• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 1 answer
  • Subscribers 62 subscribers
  • Views 1324 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.

Questions on TPS61196

FZ37
Intellectual 450 points
Other Parts Discussed in Thread: TPS61196, TPS61196EVM-600, INA226, TPS61199

The application is to drive two groups of LED strings with 180 degree phase off (each group turns on for ~100 us @~200 mA and then off for 120 us, no overlap and ~10 us in between).

1) One way is to drive different groups with different ICs. Another way is to drive group A with 3 channels and the other 3 for group B from the same chip. Either way it requires at least 6 TPS61196 in each application.

The question is which way is better? Heavy load turning on and off every ~110us or half load running almost continuously with ~12 us off time.

2) If LED shunt protectors like the one below are used in parallel with all the LEDs and one or two of LEDs fail open or short, any switching or stability issues (~100us switching on and off) other than 1-2V more voltage drop across the current sinks (IFB pins)? Are the current sinks going to be able to dissipate the extra 1-2 Watts heat? Assuming the layout is similar to the demo board -- TPS61196EVM-600.

www.onsemi.com/.../NUD4700-D.PDF

Thanks,

FZ

  • 0
    TI__Intellectual 1780 points
    Hi FZ,

    Thank you for the question. I discussed with the Systems Engineer and here are our responses.

    1. Half load On for most of the time and Off for ~12us (out of total 220us), would be a better option for the following reasons:
    - Lower Output Ripple. The output ripple would be at 10kHz due to load switching.
    - May be able to use less output capacitance
    - Boost can adjust output voltage for two groups of 3 strings at two different times. This could minimize headroom loss, depending on the LED mismatch.

    2. The TPS61196 shouldn't have an issue dissipating a few watts of power. Moreover, it will take more than a few failed LEDs to increase the headroom losses by that much.

    Please feel free to let us know if there are any further questions.

    Best,
    -Harini
  • 0 in reply to Harini Sankabattula
    Intellectual 450 points
    Harini,

    Thank you for the response. Now I learned more details of the LED strings. Each string has up to 40 same type LEDs. But different groups of strings are made of different types of LEDs. The voltage drops vary from 1.4V to 2.0V @ ~200mA. Here are more questions on the device.

    1) Is the device capable of driving ~80V output @ 600mA to 1200mA total with a 24V supply? The demo boards TPS61196EVM-600 I ordered are set to drive up to 60V output @ 787 mA (130mA per channel). I have to change R7 to ~300kohms to be able to test the device with my LED strings.

    2) Since the voltage difference could be very big (up to 24V) between different group LED strings, it's probably better driving the same type of LED strings with this device at ~5KHz PWM control with 50% duty cycle or ~10KHz with 90% duty cycle. This would require minimum voltage dynamics other than switching on and off, right?

    I am also figuring out how to make the current setting adjustable in the driver. I need a digital Rheostat connected to Pin ISET. However the current sink head rooms are set by the ratio of the two resistors on Pins IFBV an ISET. I am trying not to use two digital Potentiometers. Any suggestions?

    One last thing is the monitor the actual current flowing through LED strings. I am thinking of INA226.


    Regards,
    fz
  • 0 in reply to FZ37
    TI__Intellectual 1780 points

    Hi FZ,


    1. Yes, this should be possible, but the following need to be done:

    - OVP divider resistors R3 & R4 need to be modified. -- Page 17 in DS
    - ISNS resistor (R7 in Fig 23) need to be ~50mohm for ~8A inductor current limit (400mV/8A). -- Page 16 in DS

    R11 (in Fig 23) if set to 300kohm, gives I_LED = (1.229*3992)/(300*1000) = 16.35mA per channel.

    You mentioned several times that you need 200mA/channerl, in which case, this ISET resistor R11 need to be ~24.5kohm (Equation in Page 15)

    Best,

    -Harini

  • 0 in reply to Harini Sankabattula
    TI__Intellectual 1780 points
    Here is about your Question # 2:

    Yes, if the LED string voltage is that different then you would be better off using one device to drive the same type of strings.

    This will avoid adaptive boost voltage function from having to make large voltage transients (or) burning too much driver loss power in the case the adaptive boost could not keep up.
  • 0 in reply to FZ37
    TI__Intellectual 1780 points
    Hi Fz,

    Btw, why do you want to dynamically adjust the ISET pin Current?

    For the current monitoring, INA226 looks like it could work. You will need a shunt resistor between the LEDs and the TPS61196.

    Thank you,
    -Harini
  • 0 in reply to Harini Sankabattula
    Intellectual 450 points

    Harini,

    R7 = 300kohm (on demo board) is R3 in fig 23 on DS to set the OVP. One problem I saw here is at the high limit of ~80V output when the LEDs were switched off, IFB pins will see a high voltage about half of output voltage (~40V). However the max voltage is 40V for IFB pins on page 4 of data sheet. This is probably why 3 of my demo boards are not working any more. I have one more left. I only run it with 20 LED strings for now.

    Probably, a ~30V zener can help here.

    Regards,
    FZ

  • 0 in reply to Harini Sankabattula
    Intellectual 450 points

    Hi Harini,

    In the system, the LEDs are used as flash lights for super fast cameras. Different types of LEDs are synced with different cameras. We want the intensity of light to be adjustable, not very fast but at least it can be changed through GUI interface. I thought it might be ok to leave the head room to high enough for all different current settings. I need to do some calculations here.

    The latest test on the demo boards:
    I used the board to drive 6 strings of 18 LEDs with the default settings of 130 mA current on each string. I found that the chip is getting very hot due to the 2.5V voltage drop on 4 of IFB pins. Although they are the same type of LEDs, two strings are from different batchs. This means that TPS61196 could see a lot variations on LED voltage drops if they are not perfectly matched.

    Thanks for helping out on the questions.

    FZ

  • 0 in reply to FZ37
    TI__Intellectual 1780 points
    Hello Fz,

    R3 (in fig 23) = 300kohm corresponds to Vovp = ((300k/10k) + 1)*3.02 = ~93.6V
    For Vovp = 80V, R3 = 255kohm

    IFB has OVP during start-up. So, the demo board(s) shouldn't have damaged at start-up.

    A few questions for you so that we understand your test better:

    - How did you measure voltage on IFB pins (using DMM or scope?).
    - What is your boost voltage?
    - Can you provide Vf - If characteristic and forward voltage of the LED you are using?

    Thank you,
    -Harini
  • 0 in reply to Harini Sankabattula
    Intellectual 450 points
    Thank you Harini!

    I found out this part needs two resistors to be adjusted to change the LED current settings. One for Iset and the second one for IFB head room. I decided to look into TPS61199, which doesn't seem to need one. You saw my post on that device too.

    -FZ