Jazdzor

No! 25.xx is for removable displays. Update to 39.27, available here.

I thought I'd start a new thread about the various chipsets in the Voltronic Axpert inverters.
The main DSP chips have until about mid 2022 been TMS320F2809s, except for some low power models with 8-bit microcontrollers, and the VM series which have the TMS320F28062 chip, which has half the flash size of the '2809. Since about mid 2022, the TMS320F28066 is used. [ Edit: The Axpert King 1 probably used the '28066  DSP earlier than that. ] The '28066 is a more modern DSP chip, with only a single power supply (3.3 V), where the '2809 used two power supplies (3.3 V and 1.8 V). As a result of this, you can tell by looking at a control board whether it uses the older or newer chip: the older ('2809) chip will have two voltage regulators, and the newer chip ('28066) will have one:

There is also the confounding factor of the "experimental" main firmwares, version 9x.yy. These numbers seem to be almost randomly assigned, so you can't read anything into the major version number, except that it's experimental in some sense.
The DSP firmwares for the two chipsets are incompatible, and have different major version numbers (the digits to the left of the period). If you try to flash the wrong chipset (but get everything else right), the firmware upload will fail, and your inverter should not be bricked. It should come good after a restart.
At about the same time, the displays (removable and colour with round LED ring) started changing microcontroller. This may or may not be a result of the worldwide chip shortage, and the same could be said about the main DSP chip change. Until about mid 2022, the displays always had an ST Microsystems STM32F107VCT6, which is an ARM Cortex-M3 V7 or V7M3. Some later displays have used the GigaDevice GD32F105VCT6.
The display firmwares for the two display chipsets have also been incompatible. Unfortunately, if you reflash a display with the wrong type of firmware, there is no anti-bricking protection! Fortunately, the display is usually not permanently bricked, but it can be quite an effort to recover it, requiring patience and multiple attempts with varying timing from power-on to clicking the update firmware button on the PC. Firmware version numbers were at least segregated: the GigaDevice firmware version numbers started with a leading "1" (e.g. 122.34), and the ST Microsystems firmware version numbers do not start with a leading "1" (e.g. 22.34). But here is the kicker: the displays can only show the last 4 digits!! Fortunately, the round colour GigaDevice displays attempt to indicate the leading one with a column of dots, like extended decimal points:

I finally found a pair of comparable xx.yy and 1xx.yy display firmwares: 36.01 and 136.01. My theory, not totally confirmed, is that these only differ in whatever customisations are needed for the different chipsets; the functionality and logic would, per this theory, be identical. So I was curious to compare the firmwares.
Alas, internal addresses used in the firmware seem to be different. For example, different addresses appear to be used on the I2C bus, and extra GPIO bits seem to be used. If I'm right, that means that the display boards may have different circuits, even though at first glance they seem to be the same except for the brand of ARM chip. So that makes it impractical to take a 1xx.yy firmware and with a few patches turn it into a xx.yy firmware, or vice versa. With the difficulty of obtaining certain firmware versions, that might have been a useful ability. [ Edit: Fortunately, this is not the case; see below. ]
Edit 16/Mar/2023: One concrete difference seems to be how the two chips send and receive I2C data. I2C seems to be used for the EEPROM chip and some other functions I have yet to identify. In the ST chips, there is hardware that implements I2C; in the GigaDevice chips, this seems to be missing. However, I believe that ARM input/output pins are very programmable, so it seems likely that they were able to use the same pins for I2C data, clock, and enable pins, and emulate the I2C hardware with "bit banging" and busy wait loops. Crude, but if you can't buy the ST chips, what else can you do? It's not like the display chips are chock full of time critical functions.
Edit 2 16/Mar/2023: When I went to look at the I2C code in firmware version 12.21, which is supposedly able to run on both chipsets, I expected both hardware and bit banging code, selecting the appropriate one for the present hardware. But I just found the bit banging code. Then I realised: ST chips can bit bang just as well as GigaDevice chips, and the speed of both would be close enough for I2C, which from vague memory is more or less synchronous (each side waits for the  other to transition, except for the actual data bits). There may be more to the unification than that, i.e. it's possible that older 1xx.yy firmware won't automatically work in hardware that came with 1xx.zz firmware. But at least the I2C hardware difference is easy to address.
Edit 3 12/Oct/2023: For the Axpert King 1 and Axpert VM III display firmwares (02.xx and 102.xx), they seem to have switched exclusively to bit banging firmware (which should be compatible with both chipsets, i.e. universal) somewhere between 02.70 (not universal) and 02.81 (universal). Probably 02.73 was also not universal, since 102.73 is not universal. For Axpert MAX display firmware, the existence of 112.19 strongly suggests that 12.19 was non-universal. 12.21 is known to be universal. So with the latest display firmwares, it is no longer necessary to know which chipset the display firmware is written for; it should work on both chipsets equally well.
Edit: Added paragraph on experimental firmware versions.
Edit: Added more on the display firmware differences.
Edit: Added that the Axpert King 1 uses '28066 DSP.
Edit 3: Added a paragraph about which King 1 and Max display firmwares are universal.

👍
Yes, they dropped the leading 1 when they figured out that one of the firmwares works fine for both display chipsets.
 
Maybe. Certainly worth a try.

Hmmm. Sounds like you have a display/comms board with slow opto couplers. They operate at a very lazy 2400 bps, except when reflashing, in which case they switch to 4x the speed (9600 bps). If you're handy with electronics, you could replace them yourself, or you can try and get the inverter repaired (they will replace the whole display board).
Or it may be that your USB to RS-232 adapter isn't one that works well with these inverters.
I would try copying the 60.03 hex file to a blank USB stick, renaming it to dsp.hex, and attempt updating via the USB stick. That will at least bypass the opto couplers between the display and the RS-232 port, and also your USB to RS-232 adapter.
Firmware update instructions for models with a round display.