The Sinclair QL Forum

I removed the crystal of my most abused QL (by the previous owner!) and wanted to feed an external clock, only to note that I do not have any source capable of doing so. Even the very expensive NI DAQ hardware I have here from work couldn't meet the specs But fortunately I had also ordered some 15.36 MHz crystals and they already arrived today. So put I that one in: Then I checked out an existing QL-VGA with it, the effect is very similar, just even more "sweet spots" and more pixels missing on the left side (something I missed during inspection of the original picture): Then I applied my fix and voila, picture becomes rock stable despite the 10.24MHz pixel clock: So I'm fairly confident now that I can get US QLs to work, too, though probably not in NTFS mode (so "normal" OS like Minerva might still be required). I guess the additional board just helps with the slew rate and doesn't affect the clock otherwise. Just a pity the increased clock wasn't documented before. Or rather anything regarding the NTFS QLs...

By the way, the QL obviously boots fine and even F1/F2 are accepted, but ultimately crashes when accessing the Microdrives.

Cheers, Marcel

Doing the calculations, the crystal in US QLs should give you four of those "sweet spots" per line. You can see 3. the fourth being off the screen. The 15.36 should give something like 16.

varmfskii wrote: What does the 74HC mod do?

https://www.theqlforum.com/viewtopic.php ... slew#p7549

Silvester wrote:

varmfskii wrote: What does the 74HC mod do?

https://www.theqlforum.com/viewtopic.php ... slew#p7549

Yeah, that's what I speculated, too.

The CLA2345 of later European QLs have the NTSC mode implemented, too, just at a slightly lower frequency due to the different crystal. But this way I was able to implement support in QL-VGA for it. Not that it makes much sense, the QL only has 192 lines in NTSC mode, the rest will be black, but it's nice to have a complete solution (hope my maths works with the faster crystal, too ) This is what the JSU ROM looks like, it boots completely in NTSC mode, so F1/F2 is much higher up:

But for JSU in F1 mode you get the full 256 lines, no?

bwinkel67 wrote:But for JSU in F1 mode you get the full 256 lines, no?

Yes, character height is 10 pixels in F1 mode and 8 in F2.
Still wondering what the timings are in F1 mode, though. No hardware here to check...

I was just asking for the sake of the QL-VGA card...so it sounds like in F1 mode for JSU you get the full screen.

Yes, "monitor mode" has full 256 lines.

So on JSU QLs the F1/F2 selection actually matters as there are 4 possible modes. Issuing a "mode 8" after booting with F1 stays in monitor mode, i.e. 256 lines with 8 colours, while selecting F2 from the start gets you 192 lines in 8 colours.

Still wondering what the timings are in F1 mode, though. No hardware here to check...

Don't have any hardware either, but I think there's only one possibility: the same timings as with European QLs multiplied by 15.10489/15 = 1.007. So I expect the poll interrupt to be 50.35Hz for example in F1 mode and 59.6Hz in F2 mode (provisional data: 192 visible lines, 34 lines front porch, 4 lines vsync, 34 lines back porch = 264 lines at 15.734 kHz line frequency). The hardware is too simple to behave in any other way I think.