Noise reduction

Thread Starter

cb83

Joined Aug 9, 2007
24
Does anyone know any good noise reduction techniques for a high dynamic range signal?

I bring the signal into my system on a 50 ohm BNC cable. I've been using filters, RF sheilding and even a log amp to improve the resolution before an ADC. Signal averaging isn't an option for my application.

Any ideas?
 
The best noise reduction technique is to divorce her!!:D:D

Just kidding.......

The best noise reduction technique is to increase the voltage of the signal you want at the source.....and take it above the noise floor.

It isn't any simpler than that!:cool:
 

John Luciani

Joined Apr 3, 2007
477
Does anyone know any good noise reduction techniques for a high dynamic range signal?

I bring the signal into my system on a 50 ohm BNC cable. I've been using filters, RF sheilding and even a log amp to improve the resolution before an ADC. Signal averaging isn't an option for my application.

Any ideas?
You need to determine the source of your noise before you can reduce it.
Are you using a 50 ohm cable because you are doing high frequency work?
Are your impedances matched? If you are working with low frequency signals
a shielded twisted pair may work better.

What does the noise on your power supply look like? How are you doing your
grounds? Are you analog signals near high frequency digital signals?
How are you measuring the noise? If you are measuring the noise by connecting
your scope to ground with probe ground wire you are probably getting erroneous
readings. See my picture at http://www.luciani.org/eng-notes/ee-notes/ee-notes-index.html

There is a very good application note about grounding and decoupling
by Paul Brokaw called "An IC Amplifier User's Guide to Decoupling,
Grounding,. and Making Things Go Right for a Change". It is at

(* jcl *)
 

Thread Starter

cb83

Joined Aug 9, 2007
24
It is high frequency work so I need the 50 ohm BNC. I am also well impedance matched. My anolog signals are also well isolated.

I've been thinking it could be my grounds and power supply so I used an isolation transformer, but didn't see too much improvement. I'll look into that article because I'm sure I can do a better job decoupling.
 

beenthere

Joined Apr 20, 2004
15,819
It's hard to give advice on the data given. The frequency is somewhat high, and analog in nature, but what is the freq, and what is the nature of the noise?
 

John Luciani

Joined Apr 3, 2007
477
It is high frequency work so I need the 50 ohm BNC. I am also well impedance matched. My anolog signals are also well isolated.

I've been thinking it could be my grounds and power supply so I used an isolation transformer, but didn't see too much improvement. I'll look into that article because I'm sure I can do a better job decoupling.
How are the analog signals isolated? Do you have a ground plane between your
analog and digital circuits to prevent radiated noise from coupling to your analog
lines?

If you do not know the source of the noise why would you think an isolation transformer
would do anything?

Start by measuring the power supply noise. Use either test points or a Johnson
jack to connect to the power supply. Also make sure your scope probe impedance matches your scope and the probe is calibrated.

If possible you should post a schematic and picture of your layout. Also more
detailed information would be useful. Are you doing 1MHZ or 1GHz work?
Are you trying to measure nV, uV or mV?

I thought of another resource for high-frequency design ---

High Speed Digital Design: A Handbook of Black Magic
by Howard W., Ph.D. Johnson, Martin, Ph.D. Graham
Prentice Hall; ISBN: 0133957241

I believe the book may discuss noise as well.

(* jcl *)
 

Thread Starter

cb83

Joined Aug 9, 2007
24
The analog signals are in a RF sheilded box a foot or two from everything else and transferred on coax. The signal bandwidth is from DC to about 40MHz.
The noise is on the order of 1mV and I wanted to get measurements near 100uV.
 

John Luciani

Joined Apr 3, 2007
477
The analog signals are in a RF sheilded box a foot or two from everything else and transferred on coax. The signal bandwidth is from DC to about 40MHz.
The noise is on the order of 1mV and I wanted to get measurements near 100uV.
What is the signal level? Are you getting 1mV of noise on a 10mV signal or 1mV of noise
on a 10V signal?

How are you powering the electronics in the RF shielded box?
Are you using a switching power supply? Is the power supply inside or outside of
the box? What is the noise on the power supply lines?


(* jcl *)
 
Top