TPS53667: Maximum phase number setting

Carmen-san,

Thank you for your reply.

Our customer deigned 4 phase circuit, CSP1 to 4 pins are connected Powerstage IOUT pin and CSP5 is connected to V3R3.

However 0xE4 register value was 0x00. It is 1 phase operation.

Why isn't 0xE4 register set to 0x03? Do you have any idea about cause of this phenomenon?

We checked V3R3 rise up waveform. The rise time is  1.17ms (0V to 3.3V).
Now V3R3 capacitor is 10uF. The rise time became 320us and TPS53667 kept 1phase operation when they changed that capacitor to 1uF.

TPS53667 done 4phase operation when they removed V3R3 capacitor.
EN signal is added several tens ms after V3R3 rise up.

Best Regards,
Kohei Sasaki

Hi Kohei,

I'm looking into why that would be occurring but some more information would be helpful.

1) When the register doesn't read 0x03 is the part stuck in single phase mode or will other phases turn on when the load is increased? 

2) Can you write to this register, change it to 0x03, and have things work normally?

3) I'm a little confused about what happens when the decoupling cap is changed. Can you please correct me if I'm wrong?

- With 10µF the problem occurs and Trise = 1.7ms

- With 1uF the problem occurs and Trise = 320us

- With no cap in place the problem does not occur

Thanks,

Carmen

Carmen-san,

1) Phase number did not increase when the load is increased.

2) We could not write 0x03 to the register. The register value is kept 0x00 eveb though we write 0x03.
    Thus I think larger phase number was ignored because 1 phase setting was done by pin strapping.

3) Your understood is almost correct.

    With 10uF the problem occurs and Trise = 1.17ms
    With 1uF the problem occurs and Trise = 320us
    With no cap in place the problem does not occur. Register value became 0x03.

This problem has influenced their test process. So speedy solution is required.

Best Regards,
Kohei Sasaki

Carmen-san,

I will tell you update.

We checked CSP2 pin voltage waveform with V3R3 capacitor is 10uF. Then we observed that CSP2 pin voltage rise up along with V3R3 for about 860us and fall to 0V.

After that CSP2 voltage rise up to about 1.4V again and fall to 0V.
We are thinking this 2 times voltage rise up causes TPS53667 to recognize it as 1phase setting.

So I would like to know the reason why CSP voltage rises up.
Same one time voltage rise up waveform was observed on the other CSP pins. However 2nd rise up was not observed.

Is first CSP voltage rise up expected operation?

CSP1 to 4 pins are connected to power stage IOUT pin. CSP5 is connected to V3R3. 
12V and 5V power supply are added to TPS53667 and CSD95490 12V is added but 5V VDD is not added in this test.
They use another 5V supply to TPS53667 and CSD95490.

Do you have any idea from these information?

If you send e-mail me, I can share the waveform taken by our customer.
sasaki-k@macnica.co.jp

Best Regards,
Kohei Sasaki

Carmen-san,

Sorry to bother you again.
I have additional update.

The problem did not occur when same 5V supply add to TPS53667 V5 pin and CSD95490 VDD pin.
Should we add same 5V supply to TPS53667 V5 pin and CSD95490 VDD pin?
We could not read from datasheet. Is there a description?

And we would like to know the expected waveform of CSP pins with 4phase connection circuit at power on.

Please e-mail me if you have any question,
sasaki-k@macnica.co.jp


Best Regards,
Kohei Sasaki