Can anybody work out the frequency of this pulse?

[Andrew]

I'm designing an infrared receiver which is to receive the following
transmission:

Light wavelength: 880nm


__ ________ _________ _____
| | | | | |
|____| |____| |____|
|-10us-| |-10us-| |-10us-|

|-2000us-| |-2000us-|

The Duty cycle is 0.005

The period equates to a frequency of 500Hz (1/2000us)

What I need to know is: what centre frequency do I need to set the
filter to on the IR receiver?

Is it 500Hz (period)?

Is it 100Khz (1/10us)?

Is it 50Khz ( 1/ (2xpulse width) )

My best guess is 50KHz because each pulse acts like half a sinusoidal
which, if complete, would have a period of 20us: 1/20us = 50KHz

I have a circuit which detects this pulse using a LM567 Tone Decoder
set to a centre frequency of 500KHz! Why??? It does works though.

I'm confused.

Thanks

Andrew

 

[Jim Backus]

On Thu, 11 Nov 2004 14:55:02 UTC, a.c.bakin@lboro.ac.uk (Andrew)
wrote:

The Duty cycle is 0.005

The period equates to a frequency of 500Hz (1/2000us)

I think another poster has answered your question by saying that there
are sufficient harmonics for the detector to work at 500 kHz.

It is well known that a square wave (1:1 duty cycle) only contains odd
harmonics. However if you have a rectangular waveform with a different
duty cycle the pattern of harmonics is different. Essentially there is
a null when the period of the on or off part of the waveform contains
a complete number of cycles. So your waveform wil have a null at 100
kHz, 200kHz etc due to the on period. It will also have nulls at 500
Hz, 1 kHz etc. As the period of the waveform is 2010 microseconds
there will be a peak at 497.5 Hz. I believe that the two dominant
frequencies will be 497.5 Hz and 50 kHz. It might be worth you hooking
the waveform up to a spectrum analyser. I've used a spectrum analyser
to set up exact duty cycles based on this principle.

--
Jim Backus OS/2 user since 1994
bona fide replies to j <dot> backus <the circle thingy> jita <dot>
demon <dot> co <dot> uk

 

[Andrew]

Petrus, John, Jim

Thank you all for your replies- very helpful.

I'll try setting up my filter to 497.5Hz as recommended. I'll also use
a spectrum analyser as you suggested Jim.

Thanks again

Andrew

 

[Luhan Monat]

Andrew wrote:

I'm designing an infrared receiver which is to receive the following
transmission:

Light wavelength: 880nm


__ ________ _________ _____
| | | | | |
|____| |____| |____|
|-10us-| |-10us-| |-10us-|

|-2000us-| |-2000us-|

The Duty cycle is 0.005

The period equates to a frequency of 500Hz (1/2000us)

What I need to know is: what centre frequency do I need to set the
filter to on the IR receiver?

Is it 500Hz (period)?

Is it 100Khz (1/10us)?

Is it 50Khz ( 1/ (2xpulse width) )

My best guess is 50KHz because each pulse acts like half a sinusoidal
which, if complete, would have a period of 20us: 1/20us = 50KHz

I have a circuit which detects this pulse using a LM567 Tone Decoder
set to a centre frequency of 500KHz! Why??? It does works though.

I'm confused.

Thanks

Andrew

Sorry, you can'f filter this signal - its too wide. Maybe just a high
pass filter based on the lowest frequency component.

Luhan

 

[Pere Pala]

Andrew wrote:

I'm designing an infrared receiver which is to receive the following
transmission:

Light wavelength: 880nm


__ ________ _________ _____
| | | | | |
|____| |____| |____|
|-10us-| |-10us-| |-10us-|

|-2000us-| |-2000us-|

The Duty cycle is 0.005

The period equates to a frequency of 500Hz (1/2000us)

What I need to know is: what centre frequency do I need to set the
filter to on the IR receiver?

Is it 500Hz (period)?

Is it 100Khz (1/10us)?

Is it 50Khz ( 1/ (2xpulse width) )

My best guess is 50KHz because each pulse acts like half a sinusoidal
which, if complete, would have a period of 20us: 1/20us = 50KHz

I have a circuit which detects this pulse using a LM567 Tone Decoder
set to a centre frequency of 500KHz! Why??? It does works though.

I'm confused.

Thanks

Andrew

Could it be that your data stream is actually modulating a 500KHz
subcarrier? This is usually done to avoid interference from other light
sources.

--
Pere Pala

e-mailere-place-a-dot-here-pala "at" upc-another-one-here-es