francois

The dust covers at the top on the inverter also gets clogged with dust.

Status Update
At the end of a couple of weeks of testing various combinations of ICC versions on a Pi3 and a Pi2, with Debian Buster Operating system and with the Debian Stretch Operation system, I figured that the culprit was a SD Card with some sort of file corruption.
On creating a new SD Card from the latest Pi4 image (with the Buster OS) to test on a Pi3, ICC immediately started to run smooth without any freezing problems. When testing the same on a Pi2, the same result as with the Pi3, no freezing. However the Pi2 struggled with the screen refresh rates of ICC on Buster OS. Done the same with the new SD Card but with Stretch OS on the Pi2, also no freezing so far, but with better screen refresh rates of ICC.
As part of the tests, I replaced the power supply on the Pi2 with the power supply of the Pi3 (5.1V 3.0A), and then realize that the Pi2 was under powered. According the Raspberry Pi forums, an under powered Pi can cause SD Card corruption.  On a Pi2, the red LED must be constant On, this is showing that the Pi2 is receiving sufficient power. If it flickers or is Off then the Pi2 is under powered. This was the case with the Pi2 before changing the power supply.
My only conclusion, at the moment, is that there is file corruption on the old SD Card of the Pi2 which caused the newer versions of ICC to freeze. The file corruption was caused by an under powered Pi2.
So far, the current run time of ICC V2.9996 on a new SD Card with Stretch OS on a Pi2 with a Pi3 power supply is approaching 70:00 with out any freezing problems.

Cool - happy to hear that you got it sorted
Power and SD card problems are apparently not uncommon with the Pi. But have to admit that I also did not think of that at the time...
I've got the 4GB version of the Pi and have moved directories with frequent writes into RAM to reduce the number of writes to the SD card (see here for details and other tips on extending the life of the SD card)

Hi
I still need to read most of the posts, but i'm excited to start this journey.
 
As I'm writing this, the guys are busy installing an 8kw SunSynk invertor with 16 * 410kw Canadian Solar Panels.
I've opted (didn't really have a choice) for no batteries as I would like to start saving money first and then get the batteries in for our dreaded load shedding situation.
I'm hoping to run all my heavy loads during the day, the biggest contributors are geyser, pool and heat pump and washing machine.
 
Any important gotchas that I can ask the installers to make sure they have done?
 
 
 

Just commissioned my Easysolar GX II on Monday with two 2.4 Pylontech’s and eight 355 watt Canadian Solar panels, system working great, a huge thanks to Jaco for supplying the Easysolar II and it so happened that he was in Cape Town for a couple of installations and graciously offered to help commmision the system, which I would have spent days, also, cannot forget Plonkster’s help
ps: the WiFi setup on the Easysolar solar is basically non existent,so plan on running an eathernet cable to your router or use the WiFi dongle
   Tariq 

ROI is typically the single most important factor when considering going solar - even though being Loadshedding resilient and producing clean energy are close seconds and thirds.
The screenshot gives our savings on a 5.6kWp (320W modules) - 10KVA (Axpert inverters) - 13.5kWh (Solar MD Li-Ion Battery) for the past 3.5 years. Cost of system R200K. Annual solar savings: R25K+
What it doesn't show, is that we have further estimated savings of R25,000-R30,000 per year in avoided petrol purchases and vehicle maintenance as the system is also powering our Nissan Leaf EV. We use the electric vehicle for school runs, daily shopping and the odd meetings/outings and are able to have it on the solar powered plug for 3-4hours per day.
Standalone the solar system will pay for itself in 6-8 years but adding the car doubled the ROI of the solar system.

I am new here, been building up my system for some time now.
Started with the following a bit more than a year ago purely for loadshedding backup. System started as follows:
1. Axpert 5KVA (Mecer)
2. 4 x Vision 200ah AGM's
After a year I started having problems with the batteries and decided to upgrade to Pylontech and also at the same time bit the bullet and ordered the panels to start producing own power. Connecting the Pylontech's to the Axpert I realized that communication between batteries and inverter was not going to work, so also ordered the Raspberry PI with ICC software
System currently as follows:
1. Axpert 5KVA (Mecer)
2. 2 x Pylontech US3000
3. 6 x Canadian Solar 365W Poly KuMax
4. Raspberry PI with ICC software
Will be adding two more Pylontech US3000's (for total battery backup of 14 Kwh) as well as 6 more 365W panels in the next month or two
I will probably be asking a lot of questions as I have been lurking these forums as a guest for a long time looking for answers to all of life's power questions
Attached a few photos of my setup

Well, I finally twisted Jaco's arm into a non-blue install, and the results are beautiful...
With the whole house as a dummy load, I also got a chance to properly test the battery and assorted bits.
Load peaked at 11kW before the geyser controller managed to turn the geyser off, and then stabilised at around 8kW, with 195A draw from the batteries.  After around 10 minutes, batteries, terminals and connecting plates were all still at room temperature.  The wires connecting to the inverter were however a few degrees warmer.  Glad I did not go smaller than the 70mm^2 I eventually settled on. (I have 2x 95mm^2 wires in parallel for joining the banks - and these wires also stayed cool.)
All in all, it has turned out to be a successful experiment so far.
Been off grid for 15 hours so far, and hope to stay that way pretty much indefinitely.

This is becoming a common problem. Panel voltage is so high these days that 3S of 72-cell panels (they are nearly all 72-cell again these days, especially the panels that are over about 315 W) is too much voltage. Yes, 121 V is within the specifications, but it seems to cause problems. Another forumite found that the SCC gets much hotter when running 3S, so it may be shutting down due to over-temperature. The Synapse is basically an Axpert copy or work-alike; they may not have copied the cooling system properly.
Rewiring to 2S6P is a pain, but it will likely fix your problem.

There is your problem. A hair dryer is typically > 2kW, I think my wife's is rated at 2.3kW. The Microwave probably around 1kW.
Assuming there is another 200 or 300W base load, this is enough to trip both the battery and the Goodwe.
So the question is which one is tripping. Since you say it restarts after a few seconds, I would have to say it's the Goodwe tripping, because as far as I know if the Pylontech trips on over current, you need to manually reset it.
Edit: Sorry, it's actually not enough to trip the battery since the 74A (for 2 batteries) is not a hard limit, and they will not trip at that point. So definitely the Inverter tripping.

Those are both very big loads. Those small cheapy microwave ovens are 900W and then they go up to 2000W. The hair dryer is normally 1800W or more, expect those smaller travel types. I agree with  Stanley that it is most likely the inverter. Those 2 together will trip most inverters.
Best is not to make food while drying your hair. It's better for power usage (and hygiene  ) 

Thanks Guys.  I've got them back here later this week to fix this issue.

I groan every time I read this. 120 V is quite high, and it will get even higher in colder months.
Kings may be more sensitive than other models with respect to high PV voltage. I can't think of a good reason why; the SCC should be identical to those of an Axpert MKS model.
It might be worth isolating to just one string, and rewire to 2S for that single string, to see if it suddenly recognises the panels. Or perhaps see if the PV is recognised early in the morning or late at night, when the panel voltage is under 115 V (the maximum MPPT voltage).
I assume you're seeing the 120 V on a multimeter right at the PV input terminals.

Hi @CliveS, 
Is the attached datasheet the one for your inverter?
If so, I see the MPPT operating voltage is between 60V and 115V.
In your one screenshot where the unit is working correctly, you can see the voltage is at 107V (within the limit), but in the other screenshot where it is not working the voltage is 121v (outside the operating limit).
I am not sure if this is your problem, but maybe worth investigating.
Like @Rclegghas said, it's possible that the panels are wired incorrectly and during certain conditions the voltages are exceeding the limits of the MPPT, and then it shuts down.
Synapse 50 48V OffGrid Inverter.pdf

@RikH thanks, the build has definitely been a great deal of fun. I look back at what was a battery to extend the UPS on the computer, then after watch a zillion youtube videos then thinking to myself I will build just a small 24V system to keep the gate and alarm system powered... well I am now at a 48V system and asking myself if I shouldn't just build a battery shed to fit a small 100kWh battery, cover the entire roof in solar panels...... just a few mods that the wife will not notice!

Seeing as the 'smart' home bug has bitten me in a bad way, I've been working on integrating all the various bits of electronics in the house. One of the obvious things to get integrated is the power monitoring side and in my case, a Goodwe 5048ES inverter. I'm using Home Assistant as my automation system and there is integration for this available, but only to the SEMS portal. This won't do, as the SEMS portal isn't great, it sits in the cloud and it only updates every 5 minutes or so. 
Ideally you want to integrate directly to the inverter and get much more frequent updates, but I couldn't find anything that is integrated directly. Looking at the inverter there's a few RS485 ports available, so it might be possible to get some data off this. But it's also possible that these are only for battery integration or otherwise unused. So, I figured a more elegant way would be to try and talk to the inverter through the network seeing as it's sitting there already. Again, there's pretty much no information available on achieving this but I figured it was worth a try.
Running PV master on a mobile device and a packet sniffer running in the background, I managed to log some interesting data in order to allow me to talk to the inverter through it's network interface. I've made some progress, but this is very much a work in progress; decoding the data is a bit of a challenge but I think I'm about halfway identifying all the data. I'm posting this here as hopefully someone else has a bit more insight and can help filling in the missing blanks, or at the very least, it might help someone with a bit more smarts to crack it properly.
Basically, the communication between PV Master and the inverter is in the form of UDP packets with the inverter listening on port 8899. Because it's UDP, it's connectionless. There's no need to explicitly connect to the inverter (as would be the case for a TCP connection), it simply sits and listens for the relevant requests and then broadcasts back with the results. Nice and simple, although (because of the inherent handicap of the UDP protocol) there's no guarantee that there will be a response. In testing I've had responses from the inverter in most cases, but it did happen where the response went missing. So, you'll have to make sure your implementation is robust enough for this case and for ensuring that all the data packets do arrive and that they are complete. There is a checksum in the inverter's protocol in order to validate the data, but it's something to take note of.
There's a few requests you can send the inverter, and based on that the inverter can respond with things like it's model/serial, it's status or confirmations whens you change settings. For the most part I'm only interested in getting the inverter status. Interestingly, there's no authentication mechanism. You fire off the request, and the inverter will answer - provided your request makes sense.
I've attached a document outlining the information I was able to decode from the information request response from the inverter. Still a few blanks there but the results are encouraging. In order to get the response below, you need to issue the following bytes to inverter IP on port 8899: 0xaa, 0x55, 0xc0, 0x7f, 0x01, 0x06, 0x00, 0x02, 0x45. The 6th byte is the request type and the 9th byte is the checksum. For the purposes of this exercise I won't go to deep into the various requests you can send to the inverter. 

Good day Guys
I started a small project for a battery backup system for my home. No solar fitted at the moment, my intention is to do that in the next phase of my project.
Inverter: RCT VMIII
Batteries: 2 X OmniPower 260ah AGM GEL VRLA
This installation is 2 days old now. My inverter settings are still in default mode. I would appreciate any recommendations to fine tune it
based on the equipment that I am using.
One battery is inside the battery box & one is on top.
 
 

Hi guys,
based on the fact that many forum users are trying to understand how to use Pylontech batteries efficiently, I've extracted some of the internal configuration values. It's from a stack of 8xUS3000. What instantly catched my eye are especially these values. So I added a couple of my comments and thougths:
Total Num                 : 8                    //Number of maximum daisy-chained bricks supported by the firmware.  
Present Num            : 8                     //Number of actually daisy-chained (RS485) bricks.
Shut time                  : 72.0 H            //If the brick is not charged/discharged for 3 days, it will auto-power off.

Balance Volt             : 30 mV
Balance Start           : 3360 mV       //Looks like the BMS is able start balancing cells at this voltage it's not
                                                          //a static top-balancing, as I know that the BMS chip is able to stop balancing
                                                          //at a much higher voltage, if all the cells are at roughly the same voltage level.

Recommend chg voltage    : 53250    mV   //This is what inverters like Axpert are reading
                                                                       //from the BMS and using as CC.
Over VoltageR           : 51000 mV                 //Strange, that this is triggered at 51V, given
                                                                       //the recommended CC=53.2V.
High VoltageR           : 52500 mV                
High Voltage             : 53900 mV                 //Just 650mV of margin between recommended CC=53.2V and alarm.
Over Voltage             : 54000 mV                 //Just 750mV of margin between recommended CC=53.2V and panic.
 
A complete list goes here. Please note that the column "Battery" means actually a CELL while the column "Power" means a single US3000 BRICK:
Protect Attribution --------------------------------- Item Battery Power Over Voltage : 3700 54000 mV Over VoltageR : 3600 51000 mV High Voltage : 3650 53900 mV High VoltageR : 3500 52500 mV Low Voltage : 3050 46000 mV Low VoltageR : 3100 47000 mV Under Voltage : 2900 44500 mV Under VoltageR : 3250 49000 mV Sleep Voltage : 2500 38000 mV Charging OT : 61000 61000 mC Charging OTR : 55000 55000 mC Charging HT : 59000 59000 mC Charging HTR : 55000 55000 mC Charging LT : -9000 -9000 mC Charging LTR : -5000 -5000 mC Charging UT : -11000 -11000 mC Charging UTR : -5000 -5000 mC Discharging OT : 61000 61000 mC Discharging OTR : 55000 55000 mC Discharging HT : 59000 59000 mC Discharging HTR : 55000 55000 mC Discharging LT : -9000 -9000 mC Discharging LTR : -5000 -5000 mC Discharging UT : -11000 -11000 mC Discharging UTR : -5000 -5000 mC Charging OC : 102000 mA Charging OC Alarm : 50000 mA Charging OC AlarmR : 40000 mA Discharging OC : -100000 mA Discharging OC Alarm : -50000 mA Discharging OC AlarmR : -45000 mA OC Delay : 15000 mS OC Release : 60000 mS Charging OC2 : 200000 mA Discharging OC2 : -200000 mA OC2 Delay : 100 mS OC2 Release : 60000 mS Discharging SC : -400000 mA SC Delay : 0 mS SC Release : 60000 mS Charging Max Cur : 10000 mA Balance Start : 3360 mV Balance Volt : 30 mV Shut time : 72.0 H BUV/PUV time : 2400 S Sleep ctrl state : OFF Sleep every day StartTime : 20:00 EndTime : 08:00 HwSleepInfo HwSleepStatus : OFF HwSleepTime : 6 S HwWakeupTime : 60 S Data Save every day StartTime : 00:00 EndTime : 23:59 Save Interval : 1800 S  
Power System Information --------------------------------- System is idle Total Num : 8 Present Num : 8 Sleep Num : 0 System Volt : 49381 mV System Curr : -367 mA System RC : 352367 mAH System FCC : 585266 mAH System SOC : 60 % System SOH : 99 % Highest voltage : 3294 mV Average voltage : 3292 mV Lowest voltage : 3290 mV Highest temperature : 23000 mC Average temperature : 21875 mC Lowest temperature : 21000 mC Recommend chg voltage : 53250 mV Recommend dsg voltage : 47000 mV Recommend chg current : 296000 mA Recommend dsg current : -296000 mA Hope this helps all the young scientists that are developing their own monitoring and control solution for the batteries and inverters
Youda

Well I guess an option would be to have a big bank that won't be able to charge from 20% SoC to 100% SoC with the PV available in which case Victron's BatteryLife (if enabled) will kick in. The way I understand it (pretty much pasting from there) is that it will raise the Low SOC limit by 5% each day until it reaches a 85% SoC. It will then stay on that level, but if it reaches 95% then it will lower the Low SOC limit by 5% again.
It's not exactly the "ensure 100% once a week" state, but it ensures that it reaches 85% to 100% each day.

Personally, I'm cycling my US3000 stack between 100% SoC and 40%SoC. Once discharged to 40% SoC, my system automatically switches to the grid. Since manufacturer's specs allow cycling between 100% and 10%, I have a lot of margin here.

BTW:
Technically, it would be better to cycle these between 40% SoC and 90% SoC daily, while allowing a full charge (100%SoC) only once a week, to let BMS reset amphours counter and balance the cells. But as far as I know, no inverter on the market has ability to set such a complex rule.

Are you sure that you are not confusing DoD with the SoC?
80%DoD means that there's roughly just 20% of energy left in the battery. Such a deep everyday cycling cannot be called "wasting of a potential" as consuming that last remaining 1.4kWh won't make any change but it will kill the batteries for sure.
In the US3000 BMS the 9%SoC is hardcoded as a lower limit. Once you discharge US3000 to this limit it will auto-shutdown and log an error to its NVRAM.

Hi guys,
just an update to Pylontech BMS + Axpert/InfiniSolar data connection:
The original information that I've received from MPP Solar, that BMS RJ45 cables are the same for MPI Hybrid and PIP, was found to be wrong. Actually, the wiring of the cable is a bit different, so here's the update:

Just for the reference, the live pins on the Pylontechs' side are still the same 1&2, protocol RS485. But the RS485 pins on the Axpert and on the InfiniSolar differ from each other. That's the reason, why you need a different cable for each type of the inverter.
Hope this info will be help someone...

Hi guys,
couple of months ago, MPP Solar announced the capability of it's inverters (PIP-5048MK, PIP-5048GK and MPI Hybrid) to actively communicate with the Pylontech Low-Voltage batteries US2000 and US3000.
The main thread about the topic on the PowerForum is here: https://powerforum.co.za/topic/2519-mpp-solar-inverters-support-pylontech-some-models/?tab=comments#comment-40801
Since then, couple of people have asked on the forum about the correct pinout of the required RJ-45 BMS cable. So, I purchased a respective kit for MPI Hybrid from the MPP Solar and now we can take a look at it
RS485 BMS Card + cable for MPI-Hybrid:

I received a confirmation from MPP Solar, that the provided cable is compatible not only with MPI Hybrid, but with PIP-5048MK and 5048GK too. The IO card itself is only for MPI Hybrid, as PIP inverters have the BMS RS485 RJ-45 port already embedded.

 
Cable wiring is pretty simple:

Here's the PDF manual for the card:
RS485-Card-Box-for BMS-manual-20180927.pdf
Speaking of PIP, the compatible firmware has a "05 - Pyl" option available in the menu:

 
Here are the respective PIP user guides, which cover how to configure your PIP for use with the Pylontech battery:
http://www.mppsolar.com/manual/PIP-GK/PIP-GK manual-20190201.pdf
http://www.mppsolar.com/manual/PIP-MK (PF1.0)/PIP-MK manual-20190429.pdf
So, if you do believe that your PIP or MPI Hybrid is a newer 2019 unit and already has a firmware with Pylontech support, but you just miss the cable, then you can create a cable yourself and give it a try.

Okay thanks for confirming.
Mine looks the same.