Re: 6545 40/80 col register values

From: André Fachat <>
Date: Thu, 26 Apr 2012 07:36:34 +0200
Message-ID: <>
> Von: Rhialto <>

> On Wed 25 Apr 2012 at 09:26:43 +0200, "André Fachat" wrote:
> > (BTW: the CRTC emulation in VICE is cycle-exact in case you want to
> > test it :-))
> Too bad the rest of VICE xpet isn't (apparently): the Cursor program
> HiRes (which seems to fiddle with the character matrix while it is being
> displayed) doesn't work. On the other hand, I've never seen it work
> properly on a real PET; maybe it is a 50/60 Hz difference? (although the
> pre-CRTC PETs are all supposed to be 60 Hz I think?)

Hm, at least I think it is cycle-exact. You can "in-flight" re-program the CRTC registers and it uses them appropriately. I thought I had the memory fetch done as well, but it's been a long time. Maybe test it over the weekend...

On the other hand, changing the memory at the right time is tough. The PET only has a single interrupt on the vertical blank, and from there it is only counting cycles. And those timings - position of the vert.blank, line widths and so on - varies in an amazing range. Not only between graphics and upper/lower case but also between ROM versions.

For example the original PET (9" screen, without CRTC) had exactly 16000 cycles, making it a 62,5Hz screen refresh. The (European) 12" PETs with CRTC in general have 20000 cycles, which is good for 50Hz.

Looking at the schematics of the 9" screen indeed the screen rectifies the 9(?)VAC it gets directly. I would have thought this one reason for the different refresh frequencies.

Another one is probably that all the normal light sources in a room have a ripple with the line frequency. So if the screen would have a different frequency, it would give screen effects with the beat frequency ("Schwebung" in German, the time equivalent of the spatial Moire effect).


Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir
belohnen Sie mit bis zu 50,- Euro!

       Message was sent through the cbm-hackers mailing list
Received on 2012-04-26 06:00:30

Archive generated by hypermail 2.2.0.