Coulomb

It would help others if you told us how you did this. I assume that you reflashed the display with 22.x x or 30.xx. As far as I know, the WiFi module has its own firmware, and you can only see that firmware version under certain circumstances. I have never seen a firmware update file for updating the WiFi firmware. So I assume it's still there.

I have a dozen or so King II 5 kW firmwares, but no 6,2 kW firmwares, sorry. Only one of those is a Twin firmware, too. For some reason, these are far less common. I'm not certain, but I think that ones with a colour (or pseudo colour, with cellophane) displays are all clones. [ Edit: I mean the smaller display as shown above; the "round" colour displays (likely also using cellophane or similar) are genuine. ]

I guess that could work. It will be a bit of effort finding a suitable one, figuring out how to mount it, etc. You'll probably need to leave the bypass capacitors there, so you'll also ned to figure out strategic places to cut tracks. It would have to be a moderately large unit, since most of the idle power will be supplied from this supply (a little will be the original power supply's losses). So you'd be looking at about a 75 W unit for MAX models. It should also be able to be turned on and off like the original. And isolated input to output. Tricky.

What firmware version did you have, and what version did you replace it with? What display firmware version are you running? It seems that lately the display firmware may have to be updated when the main firmware is updated in a certain way. And of course, none of this is documented anywhere that I'm aware of.

Have you replaced the 1000μF capacitors? See for example this recent post:

Hmmm. How are your transformer winding skills? I always thought of transformers as non-repairable, but I keep coming across stories of people fixing them. Maybe someone should start a transformer repair business, if they have the skills required?

You should be able to, yes. I believe it's mks2.hex for Axpert Maxs (and of course MKS IIs and a few other models).

They can export power back to the AC-in port, yes. There are settings to manage this, which the Axperts lack. The on-grid models also have to have expensive certification, or they should.

These are patched firmware versions x11.06, based on factory firmware version 11.06 for the Axpert Max E 11kW '28066 Duplex. Do not use with any other firmware; in particular, not with any non-Duplex model. Duplex models have extra hardware for switching two AC-in sources, possibly in addition to two AC-out ports. The patched firmwares apply all of Georg594's original 90.06 patches to work around the "stuck at 90V" issue. They also fix the premature float bug, and return a version number like 211.06 from monitoring software. Use at your own risk. dsp_211.06_patched.inv contains version 211.06 forces the minimum MPPT voltage to 200V. dsp_111.06_patched.inv contains version 111.06 forces the minimum MPPT voltage to 150V. dsp_911.06_patched.inv contains version 911.06 forces the minimum MPPT voltage to 90V, as does the original, but it applies all the other patches (reduced current threshold, etc). Use only for systems with very small PV arrays, e.g. 4 72-cell panels per MPPT. dsp_original.inv contains the unpatched factory firmware version 11.06. The .inv files are just hex files, but irritatingly the Arena reflash tool looks for the .inv extension unless you override it. Firmware upload instructions for models with a removable display. dsp_x11.06_patched Max E '28066 Duplex.zip

Those caps can cause so many issues! The exact capacitance likely doesn't matter much; it will probably be the Effective Series Resistance that will make the difference. Higher value capacitors (thos with more microfarads) usually have lower internal resistance.

Yes, that's the whole idea. 99.9% of the patched firmware is the same as factory firmware version 65.08, a slightly later version of the 65.06 firmware you are running now. The handful of changes are to improve the low light PV production, and to fix the premature float bug.

Please post in English on this forum. Did you maybe start with Watchpower; that's for Axperts (off-grid models), then tried SolarPower afterwards (for "hybrid", on-grid, grid interactive models, usually called InfiniSolar models)? I have no idea if the monitoring software can handle Infinisolar models. That's maybe the problem. Solar Assistant may be the only one that can handle the Infinisolar models. You might need to exchange it for an iSolar model (these are more Axpert-like, possibly made by Voltronic, or possibly made by Easun under license). The iGrid models are grid interactive, and use a totally different command protocol. It's confusing how they use the same case for two vastly different models.

All I can think of is the buck converter; this basically adjusts the battery voltage on the bus side of the transformer. If the buck converter stops working, power can't flow into the battery. With your constant 2.2 kW load, you should still see about 2.2 kW from the panels, but maybe that's just the algorithm not sharing power optimally. The 55 V at the battery terminals might be normal for these high PV voltage models; I would not know. It may be the way that these models pass power around to feed the main power supply, when it could come from AC-in, PV-in, or the battery. Or you might have a diode leaky or shorted between the main power supply input and the battery positive rail.

The only patched firmware available for this issue on your model is patched firmware versions x65.08.

It may be that the battery is unbalanced. This is not unusual when the battery is new, and has sat on a shelf for possibly months. If so, it may eventually balance and allow charging to higher and higher SoC over a week or two. Warning code 69 merely means that the inverter has received a message from the battery BMS that it should not charge the battery until further notice. Later Voltronic firmwares deleted this warning message, as it caused more confusion than it helped. I'd rather know the reason that the inverter is not charging the battery, personally. Is the 50% showing on the inverter screen or the battery screen? In theory, the inverter should be getting the SoC% from the BMS, which should have a good idea of the actual SoC.

Is the battery voltage reading close to correct. according to a multimeter or the battery display or battery monitoring software? If the inverter thinks that the battery voltage is higher than it actually is, and you don't have many other loads, that could cause the PV production to reduce. Is the PV voltage close to correct, according to a suitable meter? (Measure this with extreme caution, and not with a cheap multimeter. Hundreds of volts DC is very dangerous! ) Perhaps you can just compare roughly with what the other inverter reports.

Watch out for damaged gate drive components, if you attempt to fix this yourself. There are plenty of pointers here and on the AEVA forum.

This is patched firmware version x63.05, based on factory firmware version 63.05 for the Axpert Max 11kW '2809 non-Twin. Do not use with any other model. The hex file is dated 04/11/2024. This firmware works around the "stuck at 90 V" problem by changing the current threshold from 0.50 A to 0.10 A. It also fixes the premature float bug, and changes the reported version number to x63.05. The internal display will continue to show version U1 63 05. Use at your own risk; I cannot test this firmware as I do not own an Axpert Max. It has been "sanity checked" by disassembling the firmware and confirming that the intended changes are present. 263.05 has a minimum MPPT voltage of 200 V. 163.05 has a minimum MPPT voltage of 150 V. 963.05 has a minimum MPPT voltage of 90 V; use only if you have very low PV voltage (e.g. 4-5 72-cell panels in series). It still has all the other patches, e.g. the lowered current threshold. dsp_original.hex has the factory firmware version 63.05. dsp.hex is the same (when unzipped) as 263.05. To use a different hex file, first delete dsp.hex, then duplicate (copy and paste) the chosen hex file in a file explorer, and finally rename the resultant file to dsp.hex. Unfortunately, this reflash tool only looks for the file dsp.hex. Do this before running the reflash tool. Firmware upload instructions. dsp_x63.05_patched_MaxII_11kW_2809_nonTwin.zip

I'm not familiar with any 1.5kW Axpert models. Does it have a high (350V+) maximum PV voltage, or around 145 V max? If the former, then it's a boost converter, and a shorted MOSFET will short the PV input (via the boost inductor). If the latter, then it's a buck converter, and a shorted MOSFET would not short the PV input. I would guess it's a high PV voltage model, and that the MOSFET is shorted. Note that the gate driver components are also likely damaged. There are traced schematics now in the AEVA forum.

Does the JK BMS talk a protocol that the inverter or monitoring software understand? I'm unclear on whether the inverter can operate on SoC% if the inverter is not connected directly to the battery's BMS. In any case, unfortunately 72.66 and its patched versions do not have the required support for SoC%. Several firmwares after 72.66 no longer have the "presumed good" MPPT logic. Factory firmware version 72.75 (the latest I know of, July 2024) has both. I have not patched 72.75. So I'd say that your best bet is to try 72.75 first. Some people report good MPPT behaviour with the "presumed good" logic, others prefer patched firmware. 72.75 is available here. However, note that the poster reports SoC% as not working. I have no idea why. I'd like to research this, but time is not available at present.

No, the on running 90.28 has a '2809 DSP; the others have a '28066 DSP. So the reflash tool will fail if you try. I have posted elsewhere my recommendations; I just wanted to make clear that they can't all be running 69.xx firmware, and don't need to be.

Ah. So they are Twins. I suggested 57.xx earlier, but that's not a Twin firmware. Sigh. I've edited the post. We'll continue this on PM, as I don't want some of my techniques to fall into the hands of clone makers, or cause Voltronic to tighten their security. You are mad as a cut snake! In a good way, that is! I would not have thought that someone with your skills should be spending time attempting to repair inverters with holes blown in them.

This is factory firmware version 69.07 for the Axpert MAX 7.2kW '28066 non-Twin. Do not use with any other model or version. Use at your own risk. The hex file is dated September 2023. That makes it ideal for paralleling with '2809 inverters running firmware version 45.13. Firmware upload instructions for models with a removable display. MAX7.2K_DSP_69.07.7z

Alas, the boot loaders are not part of the firmware update hex files. Since you have a clone, there is a small chance that they didn't use a Code Security Module (CSM) password. You can try to read the firmware from the good one. The addresses will be $3D8000 to $3F8000 (end address not inclusive). That's 256k bytes, as the C2000 word size is 16 bits (2 bytes). If the firmware is password protected, all these words will display as zeroes. The bootloader is from $3F4000 to the end (flash segment A). If using TI software like Code Composer Studio, be aware that TI has a non-standard hex format. Most software including Voltronic's reflash tools use Intel hex format with extensions for more than 64k bytes. I have liberated passwords from Voltronic firmwares before, but it's a complex process and it can't be used in your case. They seem to use different passwords for different models. So sorry, I think you'll have to buy another control board. BTW, are these Duplex models with dual AC input ports? If not, you probably need 57.xx firmware, and other possibilities exist. [ Edit: It can't be 57.xx; this is a non-Twin version, and 38.xx display firmware is for Twins. ] Edit: I would not have thought that these would have been repairable. Well done on getting them this far.

Hopefully it won't add too much resistance.