TPSF12C1-Q1: Recommended component values for use in DC/DC systems

Hello John V

Thanks for your question. I will get back to you on this

Thanks and Regards

Naresh

Hello John, 

The connections to be made in DC-DC system is same as that for AC-DC system.

Could you please share the schematic diagram of your two stage EMI filter? also, could you please share the quick start calculator that you used for our review?

Thanks and Regards,

Naresh 

Hi Naresh.

Here is the schematic for the filter: 

I used the quick start calculator found here www.ti.com/.../TPSF12CX-DESIGN-CALC

Thanks for sharing the schematic John,

Yes, thanks for using that quick start. But could you share with me your version of the quick start? so that I would understand your requirement - what do you mean by poor performance and adding noise to the system?

Thanks and regards,

Naresh

Hi Naresh,

I've attached the calculator with the values I've been using. We do not have grid-side Y-capacitors in our filter, so I estimated a low value for that.

We are testing the conducted emissions of the system, and when we compare the performance with the TPSF12C1 on/off, we see that the TPSF12C1 is only adding noise to our results.

TPSF12Cx-Q1_quickstart_calculator_B1 (2).xlsm

Hello John,

Thanks for your message. I will look into it to the question and get back to you

Thanks and Regards,

Hello John

I looked at the quick start calculator. the values look like a good start. 

When you say TPSF12C1 is adding noise, what you mean by that? Are you capturing EMI spectrum? can you please share it with us for review?

I would also recommend plotting SENSE and INJ pin waveforms of the IC for better understanding the source of the problem

Also, the following link can help you give the overview on how to go about the debug process:

(+) [FAQ] TPSF12C1 and TPSF12C3 power-supply filter IC FAQs - Power management forum - Power management - TI E2E support forums

Thanks and regards,

Hi Naresh,

Sorry for the delay. I was busy with other work but was able to redo my setup and try the TPSF12C1 again.

Using the values in the quickstart calculator I attached earlier, and attaching the IC to the filter I provided as shown in that same quickstart calculator. I see very little difference in the noise on my filter across the frequency spectrum we are interested in. The picture below shows the results with the filter powered (in purple) over the results with the filter not powered (in green). As you can see, there is little impact on the performance.


I also measured the waveforms on the pins as you asked. If these reveal anything about what may be working/not working properly in our system, please let me know.

INJ:

SENSE1:

SENSE2:

Thanks

Hello John,

Thanks for your question. I will look into it and get back to you

Thanks and Regards,

Naresh

For this device, the sense pin amplitude cannot exist more than 100mV.

Also, I made some modifications to the quick start calculator to make it stable. Increased the sense capacitors to 680pF. Can you please try with this?

Thanks and Regards,

NareshTPSF12Cx-Q1_quickstart_calculator_B1_latest_e2e.xlsm

Hi Naresh,

Thanks for the suggestion.

This seems to have made the noise worse, and the inject pin voltage has gone up. Do you have any suggestions that may help stabilize this?

Inject pin:

Sense1:

Sense2:

Hi John,

I received your question. I will get back to you on this

Thanks and Regards,

Hello John, 

The INJ pin is clearly saturating and we don't expect the device to operate in normal operation when the INJ pin is saturating. 

The second point in the following link can help you on how to reduce the saturation. 

[FAQ] TPSF12C1 and TPSF12C3 power-supply filter IC FAQs - Power management forum - Power management - TI E2E support forums

Thanks and Regards,

Hi Naresh,

Thanks for the advice. I adjusted the inject capacitor value higher (to 100nF) as advised by the link you sent. Now I see a 6.6V DC on INJ pin, with no consistent AC on it. I also see no signal or DC level on SENSE1/2. This configuration is having no affect on our noise profile, like the first set of data I sent.

At an Inject capacitor of 47nF, I also had no affect on the noise profile, but was seeing waveforms on inject and sense1, again, like the first set of data I sent.

What should I expect to see when probing these pins in a system that is working and configured properly? 

Hello John,

I received your question. I will get back to you

Hi John,

Inject capacitor cannot be increased to that high value. It cannot be beyond 22nF in the system. You should probably use other knobs to fix the saturation - maybe regulator side Y-caps. Could you please try this and share us the plots? we would be happy to recommend based on how the latest plots looks like.

Thanks and Regards,

Naresh

Hi Naresh,

I was able to test on a different DC/DC converter with more noise using the same filter topology. Here is what I found when setting the Inject capacitor to 22nF and varying the Y-caps

For all these graphs, the results with the filter on are in pink, and the filter off is in green.

Y-Caps: 47nF

Y-Caps: 10nF

Y-Caps: removed

So clearly I should have some sort of Y-caps, but varying the value seemed to not make a difference and I am still seeing no significant performance gain. Is there any other knobs I can turn here?

Thanks,

John V.

Thanks John, I will get back to you on this

Thanks and Regards,

naresh

Hi John, 

Could you please share the latest version of the quick start calculator that you have with all the values that you are using?

We can first confirm if the system is stable in the calculator with all the changes. 

Thanks and regards,

naresh

Hi Naresh,

Here is my quick start calculator. I see with my configuration here that my phase is reaching -180 past 200kHz, so I seem to at least have that in check?

Thanks

John V

TPSF12Cx-Q1_quickstart_calculator_B1 6-26-25.xlsm

Thanks John for sharing the quick start calculator. I will look into it and get back to you.

Thanks and Regards,

naresh