4
\$\begingroup\$

I measure DC pseudo-differentially over 1-2 m wires. In order to avoid any voltage drop in wires the built in ADC input buffers are used, thus there almost no current flow. I expect the wires to catch a lot of common mode noise from RF to 50 Hz power grid.

Most materials about common mode noise that I found, consider it in the context of power line that involve common-mode chokes. Howver, I am not sure if this is suitable for my measurement.

Another way is the vanilla common mode filter

enter image description here

But again, I am not sure if it would be good enough and provide adequate level of filtering.

Moreover, there might be more methods that I am not aware of.

What possibilities are there to suppress common mode noise, and what is the best method for a measurement application?

In addition, how much the application of a shielded twisted pair cable would help to ease the problem?

\$\endgroup\$
4
  • 1
    \$\begingroup\$ First step is using twisted pair cable with overall screen well grounded at the more substantial (lower impedance ground) end. Sometimes it is an option to move the ADC to the sensor location and then move the more noise tolerant digital data over the longer distances. \$\endgroup\$
    – KalleMP
    Oct 25, 2016 at 22:05
  • \$\begingroup\$ What is your source, and how good is its impedance balance? \$\endgroup\$
    – ThreePhaseEel
    Oct 25, 2016 at 22:55
  • \$\begingroup\$ Use twisted pair like everyone is saying. Also, terminate the differential signal at the receiver (ADC input). If the source cannot drive the terminated load, use a buffer on the sending side that can. Using terminated transmission line is like magic for signal integrity. \$\endgroup\$
    – user57037
    Feb 12, 2017 at 1:13
  • \$\begingroup\$ @Andrey Pro: Why do you ask only about ways to get rid of common mode noise? You should also be afraid of differential mode noise (e.g. by inductive coupling via changing magnetic fields)! See my answer below. \$\endgroup\$
    – Curd
    Aug 3, 2017 at 10:02

3 Answers 3

Reset to default
2
\$\begingroup\$

You have to take care of two kinds of interference:

  • common mode interference caused by capacitive coupling
  • differential mode interference caused by inductive coupling

Capacitive coupling is already handled well by shielding and by using a differential pair of signals because it affects bouth lines about the same.

For handling inductive coupling using a twisted pair is very important.
If wires are not twisted it is easy to induce quite a large voltage (acting like a AC voltage source in series of the signal source) caused by changing magnetic fields (possible sources are transformers, speakers, SWPS, cellphones... any AC currents). Twisting the differential pair reduces the intereferences as every other twist adds a positive/negative intereference to the signal ideally canceling out at the end (see Image below).

interference by induictive coupling in non-twisted vs. twisted pair wires

  • The upper part shows a non-twisted wire pair: interfernce by induction is large (proportinal to blua area).
  • The lower part shows a twisted wire pair: interference by induction is small (or even zero) as positive (+) and negative (-) interferences cancel out (partially).
\$\endgroup\$
0
\$\begingroup\$

I highly recommend trying a shielded cable. Using unshielded cables picks up a lot of noise from AC power lines, especially in audio applications. (I once made an unshielded guitar cable from scratch in high school, and you could definitely hear the 60Hz hum of the mains AC)

Also, make sure that the shielded cable is grounded properly, as this will make a big difference in regards to noise.

\$\endgroup\$
0
\$\begingroup\$

Using a good, properly shielded twisted pair cable and avoiding ground loops can really make a difference. Take a look at this on how to break ground loops: Shielding, Grounding and Ground loops.

\$\endgroup\$
1
  • \$\begingroup\$ "Oops! That page can’t be found." \$\endgroup\$
    – Navin
    Nov 16 at 10:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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