___

That's the entire Western Cape for you. Yes, I know I should not generalise like that, but man it can be frustrating. I mean, I'm a Namibian by birth... you stand in line in a supermarket in Windhoek, you might as well befriend the guy next to you, you're going to be there for a while... but even that is nothing to the glacial pace at which things move in the Western Cape. "Slaapstad" is indeed accurate.

If it is still under warranty, it should be repaired again or possibly be replaced. Talk to Current Automation and if necessary, involve one of the sales managers. Their addresses can be found on the Victron website.

What do you mean "solar priority"? 🙂 @jykenmynie actually did a good job of explaining it. You tell the system whether you want to use the batteries or not. If you decide to keep them charged, then the generated PV powers your loads, and you can optionally push the excess into the grid (or not). If you decide to use the batteries (down to some configured minimum), then the batteries (and any incoming PV) is used to power loads. If there is more PV than loads, then the rest charges the batteries. If there is less PV than loads, the rest is taken from the batteries. Down to the configured minimum SOC. And again, you can optionally put the excess into the grid (if the batteries fill up and there aren't loads to use the power).

Something else I learned in the mean time, is that at really low power values your signal to noise ratio (SNR) gets really bad. To understand this practically, imagine that on that ADC scale of 0 to 1023, you should read a zero (but noise causes you to read a 1), and imagine now that the range was wide enough that 1 means 250W... well you end up measuring little 250W noise spikes when there is no load... 🙂

I prefer that someone who knows that stuff better does that 🙂

Yes. You'd have to set up bus bars and split your batteries into groups, and tie each group separately to the bus bars. And then you'd have other cables from the bus bars to the inverter and the MPPTs. Since the inverter is only 8kW, the max draw from the batteries would be around 160A. If you split them into two groups of three each, each side has to do about 80A, and the standard cabling will be sufficient for that. Then you'd also fuse the battery connection to the rest of the system at 300A or so. On the batteries there is a link port, a communication bus. You link all the batteries together, and one of them becomes the master and the rest are slaves. The master makes decisions about shutting down and so forth for the entire stack. This stack can be max 8 batteries (it has to do with addressing between the master and slaves, not with the cable sizes). If you have more than 8 batteries, you divide them into groups and you combine the groups into a larger whole using a "LV hub". The DIP switches come in when you have multiple groups tied together with an LV-hub. Then each group gets their own ADDress. For 8 batteries or less (with no LV-hub) you just leave the dip switches alone. Edit: So to be clear, you'll electrically split the batteries into two groups of 3 each (because of cabling limitations), but from a data communications viewpoint it will be one group (ie all 6 will be daisy-chained using their link ports).

Hah! Yes, I used to do the same thing, but unfortunately the neighbour's generator had a mechanical failure during the last round of outages. The other option is to look for the other neighbour's WiFi AP. When that disappears, I know the power is out 🙂

I remember a friend of mine had a a very unusual way. He had an old Sony Playstation that wasn't backed up by his UPS, and he had monitoring set up for his home computers. Whenever the PlayStation dropped off the network, it usually meant there was a power outage 🙂 Edit: Hence, this could be as simple as placing a Sonoff on the non-essential side and using a stock-standard monitoring package (on a phone maybe?) to ping it occasionally. You could even integrate it with home assistant.

Use a latching relay. You need a SPDT relay. The N/C contact powers the buzzer, and the N/O contact is fed back to the relay coil (in parallel with the "reset" switch) so that once the relay closes it powers itself. Then you need a second relay wired to the grid that drops out when the grid fails, and this powers the second latching relay/buzzer setup. Or.... just buy a cheap Arduino board. Use some old wall-wart transformer plugged into the grid side as a "Sensor" and feed that through a suitable voltage divider (to get 5V or 3.3V) into one of the input pins. That's your "grid" signal, it goes to zero if the grid fails. Write a 10-line sketch and upload: if grid goes out... wait X seconds (ie debounce it bit), then activate buzzer unless the button (on another input) has been pushed yet. Find some suitable way to power said Arduino so it works when the grid is off. Which one? Well, what have you got lying around in the parts bin...? 🙂

True 🙂

Aaah ok, I see it, but note this warning text (they try to disable copying but my former web-developer backside has some skillz): So it sounds like they do the same half-double-conversion as the King.

Working at 1C discharge, it is 70A per module, and as I recall they ship with 25mm^2 cabling? Or is that that the US2500? In any case, two modules would be the max for 25mm^2 cabling, so 6 is definitely too much. If you work with C/2 currents (the recommended maximum discharge), then it's 4 modules max. Best would be to do some research on what Pylontech themselves recommends. Yup, no problem. Just make sure the maximum charge current is somewhere between C/2 and 1C of the batteries, ie somewhere between 200A and 400A, and make sure your cabling is sized for that.

I looked at the spec sheet, and it doesn't really leap off the page to me. Where did you hear this or read this? (I'm curious, I don't normally care too much about this brand 🙂 ). What I can deduce from the spec sheets is that it has a high voltage PV-input, so the PV is likely buck/boosted directly onto the high voltage DC bus. Maybe it does the same thing as the "King" model then? That is technically a kind of double conversion, in that the AC is turned into DC and back into AC, although it is not taken down to battery voltage and back up (which makes it a tad more efficient). Interesting that it can also run battery-less. I would assume it works like the SMA models that have the emergency power feature: It uses the PV directly to make AC, but limits you to a fraction of the normal maximum power.

Maybe. Installing Pylontech batteries isn't hard (it is very easy actually), but most of your installs is going to have less than 8 modules, so it is easy to forget that there are limits. The Murata batteries are even more interesting (I note these units have been discontinued now in favour of something new). You buy battery modules (1.2kWh each), and then you buy controllers. There are two kinds of controllers, a 180A one and a 50A one. Depending on battery size, you'd use one or more of these controllers (for a very small setup you'd use one 50 unit, or very large one you'd use several 180A ones. Now imagine building a battery with this flexible Lego-type kit... you have to know your stuff! 🙂

This is a safety thing, so what you want is for it to be super simple, less chance of failure. There is also the danger that someone doesn't know this DB (where the light is in) is also supplied by the grid, and then he turns off the inverter and thinks it should be dead. When there are two sources, there needs to be a light on if the thing is hot. If you want to see the status of the grid, I suggest using two lamps, one wired to the grid and one to the inverter. Both of them would be on during "normal" operation then. Then there is no confusion, ans both lights need to be dead before you can work on it.

Ahaha! I've had a support case that involved one. Customer forgot to set the dip switch to terminate the CAN-bus... 🙂

The support is there. What is missing is Pylontech support for the new CAN frames (the BMS has to send us that info). But apparently support for this feature is not far off...

That aligns with my (later) suspicions that it sits in the supply line to the main SMPS (the one I marked 2 in the image). So the "blowing up" could be because the stock SMPS used cannot take the resulting voltage spike. I remember trying to find out what chip is on that little board but I could not figure out what it is. The protection board could just be a simple power supply, possibly even linear, that just chops the top off.

Of course I am a tad bemused by the budget inverter paired with a sizeable investment in batteries, but to each their own 🙂

Truth be told, the Pylontech battery has a very good BMS. Of course I'd love to sell you a nice bit of kit, but I kinda feel like you already have enough kit, and I am concerned about warranty matters when you don't wire it up properly. My advice would be to get the LV hub. I believe ICC can read the Pylons. Edit: The protocol from the batteries and from the LV-hub is the same. At least on the CAN-bms side. I assume ICC would have no trouble reading the hub instead of the master battery. But you may have to look into that...

You're right! Small details you miss. Indeed, you are limited to 8 per group. You can combine up to 5 groups (40 modules total) by using a "Pylontech LV hub". Can't provide a link, pylontech.com.cn is unreachable for me at the moment. That little hub is almost half the price of a battery module though... But I'm not convinced this is the reason for his troubles. Or it could be... I don't know. If you connected all the batteries together, it may well be that the last 4 can't be addressed by the master unit and they open their master disconnect (a big old FET inside), so that they don't contribute to the whole. Then I would expect their SOC indication to be different from the others. Or.... they do contribute, because this is an Axpert system and the comms really isn't all that important.

Infer? With a lead acid? Well that's a good joke 🙂 The BMV simply counts amp hours, and then depending on the efficiency value you programmed into it, it ensures that slightly more Ah is returned to the battery before it considers it full again. It can do rudimentary Peukert calculations too, where it derates the usable capacity if you work the battery hard. But it relies on a programmed capacity and yes, I do see how that's a chicken and egg thing... how are you supposed to know? Well, with a new bank, you know the capacity. And from the spec sheet, you can adjust the number depending on the cycles. Some installers may even program the initial capacity to 10% less to compensate for this somewhat. But I think the long and the short of it is that once the battery is beyond a certain point, the only way to find out is to do a full discharge and see how much you get. Lithium BMSes can do a little better, because along with Ah counting it can rely on the voltage. If the lowest cell drops to 2.8V... we're done here mate, battery is empty (I have BYD charts somewhere of SOCs dropping through the floor... the current sensor is not very accurate below 2A, so it totally underestimates the Ah bit of it... and then when the voltage measures low it catches up).

Remember that the SOC on the BMV can be wrong. To use just one example, when you first power it up it always starts at 100%. The BMV has to be programmed correctly with the capacity of the battery (the current capacity... not the capacity it had when it was new), and then you have to fully charge the battery so that the BMV's 100% coincides with an actual 100%. Now if you do all that, and then find that the voltage still drops out to below 48V at what should be a good SOC, then you have one or more weak batteries. Then you go in with a volt meter and you go find that battery. This is actually what it means when we say a battery can be taken down to 50% DoD for 1000 cycles. It means that after 1000 cycles, the battery has half its capacity left. A 200Ah battery turns into a 100h battery after 1000 cycles (possibly less, I'm just using nice round numbers). A BMV programmed for the original 200Ah will report 50%, but in reality that thing is empty man...

OK. What can it read? In a case like this, I like to plot the SOC (if you see it discharge at a steady rate, and then suddenly dive to zero, that's evidence of something not being healthy). Also, I like to see the individual cell voltages, or at least the minimum and maximum (the min and max is actually all you need). When it switches off, what's the voltage of the lowest cell? Also, who is doing the switching off. Is it the inverter, or is it the BMS that disconnects the battery from its loads to protect it?

As a rule of thumb you can assume 90%. It varies according to load and not all inverter are equally efficient. 88%-92% is what I typically see in the blue world. "Transformerless" units like the voltronic should in theory do slightly better, but 90% is probably still a good rule of thumb.