TPS63011: What happens if FB is not connected? Seeing odd behavior.

Scott Brenneman

Part Number: TPS63011

A handful of PCBAs in a recent batch are exhibiting problems with the 3.3 V output from a TPS63011-based circuit. I've attached a screenshot of the schematic. We're suspecting a bare-board fab or soldering problem, but we're trying to get a better understanding. Input voltage is 4 VDC from a bench supply. The EN pin is high. Here are the symptoms:

1. Under no load, the output is 3.5 to 4 V.

2. Switching node looks bad -- noisy and jittery.

3. Under even a light load the output drops. For instance, with 100 ohms load the output is 2.5 V with noticeable ripple.

If the FB pin were to be disconnected or poorly connected, could that explain this?

Is it possible for the IC to damage itself upon first power-up if FB is floating, and thereafter exhibit the above problems?

Thanks,
Scott

over 4 years ago

0 Minqiu Xie over 4 years ago

TI__Mastermind 21565 points

Hi Scott:

May you share the waveform and layout with us?

0 Scott Brenneman over 4 years ago in reply to Minqiu Xie

Genius 3285 points

We can provide those tomorrow. However, this circuit is proven to work in previous builds, so in the meantime, could you please tell us if disconnecting the FB input can cause damage?

Thanks,
Scott

0 Minqiu Xie over 4 years ago in reply to Scott Brenneman

TI__Mastermind 21565 points

Hi Scott:

I don't think it will cause damage, as the TPS63011 has overvoltage protection. But I still need to note, it's not a proper way.

0 Scott Brenneman over 4 years ago in reply to Minqiu Xie

Genius 3285 points

Hi Minqiu,

I agree, it's not a proper condition. What will the output do, if FB is floating? If this were a buck converter, the output would rise, but since this is a buck-boost, I don't know if the output would rise, or if it would fall.

Thanks,

Scott

0 Yichen Xu over 4 years ago in reply to Scott Brenneman

TI__Expert 4760 points

Hi Scott,

The output voltage would rise too if the FB pin is static, but the it would be uncertain if there is a sudden interrupt on the FB pin since it is floating.

Best Regards,

Yichen Xu

0 Scott Brenneman over 4 years ago in reply to Yichen Xu

Genius 3285 points

Thank you for clarifying that.

Scott

0 Scott Brenneman over 4 years ago in reply to Scott Brenneman

Genius 3285 points

You asked for the waveforms and layout -- they are attached.

Thanks,
Scott

0 Yichen Xu over 4 years ago in reply to Scott Brenneman

TI__Expert 4760 points

Hi Scott,

In your schematic, VIN and VINA are connected together which is not good for powering the device. VINA should not connect VIN but need a 0.1uF decoupling capacitor to connect GND. It would be better if you add a 0.1uF capacitor at the output side too. These capacitors should be places next to the device with enough trace width. I didn't find GND pour on your layout, and without big GND pour the device could be easily interrupted.

Best Regards,

Yichen Xu

0 Yichen Xu over 4 years ago in reply to Scott Brenneman

TI__Expert 4760 points

Hi Scott,

What do these waveforms represent? Are they output voltage or input voltage? What about the test conditions? The first two waveforms have a obvious drop which is very strange. And the 3rd waveform seems to work in PSM but only has 2.45V DC. What power supply do you use? We think the voltage drop may be caused by limited power supply, and you can try with a large power supply. And you can also test the device in both buck mode and boost mode to figure out if there are any differences.

Hoping you can provide more information to solve this issue better .

Best Regards,

Yichen Xu

0 Scott Brenneman over 4 years ago in reply to Yichen Xu

Genius 3285 points

Hi Yichen,

Noted on the VINA connection -- we will fix on the next board spin. In response to your other questions:

- There is a full ground plane under the circuit.

- The waveforms are the output.

- Input is a bench power supply with 5 amp capability.

- Both boost and buck mode are not working correctly. Boost mode actually looks worse: more ripple, higher input current, even with no output load.

I want to reiterate that we believe that there's a manufacturing error, so that the actual schematic doesn't look like the one we designed. We think that one or more missed connections under the BGA could be a cause. The X-Rays look good, but those can't always show a missed connection.

Thanks,

Scott

0 Yichen Xu over 4 years ago in reply to Scott Brenneman

TI__Expert 4760 points

Hi Scott,

Thank you for your explanation. If you doubt there is a missed connection, I think multimeter could be used for verification.

Best Regards,

Yichen Xu

0 Scott Brenneman over 4 years ago in reply to Yichen Xu

Genius 3285 points

Hi Yichen,

The problem is that it's a fine-pitch BGA and thus we can't access the connections under there.

Regards,
Scott

0 Yichen Xu over 4 years ago in reply to Scott Brenneman

TI__Expert 4760 points

Hi Scott,

You can remove TPS63011 first and test if there is connection issue on your PCB. If the pad of TPS63011 is well connected to other traces or pours, solder TPS63011 back(make sure it is well soldered) to test if the issue still exists. If does, use a new TPS63011 to exclude device problem. You can check step by step like this.

Best Regards,

Yichen Xu

0 Minqiu Xie over 4 years ago in reply to Yichen Xu

TI__Mastermind 21565 points

HI Scott:

As long time no hear from you, I assume you have solved the issue. If not, just reply below. Thank you!

0 Scott Brenneman over 4 years ago in reply to Minqiu Xie

Genius 3285 points

Oh, sorry -- I didn't see the previous reply. The forum notifications don't seem to be working consistently right now. No, we have not solved the problem, unfortunately. We have removed and replaced ICs, but that doesn't always help. We are now wondering if the VINA1 connection being incorrect is causing a subset of the ICs to misbehave, along with maybe some soldering problems. We will have to do a PCB revision and try again.

Thanks,

Scott

0 Yichen Xu over 4 years ago in reply to Scott Brenneman

TI__Expert 4760 points

Hi Scott,

You can check your PCB connection like my last reply. If you need a PCB revision you can use our EVM layout as reference and separate VIN and VINA. Hope to hear good news from you~

Best Regards,

Yichen

Power management

Power management forum

TPS63011: What happens if FB is not connected? Seeing odd behavior.