| |
chatGPZ
Registered: Dec 2001
Posts: 11150 |
VSP Hardware fix revealed
click here
Many thanks to everyone involved. HAVE FUN! |
|
| |
MagerValp
Registered: Dec 2001
Posts: 1060 |
This is awesome, thank you! |
| |
White Flame
Registered: Sep 2002
Posts: 136 |
Such a simple solution to such a long-standing nuisance!
As somebody whose machines are already VSP compatible, I hope this helps make more real hardware compatible as well. |
| |
ChristopherJam
Registered: Aug 2004
Posts: 1382 |
Excellent, great to see. |
| |
CreaMD
Registered: Dec 2001
Posts: 3036 |
Revolutionary. Someone could maybe make a youtube tutorial / video? |
| |
AmiDog
Registered: Mar 2003
Posts: 97 |
Ok. Time for some stupid (?) questions. The point of the fix is to prevent the VIC from changing the address lines when it shouldn't, right?
So why is the fix applied "just before the address lines are fed into the RAM chips"?
Wouldn't it be possible to apply the fix at the VIC end instead? By creating a small board which you fit between the VIC and the VIC socket on the motherboard? I guess not since the page says "a fix would not work nicely as a customer-installable product", and installing such a board would be no harder than installing a 64KB VDC-ram board in a C128 for example. |
| |
AlexC
Registered: Jan 2008
Posts: 293 |
Awesome! Is there anywhere a list of releases using VSP? |
| |
ready.
Registered: Feb 2003
Posts: 441 |
other option would be to insert the board in between U13 and U25 (the two 257 MUX chips) and their sockets. |
| |
chatGPZ
Registered: Dec 2001
Posts: 11150 |
AmiDog: you can bet on that if such simple adapter solution would be possible, jens would have made a product out of it :)
"Is there anywhere a list of releases using VSP?"
ugh, no. that'd be a huge list =) however, the problem even exists for programs that dont actually use VSP, but simply switch $d011 without synchronizing to the beam - which may or may not trigger the bug (with a roughly 1:312 chance) |
| |
Bob
Registered: Nov 2002
Posts: 71 |
\o/ I can go hay wire on VSP now.. no limitations \o/ |
| |
Oswald
Registered: Apr 2002
Posts: 5031 |
Quote: Awesome! Is there anywhere a list of releases using VSP?
not until csdb implements tags :) |
| |
AmiDog
Registered: Mar 2003
Posts: 97 |
@Groepaz: I doubt that, since it probably wouldn't pass his "install by girlfriend test" which prevented an internal ACA500 for example. And this most likely wouldn't be any easier to install for said girlfriend... |
| |
ChristopherJam
Registered: Aug 2004
Posts: 1382 |
Ugh, phrasing. My wife's at least as capable as I am at installing hardware.
We've both dismembered various tightly packed laptops and small formfactor PCs to replace drives etc, and her soldering's a hell of a lot cleaner than mine is. |
| |
ready.
Registered: Feb 2003
Posts: 441 |
@ChristopherJam: you said wife, AmiDog was talking about girlfriend. Probably you wounldn't have married her if she did not pass the girlfriend test. |
| |
chatGPZ
Registered: Dec 2001
Posts: 11150 |
AmiDog: a much more important reason is that it will not work the same on different board revisions, and require soldering with some of them. ie: support nightmare. |
| |
Perplex
Registered: Feb 2009
Posts: 254 |
Quoting Bobo/ I can go hay wire on VSP now.. no limitations o/
In what way have you been limiting yourself up until now? :-) |
| |
AlexC
Registered: Jan 2008
Posts: 293 |
Quote: AmiDog: you can bet on that if such simple adapter solution would be possible, jens would have made a product out of it :)
"Is there anywhere a list of releases using VSP?"
ugh, no. that'd be a huge list =) however, the problem even exists for programs that dont actually use VSP, but simply switch $d011 without synchronizing to the beam - which may or may not trigger the bug (with a roughly 1:312 chance)
I never took a while to think about it but you are obviously right. Anyway I still wonder how many releases are being hit by VSP issue. In case of demos I doubt there is a lot of software not synchronizing with beam. |
| |
chatGPZ
Registered: Dec 2001
Posts: 11150 |
you can almost bet on that a random program that uses raster irq does a non synced write to $d011 somewhere at the beginning :) |
