diy solar

diy solar

Failed JK BMS

Very cool interesting stuff going on here! I am in the process of swapping BMS's Chargery to this JK 4s one that I have had for 5 or so months. Now I am a little weary of installing it but really have no choice.

Thank you all for what you are doing and if something is resolved I might have to do the upgrade / correction to mine.
 
Schematics on the 4s-8s bms looks like typical for RY3820E. Particularly there is Cff capacitor parallel ro R1 (top resistance in divider) as well as two output capacitors in parallel.
 
nvbt said:
Schematics on the 4s-8s bms looks like typical for RY3820E. Particularly there is Cff capacitor parallel ro R1 (top resistance in divider) as well as two output capacitors in parallel.
Click to expand...

Yeah, I also ordered the RY3820E - just in case. I still have to isolate the resistors to make sure which one it really is, but I'm doing that once I have the components here.
 
I just ordered 4 B2A24S20P, do we know if this is also an issue on the 24s models? I want to swap my chargery out as they don't balance worth beans and their current goes wildly out of calibration too often for my liking. The one thing I will miss is the contactor based on and off.
 
So, I replaced the converters with the RY3820E based on the resistor values and other aspects. The good news is that the BMS works, it can handle the current, no more buzzing in BMS or charge controller and no more boot loops. However, the BMS measurement values in the app are now all wrong, no more current measurement, and no control over the MOSFETs. Putting the old chips back doesn't solve things. So something got fried somewhere; I'm looking into it. However the output voltage from the converters seems to be perfectly fine: one of them is the 3.3V that also turns on the BMS for instance, and this works perfectly.
 
Ok, couldn't leave it be. I went back to my original assumption that they had placed a RY3820E on there in the first place, instead of RY3820. RY3820 is 600kHz, RY3820E is 500kHz - and that would explain why the BMS had the same behavior when I replaced the chip with an RY3820E. So I put in the STI3470 which I knew was close to the RY3820, 600kHz and all (and because the original RY3820 is pretty impossible to get). Result: no more noise in the BMS nor the charge controller. Output voltages of the converters all fine as well.

I'll run more tests later. Sleep first...
 
I would not think the difference of 500kHz to 600 kHz would make that much difference, considering make tolerance on I.C., unless switcher is being pushed close to its maximum current and/or the chip inductors are getting too close to saturation. The loop stability of the DC-DC switcher goes out the window when the inductors get too close to saturation.

Going to higher switching frequency will reduce the inductor saturation.

I still think the chip inductor's quality is root cause of issue.

Lot to lot make tolerance on switcher I.C. frequency is probably +/-20%. Of course, the RY3820 Chinese spec only specifies typical value. +/-20% make tolerance is typical on other manufacturers similar part designs. It will also vary a bit vs. temperature. The external component selection has to take these variations into account for a solid design.
 
Last edited:
RCinFLA said:
I would not think the difference of 500kHz to 600 kHz would make that much difference, considering make tolerance on I.C., unless switcher is being pushed close to its maximum current and/or the chip inductors are getting too close to saturation. The loop stability of the DC-DC switcher goes out the window when the inductors get too close to saturation.

Going to higher switching frequency will reduce the inductor saturation.

I still think the chip inductor's quality is root cause of issue.

Lot to lot make tolerance on switcher I.C. frequency is probably +/-20%. Of course, the RY3820 Chinese spec only specifies typical value. +/-20% make tolerance is typical on other manufacturers similar part designs. It will also vary a bit vs. temperature. The external component selection has to take these variations into account for a solid design.
Click to expand...

I was of the same opinion, but that was the only real difference between the two. I don't have a non-E version to try, so I can't be sure and the markings on the chips don't tell me anything. So my initial thinking was someone screwed up somewhere and put the E version (seems the non-E version is not available anywhere). It could be something else with that chip, but it seems solved with the STI3470. The only real difference I see is the frequency, so that's what I'm going with until I have some time to dig deeper.
 
So, I've had it running for about two hours now. No noise in the BMS or charge controller. All seems good. I'll try some higher power loads later, but it seems to be fixed.
 
I am curious where the Vin for buck DC-DC switcher comes from since the Vin max spec on part is 16v. Hopefully not taken directly from 4/8S battery voltage.
 
Last edited:
It sounds like @upnorthandpersonal may be getting closer to figuring out what was wrong with the stock JK BMS that I sent him. The obvious question: How do we get JK to recognize and fix it? Having experienced a failure (I started this thread months ago), I'm more than a little hesitant to order another JK BMS until and unless I know they have a new hardware version that corrects the issue.
 
RCinFLA said:
I am curious where the Vin for buck DC-DC switcher comes from since the Vin max spec on part is 16v. Hopefully not taken directly from 4/8S battery voltage.
Click to expand...

Just checked: they have a boost regulator before this, a CN5120, that provides 12V.

Edit: datasheet attached.
 

Attachments

  • 1809140219_ShangHai-Consonance-Elec-CN5120_C128237.pdf
    321.2 KB · Views: 9
You must log in or register to reply here.

Similar threads

R
JK Inverter BMS & Victron
Replies
8
Views
384
robot_zombie
R
7
Which JK 24v BMS?
2
Replies
37
Views
2K
upnorthandpersonal
upnorthandpersonal
AlpineJoe
Show me your JK BMS mounting solutions
2 3
Replies
54
Views
2K
brb58
B
F
Single JK BMS Active Balance two 4S Packs
Replies
0
Views
138
fastknight
F
B
JK or HKIVI BMS
Replies
0
Views
152
bitbricked
B
Back
Top