• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 63 subscribers
  • Views 436 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TPS92518: PKTH_DAC change with VIN/VLED duty ?

KANJI SHIBATANI
Mastermind 6205 points
Part Number: TPS92518


Hello


My customer finished their design using TPS92518  ( already in mass production ) .

Thank you for your support so far.

Now customer intend to add  smaller ILED product-line for ILED is 300mA at 100% brightness.

( So far, ILED is 1000mA )


They hope to have without H/W design change.

Customer want to have  just change  LEDx_PKTH_DAC register only and tested in their product actually.

Condition:( schematic is also same as EVM)

* LEDx_PKTH_DAC = 102 ( changed from 223 to 102 in order to adjust from 1000mA to 300mA max at 100% brightness.)
* LED_TOFF_DAC = 45 :no change from 1000mA case
* switching frequency is almost 320KHz :same .
* inductor: 330uH TZSH12080-331（DSC):same
* shunt FET dimming is used. So, cap with Inductor ( EVM manual* C18/C17) are removed and 0.1uF are inserted in place of those as EVM.
www.tij.co.jp/.../sluubm0c.pdf


Issue

ILED is changed depend on duty ( 304mA ~ 318mA) if VIN is high.

VIN=55V
ILED=318mA@VF=8V
ILED=309mA@VF=26.5V
ILED=304mA@VF=45V

VIN=24V: no problem
ILED=301mA@VF=8V
ILED=300mA@VF=13.5V
ILED=299mA@VF=19V

-----

On the other hand, when ILED is 1000mA, no problem found.
VIN=55V
ILED=1003mA@VF=8V
ILED=999mA@VF=26.5V
ILED=996mA@VF=45V

VIN=24V
ILED=997mA@VF=8V
IL ED=996mA@VF=13.5V
ILED=996mA@VF=19V

------


Inquiry ( please have your comment)


I think inductor magnetic energy is too much for stable control in case duty is too small/ and ILED is small (300mA) for TPS92518.

To reduce frequency ( bigger LED_TOFF_DAC ) seems to be corretive action, but I am not confident with theoretically.


Customer had also tested and to reduce frequency (320KHz to 250KHz) with LED_TOFF_DAC = 45 to 50 seems corrective action....


Customer want to know the mechanism in theoretically, but would you please advise your opinion about the mecanism ?


We know VCSPx-VCSNx peak current threshold tolerance is good when LEDx_PKTH_DAC.( But seems no relationship to duty..)

Do you recommend to have LEDx_PKTH_DAC = 223 or something ( bigger ) and change shunt resistor?

Best Regards

  • 0
    TI__Intellectual 2590 points

    Shibatani-san,

    First off, the TPS92518 is a hysteretic converter so it is inherently stable all of the time. 

    Because this is hysteretic (peak current sense with off-timer variant), then the converter will operate will near constant ripple over all operating points assuming you have the same hardware components and off-timer.  Only the frequency will vary, therefore in your case....I don't know what you mean by 320kHz....that will only happen at one operating point.  Over your three operating points above, you will see three different switching frequencies at the same off-time DAC value.

    Changing only the peak current sense DAC will just reduce the average for the most part as the ripple would remain the same.  This means that your inductor current ripple as a function of the average current will increase significantly at 300mA (i.e. if it was 10% at 1A...it would be more like 33% at 300mA).  The accuracy of a hysteretic type converter with this off-time approach can decrease slightly when that ripple % increases. 

    However, the most likely reason you are seeing the accuracy get worse is due to the inherent random offset of the current sense comparator.  At a peak threshold DAC value of 223, you are nearly at full scale range and have about 223mV of sense voltage.  When you change the DAC to 102, then you only have 102mV of sense voltage.  Therefore, if your random offset on a given part is 4mV (for example), then you would see a roughly 2% tolerance at 223mV sense and closer to 4% error at 102mV sense. 

    So, clearly the solution would be to change the sense resistor, and not the DAC peak current setting.  That would allow you to use more of the full scale range of the DAC and minimize the effect of the inherent random offset of the high side sense comparator.

    Hope that helps.

    -JP

  • 0 in reply to James Patterson
    Mastermind 6205 points

    Dear James san

    Thank you for your quick answer!

    Yes, I had omit detail of frequency for each case, but  switching frequency was different.

    I will talk with customer to change sense resistor.

    Thank you