TPS61032: IC Shorts When Switching on Output to PocketBeagle

Rene Carballo

Part Number: TPS61032
Other Parts Discussed in Thread: BQ2970

Hello,

I am trying to make a board that can power a PocketBeagle SBC at 5V from a 3.7V Li-Po Battery. I have 3 main components: a Charge Controller IC, a Battery Protection IC, and the Buck-Boost Converter IC. Before having soldered on the Boost Converter, I tested the Charge Controller and Battery Protection IC, and they both were working without any issues. I control the output of the Buck-Boost Converter by having a switch that connects the output of the Converter to the load I am trying to power.

In having these components soldered, I first tested the load with a simple resistor and LED combination, and turned the switch on. The output was steady at 5V, and leaving the circuit running for a few minutes yielded no overheating issues. I also flipped the power switch on and off multiple times quickly with no issues. I then connected the output to a 10MOhm load, and repeated a similar process. The output was still steady at 5V. When I made the PocketBeagle the load and turned the switch on, the board turned on, and the output was still steady at 5V, even after many minutes of leaving the system on (no overheating issues either). However, when I turned the switch off and back on again, a voltage spike is seen at both the input and output pin, and smoke is immediately smelled. After removing the IC, I found that the SW and PGND pins had shorted together.

I'm not really sure what other steps to try, so any help would be appreciated. Thanks!

Links:

Relevant portion of schematic (Note: all capacitors used are ceramic)

Relevant portion of board:

Datasheets:

- Battery: www.digikey.com/.../9560998

- Charge Controller: www.digikey.com/.../7065590

- Battery Protection: www.digikey.com/.../5973209

over 5 years ago

0 Minqi Zhao over 5 years ago

TI__Intellectual 1660 points

Hi,

Thanks for your detailed description. For your case, there're two main reasons could cause your voltage spike problem.

1. There maybe a small parasitic inductance between your boost 5V output and your control switch, since maybe you have a long line. And maybe your load is inductive. These parasitic inductances could generate large voltage spike when your control switch is quickly turned on, thus there is a large spike on and lead SW shorted.

2. Your layout need to be improved. The commutation circuit loop is too long and has large parasitic inductances, which may cause the spike as well. Please see suggestion as below. Red line is IC package.

0 Rene Carballo over 5 years ago in reply to Minqi Zhao

Prodigy 70 points

Hi,

Thank you for your quick reply, and sorry for my late one.

Hmm, I see what you are saying. Is there any way that I can verify that this is the problem with my existing board (For example, adding an inductor in parallel to some of the lines to reduce the overall inductance)?

Thanks,

0 Minqi Zhao over 5 years ago in reply to Rene Carballo

TI__Intellectual 1660 points

Hi,

There two ways to verify which one leads to the spike. Please note that to capture the spike, you should use minimum loop test, change your oscilloscope setup to full bandwidth, highest sampling rate and zoom in to see the spike:

1. You can capture the Vout when you turn the switch (between Vout and load PocketBeagle ) off-on in one screenshot to see if there is spike on Vout. If so, then the spike is from parasitic inductance of load;

2. You can capture SW with applying heavy load, to see if SW has large spike when low-side FET turned off while Vout has no large spike. If so, then the spike is from parasitic inductance between GND and Cout since the layout is not so good.

Please let me know if you have any questions.

0 Rene Carballo over 5 years ago in reply to Minqi Zhao

Prodigy 70 points

Hi,

From the tests I mentioned in the first post, I was measuring the voltage across the board ground and the Li-Po (-) (When the battery protection ic detects anomalous behavior, it disconnects a mosfet which normally connects the Li-PO - and board ground)(red) and the voltage on Vout close to the switch (yellow) in an oscilloscope. The spike below appeared immediately after turning on the switch to power the PocketBeagle (note the difference in voltage scale settings on the right) . Does this help to isolate the problem, or am I misunderstanding what you mean?

Thanks,

0 Minqi Zhao over 5 years ago in reply to Rene Carballo

TI__Intellectual 1660 points

Hi,

You need to upload your oscilloscope screenshot by inserting file instead of paste the picture. I can't see your screenshot here.

0 Rene Carballo over 5 years ago in reply to Minqi Zhao

Prodigy 70 points

Hi,

My apologies. Hopefully it should work now.

0 Minqi Zhao over 5 years ago in reply to Rene Carballo

TI__Intellectual 1660 points

Hi,

Can you zoom in to us or even ns level so that I can see the detail of the spike? Thanks.

0 Rene Carballo over 5 years ago in reply to Minqi Zhao

Prodigy 70 points

Hi,

Unfortunately that is a screenshot, and we would have to replicate the test setup in order to produce it again.

Is there anything that can be gleamed using just that screenshot?

Thanks,

0 Zack Liu over 5 years ago in reply to Rene Carballo

TI__Mastermind 34231 points

Hi Rene,

The engineer supporting this device is having Chinese New Year holiday now till Jan.31. Will reply you soon once she is back.

0 Rene Carballo over 5 years ago in reply to Zack Liu

Prodigy 70 points

Hi,

Ok, no worries. Whenever they are available :)

0 Minqi Zhao over 5 years ago in reply to Rene Carballo

TI__Intellectual 1660 points

Hi,

Sorry for my late response. Hope everything is OK.

We should start from boost circuit first. Could you detach BQ2970 and recapture waveforms that I said before? The test reference ground should be boost ground. I'm not sure about whether the board ground you said is boost ground and can't tell in your screenshot before.

Please let me know if any questions.

0 Minqi Zhao over 5 years ago in reply to Minqi Zhao

TI__Intellectual 1660 points

Hi,

Is there any problem? If you don't, I will close this thread.

0 Rene Carballo over 5 years ago in reply to Minqi Zhao

Prodigy 70 points

Hi,

I have not been able to take the screenshots yet, but I will update this post as soon as I can.

Thanks,

Power management

Power management forum

TPS61032: IC Shorts When Switching on Output to PocketBeagle