Enhanced ARX True High Resolution Display Driver

[Moggy]

I'm also wondering what the 128 char mode is.

It allows you to use the inverse chr set.

[Fruitcake]

While testing the new driver I noticed that DSTAR crashes in 192 rows mode unless I add at least one NOP instruction.
I'm not sure whether the crash is caused by a memory alignment problem or timing/sync issues, other programs run smoothly on EO.

Can you share the program with and without the NOP and I will then see if I can work out why that NOP makes a difference.

[Moggy]

Thank you. Could you test both the programs? They seem identical but one of the two includes a break keys detection

Break key didn't work on either of them.

[Moggy]

While testing the new driver I noticed that DSTAR crashes in 192 rows mode unless I add at least one NOP instruction.
I'm not sure whether the crash is caused by a memory alignment problem or timing/sync issues, other programs run smoothly on EO.

The Dark star in your sampler runs OK using UDG4ZXpand but as soon as the first screen is cleared it resets the Zeddy when used with the ZXpand overlay ROM but OK when the CONFIG "X" command is used to switch it out.

[stefano]

the wrx driver relies on ROM routines, it could be that something is different on the improved ROM.

@Paul, I will provide a build with and without the NOP.
Oddly it was only DSTAR crashing, and only in full screen mode.

The BREAK feature is rather subtle, you probably can't experiment it on the example I've given.
When a program ends, one of the two drivers simply gets back into text mode. The other one stays in graphics mode until SPACE is pressed, iirc.

[Moggy]

The Zxpand uses the equivalent of the 622 (issue 2 ROM ) ROM according to Charlie, and relies on the Zeddy ROM for the CHR table.
When switched out, in my case, Dark star is running on the 649 ROM (Issue 3. ) and works OK.

In both cases constantly pressing break just returned both files to the K cursor.

[stefano]

Well, if the picture is showing correctly in both the cases, we are going in the right direction.

The sampler includes something running in 64 rows mode.
I noticed that for some reason the interrupt timer runs twice faster which could mean the driver is refreshing at doubled rate when in 64 rows mode?


I haven't fully understood what is the 'X' excluding. Is it hardware or software stuff?

I'm not sure I can work on it this weekend, but hey, it is for fun

[Moggy]

In very simple terms, when the "X" command is used it means the ZXpand ROM is switched out and the ZXpand becomes just a 32k RAM pack with the ZX81 ROM now being in control.

[stefano]

The ROM entries I'm using at the moment:

$29B - Other drivers use $29E only. Save blank lines, start NMI, POP registers and RETURN
$29E - Update A', start NMI, POP registers and return to BASIC/programs
$220 - PUSH register, then do VSYNC and get KEYBD
$2A2 - Enable the NMI generator and return to the user program, which will be interrupted as the top border is being generated.
$2A4 - Used to get back in normal text mode when BREAK is detected (only witht if the BREAK detection is enabled)
$F2B - SLOW mode. Used while setting up the hrg mode.
$F23 - FAST mode. Almost never used (not in the programs I've shared), but it is in.

[stefano]

This is DSTAR without NOPs added.

dstar_crashing_on_zxpandplus.p.gz

(2.84 KiB) Downloaded 20 times

I think in my original driver I've put an even longer delay block, by the way here's the current fix wit a single NOP I'm using in the updated driver.

dstar_extra_nop.p.gz

(2.82 KiB) Downloaded 22 times

They are built with the "BREAK" feature, it means that if you press 'G' and exit from game, you still have to press "BREAK" to get back to text mode.