It is extremely important to gate the 38KHz modulation synchronously i.e
you always transmit whole cycles of modulation. Otherwise you will get
nasty jitter on the recovered signal.

> ----------
> From:         Martin McCormick[SMTP:martin@DC.CIS.OKSTATE.EDU]
> Reply To:     pic microcontroller discussion list
> Sent:         Monday, January 05, 1998 11:53 AM
> To:   PICLIST@MITVMA.MIT.EDU
> Subject:      Re: IR Remote control emitter
>
>         The case of the Sharp module must be _WELL_ grounded for one
> thing.
> If not, it tends to false trigger frequently.  Unless we are talking
> about
> different modules, the Sharp modules I have experimented with were
> very
> forgiving as long as they saw the near infrared light from the kind of
> LED's commonly used in remotes and the carrier frequency was
> within a KHZ or so of 32,767 HZ.  The range and quality of the
> received
> data go right down to nothing if the carrier frequency is off.  If it
> is
> way off such as 40 or 50 KHZ, you may get brief triggers from the
> Sharp
> module, but it won't follow the carrier.
>
>         In one case, I took a 33 KHZ generator I had built and gated
> it with
> a square-wave signal generator and began raising the frequency of the
> generator.  The output of the IR module tracked the square wave pretty
> well
> until I reached about 2,000 HZ at which time the Sharp receiver began
> to
> falter.  I must admit that this test was not very scientific in that I
> simply listened to the output on a speaker as I cranked up the
> frequency
> and noted the point when it got rough and noisy which occurred rather
> quickly
> above about 2 KHZ.
>
>         Most IR remotes we had lying around the house made the Sharp
> module
> respond with the kind of range typical of such things, so it shouldn't
> be
> too tricky to get it to work as long as the carrier frequency is
> right.
>
> Martin McCormick
>