• State TI Thinks Resolved
  • Locked Locked
  • Replies 17 replies
  • Answers 4 answers
  • Subscribers 32 subscribers
  • Views 1260 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LM96163: PWM output duty cycle not remains stable

Nimesh Rana
Intellectual 430 points
Part Number: LM96163
Other Parts Discussed in Thread: LM94, LM96000, LM93

Hello,

We are using LM96163CISD/NOPB IC for controlling the Fan speed using the output signal of PWM. We are setting PWM frequency using I2C command, and measuring PWM signal at the output pin of PWM (#5) using an oscilloscope. We are observing the fluctuation in duty cycle at output of PWM for example, for set duty cycle of 50%, we are getting 46 - 49% duty cycle variation measured on CRO.  

We have checked HW connection, there are no any changes compared to EVM. Is there any SW register setting from which we can make PWM duty cycle stable? Please suggest.

I am attaching here a snapshot of Schematic for your reference. Please help to resolve this issue.

Regards,

Nimesh

  • 0
    TI__Expert 6820 points

    Hi Nimesh,

    Can you please send me your register settings for 0x45-0x4D?  We do have a spec for the Fan PWM output Frequency Accuracy of ±7% so there may be some variance in the PWM output.

    Thanks,

  • 0 in reply to David Vaseliou
    Intellectual 430 points

    Hello David,

    Thanks for the quick reply. Please find a below-attached image of Register dump from that you can find the required register's values. 

    Regards,

    Nimesh 

  • 0 in reply to Nimesh Rana
    TI__Expert 6820 points

    Hi Nimesh,

    I want to clarify what you mean by 46-49% duty cycle variation?  Do you mean:

    1. When you set PWM to 50% do you see a duty cycle the ranges around 46%-49%?

    2. When you set the PWM to 50% you see variation that is 46%-49%, meaning the duty cycle extends beyond the device's Fan PWM output Frequency Accuracy of ±7% spec?

    Thanks,

  • 0 in reply to David Vaseliou
    Intellectual 430 points

    Hello David,

    46-49% duty cycle variation means when we set PWM to 50%, the duty cycle ranges around 46% to 49%. (Point 1).

    Regards,

    Nimesh Rana

  • 0 in reply to Nimesh Rana
    TI__Expert 6820 points

    Nimesh,

    Thank you for the information.  From the settings it looks like you are running at a frequency of 22500 Hz, but the PWM value (0x4C) is set to 0%.  Can you send me oscilloscope waveform with the Duty Cycle variation when 0x4C is set to 50%?   

  • 0 in reply to David Vaseliou
    Intellectual 430 points

    Hello David,

    Attached videos show the behavior of below 50% duty-cycle. Apart from that few more questions as below:

    • Can we enable smoothing function on directly drive a PWM input of fan?
    • Can we use more than one remote temperature sensor to use LUT fan control algorithm approach on lm96163?

    Waiting for your prompt reply. 

    Regards,

    Nimesh 

    Click here to play this videoClick here to play this video

  • 0 in reply to Nimesh Rana
    TI__Expert 6820 points

    Hi Nimesh,

    I do not see the variation in duty cycle on my set up using the same register settings (except 0x4C is set to a non-zero value).  Is this waveform taken at the PWM output pin?  Also is there anything external connected to the PWM pin such as a fan PWM input or transistor?

    Thanks,

  • 0 in reply to David Vaseliou
    Intellectual 430 points

    Hello David,

    We have BLDC driver IC to drive Fan, but for the above measurement we have separated that IC to identify which IC creating issue and found at PWM output of LM96163 IC itself having this type of issue. In measurement setup we are only configuring LM96163 IC over I2C and measured  PWM signals at output of IC (5th Pin). 

    Just to confim, when you are saying you are not obersrving variation are you checking it by stopping waveform on CRO? If yes than please check in Run mode only you can see the variation in duty cycle.We have checked in external signall generator and not observed this type of behaviors. 

    Regards,

    Nimesh 

  • 0 in reply to Nimesh Rana
    TI__Expert 6820 points

    Nimesh,

    I am letting the oscilloscope run and and not see the variation.  Can you see if the power supply to the device is noisy?  Do you see this on different boards? 

    Thanks,

  • 0 in reply to David Vaseliou
    Intellectual 430 points

    Hello David,

    We have provided supply for LM96163 IC from PMIC LDO. So it may not be noisy.

    We have tested multiple boards for this and observing same behavior.

    We are waiting for reply for some queries requested in earlier conversations, can you please check and update?

    Regards,

    Nimesh

  • 0 in reply to Nimesh Rana
    TI__Expert 6820 points

    Hi Nimesh,

    I was able to replicate the variation of the PWM duty cycle. This variation is a result of using High resolution mode (register 0x45 bit:4 =1).  For more consistent PWM use standard set register 0x45 bit:4 =0.  

     Also below is my feedback on your other questions:

    • Can we enable smoothing function on directly drive a PWM input of fan?
      • Smoothing ramp rate control function can be used on PWM input of a fan.  
    • Can we use more than one remote temperature sensor to use LUT fan control algorithm approach on lm96163?
      • LM96163 only has one remote channel.  If you need a multi-channel fan controller please see LM96000, LM93 or LM94. 

    Thanks,

  • 0 in reply to David Vaseliou
    Intellectual 430 points

    Hello David,

    As per suggestion,we are getting stable duty cycle on 6.25% PWM resolution [register 0x45 , bit 4=0] with 22.5KHz frequency.
     
    We have below query, hope you can answer it.
    1. What can be the causes of variation in the duty cycle on the high resolution [register 0x45, bit 4=1]?
    2. How can we get more consistent PWM on high resolution also as it needed for the smooth driving of Fan?
    3. How can we define the range of duty cycle variation? (Any configuration OR computation?)

    Regards,

    Nimesh 

  • 0 in reply to Nimesh Rana
    TI__Expert 6820 points

    Hi Nimesh,

    We do not spec the duty cycle% (DC%) of the PWM so the variation may vary form device to device, but from my observations on a single device at room temperature I have seen that variation is less on values that are divisible by 6.25% (6.25%, 12.5%,18.75%, 25%...).  I also observed that the average of the PWM will be very close to the set PWM DC%.  When running a fan, the small variation of the duty cycle at the 22.5Khz frequency averaging out to the desired PWM DC% is virtually the same as a PWM at a consistent  DC%.  

    Thanks,

  • 0 in reply to David Vaseliou
    Prodigy 30 points

    Hi David

    May I jump in on that topic. I work together with Nimesh on the same project.

    David Vaseliou said:

     When running a fan, the small variation of the duty cycle at the 22.5Khz frequency averaging out to the desired PWM DC% is virtually the same as a PWM at a consistent  DC%.  

    I am not sure if I understand your statement correctly. I see you are observing the same behaviour. But is it what we have to expected? 

    When setting a duty-cycle, one would expect exactly that duty-cycle on the output and not something that a fluctuating dc that only in average matches the set value. 

    Where in the datasheet would that to be found?

    Thanks

    Roland

  • 0 in reply to Roland Bieg
    TI__Expert 6820 points

    Hi Roland,

    This is the expected behavior in High Resolution Mode.  The duty cycle will modulate between two different steps and average over a period of time to get the higher resolution selection.

    If you want a more stable Duty Cycle please do not use High Resolution mode.

    Thanks,

  • 0 in reply to David Vaseliou
    Prodigy 30 points

    Hi David,

    Thanks for your fast reply.

    I am bewildered. Where would I find a description of this behaviour in the datasheet? I think this might be worth mentioning.

    Regards

    Roland

  • 0 in reply to Roland Bieg
    TI__Expert 6820 points

    Hi Roland,

    This behavior is not described in the datasheet.  In the next revision of the datasheet we will see if we can add a mention of this behavior.  

    Thanks,