How to measure GND noise?

Go To Last Post
10 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I use oscilloscope to measure GND noise level,of course, i am silly, i use the short probe and the long probe connected with GND which i try to measure the noise level, and it shows me about 40mV. i think this method must be wrong, but i don't know how to measure the noise level.
why i have to measure the GND noise level?
because i use ADC which refer to GND, if the GND noise is too big, i know my ADC result must be wrong.
My PCB use 50mil GND line, and GND covers large copper to prevent noise, by the way , digital GND and analog GND separated.
but My oscilloscope seems there is about 40mV in my GND

Our Martians are beginning to learn AVR

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Shouldn't the digital and analogue grounds be connected? I get good results with the on-chip ADC by filtering the analogue supply and using a common digital and analogue ground.

Leon

Leon Heller G1HSM

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Measuring noise with an oscilloscope is difficult at the best of times. Tektronix have a pcb mount receptacle that mates with their probes to get the best possible measurement. Usually the ground lead on the 'scope probe acts as an antenna and picks up noise giving a false reading, so you want to minimise the probe earth distance - this is what the Tektronix receptacle does.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

read a fantastic quote once about how ground was a convenient lie dreamed up by engineers to make understanding circuits easy. wish i could remember it, it was expressed perfectly.

having lots of ground area might increase the amount of noise you pick up from your surroundings though.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

yyw794 wrote:
I use oscilloscope to measure GND noise level
.....
because i use ADC which refer to GND, if the GND noise is too big, i know my ADC result must be wrong.

Greetings (No Name Supplied),
If you oscilloscope has two channels, and
one can be inverted, and a 'sum' setting
on the Y channel, you could measure the
difference signal from your ADC input (CH.1)
and the ADC's gnd ref (Ch.2). The ADC doesn't
care about the noise elsewhere in your PCB.

Comments Welcome!

Peter

--
Peter J. Stonard
www.stonard.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I remember some Analog Devices Application Note, where simple circuit with instrumentation amplifier was used to measure GND noise. Unfortunately I was unable to find it, but here are few links about proper grounding:

http://www.analog.com/analog_roo...
http://www.analog.com/library/an...

Best regards,
Gintaras

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
Unfortunately I was unable to find it,

This app note?
http://www.analog.com/analog_root/static/raq/moreInfo/groundNoiseMeasurement.pdf

Stan

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The signal on ground is most likely not real noise, but a high frequency signal coming from the controller or may be a switched power supply. Frequencies are typicaly to high to measure with a common instrumentation amplifier. I use the normal scope probe. Connect the scope's ground somewhere an move along the lines. A test at the point where the ground is connected gives an estimate on how much the loop recieves as an antenna. If there is significan noise, this is usually enough to see where the source of ground noise is. 40 mV signal on the ground is ususally a sign for a bad layout.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Greetings Stan,

Your link is to a differential input
adapter for use with a scope. The
scope can do this without additional
hardware (if it has a dual input with
SUM and Ch.2 INVERT features).

yyw794 wrote:
My oscilloscope seems there is about 40mV in my GND

Although the adapter in your link has
much greater sensitivity (Av = 1000),
the OP believes there is 40mV pk-to-pk
noise using their measurement technique.

I suspect that by reconfiguring the
scope for differrential input measurement
the OP will get a much different reading.

The greater sensitivity of the adapter is
not needed, and the scope's Y channel
offers full bandwidth down to DC (unlike
the adapter which is based on a bandwidth
limited IC op amp).

This is a very simple bench check and would
give us all a new data point to consider.

Comments Welcome!

Peter

--
Peter J. Stonard
www.stonard.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thank you Stan, I was talking about this Application Note.

To Peter - I am in a big doubt, that two channel oscilloscope, configured to show A-B, can do this job. Such configuration CMRR (Common Mode Rejection Ratio) is very low. That is why instrumentation amplifiers are used for and why differential probes exist. I am using such - Tektronix P6046, so oscilloscope pictures with differential probe and A-B configuration are completely different.

Best regards,
Gintaras