Coulomb

Oh. That will do it. Even 102 Ah is still a bit small for a 5 kVA inverter, if lead acid.

That's normal and expected. That charge looks pretty normal to me. But of course, that's utility charging, and 40 A is more than a fifth of the maximum charge current. Presumably, you normally want to charge from solar. As soon as a cloud goes by, such that you generate less than a fifth of maximum charge current (setting 02), and the battery voltage is not much lower than the float setting, it will terminate the charge. It won't restart the charge until the battery goes below 50 V (if the float setting is at 54 V), which is unlikely, even with a large load, unless it's real state of charge is quite low. Factory firmware; you're welcome

Welcome, Tony. That's not all that unusual, but as @Youda mentions, it might mean that the charger has gone to float prematurely. It's probably at 54 V because your float setting is 54 V. I don't believe that the absorb (misnamed bulk, sometimes called "C.V") or float settings are bad. My system, which is 16S, uses 55.2 V absorb and 53.7 V float. Yes, the battery does discharge a fraction of a percent when going to float, so it might sit at around 99.2 % SOC during the day, assuming that there is enough solar power to cover loads. It's better to stay away from high battery voltages (> ~54.0 VDC) as long as possible. Assuming that this is during the day with intermittent (solar) charging, this is normal, especially on a cloudy day. A cloud will come over, momentarily the load will exceed charge power, and the battery discharges a little. The cloud passes, and the solar charger brings it back to 54.0 V, replacing the charge lost, by charging at up to 6 A. In better weather, it could charge a lot more than 6 A for a minute or so, depending on how much sun there is, and how much charge was lost. However, when you have constant sun for a while, enough to cover the load, then the battery current really should be nearly zero. If not, then the battery isn't fully charged. The factory firmware in Axperts have since inception had two premature float charge bugs. These are serious, and can under-charge your battery, causing higher than necessary discharge, lowering life. It's especially bad with lead acid, since for maximum life, you really need minimum depth of discharge. It's so serious, a colleague and I went to the trouble of patching the firmware in some models, initially only to fix these bugs. We also provide a "LFP flavour" that has much more LFP-friendly voltage thresholds. If yours is a 4 kW model (PF0.8), or a 5 kW model (PF1.0) with the 64 V option, you can use patched firmware to fix these problems. See the AEVA thread index for details. Finally, I have to gently tell you that you got three out of four units wrong in your first post. (I have to do this, because I'm a pedant, I have no choice ). Your battery has 7.4 kWh of capacity, not 7.4Kw. (The "k" in "kilo" is always lower case; the W in watts is capitalised for units that are the names of people, and not for the unit written out, as in "watts".) Your mains charging was at 40 A, or 40 Ah/h (40 Ah of capacity was added to your battery every hour). Your constant load is 300 - 350 watts, not watts per hour. But "6 amps" was correct 🙂 These unit errors are very common, especially with beginners, so don't feel too bad.

Those "fuses" are now a stamped piece of what looks like plated copper with a narrowed region. This thing is bolted down. There should be one (1) spare fuse in the box that the inverter came in. They may have changed, but at one point they looked like the attached. I suspect that the "large diode" that @plonkster mentioned is actually the battery-side DC-DC MOSFETs' inherent diodes. In other words, they added nothing to the circuit, but still claim that the inverter-charger is protected against reverse polarity. (They always had a 200 A fuse before the stamped metal thing). But perhaps that's just me being cynical. 😮 [ Edit: "high power diode" -> "large diode".] [ Edit: Added image. ] [ Edit: "sceptical" -> "cynical". ]

Yeah. An unstable grid is a frightening thing to consider. I like Germany's idea of using ramps instead of hard edges, though it took them a while to realise that hard edges are a bad idea. I mean things like how much power to reduce when the grid frequency increases to say 50.2 Hz. If all the PV inverters with their crystal locked timing drop their power 100% to 0% a milliHertz above say 50.2 Hz, then you get the "take up" problem you've been discussing. If however, they reduce to 90% at 50.21 Hz, 50% at 50.25 Hz, and 10% at 50.29 Hz (a simple straight line ramp), then if the spinning generators happen to overspeed a little, the PV inverters can back off and that will load the spinning generators more, slowing them down. But of course, this is a massive control system with delays due to the speed of light, inductance, and so on, so it could oscillate with larger and larger amplitude: instability. Control systems were never my strength. Kudos to those that design these things. At some point (due to instability say), you do have to just trip off and start again. If the whole grid trips out, that's a big deal, worthy of their own Wikipedia pages. Restarting a large grid like in the USA, Europe, or Eastern Australia is a planning nightmare in itself. I imagine that South Africa has a relatively small grid compared to those, but that may not make it easier to control. Edit: please pardon the topic drift.

It's OK if the cables to the inverter are different lengths, as long as the distance from string A to the commoning point is the same as the distance from string B to the commoning point. With exactly two strings, you can use "diagonal takeoff" to avoid a commoning point. The important thing is that no string has "privileged access" to the inverter battery terminals.

I don't own an Infini, and haven't played with one for any length of time. However, I do collect firmwares, and I find that the exact same CRC function is present in the Infinis as in the Axperts. Don't forget that in addition to the 0x1021 CRC-16 ("Xmodem") calculation on the command or response data, you have to check the upper and lower halves of the 16-bit result for three special values. If the upper or lower byte equals one of those special values, that value is replaced by the next highest byte value. The three special values are '(' (0x28), carriage return (0x0D), and line feed (0x0A). Each occurrence of these would be replaced by ')' (0x29), 0x0E, and 0x0B respectively. That way, a CRC can never contain those special values. This is despite the fact that in most Infinis, '(' is not a special value, but '^' is. The CRC calculation on the command or response does not include the carriage return.

Yes, measure across each 12 V module before and after the load is applied. If a few sag much more than the others, perhaps give them a charge with a 12 V charger and repeat.

That does sound too much. I suspect some bad cells. Since you have two strings, it probably has to be at least one bad cell in each string. The battery doesn't seem good.

Those settings should be fine for most lead acid batteries. Certainly, I don't believe that they will have harmed your battery, certainly not in a day or two. Cheap lead acid; not a grand combination. But before panicking, I'd run down to 48.0 V again, and then switch to bypass mode, so the batteries have no or negligible load, and no charging. Let them sit like that for an hour or two, and note the battery voltage after that. A heater is a moderate load, so there would have been some sag. Use the rested voltage to get a better estimate of real SOC. A BMV won't help here, since it would need to know the actual capacity, and that's what you need to find out. Are you confident that your battery wiring is up to scratch? Perhaps too thin wiring, or a bolt or two not tight? Don't pay much attention to a voltage-based SOC measurement, like those on Axpert inverter-chargers.

No, it just means it's probably not genuine Voltronic Power firmware, or even a direct copy of one. Though it is possible that it's a V. P. firmware that I've not come across. Certainly, the firmware versions don't merely get bigger the more recent they are. As an example, version 72.20 is more recent than 72.40, and they are for different models (and not compatible with each other). Edit: yours seems to be a PWM model, so it won't be compatible with most available firmware updates, certainly no patched firmware.

Zero personally. I note that it's a Value model (hence the V in the name), so it's a little cheaper, but not parallelable. It's also a 450 V MPPT model, with all that implies: no patched firmware, may not be approved since it doesn't have insulation monitoring, yet easier wiring on the roof. The model with many of the features of the VM III but with 145 V MPPT and parallelable is the Axpert King (= PIP-5048MK). But I haven't seen any firmware for it, so there is no patched firmware as yet.

Be aware that these models are a sort of "orphan"; they have their own main firmware series (used to be 75.XX), and can't parallel with any other model. There is patched firmware for them, but it's version 75.31a (based on factory firmware 75.31). This is for the 4 kW model (PF0.8), new models seem to be 5 kW (PF1.0). So the available patched firmware won't be compatible with new models. Besides that problem, it only fixes the premature float bugs, and nothing else. There isn't even an LFP-friendly variant. Even if you can only afford lead acid now, you will likely want some form of lithium battery in the future, and right now by far the most popular lithium chemistry for home storage is Lithium Iron Phosphate (LFP, or LiFePO₄). These have a very flat voltage curve, and the Axpert firmware as it comes from the factory doesn't work well with that flat voltage curve. The MPPTs in these are only 60 A (at the battery) each, so that's a total of 180 A. Two ordinary Axperts in parallel, which have 80 A MPPTs each, will give you nearly that much: 80 × 2 = 160 A total. With all those large loads, you might want to end up with paralleled machines. I don't know if you can parallel a triple MPPT machine with a dual, but certainly you can't parallel it with a single MPPT model. The dual and triple models don't seem to be all that popular, so there is less support from the community. They also pack a lot of conversion hardware into one box, so cooling may be an issue, and the fans might be roaring all day. They also come in a physically deeper box, which may not suit some installations. Finally, you will be limited to 145 V MPPT max. This is a good thing in a way, as the high voltage MPPT models (450 V MPPT max) don't have patched firmware. But it means that you will be limited to 2S or 3S panels, and I seem to be the only user in the world running 2S. With your proposed eight panels, you'd have to go to 2S 4P and have four pairs of wires down to a combining box, or get another panel and go 3S 3P. If you don't want to re-wire your roof and stay 3S, then you'll have to run a model with the high voltage MPPTs; see PIP-5048GE, PIP-5048GK and PIP-5048MG inverters (with unsafe SCC). (The PIP models are equivalent to Axperts). Arguably an even bigger resource than this fine forum is the Australian AEVA topic PIP-4048MS and PIP-5048MS inverters. It is so large (105 pages at 25 posts each) that it has its own index, and it's now too big to read in total. But there is a lot of information there, including partial schematic traces, links to manuals, etc. So: lots to consider, even just in inverter-charger selection. The triple MPPT model may not be the most future proof after all. Good luck with your planning.

Just remember that the maximum charge current settings add (they are per inverter). So for example, you could have a maximum charge current setting on one set to 30 A, and the other to 40 A, for a total maximum charge current of 70 A. A 5 kW Mecer probably has a maximum charge current of 80 A per inverter (depending on the exact model), and the 5 kW of panels could provide about 100 A each (at 50 V battery voltage), more than the inverters can use. So it seems a shame to charge at only 30 A per inverter. Since your batteries seem to be able to take more current than average, and the usual maximum charge rate for lead acid is 0.15 C (15% of the Ah capacity), that would come to 0.15 x 600 = 90 A. So that would be 40 A on one and 50 A on the other, or you could go for 50 on both. I agree with @Jaco de Jongh that 180 A is probably too much, as three strings of batteries won't share perfectly, so one string would be getting more current than the manufacturer's limit. At a wild guess, I'd say no more than 120 A maximum, but that's just a gut feel.

The manufacturer says don't do it, so that's really the end of the story. Having said that, with some careful firmware, I imagine it could be done (without having thought it through in detail; dreaming is easier than designing 😴). One problem is that the batteries could all be at different states of charge. So one might have its low battery warning showing, so it can't contribute to the loads. Another might be just over the low battery warning level, but the Axperts don't have a good estimation of battery SOC, so you can't say OK this one has a power limit of 2 kW and that one has a limit of 4 kW, so a 3 kW load should be shared as 1 kW and 2 kW respectively. It would be a nightmare to test; so many corner cases. Plus, it's entirely possible that there good reasons I simply haven't thought of. 🤔 TLDR: don't do it 😶

The only way to find that out is to pull as much power as possible, and see how much that is. There might be clever ways of estimating it, but I don't know of any system that does this. There may well be some, e.g. those Zappi EVSEs for charging electric vehicles from "available" solar energy only. Axperts don't measure the power going into the MPPT. But they're pretty efficient, I think 98% or more, so the difference between input power and output power, at least at full load, is slight.

Um, won't the museum want those exhibits back some time? Those items are classic. Better not break that seal... R10 fine! Thanks for posting.

No. It just depends on how long the unit has been stored in a warehouse, either in China or with your reseller, or both. Voltronic Power used to be great with making all the latest firmware updates available. But the appearance of many clone manufacturers shut all that down. End users suffer as a result, because they don't get the latest firmware updates. You can easily tell what firmware you are running, from either your monitoring software, or from an LC Display option. But now it's hard to find out what the latest firmware is, since they're no longer published. Even when the firmware updates were readily available, the manufacturer never (to my knowledge) published release notes, so you don't know if it's worth performing an update or not. About all you can do is post what your current firmware version is, and ask if anyone has any later than that. If you have a specific issue, ask them if it seems to be fixed in their version. If so, you have a case to ask for a firmware update from your provider. Since the arrival of the clones, they seem to be quite reluctant to provide any firmware updates, so you'll need a good reason. My experience is mostly with the Axpert off-grid models, but I'm pretty sure the situation is the same with the Infini hybrid models.

That works great (I do it myself) if you are on-grid and have the one and only permanent connection from neutral to earth where the grid comes in. But with a generator, the "grid" can be unplugged. If you bond AC-in neutral to earth, then when you run a generator, again the generator earth wire takes some of the neutral current (or so it seems to me). Maybe that's not such a disaster. Edit: also, in places where they like to isolate the neutral as well as the active when isolating the grid input to the inverter, you can't rely on the grid's neutral to earth bonding. It seems that this is common in South Africa.

I think that there is still a problem with though, if a generator is used. A generator presumably bonds its neutral output to earth, so that some of the generator neutral current would travel through the earth conductors. I think the GFCI won't trip, though these things do my head in when I can't see the whole schematic. I don't know the answer to that one. Especially when the generator can be disconnected. [ Edit: apart from an external earth to AC-out neutral relay. ]

There are actually a few 120 V Axpert models, but you have to buy two of them to make a split phase system. This is possibly one of the reasons that Axperts are not all that common in the USA. So yes, I can confidently say that bonding AC-out neutral to earth on a single Axpert inverter won't blow it up.

That only happens in very unusual inverters with split phase outputs (for those legacy countries still using 120 V), and when they also internally bond one of the phases to the metal case and to earth. As @plonkster told me recently (thanks!).

Heh. I see you have labelled your AC out with + and —. I friend of mine told me that in countries that use black for neutral and red for active (most house wiring in Australia is like that), there are plenty of electricians that believe this to be literally true, i.e. active is always positive with respect to neutral. In Australia, we use the international colours of brown / light blue / green-yellow for active / neutral / earth, in our flexible cables. But for reasons that escape me completely, our house wiring is red / black / green-yellow. @Nuno, I hope that you realise that AC (Alternating Current) also implies Alternating Voltage, even though we never call it that. So half the time, the red wires are negative with respect to the red ones. And a hundred times per second, they cross zero.

As @Javi Martínez suggests, that covers a lot of models: older 4 kW models, newer 5 kW models, 450 V MPPT models or 145 V MPPT models, Value models and standard models, and so on. The only models I'm fairly certain that have the AC out neutral to ground relay contact are the ones that come with main firmware 73.00. If you're off-grid, you can probably get away with connecting neutral AC out to earth. The GFCI breaker relies on a neutral to earth connection somewhere.

I haven't, though I haven't looked seriously. It seems to me that the energy density is nowhere near that of batteries, and although the power density is quite good, you don't need power density for a solar energy system. There are also some scam products out there (huge thread). There is some question about self-discharge. Finally, the legitimate products seem to be very expensive, compared to batteries. But their cycle life is excellent (again, except for the scam products that are actually batteries anyway).