Best settings for Mecer 48V 5KVA with Lithium ion battery / solar and grid

[Lex de Lange]

I have recently replaced my unsuitable and all-but-dead lead-acid truck batteries (don’t ask!) with a 10kWh Lithium Ion battery from Hangzhou Kingor New-energy Technology Co.,Ltd., KG48-200FT64, ie 48V, 200Ah. to go with my Mecer 5KVA 48V and 12 Canadian 330W each panels.

My target is solar consumption and charging during the day, battery draw down during the night and thanks-but-no-thanks to Eskom (my stove and geyser are still wired to Eskom, until I get a 2nd 5KVA!).

After an initial finding of hectic charging/discharging (see below) I ran the battery for a long session with both solar and grid disconnected, performed well. The inverter settings have been changed by the person installing it and the result was charging for an hour, discharging for an hour, this 5 times until the inverter decided to stop the switching and just power the load by grid and keep the battery charged. This clearly does not work for me.

The main settings are: SOL for priority (should be correct, Solar first?) but particularly in 'back to battery' and 'back to grid' I have a question mark: both are now set at 50V, that would cause the hectic switching (solar not in play here, this was during night)? The spec sheet for the battery says ‘Maximum charge voltage 54.7V (Overcharge protection voltage not higher than 56V)’, and 'overdischarge final voltage' equal to or greater than 42V. Also the max charging (solar+grid) is set at 20A, spec sheet says max 100A, recommended 40A. I imagine charging mode should be set at ‘max solar + load’?  

I have some other questions as well, full settings and questions are contained in the attached spreadsheet.

I also received a comms cable with the new battery, RJ45 to USB. I have connected this cable between the battery and my Raspi3 running ICC, but nothing happening.

Presumably there is a software solution to manage the battery's BMS via this cable. Anybody any ideas? (it's a simple cable, whereas, when I look at images of the Pylontech cable, there is some sort of 'box' in the middle of it. Not the case here).

 

Any help greatly appreciated!

 

Axpert settings.xlsx

[Coulomb]

5 hours ago, Lex de Lange said:

The main settings are: SOL for priority (should be correct, Solar first?) but particularly in 'back to battery' and 'back to grid' I have a question mark: both are now set at 50V, that would cause the hectic switching

SOL output source priority (setting 01) switches to utility for powering loads at sunset. You want something like SBU. It's called different things on different models, and you didn't specify which Mecer 5kVA 48V model you have.

You always want a gap between the back to grid and back to utility settings, otherwise you'll get rapid switching. 

6 hours ago, Lex de Lange said:

Also the max charging (solar+grid) is set at 20A, spec sheet says max 100A, recommended 40A.

I would set it at the recommended 40A, possibly 50A. I would expect a 200Ah LFP battery to be comfortable with 100A.

[Lex de Lange]

Txs, Coulomb: I figured I need a gap between back to grid and back to battery, I guess, I'll start at the parameters from the spec sheet and see where that gets me.

I have an SBU setting on my inverter and upon re-reading the explanation you must be correct: SOL switches to grid when no solar available, or battery voltage below certain level, SBU only switches to grid when battery voltage below certain level.

I'll set the charging to 40A, if that leaves the batteries not sufficiently charged I could always go higher?

[Lex de Lange]

I'm also looking for the correct Bulk Charging Voltage (CU voltage) and Float Charging Voltage settings. In my case both are set @ 54V. I've read some scary comments on the life time reduction a Float Charging Voltage set-too-high can cause ....

[Coulomb]

On 2021/03/29 at 1:24 AM, Lex de Lange said:

I'll set the charging to 40A, if that leaves the batteries not sufficiently charged I could always go higher?

It's probably not too serious. But it's better to keep to the recommended limit, adding more battery (and therefore a higher recommended charging rate) if you need more. Of course, that's a lot more expensive than changing a setting.

[Coulomb]

15 hours ago, Lex de Lange said:

I've read some scary comments on the life time reduction a Float Charging Voltage set-too-high can cause ....

I don't see much down-side to reducing the float voltage to say 51.8 V (the figure that seems ideal for 15S Pylontechs). 54 V will also cause disconnection of a Pylontech.

[Lex de Lange]

Thanks for your help, Coulomb.

In the mean time, the importer that sold me the battery, has given me updated recommended settings for my inverter/battery. But there's something I do not understand. Ok, these settings 'will give you the best use of solar input and battery life cycle', so there would be some stretch for 'opinion' in there, but the battery was advertised to me as '3000 cycles at 80% DoD, 2000 cycles at 90% DoD'. I'm happy with 3000 cycles for the price (was definitely cheaper than Pylontech!), so one would expect to set it up towards 80% DoD. Yet, he recommends 'back to utility' to be set at 48V, with 'battery cut-off' at 47.4V, which he says is 20% SoC. At that 48V, the inverter starts charging the battery from utility early in the morning, before solar has a chance to kick in. Would it not be more logical to set 'back to utility' at 47.4V as well, to target a consistent 80% DoD (obviously, when solar is available earlier, one won't get to 80% DoD, I'll consider that a bonus!)? Or is that making a link where there is none? Same for 'back to battery' which is now recommended at 51V, with 'Float Voltage' at 51.5V. If charging stops at 51V (back to battery) it will never reach Float at 51.5V? Bulk Charge Voltage is now recommended at 51.6V. I (obviously) know nothing about solar and batteries, but logic brings me to these questions

[Shadders]

On 2021/03/27 at 10:07 PM, Lex de Lange said:

I also received a comms cable with the new battery, RJ45 to USB. I have connected this cable between the battery and my Raspi3 running ICC, but nothing happening.

Presumably there is a software solution to manage the battery's BMS via this cable. Anybody any ideas? (it's a simple cable, whereas, when I look at images of the Pylontech cable, there is some sort of 'box' in the middle of it. Not the case here).

This unfortunately is not as straight forward as plug in the cable and go. Each BMS will communicate the battery parameters via a protocol and then ICC needs to know what that protocol is. Pylontech for example does this via a serial RS232 connection so the box you see on the pylontech cable is actually a RS232 to USB converter. ICC then reads the USB data sent to it and converts it to human readable data. ICC currently supports Hubble, Revov, Pylontech, Narada and the Victron BMV and smart shunt models. For each you typically need a different interface cable such as RS485 to USB, RS232 to USB, V.E direct to USB.

For the ICC team to program in your battery they would need the protocol from the BMS manufacturer and if this is only a one off they might not be inclined to set this up... but hey you could try asking, there's nothing to lose. Easiest solution to import data into ICC would be to get a BMV or smart shunt from Victron with the V.E. direct cable.

There could very well be software to monitor your BMS (most likely on windows but perhaps on linux which could allow you to install it on the pi) the importer/manufacturer would have to advise on that. Even if you could install it on the pi you wouldn't be able to pull it into ICC.

[Lex de Lange]

Txs, Shadders, that's helpful. The set-up with my battery is actually a RJ45 plug (USB plug on the other end), that fits in a relevant port on the battery, marked RS232? Could that mean that the RS232 go USB is possibly built-in? Enticing thought! But, good advice, I'll contact ICC and see what they say.

[Coulomb]

7 hours ago, Lex de Lange said:

In the mean time, the importer that sold me the battery, has given me updated recommended settings for my inverter/battery. ... Yet, he recommends 'back to utility' to be set at 48V, with 'battery cut-off' at 47.4V, which he says is 20% SoC.

47.4 V is not far off 20% SOC for an LFP battery, depending on load and temperature.

7 hours ago, Lex de Lange said:

At that 48V, the inverter starts charging the battery from utility early in the morning, before solar has a chance to kick in. Would it not be more logical to set 'back to utility' at 47.4V as well, to target a consistent 80% DoD

I can't know what the importer intends, but if you set the battery cutoff voltage to 47.4 V, then with factory firmware, the back to grid voltage is actually at least 49.4 V for factory firmware. See FAQ #2.

You don't want the back to grid voltage to be too close to the cutoff voltage, or it might cut off when a fridge starts, just before it would have started running loads from the utility.

If you want your back to grid voltage closer to the cutoff voltage, you could possibly use the LFP flavour of fully patched firmware. You didn't say what Mecer 5 kVA 48 V model you have, so you'd have to check Can I Update my Firmware, and If So, To What? Fully patched LFP flavour firmware allows the difference between these two voltages to go down to 0.5 V (down from 2.0 V, lead acid thinking).

7 hours ago, Lex de Lange said:

If charging stops at 51V (back to battery) it will never reach Float at 51.5V? 

The idea is that at the back to grid voltage, there is enough solar charging to support some loads if needed, without nuisance switching back and forth from utility to battery. I actually keep mine only 1 volt apart, so that I use the minimum utility energy possible. That does mean that in wet weather, I have to manually set it 2 V apart (otherwise it switches between the two too much). I have utility charging set to 2 A max, just enough to cover self consumption, and a tiny buffer.

So having the back to grid voltage less than the float voltage is fine. PV will fill it the rest of the way to the aborb/bulk voltage (which is where charging stops, not the float voltage). If there isn't enough PV after all, then the utility will cut back in eventually, either charging the battery from utility if you enable that, or waiting till next day otherwise. I rarely charge from utility, and have to enable it manually if I do. That won't suit everyone.

[Lex de Lange]

Txs, Coulomb. Seems to me that you pursue the same goal I do, as little as possible utility charging! Although mine still reverts back automatically. I'm still finding out about battery capacity and allowable inverter current, to go totally Eskom-free I may have to go bigger. But at least by that time I hope to understand what I'm doing! Txs for the explanation, that's quite helpful. These may well not be my last questions yet!

[Lex de Lange]

Btw, I'm also not entirely sure which Mecer 5kVA I have (it should be about 2yrs old), I attach a picture of the product label. There is a later one, I believe, that's called 'King', I'm not sure whether that's a direct replacement or an additional model. But I'll check the link you provide, txs!

[Lex de Lange]

Hi Coulomb: I've done some reading on the link that you sent me. Btw I attach the label here that I promised in the previous response and forgot to attach ...

Is this one of those where I could update the firmware? The one thing I'm not sure about is the chemistry of the battery. This manufacturer produces an LFP model, but that looks different on their web pictures than this one (is also more expensive). This one does not state anything but Lithium Ion. 

I read that the voltage reading on the inverter and the battery pack can differ. Voltage seems to be fairly close in my case (i need to cover some distance between the two, so simultaneous readings are not very simultaneous), but SoC is way out and variable: at one point I read 98+% on the pack and around 20% on the inverter ...

Mecer inverter label.pdf

[Coulomb]

6 hours ago, Lex de Lange said:

Is this one of those where I could update the firmware?

Yes, you can update to patched firmware version 72.20e. Use the LFP flavour.

6 hours ago, Lex de Lange said:

The one thing I'm not sure about is the chemistry of the battery.

The vast majority of lithium ion batteries for solar are LFP chemistry. If it was lithium titanate (LTO), it would be more expensive and they'd be crowing about it. So if you can't find definite information, it seems fairly safe to assume LFP. It would be good to be sure about 15S versus 16S though.

Edit: the way to do that would be to observe the battery voltage at low load and low charge when it's at medium state of charge (not fully charged or nearly empty). A 15S LFP battery will be around 50V, and a 16S battery around 53.3V. The exact SOC doesn't matter over a fairly wide range, the cells settle at around 3.33V.

Edited March 31, 2021 by Coulomb

[FrancoisRabe]

On 2021/03/28 at 5:24 PM, Lex de Lange said:

Txs, Coulomb: I figured I need a gap between back to grid and back to battery, I guess, I'll start at the parameters from the spec sheet and see where that gets me.

I have an SBU setting on my inverter and upon re-reading the explanation you must be correct: SOL switches to grid when no solar available, or battery voltage below certain level, SBU only switches to grid when battery voltage below certain level.

I'll set the charging to 40A, if that leaves the batteries not sufficiently charged I could always go higher?

 

Settings.docx

[FrancoisRabe]

Hi

I have 2 x 48v, 3kWh in parallel connected to a Mecer 5kW inverter....Can I add another battery of 48v, 5kWh onto these in parallel?

Edited March 31, 2021 by FrancoisRabe
mistake

[Coulomb]

Yes, as long as the BMS is able to manage them with appropriate cables. There is no electrical problem, apart from different age batteries not sharing perfectly.

[Lex de Lange]

16 hours ago, FrancoisRabe said:

 

Settings.docx 154.22 kB · 1 download

Txs, FrancoisRabe!

[Lex de Lange]

On 2021/03/31 at 3:17 AM, Coulomb said:

Yes, you can update to patched firmware version 72.20e. Use the LFP flavour.

The vast majority of lithium ion batteries for solar are LFP chemistry. If it was lithium titanate (LTO), it would be more expensive and they'd be crowing about it. So if you can't find definite information, it seems fairly safe to assume LFP. It would be good to be sure about 15S versus 16S though.

Edit: the way to do that would be to observe the battery voltage at low load and low charge when it's at medium state of charge (not fully charged or nearly empty). A 15S LFP battery will be around 50V, and a 16S battery around 53.3V. The exact SOC doesn't matter over a fairly wide range, the cells settle at around 3.33V.

Txs, Coulomb, that's helpful. I'm learning here, but I feel still along way to go!!

[FrancoisRabe]

Believe me it's a steep learning curve...but we are here to help and assist, just like I still need assistance.

[Lex de Lange]

Txs, FrancoisRabe, I see that clearly developing. I have now set my machinery to the recommended settings and will run it to establish a pattern of how it works.

Once I can discern that, I may well have a new list of questions (likely!) and come back!

[Lex de Lange]

And then, of course, I will need to read up on, and execute, the firmware upgrade. Let's see if I can find out enough to become somewhat confident, else I will come with more questions on that too!

[Lex de Lange]

I don't suppose there's a step-by-step instruction list for the firmware upgrade?

[Lex de Lange]

Yes, there is a step by step instruction manual for the software upgrade and it's in the upgrade package!! And I found it!! That out of the way, I finally took the plunge and try to do something about the upgrade. I first decide to load PowerWatch on a somewhat older Windows 10 machine, to see if I could connect to the inverter at all. I had ordered a USB-D9 serial cable (of sufficient quality, I hope, I found the link somewhere in one of these forums, supplied by Communica). Seriously surprised that I could not plug this in! There are 2 ports on the inverter marked 'Parallel', obviously not the ones, since they also have a different number of pins, 3 rows, like a VGA connector, 1 in, 1 out?. The only port that looks like it could be serial (marked RS232 communication port) is an RJ45. So I plugged a cable in and connected the other end to the PC. Not for long, because smoke started pouring out of the inverter... I disconnected everything and switched the inverter off (very quickly!). Then, starting with battery, then load, then solar and then AC, I connected everything back and, luckily, the inverter still works. What does not work, though, is the data connection to my Raspberry Pi, running ICC, so I guess I smoked (literally) the comms board. That should not be to much of a hassle to replace? But that does not bring me further in upgrading the firmware!

First question: How do I connect (what cable, connectors either end, 1 end likely RJ45?) the PC with the upgrade package on it to the inverter to run the upgrade process? Clearly not via a CAT5 cable!

Second question: Is it likely that I did damage the comms board and can I replace that diy?

[FrancoisRabe]

[Lex de Lange]

For clarity I attach an image that shows the connections on the bottom of my inverter.

Mecer Axpert connections.png.pdf