Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Removing pulses from a signal

Status
Not open for further replies.
P

peleda

Advanced Member level 4
Joined
Jan 10, 2006
Messages
110
Helped
22
Reputation
44
Reaction score
12
Trophy points
1,298
Location
Monte Carlo
Activity points
1,934
I have signal that is being transmitted on a line. Because of interference, at random intervals pulses appear. That is, there is a sudden rise in the signal like a delta function.

I want to be able remove this sudden pulse from the signal.
The approach which I have so far is to include a capacitor in the line.
But I think the capacitor will also degrade the signal, when there is no pulse.

any idea on removing this pulse?????
 

I think you also may need a resistor in the line before the capacitor, and your cap should go between the signal and ground. The idea is to pass dc current and feed ac to ground. Also the frequency of the signal will determine the value of the cap. A smaller cap will pass a higher frequency of AC. Yes, I think your signal will be attenuated some, but it won't have the noise.
Hope this is of some help.
Regards,
Robert
 

You can use a clipper circuit if you know the levels of your signal.
The clipper will remove the pulses and allow the signal to pass.
 

peleda,
If you have a spec on the maximum slew rate of the normal signal, then an analog rate limiter circuit might help.
Regards,
Kral
 

This is a common problem with HF radio links. In addition to the above described cures, there is another one that is more complex and higher performance. This uses a circuit that detects the presence of the unwanted pulse. Another circuit delays the whole signal path and the pulse detector reduces the gain of the second signal path to zero while the pulse is present. The delay is required so that the pulse detector can have a higher probability of pulse detection and false alarm rejection.

45 years ago Collins used this method on their S-Line products. They had a broad band receiver covering 30-40 MHz to detect noise pulses. This worked because the atmospherics were very broad band.
 

flatulent said:
This is a common problem with HF radio links. In addition to the above described cures, there is another one that is more complex and higher performance. This uses a circuit that detects the presence of the unwanted pulse. Another circuit delays the whole signal path and the pulse detector reduces the gain of the second signal path to zero while the pulse is present. The delay is required so that the pulse detector can have a higher probability of pulse detection and false alarm rejection.

45 years ago Collins used this method on their S-Line products. They had a broad band receiver covering 30-40 MHz to detect noise pulses. This worked because the atmospherics were very broad band.
Click to expand...

Sorry to post a question on this thread, but i wanted to ask a specific question regarding flatuelnt's answer.

Hi flatulent,
You mention that there is a circuit to detect unwanted pulses. I'm trying to design a high speed integrated pulse detector in BiCMOS technology (pulses in the order of 700 ps should be detected on chip). Any ideas for such a circuit implementation or any references?

Bharath
 

The basic feature of the pulse plus desired signal is that the pulse spectrum extends beyond that of the desired signal. This was the basis of the Collins circuit. They looked at 30-40 MHz for the spectrum of the pulse and controlled the gain of a 3-30 MHz receiver.

If they share the same spectrum then your only choice is some form of time domain examination. A crude one would be the amplitude detector. A more sophisticated one is to find the cross correlation of the whole works with the pulse shape. The correlator output should be higher with the pulse present.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top