serial option.  The old inverter I tried unsuccessfully to decode the BLE data and was hoping to continue that project until I found this model is WiFi only.  I seen some nice looking graphana images here, something like that would be great!

Today it was stuck at 90V, and then again around the 113V / 2kW, not sure it will manage to charge today.  Almost like it prefers to just double that to ~230V

When it gets stuck, try to turn off PV (via proper disconnect switch). Give a minute or two, and turn on again.

If I recall correctly, I've read some posts in the past about the MPPT of various models getting stuck at a certain voltage. If it works as I suggested, it probably the known firmware bug.

I tried to settle the question of which is really the later firmware: 57.05 or 57.61. Alas, while 57.61 looks slightly more likely to be the latest (despite my earlier feeling that it was the other way around), there are just so many minor-looking changes I can't tell.

So you might want to try out this one, which is just one point ahead of what you have now:

and:

Please let us know the results if you try one or the other or both. The present behaviour is certainly strange, and not at all the usual one. Usually it gets stuck in shady conditions, and is fine (with a bit of a delay if not patched) in full sun.

Edited July 10, 20233 yr by Coulomb

@meetyg Thanks, I do this occasional when stuck at 90V, and it does the trick.  I dont have a proper disconnect switch (cant seem to find one here, only like a boat switch), so I am solar fuse switching... and so at higher generation levels I try shutting the whole thing down (remove all load first, then once the panels are not being drawn upon, disconnect them...  'rebooting the house' I call it   Sometimes it helps a little (today it went from a measly 1.6kW, to 1.8kW).  But we just had a 'blobby cloud' of the right type that created shade for around 5 minutes, and now she is running at 2kW.

Could it be also be temperature related? today is hotter, building for tomorrows heat-wave...  maybe increased DC volt loss on the panel cables?

Panels are 6x Ja Solar 455W, Vmp 41.82, Imp 10.88A, Voc 49.85 wired in series if not obvious.  Currently 2 strings of 3.

@Coulomb Thanks, Before I attempt a firmware update, is there a way to capture the existing installed firmware?  Just in case I get locked out and have no power to get back online   Actually it might also be quite useful in trying to continue my reverse engineering of the new App only WiFi 'feature'...

Also anyway to capture some graphs without a windows machine, eg hooking up a Raspberry PI on the serial, or any such magic?  It would be great to have a before and after.

 

 

PS that 57.61 looks very interesting! since there is mention of the 115V ceiling that is suspiciously close to my 113V values.  And just very happy to read that it isnt just me with this issue!!  I am looking forward to testing this   thanks.

20 minutes ago, whitehexagon said:

Could it be also be temperature related? today is hotter, building for tomorrows heat-wave...  maybe increased DC volt loss on the panel cables?

Certainly it could be heat related; remember that hot panels have lower voltage.

20 minutes ago, whitehexagon said:

Panels are 6x Ja Solar 455W, Vmp 41.82, Imp 10.88A, Voc 49.85 wired in series if not obvious.  Currently 2 strings of 3.

41.82 (at 25°C) x 3 = 125 V, perhaps 10% lower at 60°C so that's 113 V, towards the low end of the MPPT range. The solar charge controller has to work hard to boost that to bus voltage, making the inverter hotter and noisier as well.

[ Edit: So in the hot sun, your panels are going to drop to around 113 V under load no matter what. My suspicion is that the MPPT will find it harder to MPPT to that than to the voltage the panels sag to with 6S or even 4S. ]

In the shade, the panels will cool and produce higher voltage; that could be why your inverter is "sun shy".

You might find it works a lot better as 6S (unless you meant that you actually have 2 strings each 6S).

When you get to two more panels, you could probably go down to two strings of 4S.

20 minutes ago, whitehexagon said:

is there a way to capture the existing installed firmware? 

No. The best you can do is post what the present firmware is; there is a chance it's in my stash. [ Edit: Duh! You've said it's 57.04, and the closest I have is the 57.05 as linked. Sigh. ]

20 minutes ago, whitehexagon said:

Also anyway to capture some graphs without a windows machine, eg hooking up a Raspberry PI on the serial, or any such magic?

Perhaps Watchpower for Linux or Mac? Certainly there is Solar Assistant and the like for Pis, but these are for money. One of them is free, but Windows based.

Edited July 10, 20233 yr by Coulomb

It was the same issue with the single string of 6.  The 3-3 split was a creative way to test if the problem was a faulty panel (3 of them are not easy to reach) allowing me to see if the voltage was equally halved   ie (I can see ~113V for each group of 3 panels) I plan to put it back to a single string now that we might be on to a solution. 

They have an offer on some 550W panels this month, and maybe I will put 3 of them onto the second string facing west, and maybe a long run to a solitary panel facing east (that one might just help the 6 reach start voltage a bit sooner?( I assume the MPPT combines V from both strings?)).  Otherwise I am out of space and into partial shade territory, and I already had some old 160W panels fail due to that problem.  Although I read that these half-cell? panels are less susceptible to that issue...

Thank you so much for your response!  I was starting to pull the last of my hair out, especially after the previous install disaster.

Edited July 10, 20233 yr by whitehexagon

18 hours ago, whitehexagon said:

I assume the MPPT combines V from both strings?)

No, the Axpert Maxs have two completely independent (but not isolated) MPPTs. Each boots separately to the bus voltage.

Edited July 11, 20233 yr by Coulomb

Thanks. An interesting observation today.  Disconnecting one string when both were stuck at the 90V level, somehow triggered the 2nd string to already adjust to a higher voltage, even before I reconnected the other string.

So I like your calculation regarding possible 10% loss due to heat, and surface temperature is probably much higher than air temperature!  Although I would expect to see some variation there if that is the case, but without proper graphing it is hard to tell. 

However, one thing springs to mind, even if the V is stuck at ~113V, and we assume 10% is heat degradation, then why is the W affected more like 30-40%, that can only mean the A being pulled from the panels is also varying?  And I assume the inverter is controlling that 'tap' .  Sorry I have basic knowledge only in these topics.

Anyway I get some courage to do the firmware later, maybe I toss a coin

On my system I've noticed that on slightly cloudy days there is a lot more UV reaching my panels and they give far better performance than on sunny days with clear skies.

Thanks @Proff72 The first half of this year has been great, but unseasonably cloudy...  I wish I had graphs for it all!  Anyway today is solid blue sky, and it just briefly jumped to 125V (both strings giving my projected max of 250V), but almost like it panics at the amount of power available and winds itself back down to 113V...

Looking in the 57.61 archive I can see a 'dsp.hex' file, the readme.txt seems illegible...  The inverter manual mentions being able to do this from a USB stick (3s button press), but does not include instructions, is it just a case of copy file to FAT32 formatted USB stick and pray...  (I dont have a windows machine to run included .exe files.)  Or is the USB update approach documented please?  I seem to recall doing something on another device and having to create a binary compatible file first...

23 hours ago, Coulomb said:

tried to settle the question of which is really the later firmware: 57.05 or 57.61. Alas, while 57.61 looks slightly more likely to be the latest (despite my earlier feeling that it was the other way around), there are just so many minor-looking changes I can't tell.

I had factory version 57.04, inverter purchased one year ago; so i believe 57.05 is proximal to it, quite old;

Updated from January, this year to 57.61 

 

20 minutes ago, whitehexagon said:

Looking in the 57.61 archive I can see a 'dsp.hex' file, the readme.txt seems illegible...  The inverter manual mentions being able to do this from a USB stick (3s button press), but does not include instructions, is it just a case of copy file to FAT32 formatted USB stick and pray...  (I dont have a windows machine to run included .exe files.)  Or is the USB update approach documented please?  I seem to recall doing something on another device and having to create a binary compatible file first...

you should flash using RJ45 cable and Usb to serial adapter with 1 Kohm resistor mounted on it; instructions on forum;

USB stick flashing not work;

@Solar RO Ah interesting, had you also had similar issues that triggered your update or other reason to do so?  Do you also find the inverter a tad noisy?

@Solar RO Thanks, posted at the same time.  I dont have a windows installation here, I did notice .pdf instructions in the older zip file that mentioned serial ports, cables and windows

18 minutes ago, whitehexagon said:

@Solar RO Ah interesting, had you also had similar issues that triggered your update or other reason to do so?  Do you also find the inverter a tad noisy?

Is not noisy; after 57.61 is more efficient in cloudy day; I can't say 90 V is solved entirely, but is diminished.

 

 

5 minutes ago, whitehexagon said:

@Solar RO Thanks, posted at the same time.  I dont have a windows installation here, I did notice .pdf instructions in the older zip file that mentioned serial ports, cables and windows

indeed, is required a flashing tool, windows exe; but you can open a VM on Mac:)

sounds fun hehe, and on a hackintosh right

btw do you happen to have a pdf of the manual please, the included paper one is too tiny for my old eyes.  Model is 'MAX E 11K' (not the twin AC one).  I also spotted a technical manual during my search, but again it was for a slightly different model.  There seem to be around 3 or 4 variants of the 11K...

Since I cant update the firmware today, I think I'll try and grab a couple of these discount panels this month and see what difference they make.

Thank you all so much for your help and insights. 

Firmware Upgrade SOP via RS232.pdf RemotePanelFirmwareUpgradeProcess20190819.pdf

PS I found a RJ45/serial cable in the inverter box, is that the right one, or is it something special for the firmware updates that needs the resistor.  I also previously bought FTDI/FT232RL chipset USB/Serial connector for trying to interface to the old VM III 3kW inverter from a Raspberry Pi after the battery failure, but that is another story spoiler: the adapter melted!

see above plus below; the one in box should be ok, normally; check colors if they fit with schema and insert a resistor on white wire;

 

regarding flashing tool; pay attention where to click, update DSP or MCU ;

Edited July 11, 20233 yr by Solar RO

29 minutes ago, whitehexagon said:

sounds fun hehe, and on a hackintosh right

btw do you happen to have a pdf of the manual please, the included paper one is too tiny for my old eyes.  Model is 'MAX E 11K' (not the twin AC one).  I also spotted a technical manual during my search, but again it was for a slightly different model.  There seem to be around 3 or 4 variants of the 11K...

Since I cant update the firmware today, I think I'll try and grab a couple of these discount panels this month and see what difference they make.

Thank you all so much for your help and insights. 

MAX E 11K - i don't know about it; check the posts of Coulomb; it exists more versions: MAX 1 11KW, MAX II 11 KW with color round display; also MAx 11kw Twin; 

the 57.61 is for MAX 1 with removable display, squared one.; pay attentions to chipset versions, otherwise you risk to brick inverter; 

check more times the chipset version, appropriate firmware, flashing tool, procedure, cable, etc;

before flash  and validate with more data sources 

Edited July 11, 20233 yr by Solar RO