#
# $Id: PETdoc $
#
# This file is part of Commodore PET emulator
#      and Program Development System.
#
#
#   Copyright 1995 Olaf 'Rhialto' Seibert. All Rights Reserved.
#   You are not allowed to read this using any kind of Micro$oft product.
#
#
# This file contains some documentation for Commodore PET systems.
#
#
# Written by
#   Olaf Seibert     (rhialto@mbfys.kun.nl)
#   Jouko Valta      (jopi@stekt.oulu.fi)
#   Andre Fachat     (a.fachat@physik.tu-chemnitz.de)
#
# Thanks to
#   Brian Heyboer    (bjheyboer@space.honeywell.com)
# And
#   Mitch B. Parker  (mbpark@planetx.bloomu.edu)
#   Dave Tickle      (aiu214@freenet.mb.ca)
#   RICHARD WHITE    (richwhite2@mci.newscorp.com)
#   Veit Laule       (UKAH@ibm3090.rz.uni-karlsruhe.de)
#   Byron Gracey     (bgracey@freenet.niagara.com)
#   Dale DePriest    (daled@cds9172.Cadence.COM)
#   David Evans      (devans@ccubb.com), (dfevans@bbcr.uwaterloo.ca)
#   Martijn van Buul (martijnb@stack.urc.tue.nl)
#   Michael Brown    (mbrown@ersys.edmonton.ab.ca)
#   Simon Rowe       (srowe@ftel.co.uk)
#   Tracy M Nelson   (tnelson@fluorite.telesciences.com)
#   Dorian Garson    (doriang@microsoft.com)
#
#
# $Log$
#
#


General PET info
----------------

When talking about the PET 30xx, 40xx, or 8032, a lot of different 
models, hardware, software and their combinations are concerned.

The early models have almost everything besides the "big" I/O
chips and the CPU in separat ICs, especially the complete video circuitry.
Those machines were sold as the "PET 2001" with built-in tape recorder.
It featured "snow" on the screen when the CPU wrote to the video
memory that was driven at the same clock as the CPU.
The first board was replaced with a new revision that had a faster
video RAM and no more snow. Those models had a software from the initial 
- very buggy - version 1 ROMs over the "wait 6502,2" version 2 ROMs 
(AKA version 3 ROMs) up to the Basic 4. 
Those machines have been sold as "2001-8N" (replace "8" with the
RAM size and (probably?) "N" with a letter indicating the keyboard)
in the US. In Europe they have been sold as 2001, 3008-3032 with Basic 2, 
and (I am not certain of this) as 4016/4032 with Basic 4.
Many of the older machines are said to have been upgraded to Basic 4.

After this board then came the models with the 6545 CRTC controller.
Some of those models are vulnerable to the "PET killer poke".
Those boards were available with 1k video RAM and 40 columns 
(AKA "Fat-40") as well as with 2k video RAM and 80 columns.
The 80 columns were achieved by shifting video address lines,
an additional latch and doubling the pixel clock. Therefore the
video setting still indicates 40 columns in a 8032!
All of those machines had Basic 4.

For all those machines no memory management is involved. With the 
8096 and the 8296 Commodore then introduced some kind of bank switching 
to handle up to 128kByte RAM. The 8096 used an expansion board while the 
8296 had 128k directly on board.

Another machine with an expansion board was the SuperPET 9000
(AKA MicroMainFrame 9000 in europe). It featured a 6551 ACIA,
64k additional RAM (not 8x96 compatible) and a 6809 (!) CPU
on the expansion board.

Above the screen area at $8000, there are two (Basic 4.0) resp. three 
(older PETs) 4 kByte ROM areas that could be used for a ROM expansion. 
In hardware, there are two (three) ROM sockets on the motherboard where one 
can put the expansion ROMs. The sockets map to $9xxx, $Axxx, and $Bxxx,
Above $B000 (Basic 4.0) resp. $C000 the system ROM starts. 
Older PETs have the I/O area, where the interface chips are mapped,
between $e800 and $f000. Some newer PETs, starting with some 
8032 models, have the I/O area between $e800 and $e900 only. This gave
more place for an extended (nationalized) screen editor.

The PET came with several different keyboard layouts (see below). 
The most common are the graphics keyboard used on the 3032 and the
4032 and the business keyboard as used for the 8032. 

The normal PET have - not counting the CRTC - three I/O chips.
Two PIA 6520/6820/6521 (they are compatible to each other) are used
for keyboard resp. IEEE488 interfacing. The VIA 6522 provides a
timer and a shift register - which is used to play 'music' on newer
PETs. The standard IRQ that is used for clock incrementing, cursor
blinking and keyboard input is not procduced by a timer on any of those 
chips, though. Instead a vertical retrace signal is generated directly 
by the hardware. This is then fed into a PIA1 CB1, which is set to 
trigger an interrupt.

Some of the newer PETs are reported to have an ACIA 6551 serial port
interface chip. This may be the SuperPET 9000 as well as
the business models CBM 600/610/700/710. The latter are
not covered here. as they are not PET compatible :-)

Andre

------------------------------

The following is from the PET FAQ http://www.goldrush.com/~foxnhare/petfaq.html
and covers the PET joystick handling more to detail.
This also coincides with the article about joysticks below, if you 
say that the port is VIA PA, which is connected to user port pins C-L.
								Andre

I HAVE A PET PROGRAM THAT CAN USE JOYSTICKS, HOW DO I MAKE AN INTERFACE?

  The PET was not initially designed for joysticks but a determined user had
  devised a dual-joystick interface employing the parallel port.  This interface
  allows you to plug in standard Atari/Commodore compatible joysticks....

  Interface Diagram  (use mono-spaced font to view properly)

  Looking at back of user port         1  1  1
            1  2  3  4  5  6  7  8  9  0  1  2
            =  =  =  =  =  =  =  =  =  =  =  =
           ### ######################### ######
            =  =  =  =  =  =  =  =  =  =  =  =
            A  B  C  D  E  F  H  J  K  L  M  N
            !     !  !  !  !  !  !  !  !     !
            !     !  !  !  !  !  !  !  !     !
  port 1->  8     1  2  3  4  !  !  !  !     !
                              !  !  !  !     !
                   port 2->   1  2  3  4     8


  Joystick pin diagram =>  5  [4] [3] [2] [1]
  looking at back of                    \ /
  male connector             9  [8]  7  [6]

  The 'fire' button (pin 6) is added by connecting joystick port pins 1 & 2
  to pin 6 with diodes, solder each diode from (already wired) pins 1 & 2
  to pin 6 (with the banded ends to pin 6).


------------------------------

I/O lines
---------

PIA1: $e810
	Port A: Bit 0-3 - keyboard row select lines (decoded with a 4 to 10
			  bit decoder)
		Bit 4   - Cassette #1 sense 
		Bit 5   - Cassette #2 sense 
	        Bit 6   - IEEE488 EOI input line
		Bit 7   - Diagnostic sense input line. Pull it low and
			  the PET comes up in the builtin machine language
			  monitor. Bit 5 on the userport.
	CA1		- Cassette #1 read line
	CA2		- very old PETs can blank the screen with it. Newer
			  ones have this as IEEE488 EOI output.
        Port B: Bit 0-7 - Contents of Keyboard row
	CB1		- screen retrace detection in.
	CB2		- Cassette #1 motor on

PIA2: $e820
	Port A:	Bit 0-7 - IEEE488 Data 1-8 Input Lines 
	CA1		- IEEE488 ATN in
	CA2		- IEEE488 NDAC out
	Port B: Bit 0-7 - IEEE488 Data 1-8 output lines
	CB1		- IEEE488 SRQ in
	CB2		- IEEE488 DAV out

VIA: $e840
	Port A: Bit 0-7	- Userport
	Port B:	Bit 0	- IEEE488 NDAC in
		Bit 1	- IEEE488 NRFD out
		Bit 2	- IEEE488 ATN out
		Bit 3	- ???
		Bit 4	- Cassette #2 Motor on
		Bit 5	- vertical retrace detect
		Bit 6   - IEEE488 NRFD in
		Bit 7   - IEEE488 DAV in
	CA1		- ???
	CA2		- screen character ROM switch (graphics/text)
	CB1		- ???
	CB2 (SR out)	- Shift register plays music on userport

------------------------------

 The PET 3032,4032 Keyboard Matrix -- Graphics Keyboard

  e810  e812
   0	Right	Home	_	[	&	%	#	!
   1	BSpace	Down		]	|	\	$	
   2	9	7		o	u	t	e	q
   3	/	8		p	i	y	r	w
   4	6	4		l	j	g	d	a
   5	*	5		:	k	h	f	s
   6		1	Return	m			
   7	+	2		?	,	n		x
   8								
   9	.	.	.	.	.	.	.	.


 Values	128	64	32	16	8	4	2	1
	 7F	BF	DF	EF	F7	FB	FD	FE


  The PET 8032 Keyboard Matrix  -- Business Keyboard

  e810  e812
   0    <05>	Right		KP_8	-	8	5	2
   1	KP_9	UpArr		KP_7	0	7	4	1
   2	KP_5	\	k	;	h	f	s	Esc
   3	KP_6	[	l	Return	j	g	d	a
   4	Del	p	i	@	y	r	w	Tab
   5	KP_4	]	o	Down	u	t	e	q
   6	KP_3	Shift_R	Down	KP_Dot	.	b	c	Shift_L
   7	KP_2	Repeat		KP_0	,	n	v	z
   8	KP_1	/		Home	m	Space	x	RVS
   9	<16>	<04>	: *	Stop	9	6	3	<-


 Values	128	64	32	16	8	4	2	1
	 7F	BF	DF	EF	F7	FB	FD	FE



Here is the 8032 keyboard matrix, as it appears on the keyboard connector.
Note that the columns are scanned in the opposite order.

       A        B        C        D        E        F        G        H
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1  |   2        5        8        -        8*     Right     {14}     {5}
2  |   1        4        7        0        7*       ^       {6}       9*
3  |  ESC       S        F        H        ]        K        ;        5*
4  |   A        D        G        J       Return    L        @        6*
5  |  Tab       W        R        Y        \        I        P       Del
6  |   Q        E        T        U       Down      O        [        4*
7  |  Shift_L   C        B        .        .*      {25}    Shift_R    3*
8  |   Z        V        N        ,        0*      {15}     {16}      2*
9  |  Rvs       X      Space      M       Home     {21}      /        1*
10 |   <-       3        6        9       Stop      :        {4}     {22}




 The PET 8032 Business Keyboard Matrix

  e810	e812
   0	<-	3 #	6 &	9 )	Stop	: *	<04>	<16>
   1	RVS	x	Space	m	Home	<15>	/ ?	KP_1
   2	z	v	n	, <	KP_0	<0F>	Repeat	KP_2
   3	Shift	c	b	. >	KP_Dot	<19>	Shift	KP_3
   4	q	e	t	u	Down	o	]	KP_4
   5	Tab	w	r	y	@	i	p	Del
   6	a	d	g	j	Return	l	[	KP_6
   7	Esc	s	f	h	; +	k	\	KP_5
   8	1 !	4 $	7 '	0	KP_7	^	<06>	KP_9
   9	2 "	5 %	8 (	- =	KP_8	Right	<0E>	<05>


 Values	128	64	32	16	8	4	2	1
	7F	BF	DF	EF	F7	FB	FD	FE

    Unlike normally, ^ cannot be shifted.

 03 = Stop


----

From the PetIO.doc text file (v1.2 of 08.01.95 by Olaf Seibert) 
from ftp.funet.fi:

64 K RAM expansion in 8096 and 8296
-----------------------------------
From: UKAH@ibm3090.rz.uni-karlsruhe.de (Veit Laule)
Newsgroups: comp.sys.cbm
Date: Tue, 22 Nov 1994 16:38
Organization: University of Karlsruhe, Germany
Message-ID: <19941122163828UKAH@ibm3090.rz.uni-karlsruhe.de>

8096 exp-mem (64K):
The control register is at $FFF0/65520
You have 4 16K-banks, 0...3

$8000     $9000               $C000           $E800    $F000     $FFFF
!----------------------------!!--------------------------------------!
        Bank 0 or 2                     Bank 1 or 3
!--------!                                    !-------!
  screen                                        io

Control Register $FFF0:
bit 7:  0 normal 8032 configuration (screen, ROMs, IO, ROMs)
       80 expansion memory
bit 6:  0 RAM $E800-$EFFF (only when bit7=1)
       40 IO peek through
bit 5:  0 exp-mem $8000-$8FFF (-"-)
       20 screen peek through
bit 4: 10 not used
bit 3:  0 bank 1 $C000-$FFFF
       08 bank 3
bit 2:  0 bank 0 $8000-$BFFF
       04 bank 2
bit 1: 02 write protect bank 1/3
bit 0: 01 write protect bank 0/2
when bit7=0, all other bits are ignored

The missing 32K can't be accessed witout hardware modifications.
You can only use the 2K "behind" the screen $8800-$8FFF (exact: 34768-
36863), available in the normal configuration.
The register is write-only, and the value is written through to the
previously selected ram bank.


----
From rhialto@mbfys.kun.nl Wed May 31 04:01 EET 1995
From: Olaf Seibert <rhialto@mbfys.kun.nl>

If that is from the same set of photos that I have, then none of them
shows the "Business" keyboard. It has no gfx chars on the keycaps and
no way to type them except those that are the upper case letters.

>  This is the picture, few keycaps are missing:

 	PET  N keyboard (normal)
                                                 CLR/ UP/ LFT/ INS/
  |  !  "  #  $  %  '  &  \  (  )  <-  ??  |    |HOME DWN RGHT DEL  ?? |
  _|  q  w  e  r  t  y  u  i  o  p    ^~   |__  |  7   8    9   /  SPACE
 |RVS  a  s  d  f  g  h  j  k  l  :  Return  |  |  4   5    6   *  Ret-|
  |SHFT z  x  c  v  b  n  m  ,  ;  ?  Shift _|  |  1   2    3   +   urn|
   | @  [  ]        Space       <  >  Shift|    |  0   .    -   =  SHIFT
   ----------------------------------------      ----------------------
	'~' is 'pi'


 This picture must be from the machine with
the clone keyboard; a real N keyboard has nothing to the right of
INS/DEL /*+=.

Also a rectangular (non-tilted) as well as a business version exist.


 Each key, shifted, produces the
PETSCII code that is 128 higher than unshifted. The codes for the gfx
chars are such that they are arranged "easily" on the keys. For
instance the 8 horizontal lines are on #edc@fr$ (from top to bottom).
The original square keyboard was even better in this respect - the
space key was the width of 2 normal keys so left of it are SHIFT, RVS,
@[] and right are <> STOP and shift.


	PET  B keyboard

No 8032 were made with N keyboards (officially, at least). They all have
B keyboards that look like this

      !  "  #  $  %  &  '  (  )     *  =     LEFT  RUN/
  <-  1  2  3  4  5  6  7  8  9  0  :  -  ^  RGHT  STOP        7  8  9

  TAB  q  w  e  r  t  y  u  i  o  p  [  \  UP/DN INST/DEL      4  5  6

 ESC LOCKa s  d  f  g  h  j  k  l +;  @  ]  RETURN             1  2  3

 RVS SHIFTz x  c  v  b  n  m  <, >. ?/ SHIFT RPT HOME            0   .

           S  P  A  C  E .............

RPT = repeat
ESC = escape

All keys, when shifted, give the same character as unshifted except
letters and those with 2 symbols on them as shown here.
Supposedly it is to prevent frightening secretaries.

This adds the amount of the code for the digit and :,./ keys which get their
shifted value calculated (E550).

One nasty thing is that 4.0 was supposed to get the scan codes in
address 151 the same for both keyboards (they differ under 1.0 and 2.0
because the keyboard matrices are different for N and B keyboards), by
putting the petscii value there.  However, to get different codes for
the B numeric keypad, those keys have different codes, and some keys
don't get the petscii value after all ([]\^ <- esc, because those are
the unshiftable keys of course), and funny stuff like that. (On the B
keyboard).  And if I recall right, the 4.0 upgrade roms still use the
hard scancodes there. So the chaos became worse instead of less.
 
up, and as said, the N and B keyboards differ here..


The keyboard decoding is slightly different that on the 64: the row
number is output on the port and goes through a 4-to-10 decoder
to select the appropriate row, unlike the 64 where the rows from the
matrix are directly connected to the port.


---

  PETSCII

PETSCII table is the ame as the 64, except that the pound character is a
backslash.
The character generation ROMs are not identical though, since C=
doubled vertical lines for the 64 to compensate for bad tv output.

--

This differs by ROM version, of course.

In basic 1.0 and 2.0 there are
	cursor movements,
	clr/home,
	inst/del,
	(shift) return,
	off/rvs.
   Same codes as on the 64 of course.

4.0 adds (in the 8032)
	^g,^G: bell ^i TAB,
	^I set TAB,
	^n text mode,
	^N graphics mode,
	^o set top of window (scroll region, in vt-100 words), (8032 only)
	^O set bottom of window (8032 only),
	^p (erase to end of line) (fat40 only - bug!),
	^u insert line (8032 only)
	^U delete line (8032 only)
	^v erase to end of line (8032),
	^V erase begin of line,
	^y scroll up  (8032 only),
	^Y scroll down (8032 only),
	^[ ESCAPE: turn off quote mode and insert mode and rvs.

The 4.0 upgrade roms for 3032s misses even some 4032 codes.
The ^v vs ^p confusion: ^v=16 and ^p=0x16. Must be a typo in the
sources...

Conflicts with 64 codes are ^i, ^U, ^V and ^Y.


 [ Form PetIo.doc  by Olaf Seibert ]

Keyboard matrix
===============

There are two different keyboards: the graphics keyboard and the
business keyboard.

After analysis of the 50 and 60 Hz versions of the 8032 business
roms, there are apparently even difference between the keyboards of
those versions. The 60 Hz values are listed below the 50 Hz ones,
if they differ.

In Basic 4.0 CRTC the (PETSCII) values listed in the table below are
recorded in location 151 [this makes distinguishing between the normal 0
key and the one on the numeric keypad impossible]; in older versions (and
4.0 without CRTC) the position in the table is recorded. The position is
8 * (9 - row) + (7 - column) + 1, i.e. starting at 1, incrementing left to
right and top to bottom.

The keyboard is scanned by writing the row number into the row select port
($E810), then reading the column bits ($E812). Each bit that reads 0
represents a pressed key.

; 00 = Shift
; 10 = Repeat
; 80   flags unshiftable key
; FF = No key

Business keyboard decoding table:

----+------------------------
row |  7  6  5  4  3  2  1  0
----+------------------------
 9  | 16 04 3A 03 39 36 33 DF
    | ^V --  : ^C  9  6  3 <-	^V = TAB + <- + DEL, ^C = STOP,
    |				 <- = left arrow
 8  | B1 2F 15 13 4D 20 58 12
    | k1  / ^U ^S  m sp  x ^R	k9 = keypad 9, ^U = RVS + A + L,
    |				^S = HOME, sp = space, ^R = RVS
 7  | B2 10 0F B0 2C 4E 56 5A	^O = Z + A + L, rp = repeat
    | k2 rp ^O k0  ,  n  v  z
    |
 6  | B3 00 19 AE 2E 42 43 00
    | k3 rs ^Y k.  .  b  c ls	^Y = left shift + TAB + I, k. = keypad .
    |				ls = left shift, rs = right shift
 5  | B4 DB 4F 11 55 54 45 51	^Q = cursor down
    | k4  [  o ^Q  u  t  e  q
    |    5D]
 4  | 14 50 49 DC 59 52 57 09
    | ^T  p  i  \  y  r  w ^I	^T = DEL, ^I = TAB
    |          C0@
 3  | B6 C0 4C 0D 4A 47 44 41
    | k6  @  l ^M  j  g  d  a	^M = return
    |    5B[
 2  | B5 3B 4B DD 48 46 53 9B
    | k5  ;  k  ]  h  f  s ^[	^[ = ESC
    |    5C\   3B;
 1  | B9 06 DE B7 B0 37 34 31
    | k9 --  ^ k7  0  7  4  1
    |
 0  | 05 0E 1D B8 2D 38 35 32
    |  . ^N ^] k8  -  8  5  2	^N = both shifts + 2, ^] = cursor right
----+------------------------

Graphics keyboard decoding table:

----+------------------------
row |  7  6  5  4  3  2  1  0
----+------------------------
 9  | 3D 2E FF 03 3C 20 5B 12
    |  =  . -- ^C  < sp  [ ^S	^C = STOP, ^S = HOME
    |
 8  | 2D 30 00 3E FF 5D 40 00
    |  -  0 rs  > --  ] @  ls	rs = right shift, ls = left shift
    |
 7  | 2B 32 FF 3F 2C 4E 56 58
    |  +  2 --  ?  ,  n  v  x
    |
 6  | 33 31 0D 3B 4D 42 43 5A
    |  3  1 ^M  ;  m  b  c  z	^M = return
    |
 5  | 2A 35 FF 3A 4B 48 46 53
    |  *  5 --  :  k  h  f  s
    |
 4  | 36 34 FF 4C 4A 47 44 41
    |  6  4 --  l  j  g  d  a
    |
 3  | 2F 38 FF 50 49 59 52 57
    |  /  8 --  p  i  y  r  w
    |
 2  | 39 37 5E 4F 55 54 45 51
    |  9  7  ^  o  u  t  e  q
    |
 1  | 14 11 FF 29 5C 27 24 22
    | ^T ^Q --  )  \  '  $  "	^T = DEL, ^Q = cursor down
    |
 0  | 1d 13 5F 28 26 25 23 21
    | ^] ^S <-  (  &  %  #  !	^] = cursor right, ^S = home
----+------------------------

The original gfx keyboard reflects the matrix perfectly:

	  !  "  #  $  %  '  &  \  (  )  <-     ^s ^q ^] ^t
	  q  w  e  r  t  y  u  i  o  p  ^       7  8  9  /
	  a  s  d  f  g  h  j  k  l  : ^m       4  5  6  *
	  z  x  c  v  b  n  m  ,  ;  ? ^m       1  2  3  +
	 LS ^r  @  [  ]  SPACE <  > ^c  RS      0  .  -  =

The shift key always set the high bit, producing a graphic character.

-
CONTROL CODES
=============

F: only on 4.0 fat-40, i.e. NOT on upgrade 4.0 roms.
8: only on 4.0 80 columns.
4B: only on 4.0 upgrade, B keyboard (maybe also on N keybd but I don't
    have a ROM to verify this)

+------+------------------------------+------------------------------------+
|ctrl- | unshifted                    | shifted (unshifted + 128)          |
+------+------------------------------+------------------------------------+
| C  3 | stop                         | load/run, dload"*/run, or dL"*/run |
| G  7 | bell                         | double bell                        |
| I  9 | tab (4B,F,8)                 | set/clear tab (4B,F,8)             |
| M 13 | return                       | fake return                        |
| N 14 | text mode (F,8)              | graphics mode (F,8)                |
| O 15 | set top of window (8)        | set bottom of window (8)           |
| P 16 | erase end of line (F)        | see note (no function)             |
| Q 17 | cursor down                  | cursor up                          |
| R 18 | reverse field on             | reverse field off                  |
| S 19 | cursor home                  | clear screen                       |
| T 20 | delete char left of cursor   | insert space at cursor             |
| U 21 | delete line (8)              | insert line (8)                    |
| V 22 | erase end of line (8)        | erase begin of line (F,8)          |
| Y 25 | scroll up (8)                | scroll down (8)                    |
| [ 26 | escape: end inst & quote mode| (no function)                      |
| ] 28 | cursor right                 | cursor left                        |
|      |                              | 255 pi (converted to 94)           |
+------+------------------------------+------------------------------------+

Note: on the 4.0-40-crtc version the code for "erase end of line" is
different from the 4.0-80 code. Note that the decimal notation of the
former (16, or $10) equals the hex notation of the latter ($16, or 22).
This suggests a typo while editing the assembly source for the other
version.



  PET 8032 I/O decoding


 Address	Chip    Minterms (A7..A4)
	--------------------
 E800	00	-
		PIA 1	xxx1
		PIA 2	xx1x
		d
	------------------
 E840	01	VIA	x1xx
		d
		d
		d
	--------------------
 E880	10	CRTC	1?xx
		d
		d
		d
	--------------------
 E8C0	11	d
		d
		d
		d
	--------------------
	d = don't care


    Images at E8xx

	Chip	/CS   E800 10 20 30 40 50 60 70 80 90 A0 B0 C0 D0 E0 F0
	---------------------------------------------------------------
	PIA 1	xxx1	    X     x     x     x     x     x     x     x
	PIA 2	xx1x	       X  x        x  x        x  x        x  x
	VIA	x1xx	             X  x  x  x              x  x  x  x
	CRTC	1?xx	                         X  x  x  x  ?  ?  ?  ?


E800-0F not connected (reads E8)
  10-1F PIA 1, 4 times
  20-2F PIA 2, 4 times
  30-3F the AND of E81x and E82x
  40-4F VIA
  50-5F 4x AND 1x
  60-6F 4x AND 2x
  70-7F 4x AND 2x AND 1x, approximately
  80-8F CRTC (most registers are write-only so readouts are not very
        definitive)
  90-FF This is not exactly explainable as the AND of 8x, 4x, 2x and 1x
        (there are too many non-00 values),
        nor of the AND of only 4x, 2x amd 1x (then everything should be
        the same as in the 10-7F range but some is and some isn't).
        Especially the Fx addresses resemble the 7x addresses.
        80-BF read either 00 or 20, seemingly depending on timing.

 
The official base addresses of PIA, PIA, VIA and CRTC are E810,
E820, E840 and E880. The register number in the chip is added
to the address.

----

 
But: things like these are not 100% the same in all models. From the
4032/8032 on, I understand that the address range reserved for any I/O
chips has shrunk from E800-EFFF to E800-E8FF. This was to allow bigger
Editor roms (usually at E000-E800) with localized (read: German)
support from E900 onwards.
 
>  This turns to be very interesting.  First, it saves one OR gate. Secondly,
>  I used that OR for the CRTC as well. Hmmm...
> 
> 	PIA1  CS= A4, /CS= I/O select
> 	PIA2  CS= A5, /CS= I/O select
> 	VIA   CS= A6, /CS= I/O select
> 	CRTC /CS= NOT A7 OR I/O select


I am not sure if older models echoed for instance E8xx addresses at
E9xx as well...
 
Something you probably also want to know is screen memory (8000-).
Originally this was 1K, doubled to 2K for the 80 column models.  The
original 2001 had 4 images of the screen memory, later models only 2
(8000 and 8400); the 80-columns, I think, always had only one images.
My 8296 definitely has only one image, since it was redesigned to use
the screen memory in a "normal" ram chip also containing system memory;
all older models must have used an extra 1K RAM separate from other
ram.
 
In fact some of the basic lessons I uploaded to ftp.funet.fi use
these addresses that worked on the 2001 but not on my 8296. This
can be noticed by the program promising to display certain things
but you never get to see them.

For the 80-colums you need of course 2000 bytes of screen memory.
But it is not really that easy to determine from the CRTC table if
the machine is 40 or 80 cols. In both machines it writes 40 to CRTC
reg. 1, but the 80 cols version is clocked at 2 MHz (or so I heard)
and therefore produces 80 characters per line. I think the best way
to identity the # of cols from software is to look at the Editor rom
and compare "your" version with the ones I put on ftp.funet.fi. You
can look at the first 10s of bytes to identity the right one.

------

But whatever screen memory images there are, they are always
within the 8xxx range, since 9xxx, Axxx (and Bxxx on < 4.0 models)
are for expansion EPROMs. If no EPROM is present, it reads out as
the high byte of the address. If that doesn't happen in E900-EFFF,
then perhaps the editor EPROM is already 4K and could be swapped for
one with more code in it.


No, device 2 on the PET is cassette #2. As on later computers, devices
4 and up are considered on the IEEE bus. I suppose the IEEE modem had
a device number 5 or 7 or something like that.

-rhialto

----

In general, PEts don't have RS-232 port. Exceptions do exist, however.
PET710 has got a RS232C port, as well as the SP9000.  It has a real UART and
RS-232 port, but it's only available in 6809 mode (at least according to the
documentation).

----


	PET serial

From: Dave Tickle <aiu214@freenet.mb.ca>
Newsgroups: comp.sys.cbm
Subject: Re: Do PETs have serial ports?
Date: Fri, 5 Jan 1996 01:36:38 -0600



On 2 Jan 1996, Olaf Seibert wrote:

> PETs, with the exception of the SuperPET, do NOT have serial ports.
> You could however transplant and adapt the C64 software serial
> port for this purpose.


Just a few minutes ago, I took a look inside a SuperPET, and on the upper
board there is a female DB25 connector (just sitting at the inside edge of
the board, not adjacent to the outer case of the computer).  Not too far
away from this connector are a 6551, a 1488, and a 1489 chip, which almost
definitely means there is some kind of serial port there using RS-232
voltages. 

Now, the problem is I don't have a clue how to actually access it.  For 
all I know, it may only be accessible from the 6809 side of the machine.  
But, if that were true, then it wouldn't make sense for Commodore to have 
advertised the 8000 series as having an RS-232 port.

Anybody know how to access this port?


Newsgroups: comp.sys.cbm
From: dfevans@bbcr.uwaterloo.ca (David Evans)
Subject: Re: Do PETs have serial ports?
Date: Fri, 5 Jan 1996 16:18:49 GMT


>Anybody know how to access this port?

  No, but in Volume 4 Issue 1 (I think) of The Transactor a terminal program
for the SP9000 was included; I have Vol. 4 Iss. 2 and in the Transbloopers
section they have the "tables that follow", which look a lot like everyone's
favourite control/command setups for VIC/64 serial device 2.  I have no idea
how to use this, though.
  I believe a terminal program for the SP9000 was also included in their
"Networking and Communications" issue (V6I3? can't recalll.)  I'll look on the
disk for this issue when I next get the chance.

-- 
David Evans              (NeXTMail OK)              dfevans@bbcr.uwaterloo.ca
Computer/Synth Junkie                      http://bbcr.uwaterloo.ca/~dfevans/


From: Dave Tickle <aiu214@freenet.mb.ca>
Newsgroups: comp.sys.cbm
Subject: Re: Do PETs have serial ports?
Date: Mon, 8 Jan 1996 01:58:28 -0600


I decided to poke around with a logic probe and found that pin 2 (Chip
Select 0) on the 6551 gets toggled when accessing memory locations 59424
through 59431.  I suppose that means that the 6551's registers may be
located in that range somewhere.  Some locations in that range are also 
used by the IEEE-488 bus, so I guess you can't use both ports at the same 
time.  I'll have to investigate further.  (It's actually kind of fun!)


From: Dave Tickle <aiu214@freenet.mb.ca>
Subject: Re: Do PETs have serial ports?
Date: Thu, 11 Jan 1996 00:16:17 -0600



On Fri, 5 Jan 1996, Jim Brain wrote:

> As far as I know, my PET 8032 does not have the required serial code
> in the ROMs.

Thanks for that info.  I haven't yet had time to look thoroughly through 
the SP9000 ROMs, but I imagine it's pretty similar to what you've got.  I 
vaguely recall, though, that it was only the 8096 and SP9000 that had 
this serial port.  I could be wrong, though.  I'm not quite sure what I 
should be looking for.  I haven't positively located the 6551 registers yet.

Speaking of ROMs, I found something occupying A000-AFFF that apparently 
is something called "wordpro 4 plus".  Anyone out there know if this is 
something I can use (just out of curiosity and not any practical need), 
or would I need some program to run along with it?




From: UKAH@ibm3090.rz.uni-karlsruhe.de (Veit Laule)
Newsgroups: comp.sys.cbm
Subject: Re: Do PETs have serial ports?
Date: Thu, 11 Jan 1996 13:21


Dave Tickle <aiu214@freenet.mb.ca> writes:

>Speaking of ROMs, I found something occupying A000-AFFF that apparently
>is something called "wordpro 4 plus".  Anyone out there know if this is
>something I can use (just out of curiosity and not any practical need),
>or would I need some program to run along with it?
>
Yes, you need the "wordpro 4 plus"-program (a WORDPROcessor).
The ROM is the copy-protection.

Veit Laule

From: devans@ccubb.com (David Evans)
Newsgroups: comp.sys.cbm
Subject: Re: Do PETs have serial ports?
Date: Mon, 15 Jan 96 06:36:37

Pets have a serial port just like the C64 does.  The pets however don't have
kernal support for the serial routines.  You need to "roll your own".  The
serial port on the standard Pet is called the "User port".  It requires a
special connector be built and then its basically "normal" from that point.
Heck Compuserve used to sell a package with software which contained
instructions on how to build it!  (I don't have a copy anymore..)

-------------------------------------------------------------------------------

	PET -I/F

From: tnelson@fluorite.telesciences.com (Tracy M Nelson)
Newsgroups: alt.folklore.computers,comp.sys.cbm
Subject: Re: PET GPIB-serial interface box
Date: 11 Dec 1995 18:46:01 GMT

Hans-Christian Becker (hcb@phc.chalmers.se) wrote:
: Hi everyone!
: I have a magic box that used to interface a commodore PET to a serial
: printer (or so I'm told!). It is made by TNW Corporation and is called
: TNW-2000. The board is (c) 1979. This box has a standard GPIB connector
: together with a PET User Port card connector, and a standard RS 232
: connector.
: The design seems to be centered around a "AY-5-1013A" 40-pin chip, with
: some TTL circuitry around. 
: Unfortunately, the manuals seem to have been thrown out, together with
: the PETs. Does anyone know *anything* about this thing??? If it is a
: full GPIB-serial converter, it would be very useful to me - if it is
: not, it will have to go back into the bookshelf again (sniff).

Be warned that some PETs (the SuperPET, for sure) used a bastardized GPIB-type
interconnection for connecting to peripherals.  I don't think it was 100%
GPIB-compatible, so be forewarned that you may not be able to interface it with
a "real" GPIB-type device.
--
Tracy Nelson (tnelson@telesciences.com)
// WARNING: May contain code which is too intense for young programmers...

Newsgroups: comp.sys.cbm
From: srowe@ftel.co.uk (Simon Rowe)
Subject: Re: Routine for easy SEQ file read:
Date: Tue, 12 Dec 1995 12:39:37 GMT


pap@dana.ucc.nau.edu (Paul Allen Panks) writes:
> 8 F=VAL(F$): IF F>16 THEN PRINT"FILENAME TOO LONG! MUST BE LESS THAN

First bug. I think you mean LEN() not VAL()

		Simon.

...............................................................................

From: "Mitch B. Parker" <mbpark@planetx.bloomu.edu>
Newsgroups: comp.sys.cbm
Subject: Re: PET 2001N-16
Date: Thu, 18 Jan 1996 23:12:01 -0500

On 10 Jan 1996, Eric R. Eckert wrote:

> Is anyone aware of information which would allow me to convert my PET 2001N-16 to a PET 2001N-32? I am trying to help out my kids classroom run their one-and-only program on their one-and-only PET.


First check the chips on your MB.  From what I remember from my 4016, 
commodore drilled holes through several important connections to prevent 
a 32K upgrade from occurring.  If this is not the case, good luck hunting 
down 16K 200ns chips!  I believe you just stick in these chips (4116?), 
and you are good to go...that is if you have sockets.  If you do not, 
then you need those too ;(.


From: daled@cds9172.Cadence.COM (Dale DePriest)
Subject: Re: PET 2001N-16
Date: Tue, 16 Jan 1996 20:12:36 GMT

rhialto@mbfys.kun.nl (Olaf Seibert) writes:

|> First, please be aware that your email address is set wrong.
|> 
|> In <4d18u2$99o@newdelph.cig.mot.com> Eric R. Eckert <eeckert@shelf> writes:
|> >Is anyone aware of information which would allow me to convert my PET 2001N-16 to a PET 2001N-32? I am trying to help out my kids classroom run their one-and-only program on their one-and-only PET.
|> 
|> AFAIK, there have been lots of ram configs in various PET models. It
|> was typical that you could use various sizes of ram chips, but if you
|> chose a small size you would not have enough sockets (or spaces for
|> sockets) to max out the full 32K. If you were unlucky you had to throw
|> out all your current ram chips and replace them with bigger ones. Much
|> like the current situation with SIMMs, really. On top of that, some
|> boards were filled with the bigger ram chips but the space for the
|> required additional sockets was made unusable, requiring you to buy a
|> new machine to get more memory.
 
Actually it was common practice for commodore to punch holes in the board
so you couldn't upgrade it.  At one time I upgraded one by piggy-backing
the new chips on top of the old ones.  If I remember correctly only one pin
had to be bent straight as it was a select line. then I tied all of the
select pins together and jumped it to the correct place on the board.
I don't remember the details anymore but it can be done.  Get a set of
chips like the ones in the machine and a data book.


--
    _      _     Dale DePriest  		San Jose, California
   /`) _  //     daled@Cadence.COM		voice: (408) 428-5249


------------------------------------------------------------------------------

From rhialto@mbfys.kun.nl Wed May 31 04:01 EET 1995
From: Olaf Seibert <rhialto@mbfys.kun.nl>


If you would like to see my 8296-D (a truely
impressive thing - looks like the B-128, whith monitor, and with the
floppy drives built-in) feel fee to drop by :)

------------------------------------------------------------------------------


>With its 8kb of memory it was barely usable.

I remember the days that we told each other that you'd never write
a program so big that you'd need more than 8K! (Soon after that,
a fellow computer club member wrote such a program).

>It was call the PET 2001 which stood for Personal Electronic Transactor.
>The IEEE Interface was not working at all.

The printers worked fine with it, but the support for disk drives
was buggy. I haven't tried yet but I suppose one could still open
and read/write files with it, but not load/save.

However, Commodore's idea about IEEE-488 is not identical to the official
view, i.e. it is not really compatible with real IEEE devices.

>The CBM 3040 (1980) was a unit containing two 160kb floppy disks. This unit was 
>connected to the CBM3032 with an IEEE488 cable.

Before the 3040 there was the 2040, but they were identical. Both had
DOS 1.0, which had a few features less than DOS 2.0 which was introduced
in the 4040 (and later cut down for the 1540). The most significant
addition for 2.0 was REL files. An incompatibility was that disk capacity
was REDUCED by using one sector less in one of the 4 density bands
(I forget which, but perhaps I could look it up). 2.0 drives would read
but not write 1.0 disks. The 2.0 capacity is about 170 K (like
the 1540/41/etc), 1.0 was probably around 180 K.


>I have a CBM 3040 Dual floppy drive, but unfortunately I have no papers 
>about it. I have a IEEE-488 interface too.
>My question is: can I use this drive with my C64 and what type of cable 
>and software I need?

You say you have an IEEE-488 interface; I assume you mean it's one for
the 64. If this is right, then you can use the drive with the 64.  You
should not need any additional software, since the interface should
probably contain it already. It won't be compatible with fastloaders
and stuff like that but everything else works about the same as a 1541.

As for the cable, the interface probably has a card-edge connector.
Get a connector for that, and one that fits in the 3040, and solder a
cable between them that is just straight-through connected.  I.e., pins
1-12 on the 64 to 1-12 of the 3040, pins A-M (the bottom pins usually)
to 13-24.


Subject: Re: Question: MSD SD-2 Super Disk Drive

>You use the drive number in the commands as appropriate.
>
>To save to disk 0:  SAVE"0:filename",8
>To save to disk 1:  SAVE"1:filename",8

Commdore's double drives have drive 0 right and 1 left.

To copy a whole disk: DUPLICATE 1=0  (or D1=0)	(copies drive 0 to 1)


From: mbrown@ersys.edmonton.ab.ca (Michael Brown)
Newsgroups: comp.sys.cbm
Subject: Help with CBM 8032

Hello. I'm in the market for 
instructions for making, or a source for buying (cheap), a cable to hook
a CBM 8050M dual disk drive to a CBM 8032.
The 8050M appears to have a small 'cartridge-edge' connector (the 
cable would need to be a male cartridge type on the disk drive end of 
things). Any help would be appreaciated. I can build it if someone could 
like the connections the cable would require. Thanks again.
--

The computer end of things can be bought at Jameco or anywhere that sells
the connector most people know of the VIC/64/128 user port connector.

Dunno about the IEEE-4888 24 pin dsub.
--

Subject: Re: Has Anyone Heard of the CBM 4040 Disk Drive ?

You just found the disk drive for the CBM series computers...
It fits eg. to the CBM 8082. But it's not compatible with the C64 disk format at all :(

-Gerd


From rhialto@mbfys.kun.nl Thu Jun 29 15:33 EET 1995
Date: Thu, 29 Jun 1995 14:33:18 +0200
From: Olaf Seibert <rhialto@mbfys.kun.nl>
Subject: Re:  PET

Hi, you wrote:
>  I've encountered one very strange problem:
> 
>  E60F	iE60F	LDA #$7F
>  E611		STA $E84E
>  E614		LDX #$6D
>  E616		LDA #$00
> 
>  The above code appears on both Basic 2 and basic 4 ROMs. Do you happen to
>  know why ?  There is no command to turn on interrupts anywhere...

I made a quick look trhough the source and you seem to be correct.
I assume then that the IRQ must come then directly from the video retrace
line, not through any I/O chip. That is consistent with the 50/60 Hz
difference in European/USA models.



From: bgracey@freenet.niagara.com (Byron Gracey)
Subject: Re: Has Anyone Heard of the CBM 4040 Disk Drive ?
Date: Thu, 6 Jul 1995 05:20:41 GMT


: It fits eg. to the CBM 8082. But it's not compatible with the C64 disk format at all :(

Ahh yes, the 4040 - also can be used with a 64 or 128 if you have an IEEE 
interface such as IEEE Flash for the 64!  35 tracks per disk, single 
sided (flip for side two like a 1541) 256 bytes per sector, 683 sectors 
(blocks) formatted, 664 of which are free to use normally.

Uses a 6504 processor on the controller, a 6502 on the interface and 
these both share 4x1K of RAM.  Made of 18 guage steel, weighs plenty and 
has large footprint, requires about 50 watts to run.

I happen to have manuals for it (and others in the series) if anyone is 
interested.

And if ya need more info, such as how it works (mine works peachy), how 
to change device numbers (permanently, that is) etc, just holler.  ttyl

Byron

bgracey@freenet.niagara.com

-- 
Byron Gracey
Fido: 1:247/105
bgracey@freenet.niagara.com


From: rhialto@mbfys.kun.nl (Olaf Seibert)
Newsgroups: comp.sys.cbm
Subject: Re: CBM PET 3032 information wanted
Date: 5 Sep 1995 13:07:44 GMT

In <421bvb$3ij@tuegate.tue.nl> martijnb@stack.urc.tue.nl (Martijn van Buul) writes:
>I've bought an old CBM PET 3032 last weekend, but I don't have any kind of
>information on it. If anybody can provide me some information, I would be
>very pleased. For example :
...
>2) A kernal jumptable (the jumptable from the C64 (from the Programmers
>   Guide) isn't the same. I can't find SETFLS, SETNAM etc., needed to open a
>   file)

The top part of the jumptable is mostly the same, but I don't think
SETNAM and SETLFS are in that part. There are also other differences
that are quite annoying (meaning: they corrected some earlier design
mistakes in the VIC kernel). For instance, LOAD and SAVE are meant
to be called from BASIC. The first thing they do is  use BASIC routines
to fetch the file name and device number. To use them from ML you must
either set up CHRGET correctly or jump directly into ROM (where exactly
differs with each ROM version :-( ). OPEN however should be usable.

-Olaf.
--
___              Copyright 1995 Olaf 'Rhialto' Seibert. All Rights Reserved.
\X/    You are not allowed to read this using any kind of Micro$oft product.

From: rhialto@mbfys.kun.nl (Olaf Seibert)
Newsgroups: comp.sys.cbm
Subject: Re: CBM PET 3032 information wanted
Date: 5 Sep 1995 13:07:44 GMT


>2) A kernal jumptable (the jumptable from the C64 (from the Programmers
>   Guide) isn't the same. I can't find SETFLS, SETNAM etc., needed to open a
>   file)

The top part of the jumptable is mostly the same, but I don't think
SETNAM and SETLFS are in that part. There are also other differences
that are quite annoying (meaning: they corrected some earlier design
mistakes in the VIC kernel). For instance, LOAD and SAVE are meant
to be called from BASIC. The first thing they do is  use BASIC routines
to fetch the file name and device number. To use them from ML you must
either set up CHRGET correctly or jump directly into ROM (where exactly
differs with each ROM version :-( ). OPEN however should be usable.

-Olaf.
--
___              Copyright 1995 Olaf 'Rhialto' Seibert. All Rights Reserved.
\X/    You are not allowed to read this using any kind of Micro$oft product.



From: rhialto@mbfys.kun.nl (Olaf Seibert)
Newsgroups: comp.sys.cbm
Subject: Re: CBM PET 3032 information wanted
Date: 9 Sep 1995 02:05:40 GMT

> Once you find the hardware way to do the reset, I seem to recall that the PET
>reset sequence destructively tests memory.  Poke something recognizable
>someplace and do sys 64790.  My PET's not near me, so I can't test it for
>myself.

This is true, but there is help. If you hold down the diagnostic
sense line (user port pin 5), you'll end up in the monitor instead
of doing the full reset thingie. A PET at our school's computer club
had a smart thing with a switch wired to reset, and diag through a
condensator. If you flicked it quicky it would reset; if you kept the
switch in the alternate position a few seconds it would decharge the
condensator, pull down diag, and when switched back you came back in
the monitor.

-Olaf.
--
___              Copyright 1995 Olaf 'Rhialto' Seibert. All Rights Reserved.
\X/    You are not allowed to read this using any kind of Micro$oft product.

From: martijnb@stack.urc.tue.nl (Martijn van Buul)
Newsgroups: comp.sys.cbm
Subject: Re: PET 2001 ?
Date: 29 Sep 1995 12:26:07 GMT

Tony Jewell (tony@ns.cityscape.co.uk) wrote:

> I think you're right. Shifted keys produce graphics, unshifted produce 
> upper case. There appears to be no lower case characters.

Try  poke 59500,72     or
     poke 59468,14

That should do..

> Tony.

You're welcome!
Martijn.

From: rhialto@mbfys.kun.nl (Olaf Seibert)
Newsgroups: comp.sys.cbm
Subject: Re: PET 2001 ?
Date: 29 Sep 1995 14:15:25 GMT


>hermit@cats.UCSC.EDU (William R. Ward) writes:
> tony@ns.cityscape.co.uk (Tony Jewell) writes:
>) 	We've just got a CBM PET 2001 for our little computer collection 
>) in the CB1 Cybercafe, Cambridge (UK). However, I know very little about 
>) these machines. It's an 8K model (which I think is the rarer), and in 
>) fineworking order (except the '1' key is broke). Am I right in thinking 
>) it's the first Pet ? Does anyone know what the ports on the back and side 
>) or for ? Does anyone have documentation for it, or just any interesting 

From right to left: IEEE, User port, cassette.

>) information about it ? And is there any chance of finding any software 
>) for it ?

There is some at ftp.funet.fi:/pub/cbm/pet.

>Is this the *really* early PET with the internal tape drive and the
>impossible keyboard, or the slightly newer one with a "real" keyboard
>and external tape deck?
>
>If the former, and it's in good working order, you've got a pretty
>rare find, those are hard to come by these days.  It runs BASIC 2.0,
>which is the same BASIC as the C64, so *simple* programs should run on
>it.  I think though that some of the early PET's used a form of ASCII
>where the upper case and lower case letters were reversed, so it might
>look funny if you run them.

The character generator ROM had them swapped, yes. Many machines upgraded
from Basic 1.0 to 2.0 at the same time as they got a new char rom,
but definitely not all. There were lots of machines with 2.0 Basic but
old char roms.

The reason why the chars were swapped originally does make sense.
If you switch to lower case more, text already on screen remains as
it is. I guess that having to keep shift depressed to get lower case
characters was considered too inconvenient. The result was new
programs looking funny on old machines, and vice versa. I have a few
programs that were obviously meant for the old char roms.

>--Bill.
-Olaf.
--
___              Copyright 1995 Olaf 'Rhialto' Seibert. All Rights Reserved.
\X/    You are not allowed to read this using any kind of Micro$oft product.

From: martijnb@stack.urc.tue.nl (Martijn van Buul)
Newsgroups: comp.sys.cbm
Subject: Re: PET 2001 ?
Date: 29 Sep 1995 12:21:32 GMT

Tony Jewell (tony@ns.cityscape.co.uk) wrote:
> Hi all,

> 	We've just got a CBM PET 2001 for our little computer collection 
> in the CB1 Cybercafe, Cambridge (UK). However, I know very little about 
> these machines. It's an 8K model (which I think is the rarer), and in 
> fineworking order (except the '1' key is broke). Am I right in thinking 
> it's the first Pet ? Does anyone know what the ports on the back and side 
> or for ? Does anyone have documentation for it, or just any interesting 
> information about it ? And is there any chance of finding any software 
> for it ?

There are several good books about the PET. "Osbourne's Guide To The Pet" 
seems to be a good one, but I haven't read it.
The PET 2001 is nearly equal to the CBM 30xx computers. The major difference 
is, that the PET has a built-in cassette-recorder, and the CBM 30xx hasn't. 
(There seems to be a minor bugfix to, but I'm bot so sure about that..)
So nearly evry information for the CBM 30** is fine with the PET too

Oh yes : If you want to have software : check out 
  ftp.funet.fi/pub/cbm/pet (I'm not so sure about these subdirs..)

BTW: The PET 2001 with 8 K RAM was the only 2001 available in Europe (at least
that's what this book I'm currently reading (dutch) says).


Martijn van Buul
martijnb@stack.urc.tue.nl


From: RICHARD WHITE <richwhite2@mci.newscorp.com>
Newsgroups: comp.sys.cbm
Subject: It's Alive!  (The 8032)
Date: 17 Feb 1996 02:59:30 GMT

Hi all...

After several hours of poking and prodding around my 8032, 
I finally got the thing working!  (Complete with little 
piezo startup sound!)  

The problem was the IC socket on the main motherboard where 
the 6502 (or expansion cable in my case) would plug in 
to...  I simply whipped out the 'ol desoldering and 
soldering irons, and had a new 40-pin IC socket installed 
in no time!    

So if there are dead 8032's out there... and you're sure 
the power supply is good... chances are those cheap white 
IC-sockets are going bad.....  (The ones for the EPROMS 
aren't much better)...  Press down on the IC's and try 
booting the system...  Thats what I did... (of course, 
don't touch any pins <G>)

Thanks for everyone's help...
I now have a nice 8032 w/96k (so its a 8096 then, right?) 
complete with 8050 Dual drive, MSD-1 SuperDrive, 300baud 
modem, and even a Commodore IEEE printer!

They're quite handsome computers, aren't they?  I love how 
crisp the green-screen is....


Rich



From: rhialto@mbfys.kun.nl (Olaf Seibert)
Newsgroups: comp.sys.cbm
Subject: Re: MSD-1 w/ PET...8032 w/ 96K
Date: 19 Feb 1996 15:52:09 GMT

dfevans@bbcr.uwaterloo.ca (David Evans) writes:
>  8096s were sold as such, but I think you could get the upgrade board on its
>own as well.  Is there any indication that CBM made the upper board?

As for programming info, look in PetIO.doc on x2ftp.oulu.fi. There
is also LOS-96, the Loadable Operating System.


>RICHARD WHITE  <richwhite2@mci.newscorp.com> wrote:
>>Still curious as to why I have the EPROM "Spacemaker" thinge though...
>
>  Hard to say what's in there...you could hunt through memory for nifty
>things, I suppose.

The EPROM sockets are for addresses 9xxx and Axxx, so that is where to
look. Pre-4.0 machines also had Bxxx free as an EPROM socket.

>David Evans              (NeXTMail OK)              dfevans@bbcr.uwaterloo.ca

-Olaf.
___ Olaf 'Rhialto' Seibert      rhialto@mbfys.kun.nl



From: doriang@microsoft.com (Dorian Garson)
Newsgroups: comp.sys.cbm
Subject: Kermit for CBM SuperPET 9000?
Date: 7 Feb 1996 00:11:46 GMT

I've finally opened up my CBM SuperPET 9000, and discovered a 25-pin 
female port on the edge of the "middle" board.

Even better, a yellowed copy of SuperPET Gazette specs out a project for 
creating a serial line driver for long runs, and it has a diagram that 
indicates that what I've found really *is* a serial port!

So, the next logical question is: where can I find a copy of Kermit that's 
been ported to the SuperPET?  It would be so cool to be able to use it as 
a terminal!

 -Dorian

--
Newsgroups: comp.sys.cbm
From: fachat@galileo.rhein-neckar.de (Andre Fachat)
Subject: Re: PET 8296-D info sought
Date: Fri, 2 Dec 1994 10:23:13 GMT

Received subspace message from Olaf Seibert:
<many interesting things deleted>
> >>- What is the completely unlabeled 40-pen chip at UC9? The Video Display
> >>Controller? Does someone have a register list of the VDC?
> >It's a 6545 CRT, 99.9% compatible with the 6845.

> I'll try to find out about that one. I expected it to be switchable between
> 40 and 80 columns, so they could use the same chip in the 4032 and 8032.
> Apparently not, according to my ROM listings... weird.

Actually it is. And they used the Chip in both, 4032 and 8032 and, AFAIK
also in the old 3032.
The only thing to change is the Character Clock frequency. But if you have
the video access synchronised with the CPU memory access, it is not that
easy to simply switch the frequency. Ok, you only have to fetch a byte
in two CPU memory accesses, but without you just don't care:
CPU Phi2 hi, Video Phi2 lo...

I used the Chip to build myself a video interface for my selfbuilt
small 6502 (which is mostly 3032 compatible in emulator mode ;-)))))
and this one runs with 1MHz (40 Col) or 2MHz (80 Col)

Andre

-- 
Andre Fachat                 mail me!          fachat@galileo.rhein-neckar.de
FIDO: Andre Fachat@2:2468/6020.4           GERNET: Andre Fachat@21:100/3338.4
              Why use Windows, since there is a door?
--

Newsgroups: comp.sys.cbm
From: rhialto@mbfys.kun.nl (Olaf Seibert)
Subject: Re: PET 8296-D info sought
Date: Sun, 4 Dec 1994 00:48:35 GMT

In <1994Dec2.102313.473@galileo.rhein-neckar.de> fachat@galileo.rhein-neckar.de (Andre Fachat) writes:
>Received subspace message from Olaf Seibert:
><many interesting things deleted>
>> >>- What is the completely unlabeled 40-pen chip at UC9? The Video Display
>> >>Controller? Does someone have a register list of the VDC?
>> >It's a 6545 CRT, 99.9% compatible with the 6845.
>
>> I'll try to find out about that one. I expected it to be switchable between
>> 40 and 80 columns, so they could use the same chip in the 4032 and 8032.
>> Apparently not, according to my ROM listings... weird.
>
>Actually it is. And they used the Chip in both, 4032 and 8032 and, AFAIK
>also in the old 3032.

Not in the old 3032 (with the small screens), AFAIK. There were different
ROM sets for upgrading small screens to Basic 4.0 and for the "native"
FAT-40 models.

>The only thing to change is the Character Clock frequency. But if you have
>the video access synchronised with the CPU memory access, it is not that
>easy to simply switch the frequency. Ok, you only have to fetch a byte
>in two CPU memory accesses, but without you just don't care:
>CPU Phi2 hi, Video Phi2 lo...

The motherboard I have claims to be "40/80 columns", and there seems to
be some location where you can change some connections to switch between
40 and 80 columns. Mine is set for 80 columns, obviously.

Through experimenting I have found the working of about half the registers,
and I noticed that the one that apparently sets the base address of the
screen memory works by twos: if I change it by 1 the screen contents shift
by two characters. And the register that determines the number of characters
per line also is effectively multiplied by two.

I must try to find out what memory is actually displayed when I change the
screen memory base address from the default. I suppose most of it will
simply be non-existent.

>I used the Chip to build myself a video interface for my selfbuilt
>small 6502 (which is mostly 3032 compatible in emulator mode ;-)))))
>and this one runs with 1MHz (40 Col) or 2MHz (80 Col)

And now for something completely different: I was very annoyed that I
couldn't type the graphics characters on the 8296D's business
keyboard.  So I hacked something up: (all numbers approximations since
I don't have the machine in this room)

Insert the following code into the "print to screen" vector at $00EB.
This vector is only present on the 80 column version of Basic 4.0.
Fortunately, I think there aren't many 40 column machines with business
keyboard around.

	.: 027a  ldy $cd	; # of inserts outstanding
                 beq end
                 lda $d9	; character to print
                 eor #$80	; toggle shift bit
                 sta $d9
           end:  jmp $e206	; normal vector

(I'll repost this correctly when I write it down)

Now when I'm in "insert mode" I can type all the graphics characters
that I want, because they too go through this function when input is
being gathered.  There are some funny side effects, but if they are
unwanted you can easily cancel them with the ESC key. One of those is
that after 1 INST, the second one is actually a DEL, and an UP becomes
DOWN - everything is shifted.  So to insert multiple spaces you must
first type INST and then DEL for each additional space.

In earlier versions I used the RVS flag and the quote mode flag for the
modification, but those had much more annoying side effects.

>Andre
-Olaf.
--
___ Olaf 'Rhialto' Seibert    rhialto@mbfys.kun.nl     Ooey-Gooey-Fluffy-Barfie
\X/ I'm not weird, I'm differently perceived. D787B44DFC896063 4CBB95A5BD1DAA96
--
From: UKAH@ibm3090.rz.uni-karlsruhe.de (Veit Laule)
Newsgroups: comp.sys.cbm
Subject: Re: PET 8296-D info sought
Date: Tue, 06 Dec 1994 17:12


I'll write something about the CRT-programming next week.
(e.g. the 50-line mode on the 8296....)

> And now for something completely different: I was very annoyed that I
> couldn't type the graphics characters on the 8296D's business keyboard.

I modified the EDITOR to do this. Like the ALT-mechanism on PCs. While
pressing the REPEAT-key, you can type the PETASCII-number of the character.
 REPEAT--6--5  --> "a"


Veit
--

Newsgroups: comp.sys.cbm
From: rhialto@mbfys.kun.nl (Olaf Seibert)
Subject: Re: PET 8296-D info sought
Date: Wed, 7 Dec 1994 12:21:16 GMT

>UKAH@ibm3090.rz.uni-karlsruhe.de (Veit Laule) writes:
> REPEAT--6--5  --> "a"

Cute! I sent in my little hack to Commodore Hacking, but I think the
small listing I gave in my post was correct. (If not a real hacker
would correct it anyway). Your solution is a bit neater. Mine does not
work on any other rom version, since it is the only one that has that
vector.

Something else: At the secondary school that I visited we had a couple
of 4032s connected with a network called Hydra. Unfortunately this
network was rather unsafe.  One time after a holiday I came into the
room of the computer club to use a computer, and found they were not
working: not plugged in. So I plugged one in, and heard a terrible
noise. And soon after that, a horrible smell. Apparently, this network
card could not stand being plugged in while the power switch was on,
and I had completely fried one 4032. Literally. And worse, there had
been power spikes through the network cables to the two directly
connecting PETs which were also broken (but fried a bit less :)

That's when the 4032 got a 8032 motherboad. But since it still had the
graphics keyboard, so the keyboard decoding table had to be fixed for that.
As it turned out later, when the eproms were already made, some correction
should also have been made for the number keys, because when shifted they
still produced !@#$ etc.. I suppose we must have had one of the very few
existing 8032s with graphics keyboard.

From that moment on I made sure all my programs worked properly both on
40 and 80 column PETs. How to detect 80 columns? The standard trick
was to do

    print "[clr][down][shift-space]"
    wd = 40
    if peek(32768 + 80) = 96 then wd = 80;

This can be generalised to find any screen width, of course.

-Olaf.
--
___ Olaf 'Rhialto' Seibert    rhialto@mbfys.kun.nl     Ooey-Gooey-Fluffy-Barfie
\X/ I'm not weird, I'm differently perceived. D787B44DFC896063 4CBB95A5BD1DAA96
--
Newsgroups: comp.sys.cbm
From: fachat@galileo.rhein-neckar.de (Andre Fachat)
Subject: Re: PET 8296-D info sought
Date: Mon, 5 Dec 1994 09:18:34 GMT

> The only thing to change is the Character Clock frequency. But if you have
> the video access synchronised with the CPU memory access, it is not that
> easy to simply switch the frequency. Ok, you only have to fetch a byte
> in two CPU memory accesses, but without you just don't care:
> CPU Phi2 hi, Video Phi2 lo...
Of course you could also just set the Register 1, i.e. "Horizontal Displayed
Characters" to 40 instead of 80. But then the machine only uses half of the 
screen and the characters have the same width as before.

Andre
-- 
Andre Fachat                 mail me!          fachat@galileo.rhein-neckar.de
FIDO: Andre Fachat@2:2468/6020.4           GERNET: Andre Fachat@21:100/3338.4
              Why use Windows, since there is a door?
--
From: UKAH@ibm3090.rz.uni-karlsruhe.de (Veit Laule)
Newsgroups: comp.sys.cbm
Subject: Re: PET 8296-D info sought
Date: Tue, 06 Dec 1994 17:00


rhialto@mbfys.kun.nl (Olaf Seibert) writes:

>Right. There is another ROM labeled EDITOR inside. But there are still two
>more sockets filled with something ROM like. I'll write down the
>part numbers, perhaps that is useful.
Two PLAs (82S100), doing the master-timing and the (complex) memory-
managment.

>There used to be a program that made a 80xx into a 40xx (i.e. 40 column
>screen) by reprogramming the CRTC, and re-implementing the
>logical-screen-line stuff (where you glue two 40-col lines into a
>single logical linei - on the VIC-20 you could glue up to 4 lines of 22
>columns).
Yes, I have it.

>-Olaf.
>--
>___ Olaf 'Rhialto' Seibert    rhialto@mbfys.kun.nl     Ooey-Gooey-Fluffy-Barfie
>\X/ I'm not weird, I'm differently perceived. D787B44DFC896063 4CBB95A5BD1DAA96

Veit


From: bjheyboer@space.honeywell.com (Brian Heyboer)
Newsgroups: comp.sys.cbm
Subject: Re: 8032 Questions...
Date: 19 Feb 1996 17:20:44 GMT


RICHARD WHITE <richwhite2@mci.newscorp.com> writes:

>Picked up a PET (CBM) 8032 system today.
>Also, there are two add-on boards inside the unit:
>
>The first is plugged into the main motherboard IC socket 
>UD-11...  it has a small cable extension and terminates to 
>a small board with 2 IC's on it...  One IC is labeled 
>"JiniMicro", the other just bears the MOS #80357.
>
>The second board it plugged into UD-12 directly...  Its 
>labeled:  MICROTECH "Spacemaker II", has four IC sockets 
>(A,B,C,&D) with two of the four sockets filled with MOS 
>chips (A: "8032" and B: #901472-04), and also has a 
>three-conductor wire that comes off it and terminates to a 
>switch pasted on the outside of the case....
>
>What do these two boards do?  

UD11 and UD 12 are the "Option ROMs".  They were used by various pieces of
software either simply copy-protection "keys" or to contain all/part of the
program code.  The problem was, you could only have one in each slot.  The
SpaceMaker I am familiar with, and the switch selects which of the ROMs is
enabled.

The other one sounds similar, but since it has no switch may have another way
of accessing the two ROMs.

Now, you only need to figure out which ROMs they are.  UD11 maps at A000-AFFF
and UD12 maps at 9000-9FFF.  Look through the ROM code with the ML monitor and
see if you can figure out what they are for.  Some third-party companies did
have MOS make thier ROMs for them (my Visicalc ROM has an MOS part number, for
example), so that's not a hint unless someone recognizes the number.  

From: Rich White II <rwhite2@postoffice.ptd.net>
Newsgroups: comp.sys.cbm
Subject: Manuals for PETs!
Date: Thu, 07 Mar 1996 11:29:38 -0800

Hi!

Just got back from the guy I bought my PET from, and he located all the 
MANUALS for me!

I've got everything... including the Commodore manuals for the 8032, 
Basic 4.0, 4022 Printer, 8050 drive (its for a bunch of drives really), 
and the 8010 modem!

I also found a manual for one of those EPROMS I was trying to figure 
out. Turns out the thing is entitled:
       "COMMAND..0" (by Skyles), and is an EPROM that seats in UD3 on a 
4000 or UD12 on an 8000.  If you type SYS36864, it turns the thing on an 
gives you a few more BASIC commands.  (like auto line numbering, merge 
commands, BEEP, Execute, etc).  Pretty cool!

I also got a bunch of books dealing with BASIC, and documentations for 
the following software packages (haven't located all the disks yet 
though):
    FONTRIX,  VERSICALC,  WordPro 4,  Jinsam

Also located the manual, installation instructions, and programming 
instructions on the 64k MEMORY EXPANSION board.  Turns out it was a "kit" 
after all...   I have everything except the "TEST" disk that came with 
it.  But at any rate, I've got the whole box and all (shipped direct from 
Commodore on 3/82!)  

Needless to say... I'm a happy camper right now.  I'd like to stick some 
of this info online (maybe on a web page) some day for others to see.
Some impressive stuff, IMHO.  PET documentation seems to be quite hard to 
find anymore, too!

TTFN...
-- 
 //      Rich White II <*> rwhite2@postoffice.ptd.net <*> Reading, PA


From: brain@mail.msen.com (Jim Brain)
Newsgroups: comp.sys.cbm
Subject: Re: Manuals for PETs!
Date: 12 Mar 1996 21:52:24 -0500


Bjarke Johannesen <thygesen@datashopper.dk> wrote:
>> Needless to say... I'm a happy camper right now.  I'd like to stick some
>> of this info online (maybe on a web page) some day for others to see.
>> Some impressive stuff, IMHO.  PET documentation seems to be quite hard to
>> find anymore, too!

I was going to copy the docs at one point (I have 80XX stuff), but Ron Snyder
called and said he still sold the manuals.  Find his name in the list of
suppliers as US Commodore Support.  Just thought I would warn you.


-- 
Jim Brain, Embedded System Designer, Brain Innovations, Inc. (BII)(offline sig)
brain@mail.msen.com "Above views DO reflect my employer, since I'm my employer"
Dabbling in WWW, Embedded Systems, VR, Old CBM computers, and Good Times!  -Me-
<a href=http://www.msen.com/~brain/>Jim Brain: BII, VR, and CBM info</a>


From: crs0794@inforamp.net (Geoffrey Welsh)
Subject: Re: Sigh!  No PET schematics in existence?
Newsgroups: comp.sys.cbm
Date: 23 Mar 1996 21:54:32 GMT

Markus!  Long time no see...

I have a book called "The PET revealed" (somewhere... I think) and it includes 
the schematics for the original PET 2001 (9" _white_ screen, snow, no CRT 
controller, etc.)  It doesn't sound like the machine that you've got, but the 
book _might_ be of some help...

In article <4ifqh0$f0h@bcarh8ab.bnr.ca>, mwandel@bnr.ca (Markus Wandel) wrote:
>Someone gave me a PET that doesn't work.  Old 9" screen model, 4008, upgraded
>to 32K, pre-8032 motherboard with discrete TTL video.  I've already tried
>swapping all the LSI parts and the DRAMs.  Sometimes (one out of twenty)
>when powered on it drops into the machine language monitor, so quite a bit
>of it works, but the memory is partially corrupt -- the "AA" pattern has
>occasional other values in it.  The corruption pattern is ALWAYS the same
>and "G FD16" from the monitor restores it.  Visual inspection shows nothing
>wrong with the board, supply voltages are OK.  I'd like to debug it but it's
>hopeless without a schematic.
>
>Surely these schematics can't be completely extinct?  I mean, there had to
>be a service manual at some point.
>

--
            Geoffrey Welsh, Developer, InSystems Technologies Inc.
        Temporary: crs0794@inforamp.net; At work: insystem@pathcom.com
  At home: geoff@zswamp.uucp or [xenitec.on.ca|m2xenix.psg.com]!zswamp!geoff


From: crs0794@inforamp.net (Geoffrey Welsh)
Subject: Re: CBM PET documentation
Date: 24 Mar 1996 18:20:48 GMT


strigl@sxb.bsf.alcatel.fr (Dominique STRIGL) wrote:

>A friend of me got a PET and want to use the IEEE interface on it.
>Where can he find some documentation about the port Adresses ?
>Is there a FAQ somewhere ??

I used to have a book titled "PET and the IEEE-488 (GPIB) Interface", which 
gave all the details.  I'll try to look for it.

--
            Geoffrey Welsh, Developer, InSystems Technologies Inc.
        Temporary: crs0794@inforamp.net; At work: insystem@pathcom.com
  At home: geoff@zswamp.uucp or [xenitec.on.ca|m2xenix.psg.com]!zswamp!geoff


From: Dave Tickle <aiu214@freenet.mb.ca>
Subject: Re: 2 joysticks on a PET--how?
Newsgroups: comp.sys.cbm
Date: Fri, 28 Jun 1996 21:24:59 -0500



On 27 Jun 1996, Radioactive Warrior wrote:

> Not sure how they did it but a 4X4 bit matrix gives 16 keys...
> Or, at least, that's the way I'd do it...

I looked a little closer at the byte values one is supposed to get when 
reading the port, and it looks like it's set up like this:

           Joystick 1       |      Joystick 2
      up  down  left  right | up  down  left  right    
bit:   7    6     5     4      3    2     1     0

The fire button for either stick appears to be indicated by selection of 
'up' and 'down' simeltaneously, which unfortunately means you can't have 
up/down control while pressing the button.

By the way, this information comes from Compute's First Book of PET/CBM, 
which makes reference to the Kilobaud Microcomputing article instead of 
providing the circuit diagram.  Sheesh.  I'm thinking that the diodes 
must have something to do with the fact that the up/down/button lines are 
shared.

The book mentions that there are other joystick configurations for the 
parallel port out there; if anyone knows what they are, I'd like to hear 
about it.  (I might be able to figure them out by looking at some old 
programs.)  Personally, I would have done it either the way you (RW) 
suggested, or by using three bits for direction and one for the button.

Thanks for the reply,

Dave

From: vmuikku@yrttis.ratol.fi
Subject: Re: PET 200
Newsgroups: comp.sys.cbm
Date: 2 Jul 1996 09:33:22 GMT

In article <381.6749T793T178@academy.bastad.se>,                           
   anderse@academy.bastad.se (Anders Erlandsson) wrote:           
>What is this machine PET 200? How old? Built in basic?
>Is it a one piece machine or is there separate monitor?
>
>Is it worth anything?

  I have one of these. It says "Commodore Model 200" on the front panel.
But on the backside it says 8032-SK, and by this name it is listed in
Jim Brain's Canonical List of Commodore Computers.
  This machine has basic 4.0 and 32 kB of RAM. Basically it is just
CBM8032 in a new case. It has a built-in monitor and a separate keyboard.
  To my knowledge, there were also models 210 and 220. Most likely these
were not sold as 200 series machines in USA.

Ville


From: rhialto@mbfys.kun.nl (Olaf Seibert)
Newsgroups: comp.sys.cbm
Subject: Re: CBM 80xx
Date: 8 Aug 1996 19:14:20 GMT

In <4u7bl2$li5@tuegate.tue.nl> martijnb@mud.stack.urc.tue.nl (Martijn van Buul) writes:
>My PET doesn't have an internal tapedrive. As the matter affect, I use the
>same tapedrive for both PET and C64. However, the PET seems to have problems
>reading the C64 tapes ( I can't even transfer data bytewise. The PET reads
>the file header, but that's about it.). I *do* have a scheme how to turn
>my 1541 into a 2031 (== IEEE port), but that hack requires an EPROMmer, wich
>I haven't got...

Remember that the PET does not have a relocating loader. It expects
Basic programs to be at $0401 and it does not move them if they aren't.

Also the PET does not understand C64 files saved with SAVE "x",1,1.
It does not know that type of file headers and just moves on to the
next one.

>Martijn
-Olaf.
--
___ Olaf 'Rhialto' Seibert      rhialto@mbfys.kun.nl      The only excuse
\X/    for making a useless thing is that one admires it intensely. -O.W.


From: martijnb@mud.stack.urc.tue.nl (Martijn van Buul)
Newsgroups: comp.sys.cbm
Subject: Re: CBM 80xx
Date: 14 Aug 1996 07:27:51 GMT

James Tousignant (jtous@pcpros.net) wrote:
: Mike Neus wrote:
<Snippety snip>
: > 
: > I seem to remember hearing these types of problems from other PET users.  I
: > don't know exactly why it is, but timing sticks in my mind.  Does the PET run
: > at exactly 1MHz?  It might be a simple difference in CPU speeds as this will
: > affect software timing delays...


: The PET-C64 tape differences have nothing to do with timing. The PET
: does not
: support relocatable loaders. This means at will always load a program
: from 
: the memory area that it was saved from.
: In other words, if you try and load a C64 BASIC program into a PET the 
: program will load at the C64's start of BASIC address not the PET's.
: If you move the PET's start of BASIC (if memory serves me, it's at
: $0400) to
: where the C64's start of BASIC is, your programs will load just fine.
: You do this by POKEing the start of BASIC address on the zero page.
Yes. I'm aware of that. But! If you're correct, transfer of >>data<<
files (via put# and get#) shouldn't be a problem, since this has
absolutely nothing to do with BASIC start. However, I didn't even
manage THAT. (and YES, I know how to program in Basic :)

However, Mike's answer might be the clue to the problem. Remember : 
I'm talking about an european C64 (hey! I'm dutch :), while most of you
have an American (NTSC) C64. The PAL version isn't timed exactly the
same as the NTSC one. (If I recall correctly, the PAL version clocks
at 985 Khz, while the NTSC version clocks at about 1.12 Mhz.) I don't
think that a PET (with it's built-in CRT) has different versions. If
tapetransfer should work with american C64's, it probably won't with
European C64s (since the looped delays are different.) I know there's
a jumper somewhere inside the C64, to turn an PAL machine to a 
NTSC machine (as far as the CPU speed is concerned). I haven't been able
to try fiddling with that jumper. It might work...


Thanks for your help, all of you :)


: 			JT
Martijn

-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-
          Martijn ('PINO') van Buul, martijnb@stack.urc.tue.nl
-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-=+=-