up vote 4 down vote favorite
1

I made a bread board, this board is intended to measure AC 50Hz signal from sub uV to hunderds of mVs. The entire system is floating, powered from a voltage transformer. The front end is like this, P1 is a BNC connector, and cascaded PGA204 and PGA205 to amplify the signal, and with a 100Hz active low-pass filter following (the filter gain about 2.5, not shown), all my resistors (including the ones used in the active filter) are 0.1% metal-film resistors:

enter image description here

This is the PCB

enter image description here

I noticed yesterday when i shorted the input and probe the amplified signal with 1mV/div with gain x1, the output of the signal is this

enter image description here

When i change the gain to x100, this signal is like this. It seems the noise referred to input about 2uV

enter image description here

When i put my finger near to the Op Amp (used in the active filter) or the input resistor, about 3cm or so, the output is this

enter image description here

And, when i touch the resistor or the Opamp, this situation is worse.

I'v never noticed this situation before. It's apparent there some AC power line noise coupled into the circuit, i want to know

  1. It's this 'normal'?
  2. I wonder if my front end design has some problem, is it ok to use the 'bulk' metal-film resistor?
  3. If the resistor usage is ok, what about my PCB layout, what should i do to minimize the noise.
  4. Are there some 'all-in-one' solution to get rid of the noise?
share|improve this question
  • It could very well just be the probe picking up AC hum from your surroundings. Your probe acts like an antenna, and for that matter so does your circuit. Your body picks up this hum and when you put your finger near your circuit or probe, a very small current is induced, which is visible on a scope. – DerStrom8 Aug 31 '14 at 2:03
  • @derstrom8 You mean all are 'fake' caused by the scope's probe? The circuit or the layout can't cause this by itself ? If this is real, that's a good news to me. Can you give much proof. Thanks. – diverger Aug 31 '14 at 2:15
  • @derstrom8 BTW: If that small current can cause the scope to detect them, then my circuit will detect it too, right? I have an AD in the back end. If all are caused just by the 'probe', and when i remove the probe, the noise disappears, it's a good news. – diverger Aug 31 '14 at 2:23
  • @diverger, It looks to me like you have quite a lot of higher-frequency noise there, too. Is a fluorescent lamp running anywhere nearby? That might explain the rest of your noise. – TDHofstetter Aug 31 '14 at 3:07
  • Yes, we did have a fluorescent lamp on one working bench, but the bench is from the board about 5M, and i forget if it's on at that time. And the board has a UPS and two LCD screen on near by when i take the measurement, this can make some difference? – diverger Aug 31 '14 at 3:17
up vote 3 down vote

Yesterday I was experimenting with Rigol ds1052e myself and yes you WILL see 50Hz noise with probe without even touching anything. Try to connect ground alligator to the tip of the probe and you will pick up even more noise. This goes away when you unplug probe from BNC. So yes, this is from the probe itself, not necessarily from your circuit. I've seen people suggesting to use coiled piece of wire around ground barrel on the probe, rather than attaching ground alligator. See this link for details http://e2e.ti.com/support/power_management/simple_switcher/w/simple_switcher_wiki/2243.understanding-measuring-and-reducing-output-voltage-ripple.aspx

Also check if your scope has bandpass filter. It might help.

share|improve this answer
  • The important thing is to keep the ground lead as short as possible. Your probe is shielded except for the tip, and that's where the signals are picked up. – DerStrom8 Sep 1 '14 at 13:44
  • I've tried the coiled ground barrel, the high frequency noise get better. But the 50Hz noise is there still. Please check the PCB, it seems the two input resistors and the INA comprise a 'ring', may this cause the noise? – diverger Sep 1 '14 at 13:56
  • @user40729: My signal to measure is 50Hz, so i can't use bandpass filter to filter the 50Hz. I need to make sure when i short the input, the output is 0, at least not including the 50Hz noise. – diverger Sep 1 '14 at 14:02

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.