Yeah, please don't install any switch-gear that is not certified to the standards required in your country (or at the very least the equivalent ISO standard if you are willing to do the research and take on the risk.) Even if it is a completely off-grid installation. Something like this can quickly turn into a fire. 

It's also a good idea to check all your terminals annually and tighten them, thermal cycling can cause them to work loose, leading to a situation like this.

On a side note - there are no affordable certified ATS (that I could find) - those that were certified cost about as much as an 8kW NRS097 hybrid inverter. So my suggestion is to rather design your system so that it does not require an ATS. You can use contactors, but not in the same way as you would an ATS. IOW, you cannot use a NC and NO contactor to function as an ATS, since they cannot guarantee interlocking/break before make.

You are right @Eutechnyx, all these products are basically the same crap. Ampacity and mechanical endurance numbers are doubtful, so if one has to use these, it's always good to go for the highest possible rating within the product range.... and count that the real ampacity will be just a third or fifth of that value.

For AC, it's still okay I would say, as the current goes through zero 100 times per second. On the other hand, I would never recommend anyone to use a Chinese MCB or PV switch in the DC applications. In the DC arcing is a huge problem. Any 10x38 gPV fuse in the DIN rail holder is much better solution there.

BTW: That ManHua ATS looks nice from the mechanical point of view. Shame I did not knew about them before.

 

I did not crack-open the new Geya ATS yet, it has to die first. Then, I will check how arc-extinguishing chambers look

Anyway, it's already mounted and I can say that there's definitely a controlling chip, not the relay that was in the old types:

• The chip (MCU) is visible on the schematics.
• Accoring to manual it reacts to overvoltage/undervoltage too.
• Once the source #1 is back online, the changeover is not immediate, but the ATS waits for 3-5 seconds first and then it switches. That would not possible with relay logic.

Speaking of the old type ATS, there ARE arc-extinguishing chambers, Just check these U-shaped bits of metal:

Well, another question is, how efficient is this cheap design. In the real MCBs, for example, this chamber is made of similar metal bits, but they are being hold by the two fiber-glass-epoxy boards normally. The closest to that is the chamber on the ManHua ATS photo, I would say.

 

10 hours ago, P1000 said:

On a side note - there are no affordable certified ATS (that I could find) - those that were certified cost about as much as an 8kW NRS097 hybrid inverter. So my suggestion is to rather design your system so that it does not require an ATS. You can use contactors, but not in the same way as you would an ATS. IOW, you cannot use a NC and NO contactor to function as an ATS, since they cannot guarantee interlocking/break before make.

For example, Hager HIC line is great, but just like you said - not really affordable. Not to mention the sheer size of the thing.

There IS a proper way how to create ATS using two contactors, but they have to be mechanically interlocked, not just electrically. You can even buy all the necessary components as a set that's meant for reversing the rotation of 3-phase motors back and forth. But given my poor skills, I am a bit afraid to build it, so I resorted to Chinese crap, backed with some prayers

A proper ATS using mechanically interlocked contactors:

Mechanical interlock:

 

Edited January 9, 20251 yr by Youda

On 1/8/2025 at 12:17 PM, Youda said:

discharge it to a 20% SOC too.

could you explain why please ? 

my battery began to stop discharging at 20% and suddenly become 5% after a firmware upgrade (it is Felicity 48 200 not pylontech ) 

1 hour ago, esmail-kassir said:

On 2025/01/08 at 10:17 AM, Youda said:

discharge it to a 20% SOC too.

could you explain why please ? 

To be sure that the battery is able to supply power in the expected operating range. For pylontech US that is 100-20%. If you go even lower, to 9% if I remember correctly, the pylontech will auto-shutdown. For other LFP batteries the range might be 90-10% with autoshutdown at 5%, for example.

1 hour ago, esmail-kassir said:

my battery began to stop discharging at 20% and suddenly become 5% after a firmware upgrade (it is Felicity 48 200 not pylontech )

That is a bad sign. There are 3 common causes for this:

1) The FW of the battery is bad. It thinks that the capacity of the battery is higher, that what is the reality. For example, the FW and it's parameters are for 200Ah version, but the actual cells are just 150Ah. Manafacturing mistake.

2) One or more cells are damaged, having a much less capacity. Normally, BMS calculate SOC by counting in/out AmpHours. But when 200Ah cell has just 150Ah, then the voltage of that cell drops very quicky at the end of discharge. So the AmpHours counter shows 20%, but since this cell is empty, and voltage is under low limit, the BMS jumps to 5% or any other value hard-coded as "battery near being flat". 

3) Cells are severely unbalanced, therefore one or two of them are unable to be fully charged. Effect is similar to #2. Cell voltage drop at the end of discharging and unexpected SOC fall.

Makes sense?

 

Pylontech US3000 after 6 years of usage:

 

On 2023/02/02 at 9:26 PM, Youda said:

Model: US2000plus
FW: V2.9
FLASH: us2000b_v2.9_Crc.bin

 

 

Hi Youda.
I have 6x US2000 Plus batteries, those were on FV2.0
As per instructions flashed all of those with us2000b_v2.9_Crc.bin, but in the BateryView and "info" command I can see version v2.8:



Is this expected or is the file attached not correct?

Thank you for your excellent guide!

Edited February 6, 20251 yr by irekz

Oh yes - there was a V2.9 firmware for US2KBPL released, that had "V2.8" written inside. I bet that you flashed this version.

There should be a file named US2000BPLS_V2.9_Crc.bin somewhere in the Pylontech Firmware Bundle and this should report as V2.9 correctly. Both files are dated as 21-04-29. On the other hand, there's absolutely no difference in functionality between the two files. It's just that version number and CRC.

 

BTW: The most recent FW for US2KBPL that I have is us2000B_Plus_V3.4_Crc.bin dated 21-08-07. So maybe you want to flash this one directly.

Edited February 6, 20251 yr by Youda

31 minutes ago, Youda said:

Oh yes - there was a V2.9 firmware for US2KBPL released, that had "V2.8" written inside. I bet that you flashed this version.

There should be a file named US2000BPLS_V2.9_Crc.bin somewhere in the Pylontech Firmware Bundle and this should report as V2.9 correctly. Both files are dated as 21-04-29. On the other hand, there's absolutely no difference in functionality between the two files. It's just that version number and CRC.

 

BTW: The most recent FW for US2KBPL that I have is us2000B_Plus_V3.4_Crc.bin dated 21-08-07. So maybe you want to flash this one directly.

thank you very much!

On 2023/02/02 at 7:52 PM, Youda said:

Question: Did you use an Pylontech Hub to connect them? or just 1 master and daisy chain them up?

On 2023/02/02 at 7:52 PM, Youda said:

Hi guys,

did a couple of updates to my lab in the meantime:

1) There was a sale on batteries, so I've added some more Pylontech US3000C bricks to my setup. It's 56kWh now. Note that the rack on the left is still empty:

2) I was not satisfied with the plastic distribution board that I originally used. Main reason was that I needed to add some new components in and there was no free space left. Therefore, I ditched the plastic box and rewired everything in the new Eaton distribution board that is made of metal. This board has a space for 196 modules, if I recall correctly, so I hope that it will last.

@Erwin Bakels I have 3 CAN ports on my PLC device so I dont need LV-HUB. Got them as 2x8 and then each stack into own CAN port.

Hi Guys,

I was given a Pylontech US2000 battery yesterday with swollen cells. I typically build my own batteries out of LifePO4 single 32700 cells, and was hoping that I could convert this one to a 24V setup opposed to the 52V setup it was originally configured for.

Is it possible to change? Would just look so much neater than the crates I'm storing my batteries in at the moment. :-)

Edit: Added pic of the unit.

Edited April 22, 20251 yr by ShaunCro

Hi @ShaunCro I don't think that the conversion to 24V is feasible. The built-in BMS board needs 48V power source in order to run. Even if you would change some component to make it run on 24V, there is a bunch of EEPROM values set around 48V limits (under voltage, over voltage etc.) But AFAIK these can be changed from the command line.

BTW: there is a pylontech battery designed for 24V - model UP2500.

Thanks for the info Youda. I'll play around and see what is possible. Like I said, it was given to me, so didn't really have a choice as to the model, but the info does raise some interesting idea's that I'll explore. If need be, I'll grab a JK BMS and pop it in the box to run the cells. Just enjoying messing around and seeing what is possible with different hardware, and the solar stuff is fun and new to me.

I need to ask though, do the newer Pylon's have current limiting yet? Because from what I have seen with this battery, it's nailing the batteries in the 3rd pack, on the negative side. And, a bit of googling and a couple youtube vids also show a similar thing. 3rd pack gets nailed first, one or more cells fail there, which leads to stress on the other cells through out the battery, and having to pick up the strain, they then start to follow suit. Seems like an issue that needs to be addressed, Also, I notice this was battery 1 in how ever many were installed in the cabinet. Which leads me to believe that it's taking the brunt of the charge, and most likely charging far too fast for these foil packed cells. This is not only the batteries fault, poor setup is also a big factor, and not enough knowledge on the matter. (Or just enough to be dangerous)

Sorry @Youda didn't mean to hijack your thread, just popped up the most when I was searching the forum. Nice setup by the way. :-)

Pylontech US5000 E3 version 2.0

Pylontech US2000C and 3000C version 2.4 Different versions for E2 orC3 OR non E2 C3

US2000C_US3000C_with_E2orC3_inserial_V2.4_Crc.bin US2000C_US3000C_WITHOUT_E2orC3_inserial_US_C_V3.0_Crc.bin US5000_with_E3_inserial_V2.0_Crc.bin

Hi guys, So the whole story.. I was swapping out battery packs in one of my pylontech batteries today as it had a dead cell and though while I've got it isolated from the bank, I thought I'd update the firmware to get all the cell data available in the Victron inverter. I was completely unobservant and for some reason, fixated on just selecting the zip so as not to worry about selecting the right chip firmware, I completely forgot the fact that I do not have us2000c batteries at all 🤦‍♂️ I have us2000b batteries, So I have bricked a BMS by stupidly flashing the incorrect firmware.
I have tried the update tool several times to recover the BMS but following the instructions included in the file, it just sits at "Get into bootloader mode" for as long as I leave it. Is there any specifics as to when or how long the red button needs to be pressed on the battery to get it into bootloader mode? Or is my BMS just simply dead now?

@Sagy89 -> https://powerforum.co.za/topic/32836-pylontech-us2000-bricked/

So I picked up 3 second hand us3000c batteries this week, 1 old chip and 2 new chip. Didn't read enough of the previous posts on this thread and got confused when the new chip units had v2.2 in them. Anyway, good to know that downgrading the software doesn't hurt them 😅

Thanks for all your hard work on here @Youda Great looking setup you've got over there. Sorry to have been another hijacking your thread, this thread seems to be the best place on the internet for getting Pylontech information though.

Edited August 1, 2025Aug 1 by Sagy89
Didn't read through enough of the previous posts in the thread

On 2019/09/04 at 9:09 AM, Youda said:

The AC-DC charger idea looks like this:

The charger can be an industrial 48V DC charger, or it can be InfiniSolar Plus 10kW 3-phase, like depicted.
If that's the case, it will give me some additional MPPT inputs too

Speaking of automation of charging from the grid, I'm talking to all my Inverters and Pylontech batteries via the PLC, so it's quite easy for me to turn on/off the charger, read SOC or any other operational value:

 

 

 

Hi Yoda.

I know this is maybe a bit far back, but I'm attempting to create my own integration software for my setup using Ignition. I have two Axpert King 5 kva inverters (parallel) and two UPS 5000 Pylontech batteries. At the moment I'm using a 3rd party app to monitor my system. But sometimes the app gets stuck and then I need to restart it. This is getting me frustrated, and because my background is Automation with PLCs and SCADA, I thought it can't be that difficult......🫣. The app is connected to a Rasberry Pi, and on it, there is a MQTT broker. When I interigate it, I can see values from the inverters and batteries. According to me, and I might be wrong, the app is reading registers on the Inverters by means of a protocol (MODBUS maybe). When you configured your PLC, what type of registers did you read and if possible, would you be open to the idea of sharing it?

Hi @wernerlombaard

On 2025/08/09 at 6:38 AM, wernerlombaard said:

According to me, and I might be wrong, the app is reading registers on the Inverters by means of a protocol (MODBUS maybe). When you configured your PLC, what type of registers did you read and if possible, would you be open to the idea of sharing it?

INVERTERS:

MODBUS (or MODBUS TCP) is available only for modern inverters, like SolaX, Victron, Deye etc. Contrary, all the machines from Voltronic like Axpert, InfiniSolar etc are speaking via RS232 and proprietary protocols, that differ between the actual models. They don't have MODBUS registers at all.

Just google for Axpert Serial Protocol and you will find a number of PDFs with a different versions describing these. For example here: https://ftps.voltronic.com.tw/E/Easunpower-93CB312922AC474787C6/RS232%20%E9%80%9A%E8%A8%8A%E5%8D%94%E8%AD%B0(communication%20protocol)/Axpert%20KS&MKS&V%20RS232%20Protocol-C(20170821).pdf

There's also a lot of projects that use communication libraries where you have all the commands included. Therefore, if you find a right Axpert Protocol Library on GitHub, you can port it to your PLC ;)

BATTERIES:

Speaking of batteries, even the most modern consumer-grade batteries are not using MODBUS. Normally, they run RS485 or CANBUS, and on top of it there's another proprietary protocol. For example just search for Pylontech LV protocol and you will find CAN and RS485 versions PDFs. I already uploaded some of these PDFs in one of my previous posts here.

ALTERNATIVE:

Anyway, if you find homebrew integration being too difficult, then one of the most popular solutions is SolarAsisstant: https://solar-assistant.io/

Runs on RPi too and if you have a good PowerSupply with solid cables and a sufficient cooling then there should be no stability issues.

Hi guys, let me share the recent update of my solar lab!

Originally I had 20kWp of PV on the roof, which was great, but the available space was not utilized optimally. Namely because there was the chimney blocking part of the west roof. Not to mention long shadows it casted each winter. After a couple of years I demolished that chimney (with the big help of my friend Kobra) and built a new fireplace, plus new chimney that freed the space on the roof.

For a shame, width and height of the panels that I had was not ideal, so even with the chimney removed I was still unable to utilize space optimally:

Therefore, I was continuously looking at the PV modules market each year, in order to find new panels that would fit better. As you might know, the manufacturers of PV modules are introducing new models each year, with their dimensions derived from the physical size of the actual solar cells used. And as solar cell makers are gradually able to produce bigger silicone crystals (that the cells are cut from), the PV modules are getting bigger and bigger, even when the number of cells per module stays within the standard amounts (like 72, 144 etc.).

Finally, in 2025 I found PV modules with the size of 1961 x 1134 mm, which was perfect for me as it would fill the roof from side to side almost completely, while still allowing me to put 4 rows of panels on it. So I went for 500Wp double-glass/bifacial, all-black panels and bought 3 palettes of them. Although these panels are bifacial, they are NOT transparent as the space between the cells is of opaque black color. On the other hand, putting bifacial panels on the dark roof makes no sense anyway. The price was excellent (for me) 77USD + VAT per panel. Let's see how these double-glass panels will hold up in the long run.

Firstly, I had and idea to remove all the old panels by myself, but found it being too labor-intensive. Looks like I am getting old, to be honest. So, after removing 18 panels out of 72 (with the help of my friend Vitek) I gave-up and called a third friend of mine, David, who is running a business of building large PV power plants for factories and other commercial customers. David's guys were amazing and after a couple of days, all the old panels were down, aluminium rails repositioned and leveled. Also, the new rails for fourth row of panels were added and finaly there was 86 panels installed.

West roof:

East roof:

Now it's 43 kWp of solar, composed of 6 strings:

• East: 14s + 14s +14s

• West: 14x +14s +16s

To be continued...

Edited September 18, 2025Sep 18 by Youda

Looking good Yoda.

Great utilization of the available roof space and double the solar.

Well done

While at it, I rerouted all the PV cables from running inside the house to new metal conduits that I've slapped on the outer wall. Added some more fuse disconnects, SPDs and used almost 200m of additional PV cable.

From the start I was a bit affraid of the fact that my inverters (infinisolar) can handle just 10A of PV current per MPPT input, but after some investigation I decided to take a risk and plug these new 15A string in. Works like a charm - infini just cuts all the excess power, limiting to roughly 5kW per MPPT input. Looks like even the spikes are handled without a glitch: