• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 46 subscribers
  • Views 301 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.

ADS805 Heat dissipation

Mauritz Zondagh
Prodigy 50 points
Other Parts Discussed in Thread: ADS805

Hi,  I am using 4x ADS805 on a pcb with layout spacing of approximately 21x21mm square. With a ambient temperature of 25 degrees celcius, I measure the temperature on the IC at 50 degrees celcius after a while. The only cooling is a piece of 3mm aluminium  glued over the  ADS805 with Silicon with normal room air. (For prototypes) Is 50 degrees celcius normal?

The Clk and OE inputs is 5 Volt, with Vdrv at 3V3, and outputs connected to 3V3 FPGA . The Conversion is working fine. Running at 10Mhz constantly. If I switch off the clock, the temperature does not drop. Thanks  

  • 0
    TI__Mastermind 29220 points

    Hi,

    From looking at the datasheet, the theta JA  (junction to ambient) is listed at 50 C/W and the power dissipation at 0.3W, so the expected temperature rise from ambient to junction would be expected to be about 15C.  (50 C/W times 0.3W = 15C)  (page 3 of the datasheet near the bottom.)

    The theta JC (junction to case) is not specified, but it is usually a fairly small number. The junction is the hottest part of the device (as that is the heat source) and the case is usually just a little cooler, and most of the thermal resistance is getting the heat out to the air.  (Since this device does not have a thermal pad to solder to the board to spread heat out through the board's ground planes.

    So a 25C temp rise from ambient to the case (I am assuming a good thermal connection from case to your aluminum, so that the aluminum is pretty close to case) seems a bit high to me.  if expected temp rise from ambient to junction is only about 15C, then ambient to case should be less than that. 

    It could be that since you have four such devices in a fairly close proximity that the heat is building up a bit, especially if the ambient air is not moving.  With airflow, the heat should be carried away from the aluminum, and you may do about as well carrying the heat away with airflow and not the aluminum.  I'd compare both cases - having airflow and with and without the aluminum and see if you are trapping some heat.

    Regards,

    Richard P.

  • 0 in reply to Richard Prentice
    Prodigy 50 points

    Thank you Richard,

    I would also expect the 25 above ambient to be to high. But for now we are working with it as is, and all is OK. Will see in next release of the PCB with better heat sink and airflow if things approve. The IC is working fine, i do not see any layout problems so the heat must be because it is in close proximity and not enough heat sinking with airflow. There is nothing else.

    One thing which is still bothering me, i disabled the 10 Mhz input sample clock within my FPGA to the ADS805, suspecting the the ADS805 will get cooler, but it stayed approximate the same temperature? Wanted to add a feature to disable if we do not use it, to cool the IC as a interim solution while we are designing and testing. I do not see any other way how to disable the IC.

    Thanks

  • 0 in reply to Mauritz Zondagh
    TI__Mastermind 29220 points

    Hi,

    I do not see an Enable pin on this device, which most of our newer devices will have.  So when this device is powered up, it is on.  On our CMOS data converters, there is usually a strong correlation between clock rate and power dissipation, and next to no power dissipation with no clock.  But many of our devices designed in a bipolar technology would be designed with constant-current fully differential logic, such that current always flows and the difference between a logic 1 or 0 is whether the current flows this way or that way.  These devices will have little dependence between power consumption and clock rate.  So I would suspect that this device is in a bipolar technology, although I don't see mention of such in the datasheet.  This is a device that came to TI by way of the Burr-Brown acquisition some years ago.

    Regards,

    Richard P.