Seplos does not honor warranty (BMS stops responding on RS485)

[kenzo]

Wondering if anyone else has seen this problem in one of their battery packs. This has been happening at least twice now for one of the packs I ordered from Seplos; I'm using dbus-serialbattery to gather information via RS485, and this particular BMS stops responding after some time - only a reset resolves the problem. The BMS is not fully locked up, it can still be queried locally via the buttons and the display shows all values as well.
Seplos support isn't helpful at all unfortunately - has anybody else here seen this behavior and knows a workaround/fix?
Example statements from support:

This may be the rs485 communication is disturbed, you need to reboot the bms, after a few minutes and then turn on the bms will be fine

and

Hello, this may be the 485 communication port is subjected to strong current or strong magnetic force, you need to make a shielding cable for them on it

 

[BrunoAlves]

Actually their justification is plausible... EMI is in fact "a thing", not sure if it can or cannot cause your problems (I don't really think it does), but it would be easy and cheap to do what they say. If you are using standard RJ45 patch cables get one with metal connector and shielded cable, if you are using a self made get these metal/shielded components and build another cable.

 

[Quattrohead]

Try swapping the order of the data cables and see if the fault moves. It could be a termination issue, remember RS485 needs a 120 ohm resistor on the last device. Some devices will auto terminate but I don't think many do.

 

[kenzo]

Actually their justification is plausible... EMI is in fact "a thing", not sure if it can or cannot cause your problems (I don't really think it does), but it would be easy and cheap to do what they say. If you are using standard RJ45 patch cables get one with metal connector and shielded cable, if you are using a self made get these metal/shielded components and build another cable.

I'm not saying that EMI can never be a cause for some particular issues (even if I doubt it's true in this case) - did you see their proposed solution, though? "Just reset the BMS whenever it happens"; does that really seem like a suitable solution?
I have built plenty systems based on Seplos, using the Waveshare USB/RS232/RS485/TTL converter and they all work fine; this is a single system out of a dozen that intermittently stops working. I really don't understand how EMI on RJ45 with RS485 could affect the BMS in such a way that it locks up and needs a reset - all cables in use are high quality CAT6/7 cables that otherwise happily transmit GBit/s for Ethernet. What's the mechanism you have in mind that would possibly cause this?

 

[kenzo]

Try swapping the order of the data cables and see if the fault moves. It could be a termination issue, remember RS485 needs a 120 ohm resistor on the last device. Some devices will auto terminate but I don't think many do.

There's just one data cable from the BMS to the converter; using the 2nd RS485 port doesn't change the situation. For termination, it appears to be common ground that "long" lines need termination, "short" ones don't (which doesn't mean cable length). However, it seems inplausible that an RS485 works fine for weeks without needing special handling with regards to termination, and then stops working completely out of a sudden but goes back working fine for weeks following a BMS reset. I'd expect intermittent packet loss etc. or complete lack of communication in such a situation.

 

[BrunoAlves]

did you see their proposed solution, though? "Just reset the BMS whenever it happens"; does that really seem like a suitable solution?

Of course not... TBH I'm used to getting these kind of replies, regarding the most varied products when you are attempting support from Chinese suppliers. My conversations on Aliexpress are sometimes hilarious (not in a good way...)
Also I just exchanged some emails with Seplos and they admitted to be using GTranslate, so communication is in fact a problem :-(

I really don't understand how EMI on RJ45 with RS485 could affect the BMS in such a way that it locks up and needs a reset

I understood that the BMS does continue to work, only the RS485 communication "dies". It is not uncommon that after a certain error condition occurs the controller (firmware) shuts down that part and only a reset can make it run again. If it was a hardware malfunction it would probably become permanent after some time and a reset wouldn't fix it. If it was firmware it would happen in all devices running the same firmware. That's why I said it's plausible.

Anyway, to be clear, I don't know think that this is the case and most likely the shielded cable won't solve your problem. Nevertheless it is a cheap try and that would allow you to get back at them with evidence.

 

[kenzo]

If it was firmware it would happen in all devices running the same firmware. That's why I said it's plausible.

Anyway, to be clear, I don't know think that this is the case and most likely the shielded cable won't solve your problem. Nevertheless it is a cheap try and that would allow you to get back at them with evidence.

It is a shielded cable, and I responded with that info which makes Seplos support go silent; it seems like they want to deflect, not understand or fix a problem. I do however agree that what you're saying is plausible, and there might be a chance that the RJ45 cable (or the RS485 converter) may be at fault here and replace both to see what happens next.
I guess the reason I'm "salty" when it comes to Seplos support is that there's no indication of them wanting to work with you and do some actual and open ended troubleshooting that may result in them replacing a faulty part - "it can't be us" seems to be the starting point and that's how all conversations go.
Yes, I know - you get what you pay for. I'm learning that the hard way right now ?

 

[Quattrohead]

Oh well single connection and it was fine and now it's broken, should be case closed.
It seems like Seplos are a rather difficult company to deal with, they seem to almost make a good product and then balls it up right at the end with shit firmware and customer support. Andy at the off-grid garage has documented this extensively.

 

[BrunoAlves]

It is a shielded cable, and I responded with that info which makes Seplos support go silent;

Ahhhh didn't got that info before, sorry. Then yes, it becomes a "silence treatment"
Other than a "friendly reminder" I don't know what else you can do, overall it's impossible to force an answer from them. Good luck

 

[cajocars]

I had the same problem a few days ago after a year of 100% reliability.
It started working again after disconnecting the bms from the battery and reconnecting it.
Someone on here suggested getting the latest firmware from Seplos which apparently solves the problem. Reading this discussion, I’m not sure if it’s true as they haven’t offered that solution to you.