Douw Gerber

Your inverter seems to be a 48v unit...this is a 24v battery.  How do you plan to use it?

+1 - Runs my system perfectly - I have a very similar setup to yours and I have reduced consumption from 30Kwh/day to around 16Kwh/day.  All automated and easily changed when loadshedding is going to happen.

Hi all
Just putting this here for the next guy that wants to DIY this - I fried the first one LOL - R399 school fees 🙂
Notes:
-You need to be pretty close to you WiFi router for this to work - it obviously doesn't contain a high strength antenna - my poolpump is +- 10m away from my router and it just gets enough signal.
-When setting up the switch in their app - choose generic WiFi breaker
-Press and hold the switch button for 5 sec and release when discovering the switch in the app - bluetooth MUST also be on.
Info on Switch here: https://www.acdc.co.za/pages/atms1601-230vac-smart-wifi

These two display firmwares are about 98% identical, but there is devil in that 2% detail.
It seems to me that these two firmwares (30.06 and 130.06) are for slightly different display hardware. There are a few bits of code that use GPIOA instead of GPIOB and vice versa. This seems to be associated with I²C (I2C) or SPI sending and receiving.
So until other information comes to light, these seem NOT to be interchangeable.
They both seem to be attempting to unify many models: VM III, King, MKS II, unknown (possibly MKS III), and MKS IV. The last three have internal codes 2, 3, and 4. In particular, both will show setting 28 (output mode, pertaining to paralleling), unless the model code is 0 (for VM III). But I get the feeling that this unification is not yet complete, and they are still striving to complete the unification. I note that the VM III has a removable display, while the VM IV has a fixed, colour display, so I don't see how these can be unified. Perhaps the new colour displays are a superset of the removable display somehow.
130.06 appears to be the later of the two firmwares, as it has more code.

I always wanted that my 280Ah/16S LiFePo4 battery pack with it's JK BMS fully communicates with my MPI 10k inverter.
Because I did not find any hints regarding this in the net, I started reverse engineering the official BMS card from voltronic, designed for pylontech batteries. 
Conclusion: You simply need to send a ^D054BMS... string periodically to the inverter.
Format looks like this:

I created a script to read bms data from a mqtt server (I used a ESP32 to read from JK BMS, any others BMS should be possible with little adaptions) and create the ^D054BMS string that can be sent to the inverter: readmqttbms.php
Sending this periodical to the inverter (I adapted my script infinipoll10k) let's the display show the "LI-bAt" and battery values read from BMS.

The inverter get's informed about BMS stoped charging/discharging and can react correctly:
Reading the actual bms values from the inverter can be done with the ^P004BMS<CR> command.
There are for sure some values that I missed so far (e.g.the bms fault codes), but I do not own a pylontech battery, so reverese engineering without help from others is impossible.
 

The prolific driver tells me that the cable is counterfeit.

The supplied drivers on the mini cd that came in the package is rubbish.
It gave runtime error the moment you try to run the exe.
Tried this on both Win 11 and Win 7, but still no joy.
Downloaded latest drivers from prolific which gave me the counterfeit message in device manager.
Downloaded an older version of the prolific drivers which installed fine and the port was visible in the reflash tool, but it just failed with an error saying that it cannot open the port.

So I went and ordered the expensive cable from Solar Assist as suggested by @Douw Gerber, which arrived today.
You get what you pay for: I just succesfully flashed the firmware and am now on the latest 71.97.

Thought the display would also be updated, but it didn't.
So I found the display reflash tool for the display panel in an earlier post and will flash that when my wife doesn't need the power.
EDIT: Display is now also updated to 02.70

To confirm, are these the latest versions:

CPU: 71.97
DSP: 02.70
 

I used that USB to Serial adapter from Matrix - doesn't work - it needs to use the FTDI chip to be 100% sure it works.
This one also works https://www.takealot.com/ce-link-usb-2-0-to-rs232-db9-serial-male-a-2m-cable/PLID51855970
Cheers
D

You can buy the correct RS232 cable from SolarAssistant
https://solar-assistant.io/shop/products/voltronic_rs232

I used that USB to Serial adapter from Matrix - doesn't work - it needs to use the FTDI chip to be 100% sure it works.
This one also works https://www.takealot.com/ce-link-usb-2-0-to-rs232-db9-serial-male-a-2m-cable/PLID51855970
Cheers
D

Wouldn't the generator need to support autostart ?  Dunno if 2.2kw gennie will have this function unless its an electric start.
Also, now I am asking to learn 🙂 how would he connect the gennie and eskom - I assume these inverters have two seperate inputs or a mechanism of detecting eskom outage and starting up the gennie remotely - but this would not really work for the idea to use the gennie to fill the solar production gaps.  A battery is needed for this I would think.

It would seem that the Oliter is a 8cell LiFePO4 battery, then you can basically use the 8 X xV, where for charging, which should not exceed, lets say, 3.55V, so 28.4V charge Voltage and float should then be 28.3V or 28.35V, if you want to be kind to your battery, you can drop these by 50 to 100mV per cell. On the low end, I wouldn't go lower than 2.9V on the discharge side per cell, so 8 X 2.9V, so 23.2V absolute low before calling it... but again, you may want to be more conservative.
For the most part the min/max Voltages for LiFePO4 cells are given as 2.5V/3.7V, but do some more homework on this, I do not have a Oliter and the cells they use may have different parameters (I doubt it), but getting to the extreme edges of the cycle, 2.5V and 3.7V should probably be avoided as those edges on the charge/discharge curves don't contain much energy anyway and will only shorten the life of the cells once you dip into those ranges regularly...

See the image attached @MKRandburg
sorry @TimCam LOL I was thinking you posted OP 🙂
OP asked for:
My requiremement
Reduce eskom bill - solar can help here
Prevent feedback onto the municipal grid - inverter should have this option
No batteries (I think sunsynk requires a small one) - you will need a gennie as eskom backup
Solar PV only - PV only systems are for grid backfeed solutions where backup is not required.
Daytime loadshedding back up from solar panels - your gennie will fire up if eskom is down - somehow you will pay - cannot guarentee panels will be able to supply load always
Use the generator during night time loadshedding - feasable but you will still need a battery to carry the loads during generator start
Blend Eskom and Solar when load is too high - can be done
So bottom line is I think what OP is asking for isn't practical.
You will be able to reduce your energy bill using solar but with no battery you wont have backup.  Well you could fire up a gennie but everything would go offline while that gets started if the load is higher than wht the panels can supply at that moment in time.

yeah I assumed you meant manually - but sounds like a lot of pt 🙂 I guess an ATS could possibly start the gennie if eskom is down (battery still needed during generator startup) but that still isn't achieving what you want with the infill when solar isn't generating enough.
I tested what @MKRandburg is looking for with my Axpert Hybrid inverter - solar and eskom only - no battery - supports the loads and fills in with eskom where necessary when in SUB mode. [EDIT]

Obviously if your budget allows go for a Lithium battery - but if not and all you are really after is battery backup during load shedding, then there is nothing wrong with going for an AGM battery bank - 12v100ah x 4. You run the inverter in SUB mode.  Will cost you in the region of R10-11k. 

The battery is pretty much managed by SA using the automatic power management features. 

SA got back to me to confirm that the Dyness reports 50AH but they are working on a fix for this.  Its doesn't affect the battery runtime.

Hi all
I had to upgrade to a beta release of the SolarAssist software.  All seems fine now excep that it reads the battery as 50Ah. But capacity SOC and Voltage readings look normal now at least.

The question is, are you able to measure or monitor the Volts and Amps either on the inverter or via software?
If not, your theoretical assumption is just that. Have you checked each strings voltage with a multimeter?
How are your panels orientated and what is the panel angle  ,i.e azimuth and pitch?
 

From what I have researched you can expect a max of about 70% of your panels rated watts.
I have 8 x 455w installed and today at max I was getting 2400w - it's winter and my panels are installed flat on my flat roof house (maybe a 10 degree pitch).
So yeah a lot depends on how they are orientated, angled, time of year etc.
The rated values are in perfect lab conditions

Hi there all
Made the leap to solar with backup power in the last week so far very happy with the system performance.
Inverter: Mecer Axpert MKS IV 5.6Kw
Battery: Dyness BX51100
Solar: JA Solar 8 x 455w
Having some teething issues with the battery to inverter comms - will raise this in the other threads where applicable.

Hi there all
Made the leap to solar with backup power in the last week so far very happy with the system performance.
Inverter: Mecer Axpert MKS IV 5.6Kw
Battery: Dyness BX51100
Solar: JA Solar 8 x 455w
Having some teething issues with the battery to inverter comms - will raise this in the other threads where applicable.

Hi there all
Made the leap to solar with backup power in the last week so far very happy with the system performance.
Inverter: Mecer Axpert MKS IV 5.6Kw
Battery: Dyness BX51100
Solar: JA Solar 8 x 455w
Having some teething issues with the battery to inverter comms - will raise this in the other threads where applicable.