• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 63 subscribers
  • Views 517 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: analog dimming accuracy depend on tDEL

KANJI SHIBATANI
Mastermind 6205 points
Part Number: TPS92518

 

Hello

Thank you for your support.

My customer is designing their ES hardware/software for TPS92518 now.

The difficulty for TPS92518 is, they have one PCB but they

use different LED ( 36V/1A and 12V/1A ) with 0.1% dimming.

Please kindly see attached PPT.

Q1

We see tDEL=58ns(TYP),110ns(MAX) in the TPS92518 data sheet.

Our understanding is 110ns is worst case  including  chip(lot) variance.

So, at the point where Ta=150deg-C and typical chip, we see tDEL=73ns on Fig5.

Difference between chip(lot) should be  110-73=37ns.

So, we should think tDEL=58ns(TYP)+37ns=95ns as Ta=30deg-C in case chip variance?

Otherwise, this “ 37ns “ is much smaller in case Ta is lower?

 

Q2

If available, do you have minimum value of tDEL? (58ns(TYP)/110ns(MAX) ..and MIN??)

 

Q3

Do you suggest shunt FET dimming only, rather than analog dimming + shunt FET dimming

if  TOFFMAX were well tuned, shunt FET Rdson was small and clamp diode is used carefully?

The reason why our question is, in case ILED=200mA, ILED has variance depends on

VLED( duty/switching frequency) due to tDEL influence, even same TOFF/PKTH  .

  .. Please kindly see attached PPT file page 4 …(A).

One idea is to monitor ILED externally and change PKTH by MCU, but

Customer says it isn’t easy  to have accurate measurement .

 

 

Q4

Customer intend to use switching FET whose trr is 48ns ( attached data sheet ).

Would you please kindly  give your opinion? ( Is its trr too slow? )

 

Q5

Even though in case  same TOFF, same VLED, same Vin,

Switching frequency is changed  depends on PKTH.

It became higher/lower  depend duty also..

(please  kindly see attached PPT page 4 (B) and (c))

Would you please give your comment about reason why?

I am sorry that It is so complicated questions every time, but please kindly give your suggestion.

Best Regards

 

 

 

 

VIN_55V_36VLED_12VLED_english.pptx

AP10TN028YT_Datasheet_Package.pdf

  • 0
    TI__Mastermind 28655 points
    Hello Kanji,

    1) It means that 58 ns is the typical and 110 ns is considered the maximum that any part can have. If the delays are cause too much inaccuracy going to 100% SHUNT FET dimming may be needed.

    2) This is something not specified in the datasheet, there are many other things going on that cause delays and shifts in set-point accuracy. If it is an issue going to SHUNT FET dimming is preferred.

    3) The variance when analog dimming is more than just tdel. The inductor accuracy plays a much larger role as well as all the offsets in the sense circuitry. If the accuracy of analog dimming is not good enough at 5:1 then switching to all SHUNT FET dimming makes more sense.

    4) trr is the MOSFET body diode recovery. The MOSFET body diode is not used in the TPS92518 buck converter so it is no concern.

    5) Switching frequency depends on Vin, Vout, the accuracy of the off-time circuit. When Iout changes Vout changes, when Vout changes switching frequency changes. The TPS92518 is not a fixed frequency converter, it is a constant current ripple converter. Duty cycle changes because the overall Vout or Vin are changing. This can be the LED stack voltage changing, losses in the circuitry changing, etc. This is running at 500 KHz to 1.1 MHz so delays will be a larger part in the design.

    The 12V and 36V operation appear to have very different output voltages. If analog dimming it may be necessary to adjust values to make the currents similar if this is the case.

    Best Regards,
  • 0 in reply to Irwin Nederbragt
    Mastermind 6205 points

    Dear Irwin-san

    Thank you for your suggestion !!

    Customer and I am considering carefully,because 1A swithing by sunt FET may have large spikes.

    Anyway, thank you for your answer ,

    Best Regards

  • 0 in reply to KANJI SHIBATANI
    TI__Mastermind 28655 points

    Hello Kanji,

    If there is no output capacitor the spikes should not be very big, just the Coss of the MOSFET.  It may have to be slowed a bit if EMI from it becomes an issue.

    Best Regards,

  • 0 in reply to Irwin Nederbragt
    Mastermind 6205 points

    Dear Irwin san

    Thank you for your  advice!

    I will discuss with customer .

    Thank you again.