[Total Eclipse]

Much as I hate to revive old posts, I’ve recently been through the hassle of trying to get a PAL signal working in anything more than x2 (in this case, an Amiga, ST and a load of 8 bit machines, all courtesy of a MiST FPGA board (which should be outputting a largely identical signal to real hardware).

I’d picked up a 10 year old Samsung 55 inch TV recently, and it’s *really* picky about the signals that it receives.  However, I’ve also got a much more recent Sony TV that seems to take anything – It takes anything out of spec, like 1318×1080 and just displays it.  This meant I could adjust my OSSC settings and use the Sony TV to see exactly what was actually being output.

 

And after half an hour of messing around, I’ve got a x4 profile that gives a 1920×1080 50hz image on a screen that seemed to be completely incompatible at first glance.

 

The settings I’ve dialled in are as follows:

H. samplerate 2164.00
H. srate frac 2164.00
H. synclen 77
H. backporch 170
H. active 1920
V. synclen 3
V. backporch 31
V. active 270
H. border 91
V. border 0
Sampling phase 180 deg

 

This is using firmware 1.08.  Hopefully it might provide a useful starting point for anyone else trying to get a PAL 50hz image on a picky screen.

 

EDIT: I appreciate that some fairly similar figures were posted right at the beginning of this thread, however the settings I had to dial in are very precise.  Even setting the H. samplerate to say, 2166 instead of 2164 would have some odd consequences.  If I recall, the Samsung TV would then receive a 1922×1080 signal instead of 1920, and simply refuse to display it.

• This reply was modified 2 years ago by Total Eclipse.

[Total Eclipse]

Adjusting the coast settings doesn’t make any difference.

By way of an experiment, I’ve connected the ST to another LCD TV that has a direct RGB SCART input. For each of the problematic demos, the TV shows the full picture, however there are definite sync issues. The Cuddly Demos fullscreen displays sync tearing at two or three points on the screen, suggesting that for some scan lines the sync pulses are out of spec compared to the rest of the screen.

Similarly the Union mainscreen is displayed correctly apart from the scanlines containing the full-width scroller, which again struggles to sync properly and has random horizontal distortion. When the OSSC tries to display this, you actually see about half the screen displayed just for a single frame, then sync is lost. I assume this is the OSSC being able to display the incoming field right up to the point of the crazy sync pulses for the scroller, at which point it can’t lock on any more.

Incidentally, connecting via a RF modulator doesn’t have any screen distortions at all.

So while it’s a little disappointing that the OSSC can’t sync in these instances, it’s not too surprising either. I’m not sure whether there are any other settings that might improve the chance of getting the OSSC to lock on to the signal? The wiki manual is great for explaining exactly what each setting does, but is there a more simplistic guide to tweaking?

[Author]

Posts