Coulomb

@plonkster, you're spot on. I haven't traced that part of the comms board, but I'm 99% sure that burned part is a transformer that connects to Gnd and HFPW (High Frequency PoWer) from one of the power supplies on the main board. It provides the + and - 12 V for the RS232 signals, and 3.3 V or so for the logic. I've never come across that fault before. All I can think of is a shorted turn in the transformer, so it just barely worked but with very poor efficiency, and burned itself out from the heating caused by the short. Or perhaps the short developed over time, due to nicked insulation on some part of the coil wire. Failing after 30 minutes of service is a pretty clear manufacturing problem, so you should have no issue getting a replacement comms board under warranty. Even if you had to buy a replacement, it should be fairly inexpensive. My guess is that nothing else will have been damaged, but that's far from certain.

Yikes! I just used data from the first manual I found: https://anro-developments.co.uk/wp-content/uploads/2015/06/Goodwe-ES-Manual.pdf . It says that each MPPT can charge at 15 A (that has to be PV side although Axperts always talk battery side). Your manuals say 11 A. There are three figures from the three sources for the maximum DC power: 5000 W, 4600 W, and 5400 W. And 11/15 × 5400 = 3960, so even if the current is taken into consideration, they still differ. I have no idea why these figures change so wildly. The URL for the user manual I used has June 2015 in the date. I can't find dates for the others in the URLs or in the files. Your User manual has "340-00004-02" on the front page; mine has "350-00082-01", suggesting that mine might be more recent. The version on the Goodwe site has "340-00004-03 Version: 1.0". It states 6500 W maximum DC input power, and 11 A per MPPT. Perhaps they upgraded the MPPTs at some point and most of the on-line manuals are not yet up to date. Frustrating. [ Edit: One piece of consistency: 130% × 5000 W = 6500 W. ]

It won't produce more than the inverter can handle, except perhaps for brief transients if the control system becomes unstable. It seems that the Goodwe will draw a maximum of 5400 W (its "DC power" rating), and it's recommended that no more than 130% of that figure is connected, i.e. you would be safe with 130% x 5400 = 7,020 W.

Though I note that this page says the maximum PV power is 4500 W: https://thepowerstore.co.za/products/axpert-mks-5kva-5kw-80a-mppt-48v-solar-inverter That's the same inverter as the title of this topic, right?

I doubt that anything will blow up as a result of that slight difference, but wiser heads may contradict me. It could sometimes produce more than 7000 W, but not for long. Most of the time, they will produce about 80% of rated power near noon and with no shade, and less at other times. I'd say that the manufacturer is referring to 6500 W of nominal photovoltaic power, and the losses would be taken into account already. The solar controller is a feedback system. If they say 6500 W maximum, I think they mean that they don't guarantee the stability of the control system when it has to try to tame that much power. So it might tend to overshoot the battery voltage it is aiming for, or undershoot as well. If this went on for long, it could damage your battery, or even the inverter. But it's unlikely to happen for long, so I'd say that this minor overload is acceptable. But you have to draw the line somewhere. Voltronic Power inverters, such as your Mecer, don't have the greatest solar controllers to start with. So I would no push the PV power too much. But as long as you can think of at least one non-ideality (not quite optimal tilt, some shading, etc), then you should be fine with just 100 W over the limit. But you did mention solar tracking; do you really want to do that? Few seem to consider the results worth the effort.

In your case, with two paralleled machines, that's probably the case. But I'd point something I probably forgot to mention earlier. The AC input connection to the grid has a second purpose, in addition to allowing grid charging. It's allowing the Axpert to switch to "line mode" on overloads. This happens at my home perhaps once a day on average. We're all pretty conscious of loads here, and try to prevent "burning coal unnecessarily", but it's nice to know that if you forget or the fridge and freezer come on together at just the wrong moment, there is no loss of power, just a clean transition to grid power for 2 minutes or so. The grid rarely fails here, so I don't know what would happen if I didn't have the grid as backup. My guess is that the Axpert would reluctantly allow larger loads for longer on battery mode, but would eventually simply switch off the loads when the battery voltage reached a low enough level. That might be fine, but it has the possibility of being a little harder on the battery. So @ebrsa, it would presumably be very rare for you to switch over to line mode due to a temporary overload. But for others like myself with just the single 4 kW machine, having the AC input connected to the grid, even if there are frequent outages, has its value. I do hope you figure out a reasonable way to keep running your Axperts legally, and preferably without too much built-in functionality unused. It would be a great shame if South Africa, which was presumably a very nice market for Voltronic Power, suddenly had to abandon nearly all their Axperts (leaving only the true off-grid installations).

I think my theory doesn't hold up to scrutiny, sorry. The Axperts have to calculate battery current, as there isn't a sensor to measure that. They assume fixed efficiency, and measure PV power and AC output power. It's possible that for a small change like this, the assumptions aren't valid, and there is a small error. Ah, I've just realised that the models smaller than 5 kVA have a separate AC to DC charger; they don't use the inverter in reverse as the 5 kVA models do. So they won't have the small power supply that is capable of running just the display and electronics; it will be an actual battery charger capable of at least 10 ADC. That may have been "idling" when grid power was available, and when the grid goes down, that "idle" power disappears. I just don't know how the lower power models work for AC charging. I think it will still be interesting to do the measurements.

It's possible that the grid was supplying the losses via a small AC power supply. And that the thing that they are calling PV watts is actually PV watts into the battery. Perhaps check the PV current going into the inverter during the day when supplying loads. Then open the AC input isolator or breaker, to simulate Escom going down. See if the PV current actually goes up, down, or stays about the same. My guess is that it actually goes up, especially since the PV voltage goes down. So this is just a mislabelling of the PV current. From memory, the 5 kVA models do something similar.

Higher version numbers don't necessarily mean more recent, e.g. 72.20 is more recent than 72.70, and is for different hardware. Sadly also, the same version number doesn't necessarily mean the same functionality (e.g. I believe that there is an old 5 kVA firmware and a more recent 4 kVA firmware that both have the version number 52.30). As far as I know, there really is a difference in the hardware with the PF1.0 models. So I don't think you can safely get 5 kW from a 4 kW / 5 kVA machine by upgrading firmware. I'd love to know for sure, however.

Main and SCC firmware are available from this AEVA site; use the index in the first post: http://forums.aeva.asn.au/viewtopic.php?f=64&t=4332&p=53691#p53691 [ Edit: oops. It looks like we might not host the latest SCC firmware version 4.10. But we no longer recommend updating the SCC firmware anyway. ] If you want unpatched factory firmware for whatever reason, these are contained in the patched zip files. For example, inside dsp_LF1_73.00c.zip, there is a file called dsp_original_73.00.hex. If you delete dsp.hex and rename dsp_original_73.00.hex to dsp.hex, then you'll get the factory firmware, not the patched. Similarly, you can delete dsp.hex and copy dsp_LF1_73.00c.hex to dsp.hex, and you'll get the patched firmware next time you update. This copying is needed because the updating firmware (not patched or changed in any way) only looks for the fixed file name "dsp.hex".

It's not something I've "developed", or even my idea; it's just the application of a contactor in a situation it's more or less designed for. It's also one of those things that you don't appreciate how useful it is until you've had it for a while. I didn't realise that the Hager brand is specific to Australia. [ Edit: it's obviously not; I don't know where I got that idea from. ] Here is a model from Schneider, the first one I found with a quick search: https://www.tme.eu/gb/details/a9c20868/contactors-main-modules/schneider-electric/# @pilotfish, the point of the 230 V coil is that you run it from the output of the inverter. So no extra wires are needed. Anything that runs from a dry contact etc might not work when the inverter turns off or fails. It's so simple, there is nothing to go wrong; just what you want for a backup of last resort.

But @malek94's machine is 4 kW. I would recommend 73.00e. It has all the features of previous patches, and it is based on 73.00 which has the backwards compatibility code to parallel with various other firmwares if needed. [ Edit: The underlined text is wrong; I was thinking of 72.70, which is the one that lets you parallel with 5 kW (PF1) models if needed. Also, patched firmware have to be paralleled with patched firmware; see my parallel compatibility post. ] 73.00e has code to work with SCC firmware versions 1.24, 4.00, and 4.10. So it's not recommended to update the SCC firmware. As a point of interest, what are your existing firmware revisions, main and SCC? Looking at the sticker, it looks genuine, though I'm used to seeing "Made in China" at the bottom. Yours has "Taiwan Technology". [ Edit: some don't have anything at the very bottom. ] I don't recall seeing that before. Are you confident that it's made by Voltronic Power, and is not a clone? [ Edit: does it have the date sticker over a front panel screw near the sicker shown above? ] [ Edit: 73.00c -> 73.00e, it's just the latest version as of January 2019, when I had to edit for other reasons. ]

Yes. On one system that I worked on at a University, there was much fuss made about a particular relay having to be a "safety" relay. That was part of the design before I became involved. I never bothered to figure out what this was all about, except that it was supposed to be made in such a way as to virtually guarantee certain conditions. Of course, it cost 2-3x (from poor memory) as much as an ordinary relay. This one was only a 10-20 A model, but maybe there is such a thing as a "safety contactor", and this may be what it takes to satisfy the powers that be.

A few thoughts on my suggested changeover contactor. The contactor itself will be a well-known brand; mine is a Hager, so that's well known and respected throughout the industry, and no doubt conforms to many standards. The relevant question seems to be: can the inverter output end up connected to the mains? The Hager contactor has a pair of normally open and a pair of normally closed contacts; these are wired to perform the "changeover" function. So for the inverter outputs to connect to the mains, both normally open contacts would have to fail shorted (2 faults), and in addition, they need to do it in such a way that the normally closed contacts stay closed. To me, that requires 4 failures at once. So no single, double, or even triple failure can cause the output to be connected to the mains. (A double failure could cause the lines to be connected together, but without the neutrals connecting as well, there can be no current flow. ) That seems to me to be a totally reasonable configuration. But I don't write or officially interpret the rules.

Even that can be eliminated. My PIP-4048 (Axpert) connects to most of my loads via a 63 A changeover contactor. It has a 240 V AC coil, energised by the output of the inverter. It's wired in such a way that if the inverter ever stops producing power, them the contactor drops out and the loads are supplied by the mains. As soon as the inverter produces power, the contactor engages and connects the inverter to the loads. The inverter synchronises its output with the mains, making the changeover easier for the contactor and for the loads. I never have a computer drop out at a changeover, and rarely a fridge or freezer (and they re-start automatically after a minute or two anyway). So if I'm working on the solar energy system, I just mash the Big Red Button that drops all the contactors, everything is safe, and I know I've not lost power. After all is done, I rotate the red mushroom button, push the green start button, and the system pre-charges, then connects the contactors (all staggered; the in-rush current for each is too high to switch more than one at once), and I get clack-clack-clack (the 7 contactors), followed about 10 seconds later by a muffled click from the switchboard (DB) (the contactor pulling in). I'm lucky that where I live, there is only ever about an hour of power outage per year, if that. So the only time I'm without power is when the earth leakage breaker trips. It's done that a few times over the last 10 or so years, spanning several iterations of power system. Once, the dog chewed a power cord (he only ever did that once ), the freezer needed de-icing (causes condensation in the electrics), and a recently a contractor left a pump running continuously, which eventually cracked something and the electrics got flooded. So maybe rather than selling Axperts (which will be worth nothing if everyone is selling at the same time) and replacing with entirely new systems, you just need to implement such a change-over contactor, if not existing already, and remove the AC input to the Axpert for conformance. It's a shame not to be able to use the battery charger capability, and you might want a 48 V charger for rainy periods. But if you don't, the Axpert will automatically stop inverting, and the batteries will be maintained at a moderate SOC by whatever power you get from the panels during rain. That's assuming that removing the AC input will cause them to be declared off-grid, and hence not need to be on the list. I'm assuming that there is no chance that one of the larger Axpert distributors can't get Axperts on the list of approved equipment. It seems to me that this would be possible, but someone has to spend some money. Maybe they don't make a lot from selling Axperts, but the prospect of not selling many at all in future might make it worth some effort. This is one case where Voltronic Power's policy of branding machines with the distributor (e.g. Mecer) versus the manufacturer (e.g. Voltronic Power or Victron) might work against them. It may be that each distributor has to wear the cost of certification, rather than the manufacturer doing it once for all distributors.

The big bang means it's probably not a trivial fix. It will at least have blown two IGBTs, possibly more. I don't know the Infinis at all. A bus soft start failure usually means shoot through of IGBTs. There is the concept of "inverter soft start" as well as that of "bus soft start". The latter is merely pre-charge of the bus capacitors; in an Axpert, there is a small power supply (chip, MOSFET, and multi-winding inductor) dedicated to this. I've not had the opportunity to figure out what soft start of the inverter is. I'd say it does a series of things, designed to prevent your second loud bang. Evidently, it failed. But I think that's why there was such a delay before the second bang, after the AC input was connected. I'd have little faith that your supplier will be able to repair the unit, except to replace the main board. That might be 50-75% of the cost of a new unit, at a wild guess. As for what you did wrong, I also don't know. Certainly, I would have disconnected the battery and AC input when working on the PV wiring, or anything else. I assume that this model has a higher voltage MPPT, with some 500 V at the input. All I can think of is that some capacitor somewhere retained a nasty charge, and that somehow found its way to the "DC bus" (the ~400 VDC rail that the inverter IGBTs run off). In an Axpert, the SCCs connect to the battery terminals, so there's another system between the SCC output and the DC bus. Sorry to hear about the blow-ups, and sorry I can't be more help. As a point of interest, was the AC breaker that you mentioned tripped on the grid connection, or the AC output (sensitive loads, inverter output)? I'm assuming here that Infinis have two inverters; one for grid interaction, and one for running sensitive loads, but I could be way off with that. As I say, I don't know the Infinis at all.

Yes, the SCC has its own way to turn on the inverter, independent of the usual inverter on/off rocker switch. [ Edit: I assume that firing up a generator will not wake a newer model from low voltage shutdown, however. I never let my battery get that low that it cuts off due to under-voltage, so I can't speak from experience. ] [ Edit 2: However, it seems my assumption is wrong. ]

You can set it up so that as soon as the generator is on, it will supply loads and charge the battery. This is normally how you'd want it set up when completely off-grid. But the inverter doesn't know if you are off-grid or not. So you have to change some settings. Again, you have to make sure the settings are right before this will happen. It will depend on the battery voltage as well as settings. Others have indicated what settings need to be changed.

In AC charging mode, the load is connected to both the mains and also to the inverter output. But, all going well, the inverter is reconfigured as an AC charger, pushing power from the mains towards the battery. My understanding is that there is a theoretical possibility of the software failing to configure the inverter hardware correctly, causing it to push power from the battery and/or solar panels into the grid. I suppose it's also conceivable that some sort of hardware fault, even with correct firmware, could cause the inverter which is configured for charging, to instead somehow invert and push power into the grid. If the grid goes down, an Axpert will not attempt to perform AC charging, so it should be fine. But again, there is the theoretical possibility that a firmware or hardware fault could somehow prevent the inverter from realising that the grid is down. It's a microscopic chance, so in my opinion this is all overblown. They should take the manufacturer's word that the Axpert is simply not designed to push power into the mains, and that it is designed such that it won't even connect to the grid if the grid is down. But that's just my opinion.

I assume you mean the RJ-45 to RS232 cable that comes with the inverter. Unfortunately, it's not a standard item. You'll need to make your own; see http://forums.aeva.asn.au/viewtopic.php?title=pip4048ms-inverter&p=57914&t=4332#p57914 Or you might be able to get it as a spare from an Axpert dealer. But somehow I doubt it.

Yes, here: http://forums.aeva.asn.au/viewtopic.php?f=17&t=5486 The conclusion is that they are not supercapacitors at all, but LTO (Lithium Titanite Oxide) batteries. They're possibly OK as a battery, but why lie about what's inside them? They certainly won't have supercapacitor-like life (millions of cycles).

Spaces separate the fields with both Infini V and Axpert, so the same boxes should work for both as far as getting the data onto the SNMP network. But then it's up to the monitoring software to make sense of it. "SNMP web box that are designed for only Axpert" I assume means a web box designed only with the Axpert in mind; perhaps the Infini V didn't even exist at that point. But it should still work in the sense of getting the data into SNMP packets, since the basic form of the RS232 data (ascii commands with carriage return at the end, same CRCs, open parenthesis starts response, spaces separate ascii fields) is the same. It's just the details that are different (more, possibly exclusively 2 character commands, different number and possibly order of fields).

The protocols are somewhat similar in flavour (responses start with open round bracket, for example); some of the commands have the same name (e.g. Q1), but have more fields in them (e.g. 23 vs 17). So basically: different command protocols.

I note that the battery voltage is 49.1 V, and the battery cutoff voltage setting is 47.0 V. In the non-VM versions, the battery is considered low when it falls to less than but not equal to the battery cutoff setting plus 2.0 V. So that limit would be 49.0 V. It's not considered to have recovered from being low until the battery voltage exceeds 4.0 V more than the low voltage cutoff setting, or in your present case, 51.0 V. 49.1 V seems very low for a lead acid battery if it's being charged at all. [ Edit: so if I got this right, when the screen shot was taken, you were just 0.2 V above the low battery situation. You'll need to charge to 51.1 V or higher to remove the warning. ] Just to get the low battery warning out of your face for now, set the battery cutoff voltage a bit lower, say 45 V. The normal value for lead acid (if you don't use the User setting for the battery type) is 42 V, so that's not crazy low. I've just checked the VM II firmware version 20.44 (yours is 20.45), and it appears to use the same limits. [ Edit: "less than or equals" -> "less than but not equal to"; "reaches" -> "exceeds". ]

I also agree with others that the battery cables are way too thin. They should be a minimum of 35 mm² (2 AWG). In Australia, we would not be able to get away with using green-yellow, or even pure green cable for any purpose other than protective earth (ground). At least, that's a regulatory / maintenance issue, it won't affect performance. But from the too-thin cables you will likely get very annoying flicker as loads come on and off. Also, those cables could overheat with extended loads and even normal charging at 4+ kW (80 amps or so). It could even start a fire. Finally, I also agree that it would be very good to separate the DC and AC cables, and ideally have them supported and protected in cable troughs. Especially since the DC cables, until you get them replaced, will be getting quite hot at times. Hopefully those batteries would normally be inside some sort of box, to at least shade them from direct sunlight. Heat will cause them to degrade faster than necessary. Though perhaps you are high enough that overheating is the last of your concerns. Really cold batteries won't perform as well as warm ones.