Coulomb

But still only via USB, not RS-232? The update tool only talks RS-232. I note that Modbus is RS-485, not RS-232. You also have to make sure that nothing else is using COM3.

Those both look ok, but perhaps check the connections anyway. The connection to pin 4 or 7 doesn't matter as it's not used in this application; you need the connections to pins 1,2 and 8 as per this post: https://forums.aeva.asn.au/viewtopic.php?p=57914#p57914 It could be that you have the wrong driver for the PL2032 chip and your version of Windows. Google it. You can check the wiring by shorting pins 2 and 3 (tx to rx) on the D9 connector, and using a comms program connected to COM5. Everything that you type should be echoed back to you. With local echo turned on, you should see double echoes (e.g. type h, see two h characters).

Not what you want to hear, but I wonder if the King models (original and/or King II) are better with poor grids. They have double conversion. I suspect that means that they don't have to synchronise with the grid, and so may not care nearly as much about grid quality.

It could be unhappy with the grid frequency, or the rate of change of frequency, or perhaps rate of change of voltage. What's the grid like? The voltage seems to have a quite wide range, suggesting poor voltage regulation.

That's fortunate; I have 75.07 and 75.08. Can you first try factory firmware version 75.08? There is a small chance that it will fix the problem. You can use the QBOOT command to see if it is fixed, or whether fault code 90 will come up again eventually. Use Voltronics' communications tool, available from the first few lines of this post: https://forums.aeva.asn.au/viewtopic.php?title=pip4048ms-inverter&p=57733&t=4332#p57733 Report what it sends back. If possible, do this before you update at all. You might have to use the ATE1 command to bypass the display processor for the QBOOT command to work. Instructions here: https://forums.aeva.asn.au/viewtopic.php?p=96489#p96489

That's a good sign; it means that the display isn't bricked. Is this using a type A to micro-USB cable direct to the PC? Is that the cable that came with the inverter, RJ-45 at one end, and D9 (RS-232) at the other? And are you plugging the two cables together? It should be display -> cable that came with inverter -> USB to RS-232 adapter -> USB port of Windows computer. That should be able to run Watchpower. It's possible that the USB to RS-232 adapter isn't one that the Voltronic software works with. Check for a COM port appearing in the Windows Device Manager, and that the COM port is COM1 through COM9. I don't know about the display update tool, but early Axpert reflash tools would not work with COM10 or higher. No, the reflash tools set the baud rate automatically. It actually starts at 2400 bps, and changes to 9600 bps as the reflash actually starts. I have no idea why they put a baud rate drop-down box there. That looks OK, but as usual it doesn't say what the chipset is. The AEVA pages have suggested bands, and known good brands available from South Africa, but I don't know where you're from, and what may be available. When trying to reflash, I assume that you're following the upload instructions, and are getting the UPG YES on the display? But you probably need to get it working with Watchpower via RS-232 first. I also assume that you're plugging the cable that came with the inverter into the port nearest the PC/monitor icon.

You confused me with the two inverters. As rowesly mentioned, you need three inverters for real 3-phase loads, and they need special hardware. When you get to that stage, you will have to have effectively one battery (e.g. two 24V banks in parallel) that all 3 inverters run off. So yes as far as I know you can common neutrals of two inverters with separate batteries, but that does you no good for real 3-phase loads like compressors. 3-phase compressors usually have very high inrush current requirements, so you will probably need 3 48V inverters.

I think it should be OK to connect the AC-out neutrals together. You only need to parallel the batteries with Voltronic inverters when you actually parallel the outputs (lives as well as neutrals), or you want a genuine 3-phase configuration. I'm curious as to why you want to parallel the battery negatives only.

Yes. Interesting. The serial number is wrong, compared to what I'm used to, there is no barcode, and the QR code has only the serial number. But the rest looks totally legitimate, and sometimes bigger resellers are allowed significant variability of the sticker and serial number. I've now gone from 98% sure it's legitimate to 90%. So there is still a very good chance that updating the firmware will work. But if you don't like those odds, there is about a 98% chance that running with your existing firmwares should work. I recommend checking the voltage difference in the outputs before final paralleling, by putting one multimeter lead on each output. The reading should be well under 20 VAC, preferably under 5 VAC.

Many of these inverters are rebranded Axperts; some are clones, a few are work-alikes. A quick search indicates to me that TommaTech are a genuine Voltronic rebranding, but I can't quickly find the label on the side that usually clinches this determination. Either post a photo of the label here (preferably with some screws and the small sticker with the manufacturing date also visible), or check Do I Own a Clone? if you'd like to me more certain. The photo I saw looked exactly like an Axpert E 11 kW model, and the 73.xx main firmware and y12.xx display firmware are correct for an Axpert 11 kW '28066 non-twin, so I think that there is little doubt that the indicated Axpert firmware will work.

The latest main (U1) firmware in that series is 73.78, available here: https://powerforum.co.za/topic/20595-axpert-max-i-11kw-100v-sticking/?do=findComment&comment=197048 For unpatched factory firmware, choose file name dsp_73.78_original.inv . Or choose one of the patched versions for improved MPPT performance in cloudy weather. Edit: However, it's not essential to have identical firmware, but the closer the better, and identical is best. So you could try paralleling them with the different firmware versions if you want, or if you have trouble updating the firmware.

The VM series saves a little money by having one firmware for both 24 V and 48 V models. It decides which hardware it is running by reading the voltage on a pin at power-up. The result is stored in RAM; I think it's static RAM but it might be dynamic. In either case, a glitch in the power supply could cause the RAM contents to change, making it think that it's running on the wrong hardware. This can be very bad; I would always restart immediately if that happens. Though possibly the problem could happen due to a comms error that happened to fluke a good CRC. The display remembers what the nominal voltage of the system is from the results of the last QPIRI command. That's far less dangerous. You could find out if it's the display or the inverter itself by unplugging and re-plugging the RJ-45 cable from the removable display (not so easy if you happen to have a model with the round/colour display). Warning: this will effectively switch off the main switch, which will drop the loads. Reconnect the display before the inverter completely shuts down; it should not shut down if you have a charging source (AC-in or PV). A few screws should make this RJ-45 cable accessible. If it's the removable display firmware, then updating that may fix the problem.

The Deye seems to be a 51.2V LFP battery, meaning it's probably 16S; those settings are for Pylontech, which is 15S. So multiply the voltage settings by a factor of 16/15. The inverter has no idea of the correct SoC with battery type USEr. It is basing the SOC on crude lead acid voltage levels. For the cable to work, you first need to know what protocol (comms and hardware) that the Deye BMS is using. It's quite possible that it's not one of the several that Voltronic support so far.

Unclear so far; more photos might confirm. But all the low PV voltage Solar Charge Controllers are by ck converters, not full bridges. So not 4 separate switching devices, effectively one switch, typically made up of 2-6 devices in parallel. To complicate matters, Voltronic often use an interleaved design, effectively two buck converters. An interleaved SCC is like the below, except that for convenience the transistors are in the negative side, not the positive side: The right hand transistors in the above are synchronous rectifiers for added efficiency; they are omitted in all the real SCCs I've examined, using ordinary high-speed diodes instead. You can tell if it's interleaved by the pair of inductors, usually off-board.

What are the present (U1) firmware versions?

The second (bottom) numbers will be batch code and/or date of manufacturing. It may be that some were replaced, or it might mean that they had parts from two different orders in their parts bin. Your parts with part number NCEP85T16 seem to be OK per my crude checks, which are: * Do they have a reasonable looking datasheet online? * Is the datasheet reasonably complete (all the usual sections, graphs, etc) * Are the specs reasonable? Usually I only look at the headline figures, like max voltage and max continuous current, and a very few other figures, like input capacitance (should be under 10 000 pF (10 nF) for Axpert designs, since that's the limit for the gate drivers. For MOSFETs, I'll glance at their on resistance (RDSₒₙ). I'll also often glance at the delays (turn on, rise and fall times), to make sure they don't look ridiculously large. * Is there a real company name, usually with an address. * Do the parts have a logo that matches the datasheet? I've seen some Ali Express parts claiming loudly that they are genuine parts, yet there is no logo, and the font for the part number is very plain and uniform. Again, all these pass my crude checks. Re the Aishi capacitors, they seem to have a reasonable reputation ("second tier", not quite United Chemicon or Nichicon quality). I could not quickly find the datasheet for this exact capacitor; the RF series only went to 1500μF at 35 V working. But it's not unheard of for Voltronic to use a custom made capacitor! That makes it difficult to find a suitable replacement. These seem to be rated at 6000h @ 105°C, which is quite good (if the rating is to be believed, of course).

Those phase chopping dimmers are evil in mu opinion. I'm not surprised they upset many inverters.

Ah! Boy am I learning a lot today. So Triple A must be one of those resellers (like EASun and several others I believe) that rebrand Voltronic but also Must or other clone/workalike products. That QR code says "Originally manufactured by Voltronic Power". At last, a way of telling that an inverter is genuine! I've noticed those orange removable displays years ago. But I foolishly assumed that it was just an artistic thing to "go with" the orange blob on top. Sigh. Now I'll have to figure out how one of those buttons does up and down at the same time (at least, that's what it suggests to me). Thanks for the lesson!

Ah! Do they look like this? I can't tell from the photo if those are switch buttons or touch buttons. I note that Must inverters are not made by Voltronic, and often have non-Voltronic firmware (e.g. setting numbers different to that of Axpert, yet mostly do the same thing).

I've not heard of these. Oh! That's news to me. Just a description and perhaps a photo of these touch buttons, please. Now the multiplicity of display firmware versions is starting to make a little sense. I assume that all models with the round/colour display have touch buttons; is that correct? Thanks in advance!

Ah! That explains the 25 mm² wire to the resistors as well! Oh! So that's your only neutral to earth bond. So yes, don't attempt to undo that bond. It will disconnect if you have a suitable AC-in source. So if you ever do get a generator, or borrow one or whatever, make sure that it has neutral connected to earth, as this will be your only neutral to earth bond when AC-in powers the loads.

Ah! I thought ALL VM IIIs (old type and new type) have the removable display. * Does this VM III with the touch keys have a round (colour) display? * Is it branded Kodak? I am also very confused by the above (i.e. that 57.07 → 25.08 works for the VM III). Assuming that the 57.07 is a round (colour) display firmware, as I've just realised the 52.07 is. Is it possible that you confused 52.07 and 57.07 somewhere? It's a long shot.

I do however have 52.07 for King II Twin and others, and usually King 1 goes with VM III old, and King II goes with VM III new. Did I mention that these major firmware version numbers are confusing? So I find that 52.07 uses different GPIO pins to 25.12, or more accurately the same pins but in a different order. [ Edit: I now realise that 52.07 is for a Kodak OG-PLUS 6.2, which has a round display, yet it's a King II Twin. It seems that only Kodak Kings IIs (and possibly only the Twins of these) have the round display. ] There is a slight chance that the keys are still working, just jumbled up. So maybe the escape key is now the down key, or similar. Worth a try? If so, the solution might be tiny stickers over the key names 🤪

That might be the difference between 57.xx display firmware and 25.xx firmware: the way that the buttons are scanned. That seems crazy to me, but what do I know. Unfortunately, I don't have any 57.xx firmware to check, or for you to revert back to. Presumably, you can still use the inverter with monitoring software?

My experience is that they are usually wildly inaccurate like that, but sometimes much more reasonable, probably correct. But I could never figure out what the pattern was. So I assume it's something to do with the 72.00 firmware.