exxos's Atari, Amiga & retro forum

It does sound like the video circuit is probably a not working properly. I guess that it's character ROM based rather than bitmapped?

stephen_usher wrote: 25 Jan 2024 22:20 It does sound like the video circuit is probably a not working properly. I guess that it's character ROM based rather than bitmapped?

Yes, there is a ROM next to the video RAM called CHARGEN. If the machine does not boot it should fill up the screen with characters from that ROM.

Hope you'll figure out soon what the issue is and get your OSI working.

rubber_jonnie wrote: 25 Jan 2024 23:33

stephen_usher wrote: 25 Jan 2024 22:20 It does sound like the video circuit is probably a not working properly. I guess that it's character ROM based rather than bitmapped?

Yes, there is a ROM next to the video RAM called CHARGEN. If the machine does not boot it should fill up the screen with characters from that ROM.

Time to 'scope the ROM address and data lines then to see if it's being accessed, if not follow the trail up stream until something works. :-)

stephen_usher wrote: 26 Jan 2024 08:12

rubber_jonnie wrote: 25 Jan 2024 23:33

Yes, there is a ROM next to the video RAM called CHARGEN. If the machine does not boot it should fill up the screen with characters from that ROM.

Time to 'scope the ROM address and data lines then to see if it's being accessed, if not follow the trail up stream until something works. :-)

I did and there is activity, as is there on the video RAM address and data lines.

Somebody over at OSIWeb suggested a blind typing test to see if the system is running, but that didn't work, but there could be reasons for that, but for now I will focus on just the video section, since it needs to pass the test of 'no code running' by putting random characters on the screen.

So that test is next and I have the scope positions and expected results at each position to check.

How is the video generated on the machine? Is it all TTL logic?

Oh, I've just had a look at the UK101 manual (https://uk101.sourceforge.net/docs/pdf/manual.pdf). Interesting design. Basically having the system clock selecting the character scan line row and the output from the RAM selecting the character.

Seeing as you have sync then the sync clock is working.

From the block diagram, even if the video RAM isn't communicating and the outputs from the counters are bad the character ROM will be outputting a value, unless you're really unlucky and it's always the SPACE character being selected. In which case it would have to be the chip called the "PISO" in the diagram not translating the output from the character ROM to the screen output.

stephen_usher wrote: 26 Jan 2024 09:41 How is the video generated on the machine? Is it all TTL logic?

Oh, I've just had a look at the UK101 manual (https://uk101.sourceforge.net/docs/pdf/manual.pdf). Interesting design. Basically having the system clock selecting the character scan line row and the output from the RAM selecting the character.

Seeing as you have sync then the sync clock is working.

Indeed, sync is definitely there, but no video. I need to look at all the parts that generate the display as that part is clearly not doing anything.

Since there should be a display and characters on screen even without a CPU fitted, I'll focus on that first, it is test procedure 6 in the build manual. I should have done it sooner, but because all my parts arrived at random intervals I didn't, but hey ho.

stephen_usher wrote: 26 Jan 2024 10:37 From the block diagram, even if the video RAM isn't communicating and the outputs from the counters are bad the character ROM will be outputting a value, unless you're really unlucky and it's always the SPACE character being selected. In which case it would have to be the chip called the "PISO" in the diagram not translating the output from the character ROM to the screen output.

I think the PISO is U42, an LS165. I'll give it a check later. I really should do some work now though!

Indeed, you should at least be getting horizontal sets of pixels being displayed even if the character ROM and video RAM aren't getting their addresses from the counters. Seeing as you say that the ROM data lines are active this means that it must be the encoder down stream, which is termed "PISO" here: