160 baud ALDL data - reading on a PC, counter 0?

[Teller.John at orbital.com]

Just remember that a PC is a non-real time machine.  Your interrupts may be held
of for an indeterminate period due to some chunk of Windows/DOS having suspended
them to do something trivial (like write to the hard drive or something).  Don't
be surprised if occasionally the edge-triggered PIC work-alike in your PC
sometimes misses an interrupt or two!

The clock to that timer is based on the original PC's 14.XXXX clock which was
divided down to feed the CPU clock and video scan generator.  It has been a
while since I knew the exact numbers for all of that, but I'm not surprised your
measurements disagreed with the manual.

--- John


                                                              
                                                              
                                                              
 To:      gmecm at efi332.eng.ohio-state.edu                     
                                                              
 cc:      (bcc: John Teller/ORBVA)                            
                                                              
                                                              
                                                              
 Subject: 160 baud ALDL data - reading on a PC, counter 0?    
                                                              




Techo question for PC savvy types...

Two weeks ago I asked about the 160 baud ALDL data format.
I've since written a bit of code that uses the serial port's
CTS line as input, and DTR as a negative voltage output to
condition, with a single transistor, the ALDL signal. The code
simply uses CTS interrupts (modem change) and reads the PC's
8253/8254 counter channel 0 to generate a timestamps at each
CTS transition. It works well, but...

Does anyone know what frequency the AT clocks counter 0 at?
All my documentation says it's at 0.84 uSec (micro seconds)
BUT my measurements on a couple of Pentium and 486 machines
indicate that it's closer to 0.43 uSec AND that it counts
in lots of 2 (ie. reading counter 0 always reads an even
binary number). Now I know XTs and early ATs did clock counter
0 at 1.19 MHz, but when did they change to closer to 2 Mhz, or
am I losing the plot?

I'm going to "tidy up" this bit of code and make it available
to show the raw 160 baud ALDL data stream in real time. I'm
not going to try and decode the data stream because I'm going
to give away this code to whoever wants it. It'll be able to
run on an XT AND it should also run fine in a DOS box too. Any
thoughts on the best way to display this raw data?

regards,
--
Peter Gargano