BQ78PL114 under UN/DOT testing

Hi Randy,

Sorry to hear you are having problems wit you UN/DOT Tests.

The PL114 records and store the failure history history. You access this this information by going into the Command tab in bqWizard and View Safety History.

Could you also let us know what is involved in the UN/DOT Tests? What are tests criterion? what is the stimulus and the expected results.

Thanks,

Ben

Ben,

Storage and Transportation testing such as UN/DOT  and the similar IEC62133 test (for CE,and Europe), is something everyone who is using these BM chips for building Li-Ion, polymer packs will have to eventually deal with.

Safety history is somewhat helpful but shows only min and max values, and, the A/D has a problem with rushing to show "2000mV, 5000mV, N/A" whenever there is the slightest problem with over/under ranging with just one channel or VLAN problem (which can have a multitude of sources).

Regards,

Randy

Randy,

i got the pack through the UN-Test. Also not the first time ;-)

Biggest problem has been thermal tests:

• The BQ78pl114 balancing does not stop, if low- or high-temperature inhibits normal charge or discharge

Charge or discharge is not allowed for Li-Ion out of the moderate thermal window. At low temperature of the UN-Test  -40°C (also equal to -40°F) the cells have high impedance and charging or discharging at frozen electrolyte is strictly forbidden. Nevertheless the BQ78pl114 pumps what is equal to charging and discharging. Because of the high impedance, one pump cycle turns the voltage level of the higher cell and lower cell to the opposite. So pumping never ends -> discharge of cells due to pumping, killing cells due to charging/discharging of cells at low temp (also at high temp) -> lowered pack voltage at end of test or fuse -> UN-test failed -> 6000$ crunched.

Solution -> disable pumping during UN-test -> you have to disable pumping also during shipment -> enable after shipment -> how to do this? Could be challenging! :-))

• The BQ78pl114 has its pin-out so that hi cell voltage pins are close to low cell voltage or temp.-sensors

Humidity during thermal cycling results into completely wrong cell-voltage or temperature readings -> fuse -> UN-test failed -> another  6000$ crunched.

Solution -> insulation of PCB with polyurethane (no wrong readings due to humidity any more)

• The BQ78pl114 has an internal temp.-sensor (board.-temp) which is strongly inaccurate. Beside the calibration point, the error is > 20K at low temp.

Solution -> ignore all interlan board.-temp readings

Take the challenge, give the UN-Test one more try. No risk, no fun. With the BQ78pl114 you decided to have fun!

And don't forget: the condition you have during UN-Test, you must give during shipment.

Regards and good luck!

Erroneous pumping and Board-temperature-reading (cell temperatures are ok) at low temperature:

Erroneous Board-temperature-reading at mid-to-high temperature (cell temperatures are ok):

Erroneous board-temperature-reading (cell temperatures are ok) at low temperature (-20°C):

Andrew S

Thanks for the information.

I would certainly recommend doing your own 50x cycling with balancing on and then shut off the balancing.  UN/DOT is not a performance oriented test, but a safety test.

Make sure you use fresh cells.  I removed passive fuses (needed for UL, but, not UN/DOT). 

Although turning off the cycling, using only UN/DOT related protection for the testing phase or moving out temp trip levels just for testing,  would be considered questionable for some.  But they are not paying the $6000 a pop,and, let's not forget the $20000 for the packs that are not re-usable.

My question is still the same.  The archiille's heel of the test.  To avoid the 0V output reading after a test that needs a software reset because of a protection that kicked in that caused a 2nd level alarm.  I believe I turned off every 2nd level protection that I could find.  Yet when I read the returned packs, I had in the status the red "Fuse" on.  Yet on one pack there were no alarms set.

I still think TI should have a setup file that is validation friendly.

Randy 

Randy,

i already passed this kind of tests.

A safety test has to prove the safety during use and shipment, not only for the time of testing.

Enabled balancing with the BQ78pl114 is a strong risk, you will not pass the tests. It does not make sense, to disable balancing for testing and enabling this for the production.

So balancing should be disabled, as soon as charge or discharge is inhibited due to thermal thresholds. But this is not possible within the BQ and should be corrected in the firmware of the device or upcoming devices by TI.

The reason for your missed test could be the sensitivity to humidity. Humidity inside the pack may condense and provoke wrong and transient current- or voltage readings. This triggers 2nd level protection. Try to cool down your packs, keep an hour, warm up, shake them and look, if condensed water provoked second-level safety. The pin-out of the BQ is highly sensitive to this issue.

Hi Andreas,

Nice posting and very important, thank you.

Do you know if this problem (not disabling powerpump when cells are outside acceptable temperature range) is related to all or only some of the ICs bq78PL114, bq78PL114S12, bq78PL116, bq76PL102?

Have you had any response from TI on this issue yet?

Hi PerBear,

all of the BQ-devices have this problem of power-pumping outside of the allowed temperature range.

I asked my TI-support for help by E-mail and within this E2E forum.

If you like to see the response of TI, please read my former post of Thu, Oct 28 2010:

http://e2e.ti.com/support/power_management/battery_management/f/180/t/70967.aspx

... and an other answer, to control the BQ' s system control bit 15 (set/clear) for pumping off/on.

The problem is as follows: after production of the battery, the pack will be stored, shipped and maybe stored again with balancing disabled to avoid dangerous pumping during shipment at temperatures outside allowed thermal window. Within your end-application-device, the pumping must be enabled by clearing the system bit 15. If the battery pack is removed from the end-application-device, the balancing stays wrongly enabled also during shipment. Also it is a problem to store the battery without balancing, the BQ-chip draws some current of the lowest two cells, this causes strong disbalance after a view month of storage.

So you could add to the battery a baby-sitting-device (low power uC) for proper function. This device could poll the temperature and manage the set/clear of BQ's system control bit 15. Or you could wait for the firmware-correction of TI ;-)

Andreas

The PL116 can be placed in the Ship Mode for shipping.

Please check the TRM Page 67 (http://www.ti.com/lit/ug/sluu481/sluu481.pdf) for the Ship Mode entrance and exit conditions.

The Cell balancing is controlled by the following (page 85):

Discharge Completion Voltage Qualifier: Set in units of mV. The lower cut off for cell balancing is

governed by the following equation. It must be true for PowerPump to occur.

(Lowest Cell – Discharge Completion Voltage Qualifier / Series Cell Count) > Minimum Differential for

Balancing. There is no upper cell voltage cutoff for PowerPump.

Ben

Dear Benjamin,

BQ78PL114 exits ship mode after a couple of minutes by firmware bug, without any reason.

But also BQ78PL116 activates balancing during ship mode (see sluu481, page 67):

Exit criteria: There are two ways that the device can exit ship mode: permanent and temporary.

Permanent Exit criteria:
The permanent exit is intended to be.....

Temporary Exit from Ship Mode:
The PowerLAN Master Gateway controller will perform the following tasks on a periodic ...
(b) PowerPump Cell Balancing is performed if necessary. A cell imbalance will keep the device in
Active mode until the cells return to balance as determined by the Minimum Cell Differential for
Balancing parameter value.

And as result, power pumping will occur during ship-mode and shipping also at cold conditions and therefore the bq78pl116 will destroy the Li-Ion cell if electrolyte is cold or frozen.

Cells have high impedance if cold or frozen, so pumping  never stops (see http://e2e.ti.com/support/power_management/battery_management/f/180/t/70967.aspx).

• It should be possible, to couple the active balancing to the thermal window, where charge and discharge is allowed!

• I checked the EVM of BQ78PL114. Not only at my battery-packs, but olso at the EVM the strong error of the internal temperature sensor of BQ78PL114 is seen. Is this issue corrected in PL116?

Best regards

Andreas

Andrew S,

I haven't had this problem with balancing  because it was physically disabled for the test.   I did propose to do the 50x cell cycling at home. To turn off balancing for UN/DOT and then use it in production.)   I cannot fathon not having cell balancing in this fragile "114/102/102/102/102" environment. 

But, I did find the source of my original problem and unless someone shows me that I am doing something wrong,   it is important.  ( I may enter this topic as a new post:)

I turned every protection algorithm off that I didn't need for the UN/DOT test.  I kept UV, OC and OV , ON.  OV, ON for the over-charging test.  This OV is a safety 2nd level protection and will produce a permanent failure flag and a software reset will be needed to clear.  Not too much of a problem if the testing facility is notified to (and allows Soft resets) to pass the 0V fail-the-test condition.

Here is the rub.  Packs failed, but, before the charge test.  I have no indication that there ever was an Over-Volt condition.  But I do have an answer with screen captures to show.  My problem was that I had real cell UV condition that I got from cell group #2 and cell group #12 that were weakened by constantly being run down from storage.

Yet this shouldn't have cause Permanent failure.  UV?  What gives?  I propose this.  it is an "Under-ranged" multiplexer to the A/D converter causing the neighboring cell group channel to show an opposite high value in the reading.  When investigating the actual cell voltages, I find the UV cell to actually be undervoltage state, but the OV cell is actuall normal and shouldn't have measured the high value.  I would get a 2000mv red showing for the low cell and after a small charge to remover this "under-range" display the cell would read the actual voltage. The OV causing cell (red "5000mV") would, after the UV cell corrected, showed the correct cell value even though this cell group never was above the 4.35V OV trip level.

I've seen this time and time again working with this 114 chip.  "2000mV and "5000mV'.  Sometimes all cells would show this even though most cells were within bounds.  I don't think there is much tolerance for cells to be unbalanced and do cause A/D reporting problems.

Coupled with the 2nd cell voltage depression problems that you raised (I do really like to know more about this since it plaques our design as well) and a calling to turn off the cell balancing shows that there is a fragil environment you need to sail through.

Randy

Hi Randy,

I know this effect.

if cell a voltage is going low, voltage reading of adjacent cell b is erroneous (too high).

If balancing is enabled, this adjacent cell b is discharged erroneously into cell a until cell b voltage is close to zero and cell is destroyed.

So keep all cells well balanced all the time and the voltage readings are ok.

But take care:

take a look to the specification, where the recommended operating condition Vsup is (2.3...4.5)V for the cells 1...4.

Two rows below within the table, the recommended cell voltge difference Vin is (0...4.5)V, this is a little mistake in specification and should be ignored.

Also the electrical DC-characteristic is specified with Vcell of  (2.3...4.5)V, where all cells are balanced!

And paremetrize the bq78pl114 to shut down if Vcell < 2.5V.

And do not forget: voltage measurement is more or less correctly done, if cells are within (2.5...4.5)V and this within (0 ... 60)°C!

Altogether: this bq78pl114 is not specified to perform correctly out of the consumer range. UN-Tests are definitively out of consumer range!

"...there is a fragil environment you need to sail through..."  UN-Test environment is not a set-up to snuggle the battery-packs ;-)

UN-Test is performed to discard battery-designs, that are not able to survive the harsh conditions of shipment. I think, a battery fitted with the bq78pl114 is not the most untroublesome candidate to pass this test.

Regards

Andreas

Thank you.

The balancing algorhitm would have to be set to balance at all times.  If balancing circuit was set for during charge only and the UV limit was set low, if then cell 2 drops too low  (from powering VLD01 reg) and causes cell1 reading to climb over the COV limit, you then have a PF permanent failure that requires a software reset.  If the customer doesn't have a SMB connection to the back, he is out of luck.  This is what has happened to us.

Regrads,

Randy

"...UN-Test is performed to discard battery-designs, that are not able to survive the harsh conditions of shipment. I think, a battery fitted with the bq78pl114 is not the most untroublesome candidate to pass this test."

It seems the bq78PL114/116 has a bug that not only prevents the chip from passing the UN test:

Charging li-ion cells outside their safe temperature window is bad for the cells even when the charging is caused by active balancing. Outdoor equipment including battery tools, pedelecs and ebikes somehow need to prevent the TI chip from damaging cells by balancing when battery is operated or stored in freezing cold environment - or what we normally call winter here in Norway ;-)

It is my view that this problem can not be fixed by manually turning off balancing completely before shipping as this will not prevent damage during usage, only during shipping. This problem can only be fixed by firmware update that implement powerpump shutdown when one or more cells are outside safe temperature for charging.

It would be great to have a quick feedback on this issue from TI.

- Perbear

It seems we can never design a pack with COTS protection without still having to add a  micro-conttroller because some customer wants something different.  We have several TI designs with BQ78PL900 and BQ29330 and BQ20Z45 that are controlled by a SiLabs micro-controller.  The controller can reconfigure the pack on the fly by monitoring the cell temp and then shut off cell balancing.  (but let me check that for sure).

Randy

Perbear, Randy

i agree: without external controller, the bq78pl114 can not be used outside of buildings more or less closer than 45° of latitude to the poles. Also it is not allowed to ship the battery packs, because UN-Test normally will not be passed if you test with same parameters as you will use the pack.

So: the bq78pl114 without firmware correction (attaching balancing to the thermal window where dis-/charging is allowed) or without active (low power) controller that cares to the bq78..., the battery pack is for tinker work on your laboratory bench only.

Andreas

Well, what about the option to use the "Charge only" balance method?  Like the passive balancing circuits it would not come into effect unless customer decides to charge.  But, then that would raise my original complaint that we have to keep a "12S" system closely balanced otherwise this "Human Centipede" stacked BQ78pl102 chips will misbehave when reporting voltages and temperatures.

Does using stacked ladder method of controlling the cell stack require using an active cell balancing?  Does that mean that other manufacturers like O2 and Maxim must have bigger problems monitoring and controlling their stacks since they use the shunting method?.

Randy