True, for ordinary timers.  But, for Win95 and WinNT there
are multimedia timers that can resolve much better than
50mS. Multimedia timers are required for accurate timing
for MIDI data.  Also, in WinNT there are performance timers
which can resolve to 50uS or so.  I've used them to get
reasonable measurements to 200us.  Also, the ordinary
printer ports drivers do not use interrupts, but are polled
after a timer event.  IE, you tell your machine to print,
then the following happens:
1.  Set timer interrupt.
2.  See if port can accept data on timer interrupt.

So, what I did was set a performance timer interrupt for 100us
and check the port for data on each interrupt.
It was a fast machine, a dual PPro..

Regards
              \\\|///
            \\  ~ ~  //
             (  @ @  )
+----------oOOo-(_)-oOOo---------+
|                                |
|          Robert Zeff           |
|        rzeff@Nikola.com        |
|        http://Zapco.com        |
|        http://Nikola.com       |
|        ^^^^^^^^^^^^^^^^^       |
| Free circuit simution software |
|                                |
+----------------Oooo------------+
          oooO   (   )
         (   )    ) /
          \ (    (_/
           \_)


----------
> From: sdavidson@ITS.BLDRDOC.GOV
> To: PICLIST@MITVMA.MIT.EDU
> Subject: Re: How fast can the PC handle interrupts from LPT1
> Date: Thursday, February 27, 1997 5:26 PM
>
> Not sure how fast the PC can respond, but for timing the resolution of
the PC
> clock under Windows is about 50ms.  I would suggest using a 20MHz PIC
timing
> the pulses and passing the data to the PC.
>
> Steve
> --- On Wed, 26 Feb 1997 13:00:11 -0600  Norm Cramer <cramer@DSEG.TI.COM>
> wrote:
> I am trying to measure some pulse widths using the PC.  I would like to
use
> the printer port interrupt as my timing input.  How fast can the PC
handle
> the interrupts (i.e. can it handle them at about a 50usec interval?)
This
> will be under DOS.  Also how long does the interrupt pulse need to be to
> get seen.  (i.e. would a 2usec pulse every 50usec get seen by the PC?)
>
> The PIC relation is that the pulses are being generated by a PIC.
>
> BTW thanks to all who helped with the PC timing info.
>
> Thanks,
>
> Norm
>
>
> -----------------End of Original Message-----------------
>
> -------------------------------------
> E-mail: sdavidson@its.bldrdoc.gov
> Steven Davidson
> Dept. of Comm.  NTIA-ITS.N2
> 325 Broadway
> Boulder, CO  80303
> W 303-497-3411  FAX 5995
> -------------------------------------