Hello lovely people!

I recently joined this forum looking to get some insights around a new backup power installation & thought it was finally time to ask for some advice as the frustration is getting palpable.

I also apologise in advance for the lengthy post.

First some backstory: We recently purchased a house at a new development & we were delighted to see that one of the optional extras we could pick was a backup power installation to alleviate the load shedding woes. So naturally, we picked that option as the price was incredibly good & I felt really chuffed about getting a good deal. But as they say, you get what you pay for... The spec sheet for the setup on the extras list was not particularly detailed - all we knew was that it was a 10 KW setup, inverter + battery only, no panels. Looking at what colleagues were paying for solar setups vs the price we paid, I started to think that they were going to wire a few USB power banks together & call it a day. We moved into the house about three weeks ago & the installers came to install the inverter & battery shortly after & to my surprise, they installed two 5 KW inverters in parallel with an actual lithium battery. I was gobsmacked that we got one of the fancy batteries everyone was recommending & proceeded to do a little happy dance when the installer was looking the other way.

Fast forward a couple days & I was feeling proper chuffed. Coming from an inverter trolley setup at the old place which could only power the router & TV, this new setup was running THE WHOLE HOUSE.  I didn't even notice when load shedding kicked in as the lights didn't even dim & I could continue sucking at Modern Warfare for at least a couple of matches. All was good. Until it wasn't.

I started to feel like I was missing something. You see, all the cool kids had these fancy apps & monitoring dashboards to see how much power they were using, what the battery levels were & all that neat stuff. I felt like a caveman needing to walk to the garage & look at the inverter to see if the power was on or not. Something needed to change. I perused the manual which came with my inverter to see if there was some kind of companion app I could download. No dice. There was only a QR code which directed to an obscure spreadsheet which I quickly dismissed as confusing. Then I remember seeing some software called Solar Assistant & proceeded to check if my inverter was supported. Initially I was thrilled - looking at the pictures provided, it seemed my inverter was a Voltronic inverter. It looked like a Voltronic, sounded like a Voltronic, smelled like a Voltronic. The manual even resembled the sample one provided in the Solar Assistant docs. Oh, happy days. I followed the instructions to install WatchPower & SolarPower, spent a couple of hours hunting down a USB printer cable & was off to the races. I booted up WatchPower, plugged the USB into the port on the inverter and... nothing. Okay, I thought, seeing as my inverter had "hybrid" in its name, I shall try SolarPower instead, and... nothing. Balls. This is the part where I stumbled across this forum & learned about these things called clones. Reading up a bit on the matter, my understanding is that for this particular class of inverter, there are the original manufacturers, Voltronic/Axpert, and roughly 5 thousand distinct clones which rebrand these models & put their own spin on them. My particular model is a SAPPHIRE HYBRID 5048 PV18-5048 VPK. Googling this takes me to this spec sheet, which says it's a MUST inverter, which I understand is a clone of an Axpert. So I have a cover version of a cover version of the original. I have a LimeWire edit of a Bump remix of the original song. This did not inspire confidence. After some more investigation, I learned that the MUST inverters' RS485 & USB ports operate over a binary protocol instead of an ASCII protocol (Solar Power Monitor 2.2.81 alternative - Design Software - Power Forum - Renewable Energy Discussion), hence the incompatibility with WatchPower & SolarPower. These MUST models only work with another piece of software called SolarPowerMonitor. I didn't explore that yet because another issue crept up which felt more pressing than monitoring.

One afternoon I came back from work & spotted something that filled me with dread: the clock on the microwave was reset. This could only mean one thing - the power went off. To confirm, I also checked my router uptime & saw that the power was interrupted right around the time a load shedding slot started. I immediately initialised the only monitoring solution I had and walked over to the inverter screen to see what was up. Both inverters were up & running with no faults & I checked the battery display & the battery was fully charged. The only thing that was different was that the master/slave configuration seemed to have switched around. The inverter that was master was now the slave, and the other was now the master. A case of role reversal. I was curious as to why that would have occurred, but my initial presumption was that when the grid went down, the inverters switched roles & because they did, the transfer from grid to battery was just not quick enough and this caused the power to drop for a split second, resulting in the microwave clock reset. This was a complete guess because I am still a complete noob when it comes to how these things actually work. Nevertheless, this was concerning as there was now this looming risk that sometimes - not every time - there will be a brief power drop when the grid goes out. This is bad for sensitive electronics & meant that I couldn't reliably keep working on my PC when the power is due to go off because there's a likelihood that it will briefly cut out. Plus, I would always need to re-adjust the microwave clock & that is just not on.

After this happened a couple more times over the next few days (the increased stages of load shedding did not help in this regard) I delved into the manual again. I spotted a setting for 02 - AC input voltage range which affects the transfer time. The default APL (appliances) mode indicates a transfer time of 20ms (for home appliances) whereas the UPS setting indicates a transfer time of 10ms (for personal computers). Hell yeah, I thought to myself. More faster = more better right? Perhaps the default mode did not switch over fast enough & that was causing the brief interruption. I wrestled my way through the settings & managed to change the mode from the default APL to UPS. Job done, time for a drink. The next few days it didn't skip a beat through all the load shedding slots. I felt like a genius for figuring this out. That was until this past weekend when I went to reheat some food & I spotted the microwave clock. It read: 0:00. 😐

These last two days it happened twice again. Once during a 6:00 slot & another during a 22:00 slot. I witnessed the 6:00 occurrence where it seemed like a 2-3 second interruption at least. It is worrisome & I'm not sure as to what the cause might be.  There is also another worrisome thing I noticed whilst constantly checking the battery level during this last week's hectic load shedding: the maximum SOC as reported by the battery is 92%. I contacted the installer about this & he is yet to come around this week to investigate as he mentioned that the voltages might need some tweaking. It's worth mentioning that 14 - battery type in the inverter settings is set to USE (User Defined) with custom voltages & the inverter isn't connected to the battery via the CON port, so I'm uncertain of whether the BMS is being leveraged or not. This battery is an Esener 51.2v 5.12Kwh 100Ah.

Anyway, apologies for the long story, I just felt that some context would be helpful. My questions to this wonderful community at this point are:

1. What could be causing the occasional interruption during switch-over from grid to battery?
2. Is the interruption a common occurrence with parallel connected inverters?
3. Could the lowered battery SOC indeed be voltage-related, or could it be a sign of a bad battery or early degradation?

Any insight will be greatly appreciated. I've attached some images of the inverter/model detail if it would be helpful.

Have a fantastic day!

37 minutes ago, SellusWallace said:

Any insight will be greatly appreciated. I've attached some images of the inverter/model detail if it would be helpful.

Numero uno, you were sold a 10kW system, you say, please check and confirm this wording, since if it is supposed to be a 10kW system, they owe you a second battery, one battery will only supply a maximum of 100A times nominal 51.2V is 5.12kW and not 10kW, if one battery is correct, then its a 5kW system, never mind the two 5kW inverters, they could have even paralleled up another 2, thus 4 by 5kW inverters and still it would only be a 5kW and not a 20kW system, the battery here is the limiting factor.

As for the power reset/outage at the start of loadshed, no idea whether this is something you can fix, it should not happen and I would talk to whoever supplied this system to you and get them to try and resolve this, but, I have at this stage a RCT IFS V2 5kW inverter, similar settings as your and I see at start of load shedding a funny occasionally, I have no Microwave with a clock or so and there are a few CPU based machine running here (Mac Mini, HP PD400, some RPi's) and none of these seem to notice, but a HP Desktop multimeter (3457A) resets, maybe twice a week (it is switched on 24/7) and thus I know something went funny, but never have any real outages, none of the Mac's or PC's or RPi's reboot...

Those inverter's look super close together,  as far as I am aware there should be around 300mm between them for ventilation. 

Regarding the modes the inverter runs in, maybe @Kalahari Meerkat or @Coulomb can confirm as I think they know the Voltronic inverters much better than I do 🙂. My understanding is that the APL mode switches the load to the inverter when it detects a drop in the grid frequency, this results in a slight skip in the "beat" of the grid frequency as it switches over from grid-inverter. 

The UPS mode basically powers the inverter with power from the grid and the inverter then runs the load, so there is no switch over as such, just a drop of grid power and switch to powering the inverter from the battery.  

The disadvantage of just running the inverter in UPS mode all the time is that the inverter is not 100% efficient and as such it will consume more power than it provides to the load. Not sure of the efficiency but I would imagine/hope that it would be around 95%, so you pay for say 20kw/hr of power but only get to use 19kw/hr of that in UPS mode. 

 

Edited December 13, 20222 yr by Sc00bs

Thanks so much @Kalahari Meerkat & @Sc00bs for the input, you raise some interesting points. I hadn't thought of doing the calculation of battery discharge current x voltage & looking at the math, it makes sense that this is not a true 10 kW system. Indeed, the wording does say 10 kW, so it's supposed to be a 10 kW system. But if the supplier can't calculate the appropriate battery size for the inverters they install & they can't be bothered to space the two inverters appropriately for ventilation, this casts some doubt on whether they really know what they're doing.

I highly doubt presenting them with this information will result in an "oh shucks, you're right, we'll come & install another battery for you" but it's worth a shot.

@Sc00bs, good shout on the efficiency aspect, though I would happily deal with the slight drop in efficiency rather than risking some machines which run 24/7 with unsafe shutdowns.

Regarding the 92% max SOC, I had a call with the installer this morning who suggested I bump the charge current from 40A to 50A & the bulk charging voltage from 56.4V to 57.4V but that didn't seem to address the issue, so he needs to come by later today to have a look.

5 minutes ago, SellusWallace said:

from 56.4V to 57.4V

I bulk charge mine at 3.5V per cell which is possibly already a bit high, but lower than yours at 56V and float mine at 54.7V aka 3.41865V per cell... absolute maximum Voltage per cell for LiFePO4's is quoted at 3.6V by various sources, that I would accept that value as gospel, but would not want to go that high, since you will shorten the life of your cells.

You can cry false advertising about the 10kW claim, since the only way they can get to supply 10kW is by running two of your type of batteries in parallel... as for the SOC, does the inverter claim 92%?, if so, how hum, the discharge curves on LiFePO4's is so flat, between 80% and 20% SOC, that it is impossible to actually state the SOC with any kind of precision other than +/-  50%, so if the inverter says 92%, then I would not worry too much about it...

14 minutes ago, SellusWallace said:

Regarding the 92% max SOC

Your battery has RS232 and RS485 comms ports, it would be great if you can get the correct software to talk to the battery, this should give you much more accurate data about battery health/state of charge/cell health (are they balanced/being balanced) etc. I'd chase this up as well, if I were you...

10 minutes ago, Kalahari Meerkat said:

does the inverter claim 92%?

This is indicated on the battery LCD itself as per below image (coincidentally it was at 91% at that point) but I'll try & find the appropriate software to try & get more detail out of it.

Going to raise this with the property developer as it is their extras list which advertised this as a 10 kW system.

16 minutes ago, Kalahari Meerkat said:

and float mine at 54.7V aka 3.41865V per cell

Am I correct in calculating that your battery has 16 cells? How does one determine how many cells are in a particular LiFePO4 battery?

Sorry for the noob questions, still very new to all of this. 😅

11 minutes ago, SellusWallace said:

How does one determine how many cells are in a particular LiFePO4 battery?

LiFePO4's are 3.2V nominal and hence a 15 cell pack is quoted as 48V nominal and a 16cell pack as 51.2V nominal. If you look at the label on your battery, they have exactly these two Voltages as rated Voltages... where the 51.2V is ticked in your case...

 

as per you picture above, it seems to still be absorbing 3.61A, which is app. 200W, so this could be in balance mode and leaving it for long enough, once the cells are balanced, it may well drop down to 0.x A (way less than 1A for float) and display 100% SOC, but getting the cell state info etc. out of the battery via either of the two comms ports would give you the best information on what is happening in the life and times of your 5.12kW battery...

40 minutes ago, SellusWallace said:

Thanks so much @Kalahari Meerkat & @Sc00bs for the input, you raise some interesting points. I hadn't thought of doing the calculation of battery discharge current x voltage & looking at the math, it makes sense that this is not a true 10 kW system. Indeed, the wording does say 10 kW, so it's supposed to be a 10 kW system. But if the supplier can't calculate the appropriate battery size for the inverters they install & they can't be bothered to space the two inverters appropriately for ventilation, this casts some doubt on whether they really know what they're doing.

If your battery could discharge at 200A then they'd be able to make some sort of case about it being a 10kw system. But that is asking a lot of the battery, and the labelling on the battery clearly says 100A. So that battery can only supply 5kw at any given time.

Might this be the cause of your occasional failures? The inverters will happily back up 10kw whilst there is grid. Now suppose immediately before the load shed the demand for power is 7kw (anything north of 5kw is ok for the purposes of this). The system switches to battery, the battery senses the overload condition and trips. At this point, everything that was making up that 7kw load switches off. The system starts to reboot, and when it does not everything switches back on.

If you had a 1C 10kwh battery, the system would actually be able to provide 10kw at any given time. 

We used to have a problem with my system until I moved some loads around using timer switches, and we started learning how much the system can actually do. My inverter is marketed as "5kw" but really is 4.6 with short (10 second) bursts up to 5 allowed 
1) "Background noise" for the house is something like 280w
2) Early in the morning things start happening. One of these is that the heat pump turns on. Suddenly we're at 1.7 kw
3) Somebody goes into the kitchen and starts turning stuff on. Each one pushes up the load. So the dishwasher. Which within a few minutes starts heating water.
4) Kettle gets turned on. That guy uses about 1.7/1.8 kw all by itself.
5) Something goes into the microwave...
And so on. And that 4.6 that seems so generous quickly gets exceeded, and after a few seconds the inverter (in this case) shuts down.

When it reboots most of the pesky stuff is turned off and will stay that way. The heatpump will restart, but the dishwasher certainly won't (and you've wasted that little detergent block thing), kettle might if the switch is still down, microwave won't...

This is why the monitoring software you mentioned early on is so important when you first get your system. You can see what the loads are and how the system reacts. You will soon learn to recognise different loads by their "signature", and you can maybe "move" some loads using timer switches so that they CAN'T be on at the same time. You actually have an easier task than I do, since you don't have to worry about maximising PV, but you still don't want too much on at a time. With the monitoring software you will not only see what device uses how much, but also WHEN the power gets used. So you might "move" the geyser to a quiet time of day.

This game is easy to play at times of low load shedding. When you've got what we are getting right now, you have to restrategise all the time.

36 minutes ago, Bobster. said:

At this point, everything that was making up that 7kw load switches off. The system starts to reboot, and when it does not everything switches back on.

In fact, everything on the backed up side of the DB will turn off. So your microwave may not have been the last straw, but it will still restart as the system goes down and then comes back up again.

1 hour ago, Kalahari Meerkat said:

as per you picture above, it seems to still be absorbing 3.61A, which is app. 200W, so this could be in balance mode and leaving it for long enough, once the cells are balanced, it may well drop down to 0.x A (way less than 1A for float) and display 100% SOC

Thanks for this insight @Kalahari Meerkat, this is super useful. Might the < 100% SOC be caused by the current constant load shedding that the battery simply does not get enough time to fully balance out the cells?

 

1 hour ago, Bobster. said:

Might this be the cause of your occasional failures? The inverters will happily back up 10kw whilst there is grid. Now suppose immediately before the load shed the demand for power is 7kw (anything north of 5kw is ok for the purposes of this). The system switches to battery, the battery senses the overload condition and trips. At this point, everything that was making up that 7kw load switches off. The system starts to reboot, and when it does not everything switches back on.

Thanks @Bobster., that is useful information. I don't believe high load would be the reason for my issue as these interruptions occur when total household load is well under 1000W according to the inverter display. I've had two load shedding slots today where there were no interruptions & the load was approximately 400W - 500W at both times. So while the battery capacity definitely isn't up to par for a 10 kW system, I haven't yet exceeded 5 kW yet at any point to my knowledge.

When mentioning the occasional interruption issue to the installer over the phone this morning, his response was that that is expected behaviour because "this inverter is not a UPS, it's line-interactive, the inverter is not always running, so therefore there will always be an interruption when the power goes out before it switches to battery". I mentioned that if that were the case, the interruption would happen every single time the power goes out, not only occasionally to which he responded that he will need to come check it out. Still waiting on that visit though. 🙂

33 minutes ago, SellusWallace said:

Might the < 100% SOC be caused by the current constant load shedding that the battery simply does not get enough time to fully balance out the cells?

Yes, depending on the BMS (Battery Management System) of your Battery, the balancing could be a really weedy affair that only uses 100mA to balance cells and burns the excess energy off in some resistors, possibly, then balancing can take many hours, but not knowing the BMS involved I can't say that this is definitely the case...

 

So you have inverters capable of supplying 10kw of power but a battery that can only supply 100A @ 48v = 4800W of power. 

I don't see how anyone could claim that is a 10kw system and it may actually be more efficient if you just disconnected the 2nd inverter. 

SOC problem is probably due to the lack of comms between the inverter and the battery BMS. Is impossible to get an accurate SOC from the battery using voltages as mentioned by @Kalahari Meerkat

7 hours ago, SellusWallace said:

My particular model is a SAPPHIRE HYBRID 5048 PV18-5048 VPK.

I would call these "work alikes" rather than clones. They have different firmware, with different settings numbers, for example.

Nevertheless, the hardware seems very similar to that of Axperts.

Also, yours appears to be a PWM model, not an MPPT model. But since you don't have panels yet, that makes no difference.

5 hours ago, SellusWallace said:

Regarding the 92% max SOC, I had a call with the installer this morning who suggested I bump the charge current from 40A to 50A & the bulk charging voltage from 56.4V to 57.4V but that didn't seem to address the issue,

I would certainly not run the battery charge voltage that high. I run my 16S LFP battery (similar to yours, but home made) at 55.2 V. But 3.5 V per cell is common, which would be 56.0 V.

6 hours ago, Sc00bs said:

The disadvantage of just running the inverter in UPS mode all the time is that the inverter is not 100% efficient and as such it will consume more power than it provides to the load.

These models don't have double conversion, so there is no huge loss when in "line mode". In line mode, AC-in connects to AC-out, and the firmware keeps tabs of the phase of the utility, so that it can switch over from utility powering loads to battery mode, where the inverter powers the loads. This switch-over can be smooth or jarring depending on the utility voltage, and the state of the internal relays. I'm guessing that the South African grid isn't the most voltage stable in the world, so you might sometimes be switching over from say 200 V utility to 230 V inverter.

The difference between the APL and UPS settings (in Axperts, it may be different in MUSTs) is that in UPS, there is a board with SCRs on it (maybe they are actually TRIACs), which somehow carries the current while the relays are switching over. It's still not instantaneous, since you don't know exactly when a relay will actually open and the next one will close, since they are inherently mechanical.

14 hours ago, Sc00bs said:

SOC problem is probably due to the lack of comms between the inverter and the battery BMS

Is this remedied by simply connecting the supplied cable from the inverter COM port to the battery COM port (which was not done by the installer) or is there more to it?

 

14 hours ago, Coulomb said:

Also, yours appears to be a PWM model, not an MPPT model.

Thanks @Coulomb, how were you able to distinguish this? Also, what would this mean if we were to eventually add panels later?

 

14 hours ago, Coulomb said:

I would certainly not run the battery charge voltage that high.

Thanks for this, I set the bulk charging voltage back down to the 56.4V it was before the installer recommended I bump it. Would you advise I lower this to 56V?

 

14 hours ago, Coulomb said:

This switch-over can be smooth or jarring depending on the utility voltage, and the state of the internal relays. I'm guessing that the South African grid isn't the most voltage stable in the world, so you might sometimes be switching over from say 200 V utility to 230 V inverter.

This is super interesting. Looking at the inverter display, the utility voltage varied between 226V - 228V on the occasions I checked.

14 hours ago, Coulomb said:

It's still not instantaneous, since you don't know exactly when a relay will actually open and the next one will close, since they are inherently mechanical.

Am I correct in understanding that this is not a solvable issue as this is due to the design of the relays & that it can be influenced by many factors?

Thanks again for the wonderful feedback everyone, you all are awesome! 😁

15 hours ago, Coulomb said:

These models don't have double conversion, so there is no huge loss when in "line mode". In line mode, AC-in connects to AC-out, and the firmware keeps tabs of the phase of the utility, so that it can switch over from utility powering loads to battery mode, where the inverter powers the loads. This switch-over can be smooth or jarring depending on the utility voltage, and the state of the internal relays. I'm guessing that the South African grid isn't the most voltage stable in the world, so you might sometimes be switching over from say 200 V utility to 230 V inverter.

Indeed. Where I live, the grid voltage seldom gets to 230. On a good day it's 220ish, but it can go down to sub 200.

Just some information for @SellusWallace, the inverter will usually try to match the grid voltage, phase and frequency whilst it has a grid connection. Obviously when there is no grid, it will do what it is designed to do: Convert the DC power from the battery to 230VAC at 50Hz. And so what Coulomb is suggesting is a reality: When load shed happens, the voltage the backed up devices are receiving could increase suddenly and significantly.

43 minutes ago, SellusWallace said:

Is this remedied by simply connecting the supplied cable from the inverter COM port to the battery COM port (which was not done by the installer) or is there more to it?

Comms between the inverter and the battery is definitely preferred and the only way you are going to get a remotely accurate SOC from the battery on the inverter

On 2022/12/14 at 3:40 PM, SellusWallace said:

Is this remedied by simply connecting the supplied cable from the inverter COM port to the battery COM port (which was not done by the installer) or is there more to it?

It would be, if your inverter had a BMS port. But it does not.

On 2022/12/14 at 3:40 PM, SellusWallace said:

how were you able to distinguish this? [ PWM model ]

I can tell from the maximum charge current and maximum PV voltage.

Quote

Also, what would this mean if we were to eventually add panels later?

It means that you won't get as much power from your panels (some would disagree), and you have to use a low voltage panel configuration, since for most of the time while charging the battery from solar, the panels will be dragged down to battery voltage. So generally you need two standard "24 V" panels in series, no more. Three or probably four "12 V" panels.

On 2022/12/14 at 3:40 PM, SellusWallace said:

Thanks for this, I set the bulk charging voltage back down to the 56.4V it was before the installer recommended I bump it. Would you advise I lower this to 56V?

Yes.

On 2022/12/14 at 3:40 PM, SellusWallace said:

This is super interesting. Looking at the inverter display, the utility voltage varied between 226V - 228V on the occasions I checked.

If this was in line mode, then that indicates that you have quite good utility voltage, at least when you checked. 

On 2022/12/14 at 3:40 PM, SellusWallace said:

Am I correct in understanding that this is not a solvable issue as this is due to the design of the relays & that it can be influenced by many factors?

It's hard to say. You may have a hardware issue, e.g. a relay contact that has had its resistance increased by over-current at some point.

Edited July 15, 20231 yr by Coulomb
"to standard" → "two standard"

@SellusWallace 7 Months later, how did you solve this?

Seeing this thread a bit late, but here are some of my comments:

1. MUST inverters are now supported by Solar Assistant. I'm not sure your specific model is, but should be. I know mine works with SA, but it's a different model.

2. AFAIK, the difference between APL and UPS is that it the inverter can switch over faster in UPS mode, given that the voltage is within a tighter range than APL allows. But still, as @Coulomb mentioned, there is no double conversion (what is also call "zero conversion/switchover time"). So in APL mode, it will take approx. 20ms to do the changeover, while in UPS mode it will be like 10ms. This could still be enough to power down some devices I suppose.

3. Depending on the firmware of your model, some MUST inverters have support for BMS connection via the Pylontech protocol. You need to connect the battery via the CAN wires (description in the user's manual) to your battery BMS, if it support the protocol.

4. To help with the power down during LS, I would try to turn off most of the appliances just before LS. See if it helps. I suspect that maybe a load surge during LS may demand too much from your battery, untill things settle down.

5. I totally agree with @Sc00bs, the inverters should be placed farther apart for better ventilation. Like they are installed can cause them to heat up and reduce thier lifespan.

 

Good luck and keep us posted.