This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPS61170: Issue with 20kHz PWM signal

Part Number: TPS61170

Hi

One of our customers is have an issue with the TPS61170.

Customer issue: 

I've taken measurements with the TPS61170 and I don't think this part isn't working as it should.

We have 5 V input voltage, the output resistors are set so that about 24 V output voltage comes out.
I want to reduce this output voltage to 12 V with the PWM. PWM 20 kHz, duty cycle 50%.

Unfortunately, this doesn't work at all as desired. It comes out about 12 V, but usually switches on and off with the PWM frequency of 20 kHz.
Firstly, this leads to an enormous ripple and secondly, the maximum performance is reduced by 50%.
If you do a PMW at 95 kHz, 50% duty cycle, it seems to be working properly.

However, according to DB, a PWM between 5 kHz and 100 kHz should work. What's going on here?

Measurement of the current on the coil with 20kHz PWM, Vin 5V, Vout 12V:



Schematic:





Thanks and Kind Regards
Alex
  • Hi Alex,

    PWM is multiplied by 1.229V and then filtered by an internal low pass filter. It seems maybe this signal still has 20kHz components. Could you let customer test Vout, Voltage of FB pin, SW, and PWM waveforms? Maybe PWM drive signal causes the issue. I can check and try to solve the issue in bench test if similar board can be built.

    Best Regards,

    Eric Yue

  • Hi Alex,

    I tested this issue in bench and found Vout, Iout also have this 20kHz component. It seems internal low pass filter of CTRL pin cannot filter all PWM noises and a ripple of 20kHz is coupled in FB reference voltage. This ripple frequency also changes with PWM frequency, so 95kHz PWM's ripple is smaller because of higher frequency.

    1. Could you please use higher PWM frequency if system is allowed. 20kHz PWM's ripple is actually very large.

    2. If high PWM frequency is not allowed and system's bandwidth and transient response performance have no requirement, you can change COMP parameters to increase attenuation capability to 20kHz component. Decrease Rcomp will decrease this ripple. The compared waveforms of 1k Rcomp and 10k Rcomp is shown below.

    Best Regards,

    Eric Yue

  • Hi Eric,

    Thank you very much for the tests! Please find the reply from the customer below:

    The 20 kHz get only into the system because the booster is turning on and off with the PWM frequency.

    Of Course afterwards this ripple is visible at the output and also the FB. But the root cause of the ripple is the booster not working as specified.

     

    Using a higher PWM frequency does solve the problem, but according to the data sheet am PWM frequency of 20 kHz is perfectly fine.

     

    Pictures of my measurements are attached.

    Best Regards,

    Gokhan Kaya

  • Hi Gokhan,

    If 20kHz PWM component can be completely filtered by integrated low pass filter, IC will work well. However, because this filter isn't designed well,  reference voltage contains this 20kHz ripple and we cannot change filter's parameter since it is integrated inside chip. As you can see, waveforms of 95kHZ also contains 95kHz ripple. From application perspective, we can only increase PWM frequency or decrease loop bandwidth to optimize it because IC cannot perform well as datasheet described.

    Best Regards,

    Eric Yue