• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 47 subscribers
  • Views 519 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

ISO122U gets too hot

Michael Wartig
Prodigy 30 points

Other Parts Discussed in Thread: ISO122

Hello,

I am using eight ISO122U Isolation Amplifiers in my desing, placed on the bottom side of the PCB (placement area is about 40 x 50mm).

Supply voltage is +/-15V on input and output side.

I'm using them to transfer a slow 0...10V signal (max. 1kHz).

Problem: The ISO122U get very hot, they reach >70°C and heat up the PCB (and the components on the top side).

I cannot imagine that this is normal, although the power consumption at 15V is relatively high (ca. 0,6W according to datasheet).

Is there anything I can do? My first try was to reduce the supply voltage at the output side from 15V to 12V, but that helped only a little bit (temp. >=65°C).

What else can I do?

Kind Regards,

Michael

  • 0
    TI__Genius 9975 points

    Hello Michael,

    What is the load the ISO122 feeds? Can you measure the supply current on both sides of the ISO122 and let us know the value?

    Best regards,

    Jose

  • 0 in reply to Jose Duenas
    Prodigy 30 points

    Hello Jose,

    thank you for the quick answer!

    The ISO122U feeds an OpAmp stage with 20 kOhm input impedance.

    I measured the following supply currents:

    4,2mA at +15V input side (Pin 1)
    4,0mA at -15V input side (Pin 2)

    4,7mA at +15V output side (Pin 15)
    5,2mA at -15V output side (Pin 16)

    Input side is supplied by system voltage (stable power supply).
    Output side is supplied by Traco TMR-0523 DC/DC converter.

    I think the current values are OK...

    Regards,

    Michael

  • 0 in reply to Michael Wartig
    TI__Genius 9975 points

    Hello Michael,

    You are very welcome.

    Indeed, the current values are OK and the load impedance is large enough for the ISO122 output to handle it well. The next aspect to review is the PCB layout. How many copper layers do you have in your PCB?

    The junction-to-ambient thermal resistance of the ISO122 is 100deg.C/W; but such resistance is probably not a good representation of the resistance you see in your system for reasons explained in http://www.ti.com/lit/an/spra953b/spra953b.pdf

    If conditions in your board were identical to those outlined in JESD51-1 and JESD51-3 then one would expect a junction temperature increase of about 27 deg.C starting from 25 deg.C ambient. I wonder if the close proximity between ICs is helping the significant temperature rise since you have a high density of ISO122 in a relatively small area (I assume you have two rows of four ISO122s each with the 10.5mm dimension/side running parallel to the 50 mm side of the 40 mm x 50 mm square).

    Do you see a temperature gradient in the board? Are the ISO122s in the center of the PCB hotter than those in the edges?

    Would it be possible to test the temperature increase for a board which only has one ISO122 populated?

    Best,

    Jose

  • 0 in reply to Jose Duenas
    Prodigy 30 points

    Hello Jose,

    the PCB has 6 layers.

    You are right, I have two rows with 4 ISO122; space between the case is about 6mm.
    Case distance between the two rows is 1mm.

    With the typical datasheet values I expected a temperature rise of ca. 32°C what is close to your calculation. I knew that this high density placement would lead to an additional temperature rise, but I didn't expect such a big influence.
    A different placement for more space between the ISO122U is hard to realize because of manufacturing limitations.

    The temperature of the ISO122 measured in the edges is about 5°C lower than the center ones.

    I have not tried to depopulate the board yet, but I can compare with a 2-channel prototyping board with the same circuit and only four ISO122U placed in a 25 x 65mm area. This placement is uncritical, the temperature of the ISO122U is <=50°C.

    So I think that the high density placement is most likely the problem.


    Is there any alternative to the ISO122? I don't need the high voltage isolation and high frequency, just galvanic isolation for eight 0...10V signals with max. 1kHz.



    Best Regards,

    Michael