Post Go back to editing

AD8034 as voltage follower, heat generation

Category: Hardware

I'm designing a frequency counter with an input stage that consists of an AD8034 in voltage following setup. The purpose of that is to burden the signal source as little as possible (ideally the source should not notice the counter is there).

The circuit simulated in LTspice draws very little current (in the µA range), while in real life that is not the case. The input is a quasi-square wave AC with 8V peak (16Vpp), the desired frequency range is 0.1Hz -10MHz.

So much for the intro.

Putting it together on a breadboard I noticed the AD8034 draws lots of current (>30mA if I let it) and gets extremely hot when simply shorting the feedback from VOUT to IN-. Adding the resistors helped taming the current consumption to 8mA at load (with +-12V power supply) and temperature on the chip's surface to about 45ºC/110ºF while keeping the signal good for the LT1720 that follows.

So the question is: Is that kind of heat generation to be expected or is there something fundamentally wrong with how I'm doing this?

Thanks for any pointers!

Regards

NB

PS: Yes, I'm a rookie, so please don't laugh when you look at this:

  • Reducing the voltage on IN+ to 500mV about halfs the total supply current to 4mA, only very little more than running it idle without any signal. Then, the temperature of the chip surface is less than 35ºC (down from 45ºC).

    Works for me, but is such significant heat production at almost idle current to be expected?

    Regards

    NB

  • Hi  ,

    1st cause may be the use of breadboard. A poor PCB layout can degrade the performance of the part and cause it to draw more current.

    Another thing is that you may be violating some parameters in the op amp. What is your signal amplitude and supply? Please make sure that you do not exceed the Input Common-Mode Voltage Range.

    By the way, I'm seeing that you have used a voltage divider at the op-amp output. I suggest adjusting the voltage divider at the input instead of voltage divider at the output, the lower output impedance of the op-amp driving the next component will reduce noise pickup and other circuit effects.

    Thanks,

    Gio

  • Hi Gio,

      many thanks for your reply and suggestions.

    I followed your suggestion to remove the output voltage divider and added one at the input, so input voltage got down from 16Vpp to less then 1Vpp. The output change increased the chip temperature slightly from 32ºC to 35ºC.

    But now that the input voltage is low I reduced the supply voltage from +-12V to +-5V and that seems to make a huge difference with heating, the temperature is now down to less than 30ºC.

    I somehow falsely assumed that it's a good idea to subject the opamp to my signal directly so long as it is within the generally acceptable operating boundaries. Apparently high input and supply voltage not only causes extra heat but also makes the circuit more susceptible to noise pickup. At least the signal looks much cleaner now, even on the breadboard.

    I shall see how much difference a PCB implementation makes in terms of noise.

    Thanks again

    NB