Post Go back to editing

ADA4922-1: Power dissipation / Howto connect to a short twisted pair cable


In our system (industry electronics) we use the ADA4922-1 to convert a single-ended signal to a differential signal (a short twisted pair cable will be connected). The issue is that the ADA 4922 gets quite hot, even in steady state condition (when power dissipation shout be at its minimum), no cable connected. The package heats up to about 60°C with an environment temperature of about 25°C. When calculating the power dissipation from this temperature, you get around 300mW.

In a steady state condition, we have a signal input of 0.5V, nothing is connected to the output. There is a small copper area connected to the exposed pad. Power supply is -12V and +12V, the supply pins are blocked with 100nF ceramic capacitors. Disable pin is at 12V. No cable is connected, which leaves about 1cm of copper trace per output. I guess power dissipation should be around 1mW (input signal, internal resistors).

So the question would be, where in the IC and why is this high amount of power is dissipated? Do you have an idea? Measuring the outputs and power pins with an oscilloscope shows no oscillation.

Do you also have a hint about connecting a a twisted pair cable (length about 1m) to the device? Should it be done directly, or is it better to use a resistor/capacitor network to prevent too high capacitve load? The signal Frequency will be around 10kHz.



  • Hi Michael,

    First I would like to ask you about your +/-12V sypply requirement?

    Is that supply voltage a requirement to increase the common mode voltage?

    If not, maybe we can either reduce the supply voltage to lower the power or chose another part that meets your requirement.

    Second, for such cable, you have to have a resitor on the output of the amplifier and terminate the cable on the other side. you might want to increase the gain of the amplifier to make up the 6 dB loss.

    Please feel free to contact me via private message using my community profile for more detailed info.


    Message was edited by: Andy Rose - Removed phone number from thread.

  • Hi Charly,

    Thanks for the fast reply. :-)

    The +/-12V are not a hard requirement, but it's done since I already have this supply for other parts of the board and there's no more real estate on the board to put a +/- 5V supply (which would be enough for the signal span and common mode range).

    So the heat development is normal when using a +/- 12V supply?

    In terms of using another device we're open, we just need a singleended-to-differential converter. The ADA4922-1 seemed to be a considerably cheap und easy solution for that. If you have an alternative let me know.

    The 6dB loss using a network to terminate is ok since in the target device where the signal goes we have to do a digital signal adaption anyways..what I would have to figure out is if I loose to much SNR when my signal is half than before (but I hope that noise on my connection gets smaller with a termination network). What kind of network would you suggest?

    If it's easier for you to discuss by phone you can PM me you office number, then I call you (next week).

    Thanks again,


  • Hi Michael,

    Unfortunatly we +/-12V is the only option, that will limit other diff amp options.

    But that is ok. becaue the ADA4922 should be able to handle the work.

    All we have to worry about is the proper layout and termination so that the part is not oscillating.

    Can you please send me a schematic so I can be of more help?

    Show me what you want to do exactly.


  • This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

    Thank you,
    EZ Admin