|Re: Problem: 2nd 32k of CY suddenly missing|
Message #3 Posted by PeterP on 6 June 2008, 11:05 p.m.,
in response to message #2 by Raymond Del Tondo
Well, this is what I first thought as well, and given that there is some good and detailed info here on the forum and I have all the necessary ingredients (HP-IL card, various MCODE moduls, etc) I thought it would be rectifyable and a fnu adventure.
But the behavior of that CY is a bit more intricate...
Here are the steps that I took and the result. If anyone can make heads or tails of this, I will be all ear!
- The start-configuration is a working configuration, with WWOS a in page 8 and all pages visible and usable. This is the first 32k (Block a)
- I then put a Clonix with the ML rom mapped to Page F in
- Via HP-IL/Emu41 I copied a WWOSa into page 9
- Then I did a PG01 which activated pages 9,11,13,15 from the second 32k (Block b)
- Via HP-IL/EMU41 I copied a WWOSb into page 9
- Now I put in a Games module in Port one to mask page 8
- I confirmed that PG01 has the WWOSb in page 9 and PG10 has the WWOSa in page 9
- I now replaced the Clonix with a Zenrom so that I have a Rameditor available.
- I then copied a 4k module (XIO) into all pages in both configurations. At that point, we had OSa in page 9 for Block a and OSb in page 9 for Block b, XIO in page 10,11,12,13,14 (presumambly for both blocks but, as we shall see shortly, not so much) and the Zenrom on page 15
- I now did a PG01 to have 10/12/14 from Block a and 9/11/13 from Block b
- using MCED from the Zenrom, I now changed the word xFFB/xFFC in all pages, namely 10,11,12,13,14 and confirmed that the changes actually took place
- now I did a PG10. My expectation was that I will have OSa in page 9 and XIO in all other pages, but unaltered at xFFB/xFFC
- well, not so much. As it turns out, this assumption is correct for 11 & 13, but not for 10,12,14!!!
- so it would appear that the odd pages are switchable (9/11/13) but there is only one set of even pages visible (10/12/14)
This is truly odd, as I can not imagine any HW defect that would cause this.
The one thing that gives me hope is the odd behavior whenever I do a PG10 (or the second PG<>) - the calc goes into a strange mode and is recovered after a short while or a Enter/On. This seems like some odd code in the polling area of one of the pages.
I read all the threads that can be found here and they show quite a few tips and tricks on how to successfully recover the OS for the CY. If the OS on one Block were gone, I believe that the above sequence should have restored full access to all pages with the possible exception of page 8 in Block b (as that was always masked). And with the right OS in page 9, and page 8 masked, the Block b should be quite happy and appear and respond normally to PG<> and PG10.
So I can't make sense of this odd behavior. Any thoughts?