Post Go back to editing

AD8302 overheating

I've been using two AD8302 modules for a project where S-parameters are measured. There are multiple functions, each with different codes and commands, supplying frequency sweeps from 0 to 4GHz at 0dB. Since the AD8302 is connected to directional couplers and not directly to the frequency generator, it receives around -20dB. The outputs were displayed fine, however, after a few days of using this code, the result started to just show -31dB as the magnitude ratio and the AD8302 IC itself was hot to the touch. The AD8302s are connected to an analog to digital converter (MCP3008) using the connection shown below.

Is it possible that a power surge occurred which burnt the AD8302s? Did the usage of frequencies upto 4GHz spoil the chip. Or is the fault entirely on the signal generating side? I am scared to connect any other AD8302s to the schematic shown as I fear it will just heat up as well.

    • What is the supply voltage?
    • Have you been recording the supply current while using the AD8302?  Current should be 20mA nominal. 
      • It should go no more than 27mA
    • How hot to the touch was the device? 
      • Can you hold your finger on the DUT for longer than 1 to 2 seconds? 
      • 10 seconds?
    • Can you observe the inputs of the AD8302 on an oscilloscope to double check you signal levels?
    • Supply voltage is a 5V from a Raspberry Pi. To be more accurate, when I measured it was around 4.8V.
    • No I didn't check the current, will look into that.
    • Regarding the touch, I'd say I couldn't touch it longer than 2 seconds. 10 seconds would have burnt me.
    • Didn't try checking via an oscilloscope either, I will do that.

    If it is this hot, does that mean the IC is permanently damaged? Or is there a chance to still use it?

    • Is the 5V supply from the Raspberry Pi an LDO type?  Something different? 
      • Is it stable?  Can you check the supply on a oscilloscope?
      • Can you supply the AD8302 from a lab grade power supply?

    The fact that it is that hot leads me to believe something is fundamentally wrong with either the IC or the PC board you have the device soldered to - perhaps there is a short somewhere on the board or the IC?

    I would replace it with a new IC and re-test.  If the same behavior happens, I would suspect the PC board.