TPS54560: Inverting Power Supply High Frequency Ripple

You might consider an RC snubber from PH to IC GND (-Vout).  See the enclosed:

0535.RC Snubber.doc

Mert,

I agree with John.

-Sam

Hi John,

Thank you for your suggestion. I will try to find parasitic capacitance and taking measurements from directly SW pin of TPS54560. But I could not be sure about the waveform on SW pin. Could you please check? It is OK?

Above image is measured from SW pin at 3A load.

Below image is representing high frequency switching noise. Its period is 6.8ns and I will try to increases period 13.6ns to find the parasitic capacitance.

I do not directly support TPS54560, but that SW node waveform does not look right.  How are you measuring it?  WRT GND or -Vout?

We are measuring with respect to -Vout. But I have take measurements at wrong points.

Please find schematics below. Now measurement points are PH(SW) and GND (Which is -Vout) pins. Please find new measurements below. I think they are OK to proceed. Now period is 88ns.

Hi John,

Thank you for "tip and barrel" advice. According your provided procedure, I have found Cs and Rs values. But when I done power dissipation calculation it is 3.8W and it is impossible to put a power resistor on my circuit. Do you think 3.8W is normal?

My measured Cs is 2.35nF (2x4.7nf Series) then Cp is 0.78nF, Lp is 250nH and Rs is 17,86Ohm (10 + 4.99 Series). Vds is 75V and f is 282kHz (which is relatively low). Because of Vds power dissipation is very high.

According to that calculations I think using a RC snubber is impossible for TPS54560 circuit. Also I have tried RC snubber(2.35nF and 14.99 Ohm) for very short time (to avoid resistors burned) but I could not get any improvement:(

Is it possible to make any improvement with compensation circuit? Rzero, Czero and Cpole?

Problem may be related with layout. My -18V Inverting layout is same with the +18V provided by datasheet. Should I make any change in layout for using TPS54560 as inverting power supply? Provided +18V example layout is given below. Only Input Bypass capacitor is connected to Vout since Vout is GND for inverting supply.

Another issue that I could not understand is switching frequency. For -18V inverting supply I could not get 3A at frequencies above 290kHz. Do you have any idea?

Best Regards,

Hi John,

I am suspecting about layout and trying keep circulating current loops small, but I could not get any improvement:(

What about Cd value in my schematic posted before. It is recommended as 1uF, but I am wondering if changing this value has an improvement effect?

My layout and schematic is below. Some values on schematic has changed and I add Cd between Vin and GND pins. I have tried connectinf C36 and Cout GDN pins with cable there is ver small improvement but it is not enough. Do you have any further suggestion?

Hi Again John,

I have made a new layout for testing regarding to notes below.

https://e2e.ti.com/blogs_/b/powerhouse/archive/2016/09/28/laying-out-an-inverting-buck-boost-converter-for-success

http://www.ti.com/lit/an/snva722b/snva722b.pdf

http://www.ti.com/lit/ug/slvu243/slvu243.pdf

Please find new schematic and layout below. Do you have any suggestion or comment with new layout? I will connect C34 with a cable between vias? For test card only since only one layer pcb we can manufacture.

Any comment from someone? I do not want to make this layout bad again:(

Thank you Sam.

According to your valuable comments, I have updated layout. Is it OK?

C40 is new 1206 ceramic cap value will be choosen acording to tests.

Best Regards,

Mert,

For D6 I was thinking more like the picture below (forgive the low quality edit. Powerpoint is my Photoshop).  Other than that, I think it looks good.  I'll ask John for his final comments.

-Sam

Thank you Sam. I agree it is better now.

Hi,

First of all, thank you for your suggestions.

I have tested the above layour today. The result is about 3Vpp High Frequenct Ripple. Then I add additional input and output bypass caps, Rboost and RC snubber according to following articles.

/cfs-file/__key/communityserver-discussions-components-files/196/AN045_5F00_Reducing-EMI-in-buck-converters.pdf

/cfs-file/__key/communityserver-discussions-components-files/196/SLVA219A_5F00_TI_5F00_Minimizing-Inverting-Converter-HF-Switching-Noise.pdf

The best result that I can had is below. Is it acceptable for TPS54560 or Inverting Circuits? What can I do further? Should I consider to change TPS54560?  My customer has not any requirement for HF ripple value, but they are unsatisfied with the result. How can I convince them?

I have too many questions, but I am really stucked and I do not know what should I do. Best Regards.