1

I am using an instrumentation amplifier AD620ANalong with another amplifier TL084CN in series. Both have gains of 100 which essential for me since the input is in microvolt range. The frequency is 87 hz and note channel 1 is input while 2 is output.

The circuit diagram looks something like this [ Readings:

  1. When normal input of 1.76 volts is given the DSO shows the expected output of square wave of similar frequency as the input as shown in the image here

  2. When the input signal is attenuated by 20db the signal still tends to remain intact showing the same frequency of the input with a question mark on DSO image

  3. Now here is where the problem starts whenever I switch my function generator to 40db or 60 db the signal at the output is square wave but the frequency is not the same and again a question mark is seen in the DSO in the frequency slot. And the distortion or frequency changes even further with 60db. This can be seen in below 2 images. 40db 60db

Technically the frequency shouldn't change throughout the process. Believing in this my professors and lab assistants told me that the frequency is actually the same however the DSO is causing errors and hence the frequency shown is different by the DSO. But I don't believe this because if it was a small signal of small amplitude at the output then I would understand that the DSO is not capable of sampling such small signals but since in the images you can see that the output is in the range of 8 volts then how can an instrument show false readings?? Also initially when the input signal was at 1.76 the output was pretty smooth and as expected but after making input smaller the output is varying then definitely it has got something to be related to the circuit I have created. Am I right or wrong? Can anyone spot if there's a different problem? Suggestions please!

Update: I found out hysterisis was the problem. Thank you for your help!

user183710
  • 94
  • 7
  • 1
    The AD620, with its input offset spec of as much as 50μV, is not what I would choose as a first stage here. What's your SNR at the AD620's input, anyway? – Hearth Feb 14 '19 at 16:41
  • I'm a beginner in this so mind me but can you tell me how to calculate?? – user183710 Feb 14 '19 at 16:45
  • It's not something you calculate, it's something you measure. – Hearth Feb 14 '19 at 16:46
  • Okay so how do you do that?? – user183710 Feb 14 '19 at 16:47
  • Could you please rotate the image? Or even better - use the embedded schematic editor. – Eugene Sh. Feb 14 '19 at 16:49
  • Yes sure..hold on! – user183710 Feb 14 '19 at 16:50
  • At 40db it looks like since your circuit has no hysteresis the noise is causing the output to switch rapidly during the transitions. At 60db all you get is the transitions. This indicates that the noise you have at the input is equivalent to or greater than the actual signal. The signal to noise ratio is the magnitude of the signal divided by the magnitude of the noise. A 1 is bad a 10 or more is good. – vini_i Feb 14 '19 at 16:52
  • So how do I increase the SNR to avoid all of this? – user183710 Feb 14 '19 at 16:57
  • Is your schematic correct for the TL084, its configured as a comparator with hysteresis. – sstobbe Feb 14 '19 at 17:00
  • Yes I want it to be used as a comparator. But I had no idea that hysterisis can cause a problem at this stage – user183710 Feb 14 '19 at 17:02
  • I'll research analyze and do further testing and get back here to this post in a few days – user183710 Feb 14 '19 at 17:02
  • Anywhere below SNR of 30dB (31:1 ratio of RMS signal to RMS noise), you should expect to use hysteresis. At 6.2 sigma spread (1part per million occurrence), which is 6.2 x RMS, the noise is 31/6.2 or about 1/5th of the signal amplitude Do you expect a signal-processing system to behave properly, with an interferer of amplitude 20% of your "signal"?. – analogsystemsrf Feb 14 '19 at 17:42

0 Answers0