Maker Pro
Maker Pro

How to filter noise when it has the same frequency as the signal

R

Richard Dungan

Jan 1, 1970
0
X.Y. said:
It seems that they just has a difference on the magnitude.

P.S. It is a OOK signal on receiver.

The wave on oscillograph can be found on the following link:

http://att.newsmth.net/att.php?s.172.248797.578.png

Thanks in advance!

Well, I'm lurking while people are helpful with my UL testing post, but
I just know I'm going to regret helping with this.

I assume that OOK is on-off keying or somesuch.

I would venture that whatever is producing the signal is being slightly
overdriven, given the possible clipping on the negative peaks.

Rectifier, preferably full wave
Smoothing/lowpass filter
Schmitt trigger/threshold detector

Homework?

Richard
Richard D.

Spamtrapped at ngsATradixDASHdesignDOTcoDOTuk
www.radix-design.co.uk
 
M

MooseFET

Jan 1, 1970
0
It seems that they just has a difference on the magnitude.

P.S. It is a OOK signal on receiver.

The wave on oscillograph can be found on the following link:

http://att.newsmth.net/att.php?s.172.248797.578.png

Thanks in advance!

I assume that the signal is at a known frequency and that the
modulation rate is a small fraction of the signal frequency.

If you know that the transmitter has an oscillator in it that is at a
constant frequency, you can reconstruct the transmitter's clocking
rate with a PLL at the receiver.

You can use the output of the PLL as the carrier in the detection
stage. This will give you a significant advantage against the random
noise. The sync detection method reduces noise power by a factor of
2. The filtering you can do after detection further reduces the
noise.


A cheap method is to rectify, band pass filter the rectified output,
AGC the AC amplitude, DC baseline restore and then use a comparitor.
This puts the noise near zero volts and the signal at some positive
voltage. It assumes that the SNR is large enough that the signal
jumps up out of the background.
 
Top