TPS65131: VNEG not regulated at no-load, but regulated at load

Adam Zahanuddin

Part Number: TPS65131

Tool/software:

I designed my layout similar to as seen in the datasheet and used the recommended parts by TI for +/-10V generation;

The only differences are my inductor model, and at both of the outputs, instead of 4x4.7uF, I used 2x10uF + 2x1uF totalling to 22uF.

Enabling the IC with no load attached causes the negative voltage to fluctuating between -10V to -22V at very high frequency (my oscilloscope can only go as high as 500MSps, so the noise must be higher frequency)

But when I attach a 100mA load (1K resistor over VNEG-GND), VNEG seems to stabilise to -10V

Does it have to do with my low voltage rated capacitors (some are 16V)? Otherwise please enlighten me on this issue.
Thank you.

7 months ago

0 Zhiqian Wu 7 months ago

TI__Mastermind 31585 points

Due to Chinese New Year, our expert will reply you on 2/5

0 Adam Zahanuddin 7 months ago in reply to Zhiqian Wu

Prodigy 50 points

Looking forward for your reply!

0 Steven Li 7 months ago in reply to Adam Zahanuddin

TI__Mastermind 23190 points

Hi Adam,

Thanks for your understanding. Out expert will reply to you later.

Best Regards,

Steven

0 Patrick Yao 7 months ago in reply to Adam Zahanuddin

TI__Mastermind 34820 points

Hi Adam,

Could you please send me your schematic for double check?

Patrick

0 Adam Zahanuddin 7 months ago in reply to Patrick Yao

Prodigy 50 points

Here's the schematic for your review;

Though I just realised that I have mistakenly ordered 22uH inductors over 4.7uH inductor with my manufacturer. Could this be the problem? I have ordered replacement parts of 4.7uH inductors to be replaced with 22uH inductors. I will update you guys then but please let me know if there are any problem with the schematics, thanks!

0 Patrick Yao 7 months ago in reply to Adam Zahanuddin

TI__Mastermind 34820 points

Hi Adam,

Could you please send me the waveform with longer timing base of the oscilloscope? 2ns is too small, at least we should see one complete period of the ripple.

Patrick

0 Adam Zahanuddin 7 months ago in reply to Patrick Yao

Prodigy 50 points

Apologies for the bad oscilloscope reference;
Here I have tested with different load conditions -
Y 10V/div
X 100ms/div

No load: output swings from -9.2V to -29.6V at ~3.3Hz

R 220K: output swings from -9.2V to -29.6V at ~2Hz

R 2K: output stabilizes at -26V

R 200: output stabilizes at -13.6V

R 47: output stabilizes at -8.4V

I hope this helps!

0 Patrick Yao 7 months ago in reply to Adam Zahanuddin

TI__Mastermind 34820 points

Hi Adam,

What is the output capacitor voltage ratings? And how many ICs have you tested? If only tested one IC, I suggest to replace another one to see if the issue still exists.

Patrick

0 Adam Zahanuddin 7 months ago in reply to Patrick Yao

Prodigy 50 points

I'm using MLCCs rated at 16V. I do have other populated boards with the same chip and layout but I am waiting to replace the inductor before turning it on. I have a feeling that my capacitors have been damaged due to the high voltage output.

0 Patrick Yao 7 months ago in reply to Adam Zahanuddin

TI__Mastermind 34820 points

Hi Adam,

The inductor may have some impact, please let me know the result when 4.7uH is replaced.

Patrick

0 Adam Zahanuddin 7 months ago in reply to Patrick Yao

Prodigy 50 points

Hello Patrick,

I have pulled out a new board with the same layout and same chip. I have replaced the 22uH inductors with 4.7uH - no dice! The output waveform is slightly different but still swings to -22V.

Fortunately, I found the solution shortly after playing around with the capacitor configurations -- to which I find that removing the feedforward capacitor C10 (I used GRM1555C1H7R5BA01D) was the culprit. Removing this capacitor made the output waveform at no load to resolve to -10V. I have not tested with any load yet but I am content with the result. Are there any drawbacks for eliminating the capacitor?

Regards,
Adam