Re: Buffering prlink cables

From: MagerValp (MagerValp_at_Goth.Org)
Date: 1998-03-25 17:14:18

>>>>> "Marko" == =?ISO-8859-1?Q?Marko M=E4kel=E4?= <ISO-8859-1> writes:

MV> As I've already toasted one computer with prlink (my A1200 with
MV> surface mounted CIAs :P) , I thought I'd try to buffer my prlink88
MV> cable before I toast something irreplacable.

Marko> First a question: Is your Amiga properly grounded?

Yes, all my computers are connected to the same ground. I don't have a
death wish :)

MV> So a question for Marko and Olaf: Does PA2 reflect the direction of
MV> the data lines?

Marko> See the protocol description in the README file, I don't
Marko> remember.

That very description is what makes me ask :) It says:

  6.3 The prlink cable, true 8-bit protocol (Amiga only)
  
  This is the fastest protocol that the prlink package supports.  It uses two
  handshake signals and eight data lines.
  
  6.3.1 Sending a byte to the server
  
  Ami: put the data bits 0..7 on PB0..PB7 and drop -FLAG
  C64: wait for the falling edge of -FLAG
  C64: read the data bits, change the state of PA2
  Ami: wait for PA2 to change its state
  Ami: raise -FLAG
  
  6.3.2 Receiving a byte from the server
  
  Ami: drop -FLAG
  C64: wait for the falling edge of -FLAG
  C64: put the data bits 0..7 on PB0..PB7 and toggle PA2
  Ami: wait for PA2 to change its state
  Ami: raise -FLAG, read the data bits
 
  6.3.3 Changing the data direction
 
  As the protocol uses bidirectional data lines, the data direction of these
  lines must be reconfigured when changing the data direction.  This is managed
  on the server side.  The client side can always call input and output
  routines without any special measures.
 
  When the data flow direction changes on the server, the server first waits
  for the falling edge of -FLAG, then reprograms the lines' data direction and
  finally transfers the byte in normal manner.

and I'm not really sure how to interpret that. Hmm, a look at the
source code might help.

Marko> Meanwhile, you could use the 4-bit protocol and maybe modify it
Marko> so that one line (controlled by the Amiga, so that you don't
Marko> need to change the PET side program) reflects the data
Marko> direction if it doesn't do it already.

Unfortunately I don't have the necessary tools (ie a C-compiler) on my
Amiga.

-- 
    ___     .              +   .       .              .   Per Olofsson
 ._|___|_,       .     .  .         .       . +  .         .      o   
    o-o     +   .  +  MagerValp@Goth.Org .   .     +     +     +   .  
     -     .      .     .   o         +              .     .     .    
     "       +            +    .     http://www.cling.gu.se/~cl3polof/

Archive generated by hypermail 2.1.1.