Coulomb

I don't own a VM model, but it seems that VM III and VM IV firmwares, at least for the 4kW and 6kW models, are interchangeable. The following post suggests that version 70.00 has good fan support, at least for a VM IV Twin:

If your present main (U1) firmware version is 78.xx, then yes, otherwise no.

This is secondary (display, U2) factory firmware version 12.30 for the Axpert Max and Axpert King II (NOT the original King 1), non-Twin. Do not use with any other model or version. Use at your own risk. The hex file is dated 08/10/2024. Firmware upload instructions for models with a removable display. This firmware processes the QDOP command, so with suitably recent main firmware, it should support SoC% for settings 12,13, and 29 when connected to a suitable battery's BMS. Remote Panel_Reflash_MCU(12.30).rar

I have no experience with use of these inverters. But what it sounds like to me is that when there are clouds, there is a sudden drop of PV voltage, resulting in a sudden drop of bus voltage. At that time, the main power supply may "skip a beat" or ten. Other power supplies cope with this by having decent sized capacitors to "ride through" these periods, but the display, which relies on the "12 VAC" from the main power supply (often called HFPWR or similar, for high frequency power, may not be coping, as the capacitors are undersized, or perhaps dried out with age. Has this problem been evident since new? The Ouyads seem to be clones, and it's not clear whether the problem is the main power supply skipping too many beats, or the display not coping with the skips. Replacing the larger electrolytics in the display with decent quality parts may cure the problem. Unfortunately, there is no schematic for the display boards that I'm aware of, so it may be difficult choosing the right value capacitor, and/or making it fit. This may be one area where clones are not as good as the original.

All I can tell you is that 71.03 is the only 71.xx file for VM III/IV Twin in my collection. So the safest path of updating to a later 71.xx is not available to you. So unless you can get a suitable update from your inverter supplier, you'll have to try one of the other VM III/IV firmwares from this topic, and hope that it doesn't brick your inverter or damage it, and pick one that a reader says is quieter. Hopefully, if all else fails, you can revert back to 71.03 from here. But there are no guarantees, unfortunately. If I had time, and I currently don't because of a family issue, I might be able to figure out what the quiet firmwares do, and back patch it to 71.03. But that's also a lot of work.

@Jalil Ahmad, these appear to be low resolution thumbnails, with details hard to read. Could you repost the originals, perhaps in a zip file? Especially part 2 of the control board, which I don't have so far. Thanks in advance.

What is your main (U1) firmware version? That will tell me exactly what model of Axpert inverter you have. At a wild guess, I'd say you have capacitors drying out, perhaps in the Solar Charge Controller. How old is the inverter?

That's what the user manual says. But a quick scan of one of the King 1 firmwares seems (at a quick glance) to be associated with battery under voltage. If so, that would explain the fault code with such a low battery voltage reading. Unfortunately, at present I don't have the time to investigate fully. So then it's a matter of why the battery voltage is reading so low. It could be one or more of the battery voltage sense resistors going high value. These high value SMD resistors subject to moderate voltage for long periods of time seem to fail more often than average. If it's the -12 V rail, then that would affect many readings. Do other readings seem reasonable?

I'd say it has to be one of the transistors associated with the start circuit, or the associated resistors: So Q10 or Q9 in the above. Or possibly the SCC's opto coupler (if present) is shorted. Try removing the 2-pin cable from the Solar Charge Controller to the main board.

Are you using the latest version of Watchpower? Do the values make sense as battery voltages, or are they really SoC% displayed as if they were voltages? My vague understanding of this is that the relevant command always responds as SoC if certain conditions are met, e.g. if the DSP responds to the QDOP command. So I'm surprised that Watchpower can respond with sensible battery voltages. Edit: Was that using the front panel?

They're probably reporting output power, i.e. AC-in less losses (assuming no AC charging). Each 5 kW inverter uses around 50 W when inverting; about 15 W when not.

Ah yes. 72.75 happens to be the first one that implements the QDOP command. Yes. I see that now. But 72.75 has what I have come to call the "presumed good" MPPT algorithm. Though I see that it only ever seems to be available in the "experimental" firmwares, i.e. those numbered xx.yy with yy ≥ 60. So you were definitely getting the stuck at 90 V issue with 72.75? If so, I might as well patch 72.08, which is a "mainline" version (yy much less than 60). It also handles QDOP, and is a few months later than 7275. I also have 72.75, so I could patch that instead, but if it doesn't solve the stick at 90 V issue properly, I can't see the point. Thanks in advance for any data.

I can confirm that 81.08 and 22.25 are still the latest firmwares available for the Axpert Max II 8kW '28066 non-Twin. Firmware update instructions for models with a colour display. Factory main firmware version 81.08 for the above. It does process the QDOP command. Factory display firmware version 22.25 for the above. It does process the QDOP command. Having both firmwares support the QDOP command means that settings 12, 13, and 27 will work with SOC% rather than battery voltage, provided that you are connected to a suitable battery with a supported BMS.

This is main (U1) factory firmware version 51.04 for the Axpert MKS III (PIP-5048MGX) 5kW '2809 non-Twin. This is like an Axpert MKS II with a removable display. Do not use with any other model or version. Use at your own risk. The hex file is dated 17/Aug/2023. Firmware upload instructions for models with a removable display. This firmware does not process the QDOP command, so even with suitably recent display firmware, it will not support SoC% for settings 12,13, and 29 when connected to a suitable battery's BMS. The RAR password is "flash", without the quotes. MKSIII_Inverter_Main_Reflash_Ver71.17 pw flash.rar

So these are Axpert MKS III models, also known as the PIP-5048MGX. The 71.xx are '2809 firmwares; the 51.xx are '28066 firmwares. There should be no problem paralleling 71.xx and 51.xx as long as they are reasonably well matched with respect to the compilation date. Unfortunately, for whatever reason, there are few firmwares for this model available. I have only 71.10 (2021), 71.17 (Aug 2023), and 51.04 (Nov 23). 51.04 is available here. I'll upload 71.17 next, in case it helps.

It's not compatible with a 3 kVA inverter. For a 4 kVA Twin, it will probably work, as VM IV and VM III firmware seems to be interchangeable these days. But I like to use the same major version number, e.g. if you inverter's existing main firmware was 84.xx, that would be fine.

Please post in English on this forum. Do you mean you were updating the firmware on the inverter? You probably just need to try again and again, possibly varying the timing between when you restart the inverter and press "go" on the firmware reflash tool. See If you think your inverter is bricked.

That's off topic for this thread, but it's probably version 41.18 from August 2023. Since the higher power models came out, there has not been a lot of change to these inverters' firmware.

Presuming that you mean where does it fit in the recentness of release, the 84.03 hex file is dated 10/July/2025. So it's later than 60.10, the latest of the 60.xx series that I'm aware of (60.10 is dated 14/Jun/2024).

As for whether this is the right course of action, I don't know. It seems unlikely (to me) to be an SCC firmware issue, but it's certainly possible. I note that you have to follow the SCC upload instructions closely; there is a binary file that you have to set up in the Windows registry, and I think you have to change your serial port to COM1. It's not in our general firmware upload instructions any more, because we thought it was never necessary, and is such a hassle. This might be an exception. It's also possible that SCC updating is not possible on models with a removable or colour display. It might be possible to bypass the removable display just for the purposes of updating the SCC firmware. 02.71/102.71 is quite old display firmware; you could try updating to a later version. They might have fixed some problem with the SCC uploading. For several years now they have dropped the 102 series; it's quite safe to use the 02.xx series for xx ≥ 81. The latest 02.xx display firmware is 02.89. For other readers: don't use this display firmware on the Axpert King II, or the Axpert VM III newer models (4kW 24V or 6kW 48V).

Are you saying that your monitoring software is reporting 4.5 kW load, when you're sure you have only 1 kW actual load? Could there be a load that you're not aware of? Are you sure that you're not seeing AC input power, which would include any AC charging? Can you tell if the reported load is sharing roughly equally (assuming actual parallel connection, not three-phase)? It's possible that one of the inverters has a hardware fault that is causing it to report far higher power than it is actually supplying. This could cause problems with power sharing. The most likely cause of this over reporting of output power is a high value (hundreds of kilo-ohms or meg-ohms) resistor in the AC output current sensing circuit. If so, the replacement parts will cost well under a dollar, excluding shipping, but of course you have to know your way around electronics. Having two identical models (apart from the DSP processor variation) would make repair slightly easier. Another possibility is failure of one of the EEPROMs. Do your three inverters all report their correct serial numbers? If one of them reports something like 5533555... as the serial number, then your EEPROM likely has been corrupted. This can affect measurements such as AC output power.

Please post in English on this forum. Sadly, there are very few 82.xx firmwares about. I have not seen any new ones since 82.63.

Please quote your existing main (U1) firmware version. There are so many variants with completely incompatible firmware.

That single 50-way connector (as opposed to three separated 16-way connectors) on the control board seems to indicate a VM (Value line Mppt) model, i.e. one that can't be paralleled and has a few other features trimmed or altered. If you can focus clearly on the sticker on the DSP chip, it might give clues to the existing main firmware version, which is always the best guide to what model it's designed for, as well as the best firmware update candidates. Sometimes, probably more often with more recent boards, the firmware version is directly in the number. But other times, it's encoded and you can only compare with other labels. Searching the control board PCB number 156-050133-01G leads to a single post mentioning a Sunlife 4.2kW model, see https://cdn.enfsolar.com/z/pp/2024/3/eg504tdev693volr/sunlife-4-6kw-scaled.pdf This suggests that it is indeed an Axpert VM IV or clone thereof, either 24V 4kW or 48V 6kW, possibly with modified firmware that will claim (and possibly deliver) an extra 0.2kW of rated power. Possibly a Twin output model.

That's part of it. The high PV voltage models vary with their minimum MPPT voltage: some are 120V, others 90V, and others have 60V. The start voltage is often higher than the minimum MPPT voltage. So for example the start voltage could be 150V; once this is reached, it could be dragged down to say 90V by load and/or light conditions, yet keep working. If low light dropped the panel voltage below the minimum MPPT voltage, then the PV will switch off until the start voltage (150V in this example) is seen again. I'm amazed that the minimum MPPT voltage could be (semi officially?) set as low as 30V. This would indeed be very inefficient, so the solar converter would work hard, but I would not have expected that to cause excessive interference.