Coulomb

This is what I found when fixing a similar fault: I have no idea why my tops are so bent in the middle, and yours are so flat. The above is from this post: https://forums.aeva.asn.au/viewtopic.php?p=65282#p65282 Edit: Yours doesn't seem to be going negative; perhaps I misunderstood what you were saying. Is your -12 V supply working? What volts per division was you oscilloscope on? Is yours at the SG3525 output or at the MOSFET gates?

It will be worth replacing, but its likely a symptom not a cause. Likely, something is loading the +12 V rail that this diode supplies. The noise it likely the transformer saturating, which will cause the current to increase a lot. The chip will sense the overload, pull back the current, the transformer stops saturating, and the cycle repeats, hence the noise. It's likely a UF202, Ultra Fast 2A 100 V. You might want to buy a few of these 😧 With a bench supply on the +12 V rail, look for any chips or other parts using that supply getting hot. You may need to supply -12 V as well, as some parts like the SG3525 use both. A non-contact thermometer is useful for this. BC is the marking code. It's a short code that they write on SMD parts because if the full part number and other info would not fit. These are 5.6 V zeners, likely whatever power you can get in that small package will do. These dying is likely symptomatic of high voltage getting through the gates of the IGBTs. Check all the gate driver parts, obviously as well as the IGBTs for gate shorted to drain or source. This one looks like one of those nightmare repairs, sadly.

Some inverters like the Voltronic VM III have a facility to set a different output source priority (like what you call Utility mode) for each hour of the day. You may have to check directly on the inverters' displays, if Solar Assistant doesn't handle that feature. Has your system been running more than 60 days? With serial numbers like that, I'd worry about fault code 90. But perhaps the manufacturer has disabled that check.

Think about each of the three delta loads separately. For each, current flows from one inverter's live output and back through the other. Then sum the currents for the other two delta loads. Each inverter sees a load equivalent to a single phase load. The three inverters' neutrals must be connected together; they have full load current but they all add up to zero. If you like, each inverter's neutral is supplied by the sum of the other two inverters' neutrals. It tends to do your head in, until you finally get used to it. If you're that way inclined, sone vector diagrams for the various currents may help.

Thanks, but please post in English.

I already have it, thanks. I believe that it will. That's certainly how my ancient 5 kVA model would work, but it has a very different solar charge arrangement. But I believe that they attempt to replicate the same behaviour on the high PV voltage models. The battery also has to co-operate, if it's one with a built-in BMS. Again, I suspect that it will, but don't know for sure.

It's exactly the same MPPT algorithm as for 81.07, if you know that one. It's not the "presumed good" algorithm, but I no longer think that the "presumed good" is the best. So sorry, unless you know 81.07, you'll have to try it and see. The MPPT algorithm is also the same one as in 81.05, which I have patched, so if necessary it should not be hard to patch 81.08 as well.

Oh, silly me, I found it after all. Sorry. Now that I compare 73.76 and 73.78, I see that they do indeed access the EEPROM differently. Strangely, it seems to be 73.76 that is the unusual one, at least at first glance. I don't have time to look into this in detail right now, but my preliminary theory is that 73.76 has unusual EEPROM accessing, so if you have an inverter that is running happily with this firmware, then you should be very careful updating to other firmware versions. I believe that I can now identify fairly quicky if a firmware has unusual EEPROM accesses, so hopefully we can avoid problems with serial numbers, calibration values, and so on in future.

Heh. When I started the Pylontech recommended voltages table, ICC was the dominant monitoring software. I guess that was a while ago now. I suspect that ICC worked a little differently than Solar Assistant does now. I also have to come clean that I don't use a Pylontech or any other battery with a factory built-in BMS. My battery has a BMS, but it's a custom one totally different to any factory BMS. It's also been a long time since I roughly knew how the BMS control via RS-485 or CAN bus works. I note that your battery type above is Pylontech BMS. That means that the Pylontech BMS is calling the shots. Was the battery type that way when the damage occurred? It's really unfortunate that the BMSs all seem to work with very high voltages, and seem to assume superhuman control systems that Voltronic inverters in particular can't live up to. When the inverter's battery type is Pylontech or one of the other several BMS types, settings like the bulk/absorb/CV battery voltage and float voltage are completely ignored. My table, as I noted in that post, is for people who aren't using the battery BMS connected to the inverter via the RS-485 or CAN bus cable. Are you connecting the BMS to the inverter directly, or to Solar Assistant? If the latter, I'm afraid that I don't know how it really works. It may be that SA tells the inverter that the battery type is USER, and manages the battery using messages from the battery BMS as a guide. The inverters themselves have two battery charge current settings: maximum total charge current, and maximum utility charge current. I think that the idea is that you decide the total charge current based on the battery manufacturer's recommended maximum charge current, but the maximum utility charge current can be lower than that, in case you don't want to spend too much money on utility charges, or the generator has a limit, etc. I note that maximum total charge current is one of the settings that gets taken over by the BMS: "If selected, programs of 02, 26, 27 and 29 will be automatically set up. No need for further setting." Setting 02 is the maximum total charge current, 26 and 27 are the bulk/absorb/CV and float battery volage settings respectively, and 29 is the low DC cut-off voltage. That's resumable because as above, the BMS takes over these settings. What is setting 05 on the inverter itself? Presumably either PYL or USE. Sorry I can't be more help.

Welcome. We need far more details about what you need, preferably including an existing firmware version number, and what you hope to get fixed by the newer firmware.

Sorry, I didn't notice this post somehow. 57.xx seems to be the '2809 version of the 73.xx '28066 firmware. If you try to load 57.xx firmware onto a model that came with 73.xx firmware, the reflash tool will give an error, and the firmware won't be touched. It's not clear to me what firmware and model combination you are talking about. It should be possible to parallel inverters with 57.xx firmware with inverters with 73.xx firmware (beware, there are other totally different firmwares with 73.xx firmware, such as my ancient MKS 5kVA 4kW models now running patched versions of 73.00, and some VM IIIs. As long as they are both 11 kW Max E non-Twins, they should parallel OK, despite the different firmware versions.

Yes, it's certainly possible. This incompatibility with existing EEPROM contents is concerning. I hope to investigate when time becomes available. I don't have the 73.76 firmware update file. I need that to develop a patch, obviously. Could you post a link here, or PM me with it, please? It's also possible that I've screwed up the patch for 73.78, but usually that would not lead to working firmware at all, i.e. far worse problems than serial numbers and calibrations being wrong.

Is the serial number 55355535553535 by any chance? Do you have 3 separate Pis talking to the 3 inverters? Otherwise, how do you combine the command redponses? Or does Solar Assistant drive 3 separate serial ports somehow? Do these BMS talk to one of the inverters, or one of the Pis? Does Solar Assistant talk the JK BMS protocol? Ut sounds pretty sophisticated. Sorry, most of these questions are for my curiosity, though I might be able to help in the end.

I get this sort of thing from on-line warehouses like DigiKey, Mouser, element14/Farnell, and RS Online. I believe that as long as they take South African currency and post to South Africa, they should be the same parts at the same quality, and at a very similar price (possibly excluding shipping) once the currency is converted. Th days of walking into a store and picking exactly the parts you need off a shelf are long gone, I'm afraid. You might be able to source these at a store like Altronics, but frankly I doubt it. They will probably have capacitors with very ordinary life and quality, which is not what you want when refurbishing something. My post on capacitor replacements: https://powerforum.co.za/topic/14666-repair-of-axpert-inverters-a-journey-started/page/4/#comment-150549 This post has the power supply replacements, bus capacitors if ever needed (these are large and expensive, and don't seem to fail often), and Weber's recommended battery capacitor replacements. The latter are mainly for longer life, rather than getting something going at minimum cost. They may no longer be needed at least as often, since the MOSFETs have typically been upgraded from 75 V to 80 V (a dizzying 6.7% increase!). There are complete part numbers and inventory codes there, with links to Australian versions of the parts warehouses. The Australian version merely means that the prices are in Australian dollars, and the shipping is expected to be to Australia. Note that it often takes 1-2 weeks for parts to arrive; it's rather variable. Hopefully the recent dock strike won't affect this; they should be shipping by air. If you want to order for arrival in South Africa, you can probably change the URLs by changing ".com.au" to ".co.za" .

81.08 is now available and processes the QDOP command, so it should support SoC% for settings 12, 13, 29 if the display firmware is recent enough. 22.21 is recent enough. This is factory firmware version 81.08 for the Axpert Max II 8kW '28066 non-Twin. Do not use with any other model. Use at your own risk. The hex file is dated 2024-10-03. It processes the QDOP command, so with suitably recent display firmware, it should support SoC% for settings 12,13, and 29 when connected to a suitable battery's BMS. Firmware upload instructions for models with a removable display. MAXII_8K_81.08.7z

No ECU specialist here, but from your traced schematic, it's likely that all the resistors have the same value. It's unclear in the above photo whether you can read any of the lower ones. They will all be very roughly one ohm, but since they set the current through the loads (coils or solenoids), it's critical to get the right value. Sorry I'm not much help here. Edit: BTW, I'm guessing that the emitters of the IGBTs will be separate; perhaps a fault is making them appear to be paralleled. I would think that you would want independent control of each IGBT, unless the collectors are also paralleled.

The display firmware supports it, the DSP firmware does not. Sadly, I don't have any 73.xx firmware that does.

Yes, 61.03 is too old (July 2023) for that. Alas, that's the most recent 61.xx firmware that I have; even 61.65 is February 2023. So we have to wait for a later 61.xx firmware update to turn up.

Sorry, big motors are just brutal with their start-up current. For what it's worth, I don't like either of the above two suggestions. Switching on an off rapidly will confuse the control systems. The phase control of "dimmers" will also confuse the control systems, and doesn't actually reduce the peak current all that much, unless there is filtering, which would not fit into that small box. The proper way would be a VFD (also called VFC I think), which chops the power (like the phase control dimmers, but at a higher frequency), but also has filtering. But this is quite bulky and expensive. There's no easy way around it: to start a large motor, you will need a large battery.

You have to tell us what firmware you are running now, since that tells us what DSP chipset you are running, and sometimes other important details.

Alas, phases don't "merge" without expensive conversion equipment. Probably the easiest thing, still not trivial, is some way to swap each house onto the next phase. When this happens, two houses share the same phase, so they have to be careful not to blow the street fuse and/or overload the conductors. Then you still need to swap back when power is restored to the missing phase, this time under load. The switch would have to be rated for at least 450 VAC (phase to phase voltage). It's not trivial.

Alas, no. I have no 75.xx for Axpert Max II 11kW '28066 Twins, and only one Max II 11kW firmware at all (65.08, dated July 2023, older than your inverter). One day I'll try again to make a '28066 firmware from a '2809 firmware; I tried once but I don't think it worked well. I'm suspecting a difference in clock speed now; that should be easy enough to check. But I have way too much on my plate at present, sorry.

It should not make any difference, but it might anyway. When a source is not available, the firmware is supposed to take the next available option in sequence.

I believe that's an example of integral windup. When a control system is at maximum output for a while, if nothing is done to prevent it, the Integral control term (the "I" in PI or PID control) increases without limit, and when circumstances change, the control system takes a long time to reduce the large integral term to a reasonable value. The result is typically overshoot, but in this case it's undershoot. It's a real shame as control systems are a huge part of the inverter firmware's job. Many things need controlling: solar and utility charge current, bus and AC output voltage, DC->AC converter amplitude and phase, and so on.

That firmware (46.71) is about two years old, though it does have the "presumed good" MPPT logic that some later firmwares lack. But the "presumed good" logic is back in 46.82, dated October 2023, so that's the one to go to if you want the latest available that doesn't have the "stuck at 90 V" issue too badly. I'm not aware of any particular issues with 46.71.