Oops - recovering from a dumb mistake: reversed BAT polarity (ribbon cable)

[Ben Low]

Here's something you don't want to do, but if you do, how it might not be the end of the your SBMS0...

To my great humiliation I foolishly connected one of my SBMS0's to a new battery installation with the 12-pin BAT IDC cable completely reversed - i.e. pins 10-12 connected to battery negative and 1+2 to battery positive, and so on (with a 4S battery). Being over-confident, and perhaps a little tired, a cursory "double-check" failed to note the reversal.

On plugging in the SBMS0 there was immediately that smell of the magic smoke escaping, and the SBMS0 was dead.

Connecting the power backwards fused the track from the IDC header socket to the transient voltage suppression diode TVS1 and blew the TVS. Fortunately the PCB track was repairable. I ran a fine strand of wire atop the blown PCB track and was careful to not dislodge the component adjacent to where TVS1 was. Dacian has advised the TVS doesn't need to be replaced as it has done its job (with the unspoken presumption that I won't go and reverse the power connections again...)

Attached are pics of the failed board, and repaired. The repaired SBMS0 appears to be working fine!

 

[Bernard Kobos]

Thanks for sharing. Made the same mistake today (ouch!) and this helped me fixing it in a no-time :)

Cheers!

[Dacian Todea]

Not sure why this happens relatively often as red wire should symbolize positive not negative ? 

I also got an email today from somoanelse that connected wires in reverse. Someone some months ago also has done the same to a new SBMS0 model so I will post the photo here as the new one will fail in a bit of a different place where I made a so called PCB fuse so damage is a bit more minimized compared to olde SBMS0 model.  The TVS diode did not failed in this incident with the new SBMS0 version.

[Oberon Robinson]

With everyone's permission, I'll add these pictures to the Beginner's Guide as a cautionary tale.  And the fix. :)

[Dacian Todea]

Oberon,

I do not think someone that will make this mistake will first read the user manual or your beginner guide :)

But I could be wrong.

[Oberon Robinson]

You could be right. :D  But more information is usually better. :)

[Ben Low]

In my case, I know very well how the device should be connected (i.e. have read the manual :-) ). In haste, I was careless, made a silly mistake and didn't properly check. I'm just pleased the design is such that the consequences were relatively easy to fix; the PCB fuse in the new version is better again.

Oberon: permission for my photos: go for it, all yours.

[Dave McCampbell]

Oberon.  I agree that more detailed info the better, especially on how to fix things that might not be covered elsewhere.  Small pics with details should not take up much room.

[Daniel Fischer]

  Same thing happened to me yesterday. In my case, I used a breakout board https://ebay.us/4ee0IU to a longer (and fused) 18 gauge cable and didn't notice that the numbering of the pins was reversed.... really a stupid mistake, but it was late in the night and I wanted to finish the job ...

I applied the fix to the PCB suggested above (so I just shorted the broken trace), changed the order of the sensing/balancing cables, and ... it happened again (see photo). This time, also the 7 Amp fuses of the 18 gauge cable blew. I also noticed (afterwards) that there was a short between pin 1 and 12. I am sure, I have overlooked something so I'd rather ask before I make more stupid things ... so is there anything else I can try before ordering a new one? I suspect, the diode (red arrow) is broken. Can I just remove it?

[Dacian Todea]

Yes it is possible that the TVS diode is internally shorted so just remove it as it will work without it.

[Daniel Fischer]

Another try. Now the resistances between all the pins are hundreds of kOhm (corresponding to a few mW of power at 26V), except between pins 11 and 12. Is this okay? They are on the same potential anyway...

 

[Daniel Fischer]

It works now. Thanks for the help!

[Dacian Todea]

Glad to here it works. Likely you installed the TVS with reverse polarity.

[Jhon]

"I do not think someone that will make this mistake will first read the user manual or your beginner guide :)"

I read both, wired it correctly 3 times before and the 4th time I still made that mistake.  As Douglas Adams wrote "a common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools"

I tried reversing the diode and the track got hot (I could smell it) when I briefly plugged in the ribbon cable.  Removing the diode all together did not fix the problem for me.  I checked the resistance end to end and the track seems to be complete.   I will patiently wait for the replacement as I still think this BMS is the best solution on the market.

[Jhon]

[Dacian Todea]

John,

Is next to impossible to design something fool proof and attempting to do that will not only increase price will also have some other downsides like maybe reduced accuracy or similar trade-offs for robustness.

It seems this inverted polarity connection always happened with those aluminium cells they have the positive terminal marked with black. Was that also you case ?

Double check your battery connections as there is no reason for the SBMS not to work since trace seems intact and TVS diode is removed.

Why did you needed to wire things up so many times ? Are you testing multiple batteries ?

[Jhon]

Hi Dacian, not blaming you work, I am blaming myself, I knew better and I still made a mistake.

And yes, the positive terminal is marked black on my Chinese made prismatic cells.

First time I wired them in series was to charge the cells, trying to charge new cells in parallel first with a 5Amp charger takes way to long, hooking them up in series first and relying on the BMS to stop the charge when one cell reached the peak is much faster (hours instead of days)I then top balanced them in parallel before reconnecting them in series.  I had one cell that was not behaving, so I top balanced again in parallel before connecting them in series again before complaining with the vendor.

I ordered and got a second set of cells to make a 2P8S pack and I was going to charge those 8 in series before putting all 16 in parallel and then finish the top balance with the 5Amp charger for a day or two.  unfortunately I connected the new set with the wrong polarity.

I checked the ribbon cable connections, the red and the two adjacent wires go to the most positive connection on the pack and each subsequent one to the positive of the adjacent cell, with the last two going to the most negative cell.  I double checked with a volt meter.

When the ribbon cable is plugged in for a minute, the screen feels slightly warm.  I do not see anything on the display and I can't connect via wifi either.  I did notice a set of dip switches under the ribbon cable that connects the wifi module and maybe I knocked one over by mistake?

[Dacian Todea]

It sounds that more things where not connected correctly as it should so likely the microcontroller may be damaged.

This could happen if the current shunts where not connected correctly. Where they connected and how.  Disconnecting the shunts form the battery positive while under load either charger or discharge current will damage the current shunt amplifier and likely the microcontroller.

You can just check if you have 3V between GND and that thicker trace connected to the inductor (the commonent close to one of the corners above the 2x8pin connector where WiFi/USB module is connected) and that DIP switch is not relevant as it only affects the WiFi/USB.  The damaged microcontroller may explain the heat but if it where not for that it will also be possible that you damaged the ribbon cable connecting the LCD when you worked on soldering the wire to repair that trace.

Hard to diagnose a problem without access to the system.

[Jhon]

The current shunts where not connected correctly either.  I literally reversed the battery.  It is ok, I will just wait for the replacement to arrive.

[Jhon]

I ordered and got the SBMS0 without the wifi module, assuming I could simply add the old one, but now I am questioning my decision.  Can any one confirm if I can?

[Dacian Todea]

You mean you have the old model WiFi ?  The 16pin connector will be fine but physically is different due to different shape/size of the new SBMS0 and on top of that the firmware on the new SBMS0 will not handle the old WiFi firmware so WiFi firmware will need to be updated and then it can maybe somehow be connected and work but maybe just hanging in there from the 16pin ribbon cable.

[Jhon]

I just installed the wifi module from the one I damaged when I reversed battery polarity  and it works fine, it even remembered the wifi settings.   Charging the new batteries right now, bu the delta is all over the place, it jumps between 12 mV and 200 mV.  I checked the connections and all seems fine, am I missing something?

[Soggy Paws]

Jhon, are you balancing?  (noted by < symbols next to some cells).  What happens is, when balancing is going on, the high cells get suppressed, and their neighbor cells get increased.  So it makes the delta very big... 100 mv is not unusual.  You have to watch the wifi /legacy page... when one of the balanced cells (denoted by < next to it) goes white (showing it is the highest cell), then you are getting nearly accurate readings.  The Electrodacus by default balances for 3 seconds, then pauses for 3 seconds.  During the pause, the cell voltages return to normal and you can see about what your real delta is.

[Jhon]

Yes BMS is balancing, this is a new batch of cells which I have yet to top balance.  What I do not understand is why the delta is showing as low as 12mV one second and nearly as high as 200mV a second later, it is going up and down every few seconds..  I have checked the connections to make sure they are tight.  (Later tonight they should be close to fully charged and I will finish them off in parallel (with a 5A charger)  I am not sure what could be wrong.

I am charging them in series with solar first, it is a cloudy day (so charging nice and slow) as top balancing in parallel with 5A would take a week.

[Soggy Paws]

The balancing algorithm balances the cells that need balancing for 3 seconds, then off for 3 seconds (to be able to accurately measure the voltage of each cell).  When it is balancing (applying a resistive load to that cell), the highest cell(s) will go low, and an adjacent cell will go high.  So if you are monitoring via the wifi screen, or the Electrodacus screen, you will see some erratic voltages.  (I only ever use the wifi screen, because my eyes can't handle the tiny Electrodacus screen).  My wifi "legacy" screen updates every second.  You can tell whether the balancing is on or off by looking at the legacy screen and if a cell with a < symbol next to it is showing as the lowest cell, then balancing is happening on that cell.  Wait a second or two, and that cell (or another one with < next to it) will go high.  On the legacy screen, lowest cell is yellow, highest cell is white).  When a cell that is being balanced is white, you can believe the delta is close to reality. (I usually see 2 or 3 updates (once a second, approximately) before the high cell starts getting pushed lower by the balancing.  So yes, you see the cells and the deltas changing every few seconds.  That is normal behavior when balancing is going on.  Just watch and observe, and you can tell your true delta.