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.

TLC5927-Q1 short/open detection and current variation

Other Parts Discussed in Thread: TLC5927-Q1, TLC5917-Q1

I would like to confirm the short/open detection of TLC5927-Q1.

1. In the datasheet, there is the specification of the open detection (IOUT,th).

   I don't know 0.5*Itarget (%). If Itarget is 87.5mA, how do I caluculate it?

  (0.5 * 87.5mA = 43.7mA or 0,5% 87.5mA = 0.437mA, ???)

2. The customer will make 350mA * 4 strings. They can relize the connection 4ch (87.5mA/ch) at anode of LED, I confirmed this e2e.

    In this case, can TLC5927-Q1 keep the variation of the current 350mA +/- 10%?

3.  When Iout is 87.5mA, they need set the cathode voltage (Vo) = LED VF + 1.0V.

     Is this correct?

Best regards,

Atsushi Yamauchi

   

  • Hello Atsushi,
    The appropriate Applications Engineer has been assigned to answer your question(s). You can expect a follow-up response within a few days.
  • Atsushi-san,

    1,The open detection is 0.5%*87.5 = 0.4375mA
    2, The current accuracy can meet the 10%.
    3, For 87.5mA output current, the anode voltage should be larger than (Vf+1V), rather than cathode.
  • Dear Anda-san, 

    Thank you for your reply.

    About No.2, how much is the accuracy the customer can estimate?

    They want to estimate <10% acurracy with all temperature range.

    Best regards,

    Atsushi Yamauchi

  • Atsushi-san,

    Recommend to use 10mA to 50mA current to achieve 6% accuracy, which needs to divide the current into more channels. For example, 7ch*50mA. By the way, what is the application of customer?
  • Anda-san,

    Their application is the turn signal for motor cycle.

    They use it 87.5mA /ch. In this case, they want the accuracy <10% and want to estimate how much the accuracy is.

    Best regards,

    Atsushi Yamauchi

  • Anda-san,

    We got the questions from the customer.

    - How much is the pull-up resistor at /OE and the pull-down resistor at LE? They want to know the influence of the series resister 1k ohm 

      they add.

    .

    - Can they be flushed LEDs by ON/OFF with OE pin? They want to flush 1.5Hz and 2.75Hz.

    Best regards,

    Atsushi Yamauchi

  • Atsushi-san,

    Our test engineer is helping to find the current accuracy data. Once we get the data, we will give updates to you. 

    The pull-up resistor and pull-down resistor are 500kohm typical, min 250kohm, max 800kohm. There is little influence of the series resistor. 

    Yes, customer can flushed the LED by OE pin.

  • Atsushi-san,

    Checked with our test engineer, we can guarantee 6% current accuracy from 10mA to 50mA load current.
  • Anda-san,

    I have understood 6% current accuracy with 10 to 50 mA you guarantee, because there is the specufication in the datasheet.

    The customer wants to know the current accuracy with 82.5mA and -40 to 125 degreeC.

    Could you please estimate it?

    We don't need your guarantee for the accuracy.

    Best regards,

    Atsushi Yamauchi

  • Hi Atsushi-san,

    Sorry for the late response. As you know, the data estimated in the lab is very random as it is only based on the few samples. We can test the data for you if you need it. But before we do the test, could you share the system information of the customer? If it is sensitive, you can also shoot me email directly which is in my profile.
  • Dear Anda-san,

    The customer's application is the cluster with turn signal for motorcycle. They use this device for turn signal.

    LED strings are for each turn signal (front of left and right, back of left and right) 350mA and total is 1.4A,

    They use TLC5927-Q1 by gathering 4ch for one line ( One line is 87.5mA. I have confirmed it can in this forum.).

    If you want to know more information, could you please inform it and I will send email to you.

    Best regards,

    Atsushi Yamauchi

  • Atsushi-san,

    Sorry for the delay response. Have you estimated the thermal performance of the TLC6C5712? Because if customer wants to get 87.5mA output current. The voltage drop on the mos is at least 1V. So the power dissipation on the device is at lease 1.4W. Assume customer uses the PWP package, the temperature will rise 1.4*63.9 = 89.46C. As the min TSD value is 150C, so the operation temperature will no bigger than 60C.
    Do you think there is any problem?
  • Dear Anda-san,

    Thank you for your advice.

    Turn signal has 50% duty cycle (it flashes on and off). So, the average power dissipation is 0.7W. I think it is OK.

    If the chip temperature is very high, they will use TLC5917-Q1 * 2 pcs.

    Could you please continue to test the variation (the accuracy)?

    Best regards,

    Atsushi Yamauchi 

  • Atsushi,

    It will take a long time to do the test on the bench. As we need to build a new test board. I discussed with our designer, he can help to do some simulation on this device. Do you think it this will be ok?
  • Dear Anda-san,

    It is OK for your simulation result.

    Could you please get the results?

    1. the result including temperature shift (-40 to 125degreeC) with 87.5mA

    2. the accuracy of open detect current with 87.5mA (0.5% * 87.5mA +/- ?%)

    About 2, if it is same accuracy you can explain, we don't need the data (we need your comment only.)

    Best regards,

    Atsushi Yamauchi

  • Hi Atsushi-san,

    Checked this with designer. The current accuracy under 87.5mA is better than 6% under full temperature. And the accuracy of open detect current is the same as you write above.
  • Dear Anda-san,

    Thank you for your reply.

    I understood it and sent it to the customer.

    And they have a question.

    1. Can they separate the detection for open and short?

        I think they can not it because the output for the detection is the error status code only (in the detection, output is code "0"). 

    2. Can they find the detect separately with adding the resistor at OUTx  to GND? 

        In this case, when they check the voltage from OUTx to GND, they will be able to find each detect.

        I understood it is not correct method for this IC, because the current setting may be shifted by the resistor (in parallel against inside current source circuit).

        But I would like to know whether they can find the detection separately or not .

    Best regards,

    Atsushi Yamauchi

  • Atsushi-san,

    You are correct! The device can't distinguish the error type. It is an interesting way, in theory it would work, do you have the detailed detection circuit? What detection components does customer plan to use?
  • Dear Anda-san,

    I attached the circuit.

    The parameters is not dicided, because the resistors gives the impact to the current value.

    Do you have good idea for the circuit (the resistor value) without the current value?

    Best regards,

    Atsushi Yamauchi

  • Atsushi-san,

    Sorry for the late response.I think a resistor divider should be enough for this circuit, the resistor on the right side is unnecessary. Try to select a suitable ratio to make the voltage is compatible with the MCU. Use a large value resistor to minimize the current flow through the divider resistor.