Coulomb

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).

All LFP cells look close when the cell voltage is at the 3.33 V plateau. It's mainly at the high and low ends that you can see an SOC difference as a significant difference in voltage. So your cells might not be as close as you think. I'm not saying you do this, but it's a classic beginners mistake to say "Wow! My cells are really well balanced! They're all within a millivolt of 3.333 V! It must have been my awesome balancing technique."

Since the charging LED is flashing, you're still in the bulk or absorb stage. So the relevant setting is the absorb (misnamed as bulk/CV) setting. I presume that it's at least 56 V. The float voltage setting only becomes relevant when charging is over (and the charging LED is on solid). The inverter's reporting of state of charge (SOC) is very rough. Monitoring software will report the same SOC as the inverter, unless you have something like a BMV external battery monitor that counts coulombs, and the software is set up to read that. Even then, the external battery monitor relies on the battery getting full regularly, so that it can reset its SOC counter to 100%, and you don't seem to be getting there. What voltage is the battery reaching during charging? Certainly, 54.21 V when charging is nowhere near 100% SOC. What is the capacity (in amp-hours) of your battery? What is the nominal power rating (in watts or kilo-watts) of your solar panels? What is the voltage of your individual batteries making up your main battery? Often, it's 4 nominally 12 V modules in series; please measure each one. There is a chance that you have a shorted cell, in which case one or more modules will have a voltage about 2 V lower than the others. That would cause your symptoms of never stopping charging, since the battery voltage would never reach the absorb setting. Finally, is there a chance that your inverter is a clone? Some clones seem to have problems with dropping out the Solar Charge Controller. Is the solar panel indicator on the LC Display always on when the sun is shining?

Voltronic Power are the manufacturers of the genuine Axperts, which come by a wide variety of names. Must Power is the first of the clone manufacturers. I think your only hope is to convince them that it's an off-grid system. They are in fact sold as off-grid systems. The model PV1800 ususally indicates that it's a clone manufactured by Must Power. See my page Do I Own a Clone? Clone manufacturers usually don't provide technical documents of any sort, let alone a schematic or even a block diagram. If even genuine Axperts, say a Mecer, has to provide a block diagram from a service manual with the Mecer name and model on it, that's unlikely to happen, and a lot of owners are going to be in a lot of trouble. Many of the service manuals are completely unbranded. The following block diagram, which sounds like what you want, is from a service manual that is branded Voltronic Power: The above suggests that it's one of the first clones ever made, possibly serial number 0044, made in 2015/June/22. The first warning about the Must Power clones appeared in November 2015. Good luck finding your technical documentation; I think you're going to need it (the luck). At least, Must Power is a fairly large manufacturer in their own right. Perhaps start at this page, with the "downloads" tab, or talk to one of the online assistants: https://www.mustpower.com/pv1800-mpk-series-high-frequency-solar-inverter-1kva-5kva/

[ Sorry for the late reply. ] Yes, use a pre-charge resistor! It can even be wired to a switch, or for the ultimate luxury, to a contactor or relay controlled by a small computer. All rated for at least 60 VDC, of course. I prefer a smaller valued resistor, like 3.3 Ω, for a nice quick start. It needs to be about a 50 W, high pulse power model (aluminium cased). Regardless of resistance, they all dissipate the same energy, the same amount of energy that is stored in the capacitors. So you may as well have a fairly quick pre-charge, as long as your switch / relay / contactor can handle the current. As well as being bad for the capacitors, the high surge current is also hard on the fuse contacts, or wherever the arc happens. Eventually, it can tarnish and lose metal and the contact can become high resistance. When I press my green start button, no less than seven contactors come on in quick succession (clack clack ... clack). One of those is for pre-charging the inverter capacitors. It's not good for the capacitors to charge at thousands of amps, as well as the other hazards. When I mash my big red switch™, there is a big clunk and they all drop out together, isolating AC in, AC out, PV in (both poles, both solar charge controllers). At the same time, a contactor in the main switchboard (DB) drops out and my house switches over to mains. All automatic. It did cost a bit for all the contactors, and took a while to wire up.

Yes, assuming you keep this model, and get it going. You could even add one more panel and just wire it in series anywhere. High voltage MPPTS do simplify things on the roof. But there are safety considerations; these inverters don't do insulation monitoring.

I find I often have to select the "paste as plain text instead" (or similar) option that turns up after a second or so. Especially links to pages in the files section. It does seem silly not to allow fancy linking to pages on your own forum.

I've seen the sticker; this is a genuine Voltronic Power inverter. It irritates me how Voltronic Power encourage their resellers to make up model names. As mentioned earlier, it's also a 450 V MPPT version.

I wish I saw this post in time. Because it's an Axpert MKS II, as opposed to an Axpert MKS (non II), it's a model that has the higher voltage (450 V max Voc) MPPT. These models have incompatible hardware compared to the 145 V MPPT versions. There is no firmware update for the 450 V MPPT models, patched or otherwise, that I'm aware of. Unfortunately, the reflash tool doesn't stop you from updating with 73.00 (patched or not), and in this case the results are disasterous. I received a frantic PM from @kobus joubert this morning (my time). For the sake of other readers: please read the download instructions carefully! Usually, the major firmware version numbers have to match. The exceptions are the 72.XX series (XX ≥ 40), which replaced early 52.YY series, and are replaced by 73.00 (so far). 71.XX is for high voltage MPPT models. There are no firmware update files for these models. PF1 models and PF0.8 models (see below) have different, incompatible firmware. There is patched firmware for PF1 models (with all the patched firmware features, such as AussieView™, Dynamic Charge Control, Dynamic Load Control, and KettleKomp™), but ONLY (so far) for those PF1 models with the 64 V option. If you want patched firmware, double check when ordering a new machine that it has the 64 V option (whether you actually need that feature or not). It may cost a few percent more for this option. Weber and I are debating how to upgrade the warnings that precede the download files, to try and minimise this misfortune in future. This is normal. Nearly everything in the world, of all quality grades, is manufactured in mainland China. Voltronic Power have some 100 people in Taipai for Research and Development, but some 3200 people in China manufacturing the equipment. Yes, all the new 5 kA models are capable of 5 kW now; these are so-called PF1 models (unity Power Factor). The older models were PF0.8. So the battery-side converters are no longer the bottleneck. Voltronic seem to tweak the fan control logic frequently. That's probably because it's designed for higher voltage panels, at less current. I wondered why you got a high voltage MPPT model to replace your 145 V clone. Was this supposed to be a direct replacement, only genuine? In that case, you've really had a bad run.

When you three-phase rectify (a bridge with 6 diodes) a three phase AC source, you do get close to DC, with a peak value of √2 (1.414) times the phase-to-phase voltage. But there is about 6% ripple (13% peak to peak); that's small enough to be neglected in some situations. But the power factor has to be terrible. Other than at the instant of phase "commutation", one phase always has zero current (the one phase that isn't the most positive or the most negative at any instant won't have either diode conducting). That can't be the best way to extract power from the wind turbine, and extracting maximum power is what a wind turbine is all about, right? So I would hope that a "real" 3-phase charge controller would actually have three boost converters that work on one of the three pairs of inputs. Each of these would be like the "PFC" stage (Power Factor Correction, but it's a terrible term for what it does), so that the current is nearly sinusoidal. It seems to me that these would have to be isolated from one another, so three high frequency transformers would be needed. This would seem moderately expensive. The big wind turbine farms would have to do this, I would think, but what about the smaller controllers? A quick web search seems to indicate that they just 3-phase rectify the output, with all the problems that this entails. Does anyone happen to know how it's actually done? Perhaps having current flow 2/3 of the time isn't considered so bad. Even though the current will fall from 86.6% of peak to zero for 120° of the cycle, then step to -86.6% of peak. I saw a paper suggesting that LC filter networks and a single boost converter (for the 6% ripple) can dramatically improve the power factor, so perhaps that's how it's done. But that was for fixed frequency, and high power low-frequency inductors are bulky and expensive, and with the variable frequency... it doesn't seem practical. As a total off-topic aside, we used to have trams (they would be called light rail these days) in Brisbane, Australia. These worked on nominally 600 VDC. We have 240 VAC power here, so three phase is 240 x √3 = 415 V phase to phase. They used to use mercury arc rectifiers to convert three phase 415 VAC to DC for the trams. 415 x √2 = 587 V, which rounds to 600 VDC. There is a tram museum here where enthusiasts still run a few trams over a short track, and the tour guide gives a history of the trams. There is even a "Who Killed the Electric Tram" side to the story: there was a suspicious depot fire near the end of the tram era, that soon saw the tram tracks ripped up and diesel buses took over. Back to the topic a little: they must not have cared much about power factor back in those days.

And then the fuse won't be lonely any more. Sorry if I was a bit too subtle with my wisecrack.

I guess the clones are surprising some of the suppliers as much as the owners. If only all suppliers were as accommodating. Well done.

I think it's dying of loneliness Perhaps a fuse per string would be in order.

+1. There are a bunch of small power supplies in there. One of them is the "main" power supply that provides +15 V, +5 V, and ±12 V. It's supplied on the primary side via three diodes (D9/D11/D14 in the SS 3 kW+ service manual, Figure 4.3). These diodes connect to two other power supplies, to derive power from the battery or AC in, or from BUS+ (which includes PV power). I don't know if either of these three power supplies is part of the Battery/DC-DC board or the AC board, and two of these small power supplies is all you need to power the processor and display from AC input. So as @Jaco de Jongh pointed out, running the display isn't a clear indication of the health of either of the two big power boards. In the model I worked on briefly, there seemed to me more than one "main" power supply as well, just to add confusion to the mix.

This service manual for a 3 kW machine has pictures of the various boards. It should help, even though it's not quite the right machine. I had a 4 kW model for inspection for a while, and it seems to me that there are two big boards, the main board and the battery (perhaps also called DC-DC) board. Between them, these two boards are 90% of the electronics. The comms board is very small. There are many sub (daughter?) boards on both the larger boards, not plug-in but rather soldered in. I assume that these would come with the larger boards that they connect to. https://powerforum.co.za/files/file/10-1-phase-3kw-plus-hybrid-solar-inverter/

All Axpert models that I'm aware of have only one AC input, so the question (as it relates to Axperts) is moot. The AC input would connect to either the utility or a generator, never both.

That sounds awfully generic. Does it have a brand name? As Javi mentioned, no. You will need at least 60 VDC from the panels. A single 145 W panel sounds like a totally impractical sales gimmick.

There are no dip switches or passwords. What brand and model of inverter do you have?

His machines (Axpert VM II then changed to Axpert MKS II) all have the higher voltage MPPT (450 or 500 V). The three panel limit (sometimes two) is for the models that have the 145 V maximum MPPT.

We've established that your machine is a clone. Unfortunately, it's looking like the ones that come with main firmware version LC1 72.70c can't be updated, possibly due to a lack of bootloader code. So the good news is that you have the premature float bugs fixed. The bad news is it looks like you'll have to live with the other glitches that are presumably a result of imperfect copying. Sorry.

Ok, so it won't start bulk charging again. But as long as it stays in float mode, it should still charge the battery, just not to as high a voltage. Sorry, I wasn't clear. I meant on the Axpert. Ok, I meant the middle LED on the Axpert, labelled "CHG". But the lack of the PV icon is a problem (I assume we're talking about daytime, after a full charge, say mid afternoon). I've put a lot more "clone detection logic" into an AEVA post: Do I Own a Clone? Some clones seem to have a lot of trouble with PV charging, so perhaps make one last check that it's not a clone. Does it bring back the PV icon and start using PV power if you just remove the PV input for a few seconds? Leave the Axpert running for this test. This would in effect reset the Solar Charge Controller without resetting the main DSP. It takes the SCC a minute or so to fully get going so be patient. If that succeeds, it means that the problem is with the SCC, or perhaps your PV voltage is marginal (either too high or too low). What PV voltage do you see when it won't charge? I suppose it might be zero if the PV icon is off, so you might have to use a multimeter.

Yes, I believe that's what is happening. In unpatched firmware (and patched firmware before the latest revision "e"), the battery voltage used was not load compensated. In other words, it just uses the raw battery voltage, regardless of whether that's because of low SOC or because of a load. Patched firmware 73.00e and 72.20e use a rough estimation of the battery's internal resistance (it can be adjusted in 5 steps) to compensate for the internal resistance. If this worked perfectly, there would be no change in compensated battery voltage for a sudden load like a kettle, so there would be zero reason to switch to bypass mode. In practice it's not perfect, but it doesn't have to be; it just needs to "blunten" the sag enough so it doesn't go to bypass. I've been running it for a while, and it does help a lot, especially at night as you've noted. We call this feature KettleKomp™ (of course, the ™ is tongue in cheek, we don't trademark it at all). If your machine is eligible for patched firmware, perhaps give it a go. It's a lot cheaper than another battery.

I'm not used to these newer models, which seem to have a powerful AC-DC converter, separate from the main inverter (an AC to DC converter). So they can combine mains and PV power, at the DC bus level (around 400-500 VDC). This is facilitated by the 450 V PV max models, which boost the PV voltage to 400-500 VDC, rather than buck it down to battery voltage. It would seem that if there is plenty of PV power, there could be zero mains power. But maybe that's difficult to control, and maybe that's why PV power is less than load power in all the examples so far. I doubt that the PV power could ramp up to cover a sudden load increase, and maybe there are limits on how fast the AC to DC converter can ramp up the power as well. So maybe it's easier to have a few hundred watts of AC to DC converter power, so it's easier to ramp up or down in case of load changes. Perhaps sudden large load changes have to be handled by switching to pure bypass mode, as in a relay kicks in. But then, if it was a sudden large load decrease, and there was significant PV power which can't quickly be ramped down, where does the excess energy go? The DC bus capacitors would not have enough storage. If a battery is present, it can provide the instantaneous power increase, so there is no need to keep some AC input power flowing, and no need to switch to bypass mode, unless of course the inverter is now overloaded. But the power flow diagrams on the LCD (as depicted in the manual) seems to suggest that in line mode, the relay is always on (there is a direct straight line marked "BYPASS" from AC in to AC out). This suggests that the power blending occurs at AC (at the inverter output), not DC (at the inverter input), so it's a true AC coupled system. But then there is the usual problem where sudden decreases in load might cause brief exports of power, which I would imagine would not be allowed in this type of inverter, and in some locations it might be strictly prohibited (perhaps by a power meter tripping). I'd be interested to hear the experiences of anyone running a VM III or other double conversion model without a battery; what is the experience like.

Ah, I misinterpreted "have not installed any battery yet" somehow as "the installation isn't quite finished" rather than "I don't intend to run it with a battery". [ Edit: I'm not used to an inverter-charger that can operate without a battery. ] Without a battery, and without the ability to blend mains and PV power (correct me if I'm wrong there), [ edit: I am wrong here; see next post], there is nowhere to store PV energy. So you will always be limited by the load at any point in time. I don't understand how the PV power seems to be less than the load, maybe it is some sort of lag effect. [ Edit: the difference is provided by the mains. ]

Oops, my bad. I was misreading how the version number displays. So you've updated from 20.58 to 20.59. Excellent! So you're pressing the enter (⏎) key for at least 3 seconds? Restarted the inverter (battery power off, battery power on)? It may be necessary to revert to default settings, using the supplied software. Or is this from your phone? I'm not familiar with these new fangled machines. Or from ICC?