Problems with DOS using an Y-type connection
Re: Problems with DOS using an Y-type connection
Thanks guys. I think I understand a bit more now.
Re: Problems with DOS using an Y-type connection
I can confirm that Pere's streaming programs do work with the CocoSDC and WordPak 2+. So it is obviously something environmental being introduced or exacerbated by the DMSX2+ or more specifically the CPLD.
Re: Problems with DOS using an Y-type connection
I can't resist showing off the difference in image quality between the WordPak 2+ and the DMSX2+!!
I think the WP2+ unfortunately suffers from VRAM interference, hence the jail bars.
I think the WP2+ unfortunately suffers from VRAM interference, hence the jail bars.
- Attachments
-
- 20210105_213542.jpg (2.79 MiB) Viewed 8183 times
Re: Problems with DOS using an Y-type connection
Hello John,
this is strange, I have not this problem.
What mode where you looking at? Let me know and I will try it tomorrow...
The best images use to be for BASI12 which uses two frames for each image.
But the other resolutions give very acceptable images too and no vertical noise bars at all ...
cheers
pere
this is strange, I have not this problem.
What mode where you looking at? Let me know and I will try it tomorrow...
The best images use to be for BASI12 which uses two frames for each image.
But the other resolutions give very acceptable images too and no vertical noise bars at all ...
cheers
pere
Re: Problems with DOS using an Y-type connection
It may just be my cable. I wouldn't worry too much. But are you sure you get no bars at all? This is with BASI12I.
Re: Problems with DOS using an Y-type connection
Hi John,
most probably I will not have the time untill late evening ...
But then I will take some photos of a few images in this resolution
BASI12 is using YJK color mode that allows for 19268 colours on screen
with 256x424 pixels resolution
cheers!
pere
most probably I will not have the time untill late evening ...
But then I will take some photos of a few images in this resolution
BASI12 is using YJK color mode that allows for 19268 colours on screen
with 256x424 pixels resolution
cheers!
pere
Re: Problems with DOS using an Y-type connection
Hello John,
sorry for the delay, but it has been impossible to take the photos earlier ...
I attach here the eight images we use for testing and I can asure you that vertical bars do *not*
appear on the screen on any of them.
But taking a photo opens a door for camera-screen effects and you will notice some vertical bars
on Image1 but not in the others ...
These have been made running BASI12.BAS too
By the way, using the special YJK modes that mix data from 4 vertical dots, we know that there will
be artifacts and some curved lines may appear as ladder steps but not too much.
If an image gets this effect in excess, then we should choose another resolution with just 256 colours
cheers
pere
sorry for the delay, but it has been impossible to take the photos earlier ...
I attach here the eight images we use for testing and I can asure you that vertical bars do *not*
appear on the screen on any of them.
But taking a photo opens a door for camera-screen effects and you will notice some vertical bars
on Image1 but not in the others ...
These have been made running BASI12.BAS too
By the way, using the special YJK modes that mix data from 4 vertical dots, we know that there will
be artifacts and some curved lines may appear as ladder steps but not too much.
If an image gets this effect in excess, then we should choose another resolution with just 256 colours
cheers
pere
Re: Problems with DOS using an Y-type connection
It must be my WordPak 2+ cable at fault then.
Re: Problems with DOS using an Y-type connection
OK - the immediate issue is now solved. Many many thanks go to Pere and Darren Atkinson for their assistance with the random DOS failures issue.
The fix has been to add pull-up resistors (4k7) to the address bus and RD/WR. I made the decision to mod the Oojamaflip rather than the DMSX2+, as the Oojamaflip can then function as not only a splitter, but also a clean-up device.
I'll be creating a new version of the Oojamaflip board which will have space for these pull-up resistors to be added somewhat more neatly. .
The theory behind this solution was explained by Darren Atkinson as being down to the various DOS versions using the SYNC command in different ways. Essentially, when the 6809's SYNC command (0x13) is issued, the address bus, data bus and R/W signals are set to open-collector/high-impedance. For some reason which is beyond my means of investigation, the combination of both the CocoSDC and the DMSX2+ is causing signal levels on the address bus and/or the R/W signal to fluctuate, causing spurious writes at a time when the main system should be idle, just waiting for an interrupt to arrive from the DOS hardware.
The Dragon, Coco 1 & Coco 2 all seem to suffer from similar issues (though in the latter examples, this is Darren's experience, and not my first-hand knowledge). Studying the schematics, it can be seen that for the Coco 3, Tandy did indeed add pull-ups on the address bus and the R/W line. https://colorcomputerarchive.com/repo/D ... Sopko).pdf.
Thanks all for your input on this issue.
The fix has been to add pull-up resistors (4k7) to the address bus and RD/WR. I made the decision to mod the Oojamaflip rather than the DMSX2+, as the Oojamaflip can then function as not only a splitter, but also a clean-up device.
I'll be creating a new version of the Oojamaflip board which will have space for these pull-up resistors to be added somewhat more neatly. .
The theory behind this solution was explained by Darren Atkinson as being down to the various DOS versions using the SYNC command in different ways. Essentially, when the 6809's SYNC command (0x13) is issued, the address bus, data bus and R/W signals are set to open-collector/high-impedance. For some reason which is beyond my means of investigation, the combination of both the CocoSDC and the DMSX2+ is causing signal levels on the address bus and/or the R/W signal to fluctuate, causing spurious writes at a time when the main system should be idle, just waiting for an interrupt to arrive from the DOS hardware.
The Dragon, Coco 1 & Coco 2 all seem to suffer from similar issues (though in the latter examples, this is Darren's experience, and not my first-hand knowledge). Studying the schematics, it can be seen that for the Coco 3, Tandy did indeed add pull-ups on the address bus and the R/W line. https://colorcomputerarchive.com/repo/D ... Sopko).pdf.
Thanks all for your input on this issue.
- Attachments
-
- Oojamaflip_2021.jpg (148.03 KiB) Viewed 8081 times
-
- Oojamaflip_2020_modded.jpg (205.01 KiB) Viewed 8081 times
Re: Problems with DOS using an Y-type connection
we are happy enough to have a hardware guru like Darren Atkinson helping us to solve awkward problems
Thanks a lot, Darren, you are vey kind!!
cheers
By the way, I could not resist and have taken two resistor arrays of 4k7 and a std resistor 5k6 .. the only components I have right now
and have added the pull-ups to the oojamaflip.
Now I can execute all of the programs without any issue! Really great news!!
I attach a photo of the modified oojamaflip
cheers
pere
Thanks a lot, Darren, you are vey kind!!
cheers
By the way, I could not resist and have taken two resistor arrays of 4k7 and a std resistor 5k6 .. the only components I have right now
and have added the pull-ups to the oojamaflip.
Now I can execute all of the programs without any issue! Really great news!!
I attach a photo of the modified oojamaflip
cheers
pere
Last edited by pser1 on Sun Jan 10, 2021 11:02 am, edited 1 time in total.