CBA 2 disaster - big spark frying computer!

I wanted to test the capacity of the worst affected cell in my Vectux today.

First I charged the entire battery as usual with the stock built in charger; then I topped up the weak cell 103-003 through the Manual-BMS, at 3A, putting in just over 6Ah until the voltage reached 1.472V during charging and 1.45V when the charger pauses for measurement. The cell was nicely filled.

Then I connected the CBA2 to the cell through the M-BMS. The banana plugs were already in the correct positions in the M-BMS and the other end of the cable was connected to the CBA2.

In the middle of the cable is a Deans connector, which was still apart.

When I tried to plug the Deans connector in, there was a loud spark, and there might have been a noise from the CBA2 or the laptop, not sure.
The negative pole of the Deans connector got burnt;
the laptop showed a blank, white screen and the external monitor which was connected was still showing the previous picture of the CBA2 software screen.
The laptop did not respond to anything other than turning it off with the main power button. It does not boot any more, and the external monitor does not seem to receive a signal from it, either.

The polarity was correct.

Has anyone got an idea what happened there?

I think I have figured out what went wrong!

It has to do with the Auxiliary Battery Cooling (ABCool) system I built. I changed the design from a relay to a diode OR-ing cable - that is behind the zapping of my laptop!
See: http://endless-sphere.com/forums/viewtopic.php?f=14&t=8794&p=200096#p200096
I built this:

(All the diodes shown as Zener diodes are actually Schottki diodes; I could not find a proper Schottki diode symbol.)


And it works well, until you connect a CBA II to the M-BMS to find out how the capacity is holding up.

At that instant it destroyed my computer!

Here you can see the schematic for the M-BMS: http://visforvoltage.org/sites/default/files/Vectux%20M-BMS%20Switch%20Array%202009-05-10.pdf
http://visforvoltage.org/sites/default/files/Vectux%20M-BMS%202009-05-10.pdf

The old ABCool design with a relay instead of diode OR-ing would have prevented the disaster!
Like this:



.

For reasons which I do not fully understand the diode OR-ing device causes a DC voltage between Earth and each single battery cell tab in the M-BMS!

And as I know now, USB cables are earthed. So when I connected the CBA II to the battery tabs 3 and 4, a 42V DC shock went through the CBA II to my laptop. I don't know yet if it has also fried the CBA II on the way through. I'm a bit apprehensive about plugging it into my last remaining computer to try it out!

The about 40V DC between tabs and Earth occur as soon as the plug of the auxiliary power supply is plugged into a power outlet, even if the outlet is turned off. Turning the power point off disconnects the Active pole, I think. So it must be due to Neutral or Earth making contact.

Unplugging the connector going to the impellers does not change this, the effect happens because of the negative cable from ABCool to stock (charger) electronics. As soon as I connect just the negative pole of the auxiliary power supply to the diode OR-ing cable, the 40V DC appear between tabs and Earth. It does not matter if the auxiliary power supply is connected through the box of diodes and resistors, or straight to the cable with the OR-ing diodes.

Isolation + Computer = Good.

I don't really do it though... but, I typically have several layers of "protection" (i.e., a few components to sacrifice to the magic smoke gods).

Mr. Mik said:

And as I know now, USB cables are earthed. So when I connected the CBA II to the battery tabs 3 and 4, a 42V DC shock went through the CBA II to my laptop. I don't know yet if it has also fried the CBA II on the way through. I'm a bit apprehensive about plugging it into my last remaining computer to try it out!

Click to expand...

Ouch! Though it's awful that your laptop got fried, I'm glad you found out what happened.

Measure the resistance between the two battery leads of your CBA. If the resistance is anything less than a few megohms, you'll need to replace the MOSFET. Take off the red bottom "case" to see if any of the other components have fried too. Occasionally the op-amp driving the MOSFET also blows.

Hmm...wait a sec. The 42V went thru the USB port so I guess just about anything that shares the +5V USB line (or ground) in the CBA could have gotten fried. Unfortunately, that's everything in the CBA. I might be overreacting but I wouldn't plug it into any computer without an isolating USB hub between them.

CamLight said:

Mr. Mik said:

And as I know now, USB cables are earthed. So when I connected the CBA II to the battery tabs 3 and 4, a 42V DC shock went through the CBA II to my laptop. I don't know yet if it has also fried the CBA II on the way through. I'm a bit apprehensive about plugging it into my last remaining computer to try it out!

Click to expand...

Ouch! Though it's awful that your laptop got fried, I'm glad you found out what happened.

Measure the resistance between the two battery leads of your CBA. If the resistance is anything less than a few megohms, you'll need to replace the MOSFET. Take off the red bottom "case" to see if any of the other components have fried too. Occasionally the op-amp driving the MOSFET also blows.

Hmm...wait a sec. The 42V went thru the USB port so I guess just about anything that shares the +5V USB line (or ground) in the CBA could have gotten fried. Unfortunately, that's everything in the CBA. I might be overreacting but I wouldn't plug it into any computer without an isolating USB hub between them.

Click to expand...

Thanks!
I inspected the CBA II PCB, both side look OK. That does not necessarily men anything, of course!
The resistance between pos and neg CBA II lead is about 7kOhm in one direction, and 5kOhm in the other polarity.

And yes (!), definitely an isolated USB hub is a must from now on.

At least I was smart enough to buy the opto-isolated version of the CAN to USB adapter, although officially it is not needed.

But when you are experimenting, it is necessary!

Does anyone know if there is a chance that the laptop might be repairable? May it have a fuse against idiotic attacks like mine?
What part would fry first, how do you check it out?

Sounds like the FET is OK as the 7K/5K measurements are close to my two CBA's 7.1K/7.3K readings. I was wrong about the MOSFET megohm resistance in my earlier as that only applies to the FET when not in a circuit, and only in one direction.

If the laptop won't boot, and it took 42V through the USB port, the power supply might have fried. Hopefully, there's a fuse. I'm worried about where the popping sound you heard came from though. Still worth exploring with a meter!

The problem seems to have a name: "Ground Loop"
http://en.wikipedia.org/wiki/Ground_loop_%28electricity%29

Here is one USB isolator I found so far:
https://iftools.com/isousb/index.en.php?gclid=CLzKjNPxs50CFYctpAodRC7fig

I wonder if the CBAII will run OK with this.

Because I sometimes get power outages, I would really like it if I can run battery tests reliably from a laptop, even when the power drops out during the test. For example, when checking self-discharge rate on a battery after a 2 week wait (or even much longer), it would be really annoying it the power drops out and the test is ruined!

Has anyone here some experience with a CBA II used through an isolated USB connection?

How much current does the CBA II draw from the USB cable? The above isolator can only provide 250mA without auxiliary power supply, 500mA with it.

.

ADDED: Here is another one:
http://www.bb-elec.com/product.asp?sku=uisohub4

The CBA should be OK with any isolated USB hub but I don't know how much power it draws. Probably not much as there's only a microprocessor, op-amp, a few discrete components and a LED. That 250mA unit should work but if you need more, any isolated switcher supply could be the auxiliary supply and bring you up to 500mA.

CamLight said:

The CBA should be OK with any isolated USB hub but I don't know how much power it draws. Probably not much as there's only a microprocessor, op-amp, a few discrete components and a LED. That 250mA unit should work but if you need more, any isolated switcher supply could be the auxiliary supply and bring you up to 500mA.

Click to expand...

That's one thing I don't get yet: The laptop was connected to the grid via a switchmode power supply. Does that not isolate it against a ground fault?

Regarding power consumption by the CBA II - don't forget the cooling fan!

If an auxiliary supply is absolutely needed, then I might have to rig up a UPS. I'l probably get a small inverter and a solar panel, and turn some of the NHW10 batteries into a very nice UPS!

Nice project!

Mr. Mik said:

That's one thing I don't get yet: The laptop was connected to the grid via a switchmode power supply. Does that not isolate it against a ground fault?

Regarding power consumption by the CBA II - don't forget the cooling fan!

If an auxiliary supply is absolutely needed, then I might have to rig up a UPS. I'l probably get a small inverter and a solar panel, and turn some of the NHW10 batteries into a very nice UPS!

Nice project!

Click to expand...

Hmm...not sure what happened if you were using a switch-mode. Perhaps a shared 3rd wire ground? To be honest I have not looked closely enough at your setup to know what might have happened regarding grounds and current flow.

Doh...I forgot about the CBA's fan. IIRC, it draws about 150mA so, I think, you'll be OK with a 250mA hub. Plan for adding an auxiliary supply, just in case though.

I had a similar incident in the recent past. I was testing a 280 cell Nicad pack. I was using the CBA II to discharge a group of 20 cells (pack is organized as 14 groups of 20 cells) when I needed to unsolder a connection between a different group. Because of the high voltage (336 to 400v), I'm usually really careful around the cells (wear lineman 600v gloves, etc) but this time I was in a hurry. I did not realize the CBA II grounds through the USB cable (although now that I think about it, it is obvious) so when I touched the hot soldering tip to the pack, I got a big spark (the soldering iron is grounded!). The nicad cells are capable of melting battery tabs when shorted. I was lucky. The pack design puts a 16amp 600v ultra fast blow fuse at the mid point of the pack. I was also lucky in that I was grounding one side of the pack and touching the other side of the pack so the current had to travel through the fuse (50/50 chance there). The fuse gave up its life, saved the pack and the CBA II. No damage other than the fuse. I was able to easily replace the fuse.

I learned multiple things from this event: if anything is attached to your pack (and the voltage is >60v) it might ground it and if you are using anything else that is grounded, the potential for shock/death or starting a fire/destroying things is large. Use mid pack fuses of the correct type and size, isolate all equipment if possible (buy an isolated USB hub), wear proper protection (gloves if high voltage), and try to only work on one thing at a time (don't discharge one part of a pack while doing something with another. Measure potential voltage differences before connecting anything (if it is not what you expect, figure out why before proceeding).

Even at a lower voltage, if the current capability is high, a direct short can melt things, start fires, etc.

Lesson learned but not as painful as it could have been.

webfootguy said:

I had a similar incident in the recent past. I was testing a 280 cell Nicad pack. I was using the CBA II to discharge a group of 20 cells (pack is organized as 14 groups of 20 cells) when I needed to unsolder a connection between a different group. Because of the high voltage (336 to 400v), I'm usually really careful around the cells (wear lineman 600v gloves, etc) but this time I was in a hurry. I did not realize the CBA II grounds through the USB cable (although now that I think about it, it is obvious) so when I touched the hot soldering tip to the pack, I got a big spark (the soldering iron is grounded!). The nicad cells are capable of melting battery tabs when shorted. I was lucky. The pack design puts a 16amp 600v ultra fast blow fuse at the mid point of the pack. I was also lucky in that I was grounding one side of the pack and touching the other side of the pack so the current had to travel through the fuse (50/50 chance there). The fuse gave up its life, saved the pack and the CBA II. No damage other than the fuse. I was able to easily replace the fuse.

I learned multiple things from this event: if anything is attached to your pack (and the voltage is >60v) it might ground it and if you are using anything else that is grounded, the potential for shock/death or starting a fire/destroying things is large. Use mid pack fuses of the correct type and size, isolate all equipment if possible (buy an isolated USB hub), wear proper protection (gloves if high voltage), and try to only work on one thing at a time (don't discharge one part of a pack while doing something with another. Measure potential voltage differences before connecting anything (if it is not what you expect, figure out why before proceeding).

Even at a lower voltage, if the current capability is high, a direct short can melt things, start fires, etc.

Lesson learned but not as painful as it could have been.

Click to expand...

Good points.
The pack voltage was 138V when my ground loop happened. Lucky I did not fry the Vectrix charger (at least I think it's still OK, time will tell).

I measured the ground loop with an analog meter as well, and it barely shows up. It might "just" be a capacitor that gave the spark on the negative battery lead (it was tab 4 of 103 tabs).

I think I also had the RT808D charger connected to the same leads with the ABCool connected the same way as later the CBA II, but the charger did not get zapped.

It's the second time these resistorless tabs (1-15 and 28) have given me trouble. All other tabs have 15kOhm resistors in line, all tabs have 20A fuses. The 16 tabs without resistors have an additional 10A fuse at the end where I can get to it easily.
The first time my charger got fried, but that was my own stupidity; I did not understand that a Buck charger is not isolated and tried to charge part of the battery string with energy from the same battery string.

I opened up the fried laptop to have a look ...

Not much of a chance for me to fix it; the Ground connection fom the USB cable is connected to multiple parts of the laptop - practically anything could have fried. I cannot see a fuse, but I don't think it would be able to fuse the laptop against this sort of event.

The stock Compaq power supply has continuity between the ground input pole (100-240VAC) and the negative 18.5VDC output pole. From the 18.5V input jack at the laptop the negative is directly connected to the "negative busbar" (not sure if that is the right term). The USB ground is also directly connected to this negative busbar.

So much for switchmode power supplies being isolated! HAHA!

Then I tested a bunch of different switchmode power supplies (SMPS) which I have lying around, in use or left over.

There are several types:

Type 1: Has no ground connection on the input side, just Active and Neutral. There is no continuity between Ground and negative or positive output when the device is switched off (put plugged into the power outlet). When the power is turned on, then the DMM beeps with (I believe) 50Hz frequency when measuring continuity between ground and pos output or neg output.


Type 2: Has Active, Neutral and Ground at the input side. There is no continuity between ground and negative or positive output, either turned on or turned off. My old IBM Thinkpad PS behaves like this.


Type 3: Has Active, Neutral and Ground at the input side. There is no continuity between ground and neg or pos output when turned off, but when turned on then there is the same 50Hz beeping as with type 1.
I believe this particular SMPS might be faulty, because I have four identical SMPS's, but the three others all behave like type 4.

Type 4: Has Active, Neutral and Ground at the input side. There is continuity between neg output and ground when switched off, and no continuity between pos output and ground. This remains unchanged if the power is turned on.

.

Two of these are Type 4, one is Type 3:
The fourth one is installed in the Vectux and is also Type 4.

.

All the SMPS's I have looked at have just 2 poles at the output side.

.

I have not tested the particular SMPS which was driving the RT808D charger whilst it was connected to the same cell in the Vectux battery and under the same conditions which caused the spark when I later connected the CBA II. (That is because it is running a 3 x charge/discharge test which I do not want to interrupt). This SMPS has only 2 input poles, but it might be different from Type 1. Maybe it does not cause the 50Hz beep, but remains isolated from ground at all times, like the IBM Type 2.




.

Can anyone explain how this works?

.

Could a Type 2 SMPS be used with a laptop to guard against ground faults?

The power supply that was connected to the RT808D charger is a type 1. I don't understand why it caused no trouble.

I bought anew laptop, Dell, the power supply is the type wich constantly has continuity between negative output and Ground.

When I measured and re-measured, I found that there is also continuity between positive output and Ground on some SMPS's, but only when turned on and only in one direction; probably a diode in the way?

Bad news: The CBA II is fried, like the laptop.

I got an isolated USB connector now, this one: https://iftools.com/isousb/index.en.php?gclid=COLA_ZmN8Z0CFYctpAodxFpdLA

Neither with or without auxiliary power supply does the CBA II get detected or turn it's lights on (when the ISOUSB is used).

Just to be sure I plugged it in "straight" as well: Nothing! It did not cause a computer malfunction.

I guess it means I'll have to get a CBA III....

Awwww, man. I was hoping it was just the MOSFET that blew.
If you do get a CBA III, I'd be very, VERY interested in knowing what the two MOSFETs are.
I'm curious as to whether they just doubled up on the IRL2910 used in the CBA II or if they used a different part number.

Hey Everyone,

wow .. i never thought a cba could blow a laptop lol!!

camlight .. i'm still using those mosfets you set me to fix my cba 2 ... and it still works like a charm!!!!!!

-steveo

CamLight said:

Awwww, man. I was hoping it was just the MOSFET that blew.
If you do get a CBA III, I'd be very, VERY interested in knowing what the two MOSFETs are.
I'm curious as to whether they just doubled up on the IRL2910 used in the CBA II or if they used a different part number.

Click to expand...

I've got one on order!

steveo said:

camlight .. i'm still using those mosfets you set me to fix my cba 2 ... and it still works like a charm!!!!!!

-steveo

Click to expand...

Excellent!
So far, knock on wood, there have been no failures (IPP048N06L MOSFETs) in about two dozen CBA II's.

CamLight said:

Awwww, man. I was hoping it was just the MOSFET that blew.
If you do get a CBA III, I'd be very, VERY interested in knowing what the two MOSFETs are.
I'm curious as to whether they just doubled up on the IRL2910 used in the CBA II or if they used a different part number.

Click to expand...

Looks like they doubled up!

Thanks for posting those photos!
I think I'll run some of thermal equations again to see how much cooler the CBA III runs compared to the CBA II and if the operating temperatures of those two FETs doesn't exceed their rated max at 100W (like it does in the CBA II, max'ing out at about 85W).