Author Topic: Problem with BQ29700 battery protection IC [SOLVED]  (Read 3558 times)

0 Members and 1 Guest are viewing this topic.

Offline UnixonTopic starter

  • Frequent Contributor
  • **
  • Posts: 400
Problem with BQ29700 battery protection IC [SOLVED]
« on: July 31, 2019, 09:11:44 am »
I have a design powered from a single 18650 battery and I've been asked to integrate a battery protection circuit in the design itself so that a user could make use of unprotected 18650 cells which are easier to put in the holder because these cells are a bit shorter.

Without a battery protection, the power circuit looked like this:
[(+) bat. terminal]---[P-FET switch]---[DCDC+Load]---[(-) bat. terminal]

After inserting BQ29700 circuit the new power circuit looks like this:
[(+) bat. terminal]---[FUSE]---<BAT+=PACK+>---[P-FET switch]---[DCDC+Load]---<GND=PACK->---[BQ29700 Charge N-FET]---[BQ29700 Discharge N-FET]---<BAT->---[(-) bat. terminal]
Power to battery protection is constantly delivered from the node BAT+=PACK+ as soon as a battery is inserted.
Charging circuit based on BQ24090 is also connected to the same node.

The problem is BQ29700 keeps discharge N-FET off (DOUT is always low).
Shorting BAT- to GND=PACK- or connecting charger across PACK+/PACK- as prescribed in the datasheet does not get BQ29700 out of this half-disabled state.
Is there any way to enable protection IC in this setting?
« Last Edit: September 15, 2019, 11:46:45 am by Unixon »
 

Offline Daixiwen

  • Frequent Contributor
  • **
  • Posts: 367
  • Country: no
Re: Problem with BQ29700 battery protection IC
« Reply #1 on: July 31, 2019, 09:25:28 am »
Do you have a schematic? Is everything connected exactly as shown in the datasheet? Is there any load connected to the pack terminals (i.e. is the P Fet on)? What is the cell voltage? Is there any inrush current that would trigger the overcurrent or the short circuit protection?
 

Offline UnixonTopic starter

  • Frequent Contributor
  • **
  • Posts: 400
Re: Problem with BQ29700 battery protection IC
« Reply #2 on: July 31, 2019, 09:46:53 am »
Do you have a schematic?
Yes, but I'm not sure if I'm allowed to publish it. Well, OK, I guess this would be difficult to solve this problem without an actual schematic.

Is everything connected exactly as shown in the datasheet?
I've checked that and I think it is, but maybe I'm missing something over and over...

Is there any load connected to the pack terminals (i.e. is the P Fet on)?
Only DC-DC boosters and a power LED, the entire circuit draws about 50mA.

What is the cell voltage?
I've tried 3.8V-4.2V, no difference.

Is there any inrush current that would trigger the overcurrent or the short circuit protection?
Only capacitance in DC-DC circuits, but this must not trigger protection again after shorting grounds because all caps are charged.
 

Offline Daixiwen

  • Frequent Contributor
  • **
  • Posts: 367
  • Country: no
Re: Problem with BQ29700 battery protection IC
« Reply #3 on: July 31, 2019, 11:15:13 am »
You could have just shown the part of the schematic with the BQ ;)
I'm sorry but I don't see any problem there. You may have to start looking for less obvious problems, like components mounted in the wrong direction, or pin-out/footprint errors (The pinout of the BQ29700 is correct, but maybe the MOSFets?). It's probably a stupid question but did you check that the cell was inserted with the correct polarity? And by the way if this is a user serviceable cell, maybe you should check that your circuit can survive a cell inserted with the wrong polarity.
Did you check that you don't actually have a short circuit on your PCB? Between VBATT and ground?
Unfortunately I can't think of anything else. Maybe start probing around with a scope the pins around the BQ to see if anything is happening right when you connect the cell and/or charger.
 

Offline UnixonTopic starter

  • Frequent Contributor
  • **
  • Posts: 400
Re: Problem with BQ29700 battery protection IC
« Reply #4 on: July 31, 2019, 09:18:34 pm »
You may have to start looking for less obvious problems, like components mounted in the wrong direction, or pin-out/footprint errors (The pinout of the BQ29700 is correct, but maybe the MOSFets?).
No, this is not the case. I will double check alternative MOSFETs (I've built 3 devices for different current thresholds, hence different Rds.on).

It's probably a stupid question but did you check that the cell was inserted with the correct polarity?
And by the way if this is a user serviceable cell, maybe you should check that your circuit can survive a cell inserted with the wrong polarity.
I'm testing with either a smaller LiPo battery, directly soldered with color wires to correct pads of 18650 socket, or just a lab power supply, not an actual 18650 cell.
However, I think I need to put an additional P-FET for polarity protection of BQ29700 (this will be 3rd polarity protection circuit onboard).
Alternatively, I could power BQ29700 from a node after polarity protection P-FET for the load, but this could bring some other problems if BQ29700 is powered off along with the whole device.

Did you check that you don't actually have a short circuit on your PCB? Between VBATT and ground?
Yes, I did. If I short BAT- to GND or connect PSU after the protection circuit the device powers on without issues, all voltages are OK.

Unfortunately I can't think of anything else. Maybe start probing around with a scope the pins around the BQ to see if anything is happening right when you connect the cell and/or charger.
I tried that, but couldn't catch anything significant yet.

My current decision is to put this protection circuit on a separate board with all requisite jumpers and test points and test against different conditions and state transitions.
 

Offline Audioguru again

  • Regular Contributor
  • *
  • Posts: 115
  • Country: ca
Re: Problem with BQ29700 battery protection IC
« Reply #5 on: August 01, 2019, 02:32:40 am »
Why is your schematic a negative image (white and grey lines and text on a black background)? All datasheets and most other schematics are a positive image (black lines and text on a white background). Some of your lines are almost invisible.
 

Offline UnixonTopic starter

  • Frequent Contributor
  • **
  • Posts: 400
Re: Problem with BQ29700 battery protection IC
« Reply #6 on: August 01, 2019, 06:46:51 am »
Why is your schematic a negative image (white and grey lines and text on a black background)? All datasheets and most other schematics are a positive image (black lines and text on a white background). Some of your lines are almost invisible.
Zoom to 50% or 100% and you'll see it is perfectly readable.
« Last Edit: August 01, 2019, 07:12:23 pm by Unixon »
 

Offline UnixonTopic starter

  • Frequent Contributor
  • **
  • Posts: 400
Re: Problem with BQ29700 battery protection IC
« Reply #7 on: August 01, 2019, 07:24:23 pm »
Today I've assembled another protection circuit on a separate board with 1k resistor and a LED for the load and powered it from a lab PSU.
This minimal example of a circuit behaves exactly the same way: DOUT is low, looks like floating at 100-150mV, COUT is at Vbat.
Shorting grounds does nothing but temporarily powering on the load.
I wonder what am I missing here...

BQ29700 datasheet shows optional 5.1M pull-down resistors from gates connected to COUT/DOUT to BAT- and PACK- nets respectively.
Are they truly optional? Well, COUT and DOUT are push-pull outputs and these resistors pull down N-FET gate to close it, not to open, so they have to be optional.
I don't know anymore...

What am I doing wrong? People even sell this stuff... https://www.tindie.com/products/ambergarage/guard18650-protection-for-18650-lipo-batteries-2/
 

Offline Daixiwen

  • Frequent Contributor
  • **
  • Posts: 367
  • Country: no
Re: Problem with BQ29700 battery protection IC
« Reply #8 on: August 02, 2019, 08:07:35 am »
COUT high means at least that the BQ is powered on, but DOUT low shows that it is considering that there is a fault in the discharge path, either a cell voltage under the UVP level or an overcurrent condition (OCD/SCD). I'd suggest probing directly on the BQ pins, to eliminate problems with the other components around the board. Check that the voltage between pins VSS (4) and BAT (5) is over UVP (2.8V for the BQ29700). Then with a scope probe the voltage on pins V- (6) on one channel and DOUT (3) on the other, using VSS (4) as common ground.
First connect the charger. The voltage on V- should become negative, and DOUT should become high. If DOUT stays low, check again the cell voltage, or whether there is a short between DOUT and ground somewhere. If there isn't any short I would start to doubt the BQ29700 itself.
With DOUT high, disconnect the charger and look at the two signals. While DOUT is high, the voltage on V- should always stay within the OCD/SDC window (for the BQ29700, below 0.5V after 250μs and below 0.1V after 20 ms). Of course once DOUT goes low the voltage on V- will go all the way up to BAT anyway. If the voltage on V- stays within the window and DOUT goes low anyway, then again I would suspect the BQ29700. If the voltage on V- exceeds the window then the BQ thinks there is an overcurrent protection. If the actual current on the load is not that high, then have a closer look to the MOSFets. It could be a bad soldering, or that they have a much higher RDSon than expected (wrong reference? fake parts?).

Oh and I almost forgot... before definitely suspecting the BQ29700, you can have a further test while probing with a scope the voltage between the BAT pin (5) and the cell positive terminal. It should always stay negligible. We had a prototype once where they mounted the wrong value for the series resistor on the supply pin, something like 10k instead of 100 ohms. The protection circuit would seem to work at first, but as soon as it tried to draw some current to change the state of a MOSFet its supply voltage would drop and cause all kind of weird problems. If the DOUT output is shorted to ground you would also see there an abnormal voltage drop there, due to high current consumption.

I don't really know what the pull downs would be for.

By the way is it truly a BQ29700, or another one in the BQ297xx family with different thresholds?

If you are suspecting the BQ29700, contact your local FAE, they can be helpful in finding the causes and have access to information and contacts at TI.
 

Offline Daixiwen

  • Frequent Contributor
  • **
  • Posts: 367
  • Country: no
Re: Problem with BQ29700 battery protection IC
« Reply #9 on: August 02, 2019, 08:13:55 am »
Oh and I almost forgot... before definitely suspecting the BQ29700, you can have a further test while probing with a scope the voltage between the BAT pin (5) and the cell positive terminal. It should always stay negligible.
And something else that I forgot... when doing this last test, make sure the scope is isolated, that you are using a differential probe, or that the charger is isolated from earth. If none of those are true you will short circuit the cell through the scope's earth and it won't be good (no it never happened to me  ::) )
 
The following users thanked this post: Unixon

Offline UnixonTopic starter

  • Frequent Contributor
  • **
  • Posts: 400
Re: Problem with BQ29700 battery protection IC
« Reply #10 on: August 02, 2019, 08:31:12 am »
Daixiwen, thanks for such a detailed reply. I'll check for these particular issues.
I didn't get my hands around to do all required testing yet, just confirmed the problem with a lesser circuit.

I don't really know what the pull downs would be for.
The datasheet says they are for discharging MOSFET gates.
My suspicion is that they may have side effects on state transitions because they are referenced to different grounds.

By the way is it truly a BQ29700, or another one in the BQ297xx family with different thresholds?
No, I have only *00 parts. I might need *01-*03 if *00 OCD level will be too low for a particular load, but I don't have them at the moment.

If you are suspecting the BQ29700, contact your local FAE, they can be helpful in finding the causes and have access to information and contacts at TI.
Currently I suspect myself of not understanding something very basic about this IC.
However, I'm thinking of getting other ICs - another batch of 29700, mods 01-07 of it, alternative ICs like S8261 and DW01A.

I don't know... the circuit is so simple that there's just no place to be wrong. And yet something is wrong.
 

Offline Daixiwen

  • Frequent Contributor
  • **
  • Posts: 367
  • Country: no
Re: Problem with BQ29700 battery protection IC
« Reply #11 on: August 02, 2019, 09:01:47 am »
I tend to avoid TI battery management chips because there are often gotchas in the specs that make them hard to use. But I agree that for a one cell protection IC there isn't a lot that can go wrong.
The S8261 is a good choice. Very robust. About 15 years ago we had a battery with several of them and our customer wanted us to run a RS105 test on it (EMP pulse, 50 kV/m). The circuit was unshielded and I was sure the test would fry everything, but the Seiko chips worked fine during and after the test.... indestructible!
I'm more familiar with the S8261 than the BQ29700 but the principle should be the same.
 

Offline UnixonTopic starter

  • Frequent Contributor
  • **
  • Posts: 400
Re: Problem with BQ29700 battery protection IC [SOLVED]
« Reply #12 on: September 15, 2019, 12:29:32 pm »
Finally the problem has been solved after trying almost all available alternatives.
Some of boards for BQ29700 failed due to poorly soldered QFN MOSFETs and were fixed by resoldering them after manually tinning pins, some other were fixed by replacing the protection IC itself. These boards were a small upgrade and they were designed straight for production without additional test points which made it difficult to understand what's even happening.

So, although BQ29700 boards are now up and running, eventually we decided against TI's part for two main reasons:

1) Unique package and pinout, incompatible with alternative ICs, low availability of ICs with different overcurrent detection thresholds;
2) BQ29700 starts in locked state and requires manual resetting which makes bad first experience for customers;

Other protection ICs DW01A/S8261/AP9101 all share the same SOT23-6 package and pinout plus many more options available in terms of timings and thresholds.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf