TPS552892-Q1: Schematic and layout review

Hi Jeffery,

Thanks for reaching out.

For the schematic, recommend to change C21 and C22 to 33pF to reduce the RC snubber loss and decrease the temp on R14 and R15. Others look ok for me.

But the layout looks poor especially for power trace routing, the Vin and Vout traces are too thin and there is no other copper plane and vias to help thermal dissipation. And in 2.2MHz switching frequency, the temp will be higher than 400kHz operation.

Does the customer accept to change the switching frequency to 400kHz? And it is strongly recommended to modify the layout as the attached layout guideline suggested.

0842.TPS55289 Layout Guide line.pdf

BRs,

Bryce

Hi Bryce,

Thanks for your kind response. I think that it is no problem that they change Fsw from 2.2MHz to 400kHz. Let me check your feedback with customer, and then I will let you know if they have the additional questions about this issue.

Regards,

Hi Jeffrey,

Thanks for feedback, the another concern is the layout is poor, and recommend to modify the layout as the attached layout guideline suggested.

BRs,

Bryce

Hi Bryce,

I recommended them to modify their layout as layout guideline you sent me, so they will modify it and verify to change Fsw from 2.2MHz to lower frequency. I think that it will be improved if your recommendation is applied on their board. Thanks for your support.

Regards,

Hi Bryce,

They are verifying to change switching frequency from 2.2MHz to around 500kHz, but there is audible noise issue on inductor and output capacitor of their board. Is there any point to improve or resolve this issue?

Regards,

Hi Jeffery,

Can you help probe the waveforms of SW1, SW2, Vout and inductor current at one page?

It is more recommended to select 400kHz switching frequency due to AM band and EMI test, and recommend to select 4.7uH inductor with saturation current > 8A. 

BRs,

Bryce

Hi Bryce,

Let me check the requested waveform with customer and let you know next post. And customer asked me about audible noise of 35.77kHz on Vout pin like attached waveform below. Do you know the source of 35.77kHz in this device? 

AC ripple waveform on Vout pin.pdf

Regards,

Hi Jeffery,

Do you have more waveforms including Vin, SW1, SW2 and inductor current? 

I am doubting about the loop compensation, what is the inductor value and the switching frequency on the current board?

What is the part number of the output capacitors (C27-C31)?

Recommend to change the inductor to 4.7uH with 400kHz switching frequency, and change C1005 to 5.2k.

BRs,

Bryce

Hi Brryce,

Please refer to my answers below.

Do you have more waveforms including Vin, SW1, SW2 and inductor current? They are checking it, and I will send you it if I get it from customer.

I am doubting about the loop compensation, what is the inductor value and the switching frequency on the current board? Inductor value is 2.2uH and switching frequency is 400kHz.

What is the part number of the output capacitors (C27-C31)? C27 is CL21A475KBQ, and C28 to C31 are CL31B106KBH.

Regards,

Hi Jeffrey,

With 400kHz switching frequency, it is recommended to select 4.7uH for stable loop.

And recommend to change R17=5k, C25=10nF first to see if any improvements.

BRs,

Bryce

Hi Brryce,

Please check the waveform you requested and the additional waveforms like an attachment below.

5672.Waveforms.pdf

Regards,

Hi Jeffrey,

From the first picture, it did shows unstable of unsuitable compensation parameters as we assumed.

But what is the signal of the below two charts? Can the customer probe SW1, SW2, inductor current and Vout at one page as well?

BRs,

Bryce

Hi Brryce,

They modified the compensation parameters as you comment and finally got the stable waveform like an attachment below. Are there any problems on modified parameters? Please check it and let me know.

Waveforms_Rev.pdf

Regards,

Hi Jeffrey,

The current waveform still shows the inductor is oscillating, recommend to change R17 to 5k and see if it improves.

BRs,

Bryce

Hi Bryce,

Please refer to the attached waveforms you requested. And I swapped C25 value with C26 value due to miscommunication with customer, so C25=10pF and C26=10nF are correct on previous waveforms(waveform_rev.pdf).

Waveforms_Rev2.pdf

Regards,

Hi Jeffrey,

The waveforms on Rev2 still show unstable loop. Please double check the capacitance on COMP pin, and the recommended compensation parameters are R25=5k, C25=10nF, C26=100pF. Not the values in the Rev2 showed.

BRs,

Bryce

Hi Bryce,

I attached the waveforms you requested. Please check it below.

Waveforms_Rev3.pdf

Regards,

Hi Jeffrey,

The result still shows the inductor current is oscillating, is the output current constant or sine wave of 50kHz?

And why Vout is higher than SW2? In Buck mode, Vout should be SW2-Iout*Rdson(Boost high side MOSFET), can you double check the component parameters on the customer board? 

BRs,

Bryce

Hi Bryce,

Customer verified their board and TI EVM, and it seems that LED driver operation effects on TPS552892. Also there is power good issue. Please check the attachment below and let me know the reason why.

TPS552892 Issues.pdf

Regards,

Hi Jeffrey,

Due to bandwidth limit, I will reply to you later today. 

Regards,

Mulin

Hi Jeffery,

Can you provide updated sch and layout file? When connecting LED load, is there any capacitor in V-LED net? This has large influence in loop stability.

Regards,

Mulin

Hi Mulin,

Please check the updated schematic and layout file below. There are four capacitors(22uF/25V, CL32B226KA) in V-LED net.

Schematic_TPS552892_Rev.pdf

Layout_Rev.pdf

Regards,

Hi Jeffery,

R7 C9 is connected to VCC cap GND directly but C10 is connected to an AGND layer. This AGND panel only connect to AGND pin with two vias, this will cause large parasitic parameter. Please try to connect C10 to the VCC cap GND directly as below red line and see if oscillating is better.

About the PG issue, I find that via is directly in the R5 pad. This might cause instability. Try to reinforce the welding here to see if it is better.

Regards,

Mulin

Hi Mulin,

There was a debugging error on PG issue, so now this issue is cleared. And let me check with customer about C10 connection. Customer said that it seems LED driver operation effects on TPS552892 like an attachment below. Is it right?

8535.TPS552892 Issues.pdf

Regards,