Post Go back to editing

AD734 as a Lock in amplifier noise spikes

Category: Hardware
Product Number: AD734

Hello,

I’m trying to use the AD734 as a ‘Lock in’ amplifier.

I have the AD734 on a breadboard connected as in fig. 24 page 20 of the datasheet (basic multiplier circuit).

The denominator is set to 1V with the resistors specified in table 5 page 20 of the datasheet.

X1 input is a sine wave with an amplitude of 300mV (600mV p-p) and a frequency of 350 KHz.

X2 is grounded.

Y1 input is a square wave with an amplitude of 1V (2V p-p) and the frequency of 350 KHz.

Y2 is grounded.

At the output I have a the fully rectified sine wave but small spikes appears (FIG. 1 below) at half the frequency  (350KHz /2).

fig.1

If the amplitude of the square wave is increased to just 3V p-p the spikes exponentially increases (fig.2 and 3).

fig.2

fig.3

I can’t get rid of them.

Any suggestion?

Thanks

  • Hi  I may try to find a better forum for this question. But in the meantime - what sort of breadboard are you using, and what are your supply voltages? Even "medium speed" devices like the 10 MHz AD734 can often misbehave if they're not on a properly laid out circuit board with a ground plane and well-placed supply bypass capacitors.

    How are you probing the square wave? Are you using a probe with a ground clip? If so, the overshoot / undershoot may not be real. That still wouldn't explain the spikes, but it's one thing to look at as you dig deeper into this issue.

    Just as importantly (maybe more), how are you connecting the sinewave and squarewave to the circuit? Are you using coaxial cables, and is the inductance of the ground return (shield) low, with a low inductance connection to circuit ground?

    A photo of your circuit would be helpful, too.

    -Mark

  • Hello, thanks for the prompt answer.

    I'm using a prothotyping breadbord DKS-BBOARD3.3 by Digikey and so there is no ground plane just wires and decoupling capacitors connecting to ground.

    The supply voltages are +12V and -12V.

    I'm using probes with ground clips. I would have expected some misbehaviour at high frequency but not at 350KHz.

    The waves are connected via coaxial cables (BNC to Clips).

    It is probably better to build a PCB for testing.

    Thanks

  • Indeed a PC board would be better. Another alternative is PermaProto or DigiKey's "solderful" breadboards, which are quite useful for simple circuits. You can create a "ring" of ground around your circuit by making connections from top and bottom supply rails to the vertical rows, with short pieces of wire. Arrange bypass caps around the corners of the board, and another couple close to the AD734.

    Or just lay out a board in KiCad or your favorite software, there are lots of low-cost PCB manufacturers around, 1 or 2 dollars per board is possible, depending on size and complexity.

    On ground clip misbehavior - the sinewave is a 350 kHz signal, but the frequency content of the square wave is much higher, depending on the edge rate. Actually I was going to suggest this as well, does your square wave generator have adjustable edge rates? If so, try reducing the rate a bit.

    -Mark