Failed JK BMS

[-g-g-g-]

I had tried plugging in and unplugging the button (mine came with a button instead of the display), and unplugging / reseating the balance lead JST connector.

It sounds like the same issue I have. If I try to connect via Bluetooth it goes into chaos mode and sounds U in morse, dot, dot, dash, for ever. I disconnected everything and it did the same on reconnect but after fiddling with the balance connector and replacing the display with the switch, it finally went back into working mode.

It thinks that the cell count is wrong, which reinforces my impression that it could be a connection issue.
I wonder if connection to Bluetooth momentarily disconnects the power and the reset sees a cell fault & goes to the alarm signal, which sounds like failed attempts to reconnect dot, dot, dash rather than dot, dot, dot

One morning it had started over with dot, dot, dash again probably because a cell had dipped below 2.5V and the BMS had switched off overnight.
That's when I discovered the 10 pin connector wasn't making full contact.
Not wishing to start squeezing connector pins without a replacement, I settled for "it's working - leave it alone" and reached the point where I am at today.

 

[upnorthandpersonal]

or the bloody rose emoji.

?

 

[Delphinus]

Guys, confirm the connectors, they are not good, unlike most that cut the cables for the extra cells that are not in use, I removed the terminal with the cable and notice that the terminal/pin have a loose connection, my bottom terminal in a 16S configuration only has 3 terminals and I need to use some hot glue to keep it in place. Not a good design for a 2A BMS.

 

[hwse]

Thing is, Hankzor was really good, fast replies, etc. and then basically from one day to the next it wasn't. I've done extensive tests and analysis on the older BMS versions they have and they're really good - I've been running my house on them for three years now without any issue what so ever. It seems that something changed in the mean time, but I don't know what. Hopefully I can do some more analysis and draw some conclusions based on the one @Horsefly sent me - once that arrives.

All of the issues that I have heard of are in a 12v application which was not possible before because the system needed 24v to function. those BMS's could be used for 12v by connecting the last + wire on the plug to a 24v boost converter.

I had always assumed that they model that can be used on 12v just had the boost converter built into it so that the end user would not need to do that.

Once you get the failed one from @Horsefly it will be interesting to see what it is actually doing. I am on my second 4s JK and it is currently still working. If it ends up failing, I might just go with a B2A20S20P model with a boost converter.

 

[RCinFLA]

The original units had hand soldered, total end solder tinned wires between the interlaced MOSFET's. Newer units have a reflowed bare copper rectangular jumper bar.

The new reflowed bars are only solder wetted at the bottom surface. This has more interface resistance and is less reliable a solder joint. Solder reflowing bare copper bars is not reliable. Copper will oxidize over time so solder joint quality depends on how long the bars have been in stock before sent for assembly reflow. Any bending of jumper bars before installation can lift the end off of solder pad reducing solder interface when solder reflowed.

The copper bar will expand and contract in length with BMS heating. Very tough on marginal reflowed solder joint.

I am sure this was done to reduce the hand soldering labor cost.



Would you trust these high current solder connections? Looks like the jumpers are just cut from a copper sheet with a cutter shear.
Pretty much guaranteed to fail after some BMS temp cycling.

 

[Steve_S]

Thing is, Hankzor was really good, fast replies, etc. and then basically from one day to the next it wasn't. I've done extensive tests and analysis on the older BMS versions they have and they're really good - I've been running my house on them for three years now without any issue what so ever. It seems that something changed in the mean time, but I don't know what. Hopefully I can do some more analysis and draw some conclusions based on the one @Horsefly sent me - once that arrives.

If you'd like I can also send you a JK B2A24S15P which up and failed with it stating that the MOS Temp IS 122C even though it and nothing else is warmer than 15C Actual. I have a replacement B2A24S20P en route by air so it should arrive quite soon now. If you want it, send me a PM with the ship to address.

For anyone following: I purchased my BMS' in a batch, all with sequential serial #'s from Luyuan and Luyuan is HONOURING the warranties. Even though UpNorthAndPersonal & I edited the manuals for Hankzor to help everything they are not forthcoming to us with anything other than deflections or just plain ignoring us.

I can no longer with conscience advise anyone to use JK's in their current state of stoopid and am quite upset that I spent so much time on their damned manual.

 

[upnorthandpersonal]

If you'd like I can also send you a JK B2A24S15P which up and failed with it stating that the MOS Temp IS 122C even though it and nothing else is warmer than 15C Actual.

Yes, please. I'll send you a PM.

 

[hwse]

I can no longer with conscience advise anyone to use JK's in their current state of stoopid and am quite upset that I spent so much time on their damned manual.

Is this in regard to just Hankzor or all JK suppliers?

 

[RCinFLA]

Is this in regard to just Hankzor or all JK suppliers?

Whomever the source, I am sure they farm out assembly to cheapest assembly service provider. Next month's production may be from a different assembly house. Quality control is not a major priority. Production cost and volume is.

If I was to buy more JK active balancer BMS's I would take the cover plates off and resolder all the jumper wire reflowed solder joints and check all the shunt resistors and MOSFET's source lead solder joints for sufficient solder volume.

 

[hwse]

Whomever the source, I am sure they farm out assembly to cheapest assembly service provider. Next month's production may be from a different assembly house. Quality control is not a major priority. Production cost and volume is.

wouldn't this statement be true of all BMS manufacturers?

 

[Steve_S]

Is this in regard to just Hankzor or all JK suppliers?

I've given up on several suppliers. Amy Wan of Luyuan Tech supplies the vast majority of my gear that I buy for my solar stuff and they have always been excellent. Fact is that they sell the JK's at a much more respectable price too. I helped Nami @ Hankzor because Nami was here and interacting with us to only be fooled by them. I'm working with Luyuan again with regards to BMS' etc and will be testing & "beating" on some products to find qualifiers which may actually end up replacing my fleet of BMS' for the "Last Time".

 

[rhino]

replacing my fleet of BMS' for the "Last Time".

Yes, this is getting ridiculous. Getting to the point where you really want an "open source" BMS so people can see exactly what is going on and can even fix stuff themselves if they know enough. Almost makes you want to try https://github.com/stuartpittaway/diyBMSv4

 

[-g-g-g-]

Hopefully I can do some more analysis

Power to the inverter died last night. This morning, the faulty BMS was back in panic mode.
The process to reactive it, which I hadn't taken note of the last times I got it back online, is to cut the main I/O mcb and partially detach the balance lead, push the start button and snap the balance lead back. Then it initialises and presumably goes into normal working mode.
The battery is still below the inverter recovery point but it shows as charging and DC power present. So the BMS is working to some degree. I couldn't be bothered metering it for V & A. It's a bit too damp and chilly to be playing with large surface areas of heatsink in a narrow cupboard.
I still suspect the balance circuit as that's why it is reporting wrong cell counts & that sounds very much like Horsefly's issue as well.
Anyway, just a bit of information for your imminent investigation.
I still haven't ordered a replacement yet. It will take a few weeks to arrive anyway so hopefully I'll get one organised by the time we reach a bit more daylight hours and can investigate what's going on with more detail. And no worries about not having something to replace it with.

 

[jay_jay]

Hello everyone,

I am seeing similar issues with my b2a8s20p (hw version 9) BMS. Whenever the temperature goes below 5°C (41°F) the BMS misbehaves. It goes into panic Mode, that means dot dot dash and repeat. I checked all the connections and double triple quadruple checked the balance connector and cables. No luck there. The first few times I thought I got it back working by disconnecting P- and balance cable, let it sit for a few minutes, reconnect everything and apply +5v via P-/B-. That worked but now I am convinced it was only coincidental. Since then I figured out that it boots up on its own as soon as the temperature in the solar shed starts rising above roughly 5°C. That means if I just let it panic it'll resolve itself eventually as soon as temperature goes up.
Also, I noticed that when the temperatures go near this point the BMS sometimes goes into panic when I connect using the android app. Getting data via esp using BLE I haven't seen this.
[edit] This happens only on the smartphone using app version 4.9.7. When I use the tablet which has 4.6.5 this hasn't happened so far. So probably a software bug not directly related to the issue at hand [/edit]
Also, in the first few months the current reading was pretty accurate on average. But since it is colder, sometimes going below 10°C, the current sensor acts up. If I calibrate at a relatively high current, e.g. 60A, it is accurate in the range 30-90A but it gets grossly inaccurate outside that range. Same thing happens if I calibrate to a low current, it stays accurate on low current but gets extremely inaccurate at high current.
Please note the power on counter. I guess I let the BMS sit for a few hours trying to start at least a couple of times.
I have a second battery with b2a8s20p (hw version 11 without heating) that works flawlessly so far.
So now what? I wrote hankzor beginning of January. So far no reply, now it's Chinese new year and nothing is to be expected at least until beginning of February. After reading some posts here I don't really expect an answer, let alone a replacement BMS. We'll see. I did order a new b2a8s20p without heating. So far it hasn't been shipped and now after reading some posts here I am considering canceling and getting another BMS instead. Maybe JBD. I don't know yet. What do you guys recommend instead of jk?
Cheers and sorry for the wall of text. Have a good one everyone

 

[hwse]

Yes, please. I'll send you a PM.

Have you gotten hold of one of the failed 4s JK's to see what is the failing component?

 

[upnorthandpersonal]

Have you gotten hold of one of the failed 4s JK's to see what is the failing component?

One of them arrived today at the post office - I need to go pick it up still. I'll try to make time to start my investigation this week.

 

[RCinFLA]

FYI, there has been a few reports of defective small six pin SOT buck switchers located below the balancing boost switcher between uC and MOSFET's. On 20S model, there are two small six pin SOT buck switchers, 8S model has three small switching regulators, close together with small chip inductors.

 

[upnorthandpersonal]

@RCinFLA Thanks - I read about that as well. One of the goals will be to find out what they changed there. I have some suspicion that this is due to component availability issues, and someone not taking time to validate the new component choice.

 

[hwse]

One of them arrived today at the post office - I need to go pick it up still. I'll try to make time to start my investigation this week.

Any update on the testing of the 4s failures.

 

[upnorthandpersonal]

I've done some basic testing on it - lack of time has prevented me from doing more these past few days.
There is an issue with the BMS (and the charger) buzzing when the BMS is charging a battery. I do suspect some bad solder, so I'm going to re-solder the MOSFET connections. After that, I'll get more concrete testing done.