QL Hardware Pause/Slowdown

[bwinkel67]

I saw this device on one of Rob Smith's videos (he's the maker of Amiga Drawbridge). Seems to be from the late 80s and was able to slow down and stop the Amiga. Also seems to work on a Sega Genesis.

Would something like tis work on a QL (noting that the HALT pin is 36 instead of 17...and 5V and GND are also different). Would be handy for debugging, I suppose.

It's pretty neat and you can see his video on it here where he breaks down how it works.

https://www.youtube.com/watch?v=03vjAGR6exU

[martyn_hill]

Hi Bwinkel

Unfortunately, that wouldn't work in the QL as Sinclair tied both CPU signals HALT and RESET together on the motherboard such that any attempt to activate the HALT line (whether externally, or via software) will trigger a full CPU reset.

[Derek_Stewart]

why slow down, its more speed I want.

However, it hasbeen proved with QLInvaders and Pitman, that with the correct programming, the speed of the processor does not matter.

[t0nyt]

why slow down, its more speed I want.

Testing/debugging?

[bwinkel67]

Unfortunately, that wouldn't work in the QL as Sinclair tied both CPU signals HALT and RESET together on the motherboard such that any attempt to activate the HALT line (whether externally, or via software) will trigger a full CPU reset.

Ah, thank you for the clarification. It's still neat how they did the slow down by pulsing the HALT signal controlled by the potentiometer. Curious if repeatedly halting the system like that doesn't cause other issues.

[Peter]

Unfortunately, that wouldn't work in the QL as Sinclair tied both CPU signals HALT and RESET together on the motherboard such that any attempt to activate the HALT line (whether externally, or via software) will trigger a full CPU reset.

Since the PCB plugs into the CPU socket, it would be possible to isolate the CPU's HALT pin from the motherboard.

[dilwyn]

why slow down, its more speed I want.

However, it hasbeen proved with QLInvaders and Pitman, that with the correct programming, the speed of the processor does not matter.

For new software, that's right. Plenty of older software around (for which sources no longer exist) which runs too fast on modern systems which even software solutions like SLUG and SLOWGOLD cannot always manage.

Like Derek, my main focus would be on ever faster systems. Tony might have a little point about testing and debugging, although again as Derek said there are easier ways around it when programming brand new software.

There was no way the authors of all the 1980s software could have foreseen systems with the speed of some modern systems.

[bwinkel67]

Since the PCB plugs into the CPU socket, it would be possible to isolate the CPU's HALT pin from the motherboard.

Oh, so the fact that HALT and RESET are tied together isn't critical to QL operations? If that's so, then this little board would be a pretty neat (and easy to implement) addition for the QL to be able to slow execution down if that would help to debug something (or for games since that was its original purpose). The key seems to be the TC4093BP chip, if I understand Rob's video correctly.

On a side note, I'm leafing through Andrew Pennell's QDOS Companion book on one of my last summer days before school starts going again. Did I post here that he seems to be active on one of the Sinclair Facebook groups (I think it's one of the Spectrum ones). He had an interesting quote on his forward:

"I would like to express my sincere thanks to Sinclair Research for their technical assistance, and Apple for teaching Macintosh about Me."

That last part was odd...

[Peter]

Oh, so the fact that HALT and RESET are tied together isn't critical to QL operations?

The QL would still work if they were not connected. If you have a spare 68008, you can try. Just bend pin 36 (HALT) to the side and connect it via a 1 K to 4.7 K resistor to pin 13 (VCC) on the opposite side of the IC case.

If that's so, then this little board would be a pretty neat (and easy to implement) addition for the QL to be able to slow execution down if that would help to debug something (or for games since that was its original purpose).

This barbarous method to generate the CPU HALT signal certainly has side effects. Various QL peripherals are CPU timing dependant.
But the risk to do actual damage except data corruption is low, so maybe just try it out.

[Derek_Stewart]

Hi

I thought about bending CPU pins out and soldering directly to the CPU. Sounds straight forward, but prone to damage if too much heat is applied to the pin.

Maybe wants a dedicated circuit board.

The point about testing and debugging, is it not better to use software debuggers to set breakpoint, and analyse variables?