• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 64 subscribers
  • Views 806 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.

TPS92661-Q1: Current source alternative ...

Joerg Beutel
Prodigy 125 points
Part Number: TPS92661-Q1
Other Parts Discussed in Thread: TPS92515, LM3409

[This part number forces this question into the Automotive Forum whereas I would rather expect to have it in the LED Driver forum ... hope it can be found here :-) ...]

Going through the datasheet of the TPS92661-Q1 and the schematic of the Eval kit I only find quite "unpleasant" versions of the required buck converter (constant current driver) to source it. I would rather like to use a chip with integrated FET and smaller inductance. Something like the SCT2932 would be nice ... http://www.starchips.com.tw/pdf/datasheet/SCT2932V02_01.pdf ... It also uses a current sense resistor above the LED chain and protects against voltage over- and under-runs ... with a lot less components.

My question: Could that work, maybe by syncing the PWM frequencies ? I would like to reduce the current through the chain when the overall brightness goes down. To avoid power loss and heat the idea is to have as much efficiency as possible so a non-switching current source is not good enough.

To Ti:

In a perfect world there would be a companion current driver chip to the TPS92661-Q1 that would automatically cooperate in sync with the PWM ... I also wish the LED chain had been split in two in the middle (one connection between drain and source to go through two pins as there are unused pins) so that two chains of 6 LEDs could be made (an option the LT3965 provides while lacking the internal gate supply). Such splitting would allow us to leave the voltage low for security and easier certification.

  • 0
    TI__Mastermind 27920 points
    We're moving your post to the appropriate forum.
  • 0 in reply to Akeem Whitehead
    Prodigy 125 points

    Thank you, Akeem !

  • 0
    Prodigy 125 points
    To answer my own question to some extend ... the TPS92515 seems at least to solve the requirement for an internal FET. Still needs some external parts and the integration with the TPS92661 is less than perfect, but it's a step forward.

    I wonder if I can find a way to leverage the 4 unused outputs to control the buck converter. The idea is that I would like to have an atomic serial write operation which sets the PWM of the 8 LEDs AND in parallel the PWM of the buck converter. That way if low light levels are required on all LEDs I would want to use the buck and the TPS92661 would define the relation of the LEDs one to another. If I could find a way not just to make a short connection over the unused outputs but to read them somehow I could change the buck to 16 current levels and in the same moment change the LEDs accordingly. :-)
  • 0
    TI__Prodigy 710 points
    Hi Joerg,

    The TPS92515 is the successor to the LM3409. It was unreleased at the release of the TPS92661, but it would be a good choice for matrix systems. The inductance on the TPS92661 was chosen to limit ripple current with the very small currents shown on that board. The TPS92661 EVM is intended to show how the programming interface of the part works, and give a good starting point for a matrix system design, but it is not intended as a reference design. The inductance should be chosen based on your particular system needs.

    You want to synchronize the TPS92661 PWM frequency to the switching frequency of the converter? Can you tell me more about your application?

    Thanks,
    Mike
  • 0 in reply to M. Drake
    Prodigy 125 points
    Thanks Mike,

    I just wrote more about the application in the other thread. The idea to sync the two frequencies may be silly. While I like the basic concept of the 661 there's one situation where it runs into some trouble ... when all LEDs are more or less off. In that case the whole current runs through all the closed switches, heating the chip and forcing the buck way down with the output voltage. Staying in CCM mode is a problem and so on. That's why I try to find a way to control the buck in parallel to the LEDs in order to achive maximum efficiency and leave the whole system in a sweet spot independend of the over-all brightness. Looking at the simplicity, efficiency and economics of something like the small SCT2932 V2 makes the 515 look "heavy" :-) ...