• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 2 answers
  • Subscribers 48 subscribers
  • Views 362 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.

TLV3801: Thermal Pad Connection

VR27
Prodigy 50 points
Part Number: TLV3801

I am using the TLV3801 as a signal conditioner into the PCB however the input signal has changed.

Previously I was using to convert LVPECL signals to LVDS however now I need to convert CMOS (0 to 3.3V) to LVDS. This, based on the datasheet, means I now need a Vee = -1.5V since i am converting 0 to 3.3 to 0 to 1.5V and using a DAC output to provide a reference of 0 to 1.5V to the negative input.

However, my problem comes that my current board is designed with the Vee and Thermal Pad to be connected to ground. I would like to prototype this on this current board, as a re-spin will add significant cost and I would like to de-risk this complete circuit operation. My question is what a better option is since the thermal pad I can't easily connect it to Vee as it has vias in pad:

1. Connect Vee to negative -1.5V and leave thermal pad connected to ground - I know this can cause some leakage?

2. Connect Vee to negative -1.5V and back drill the vias into the thermal pad to remove ground connection?

I understand neither solution is optimal but i would like something that work for now and in the next re-spin will go for performance.

  • 0
    Prodigy 50 points

    A third solution would be to fix the -IN to 1.5V and let the +IN to switch between 0 to 3V ( via a resistor divider) to achieve the differential input specifications however it is unclear what state will the comparator be if the -IN is Vee+1.5 and +IN is below Vee+1.5V. Maybe this is a better solution?

  • 0
    Guru 284800 points

    In general, the pad is connected to the substrate (which is at VEE potential) with glue that might be conductive. So if you connect a voltage higher then VEE to the pad, a current might flow, and affect everything on the die.

    As far as I can see, the only solution guaranteed to work is 2, i.e., let the pad float.

  • 0 in reply to Clemens Ladisch
    Prodigy 50 points

    What does it mean 'might be conductive'? What does it depend on? 

    The only reason I am asking is because I only need around 8 of these modifications to work for now to allow the concept proof. 

  • +1 in reply to VR27
    TI__Genius 11701 points

    Hi,

    The die mounting compound is conductive, so the pad is connected to the bottom of the die. 

    I like your 3rd solution. We were able to see that as long as one input is within the common mode range (so 1.5V above VEE), the other input can swing the full range (VEE to VCC) and the output behavior will be correct. So you can connect VEE/Thermal pad together to GND and have a reference at 1.5V. 

  • +1 in reply to Chi Nguyen1
    Prodigy 50 points

    Hi, thank you for your reply, I have already implemented the third solution and it seems to work reasonably well and this means no board modification. I have set the negative to 1.51V to allow the brief condition where both inputs are in the correct voltage range and hence a logic '0' is clearly asserted. I already have a series resistor to limit the current.

    One more clarification for your recommended output voltages - do you have a feel on what the tolerances are for such voltages? I.e. the datasheet mentions Vee+1.5V so is this 1.5V with 1%, 10% tolerance or is it an absolute minimum?

  • 0 in reply to VR27
    TI__Genius 11701 points

    Hi,

    For common mode range, it is a min/max spec not typ so it is the absolute min mix range across temperature. It does not mean the part is damaged if beyond the range, but some specs may be degraded or not warrantied to EC table specs.