Custom Profile (Osborne 1)

[dadecoza]

I'm having an issue creating a custom profile,

According to the Osborne 1 Technical Manual the settings should be as follows ...

Width in pixels: 416
Height in pixels: 240
Dot clock: 8Mhz

However the output I get looks like ...

Any tips?

Thanks!

[IanSB]

@dadecoza

According to the manual you linked, the actual crystal frequency is 15.9744Mhz:

So the "8" Mhz dot clock is 7987200 not 8000000 which might fix the mis-sampling but it looks like the image is shifted which can happen if the Hsync is wider than than the blanking period and not a normal hsync.
After tweaking the clock, can you post screencaps of the source summary page in the info menu and the geometry and sampling menus.
Press the Up and Down buttons together to screencap a menu, files in /captures on the SD card

[IanSB]

@dadecoza

Looking at this further, the hsync signal is, as I suspected, not a proper hsync which usually lasts for 4.7uS and it is instead a square wave:

The profile is currently triggering off the edge in the middle of the screen so try changing the sync edge setting in the sampling menu which will make it trigger off the other edge.

[dadecoza]

Progress!

Changing the edge to trailing fixed the alignment.

However it is still a little noisy.

[IanSB]

@dadecoza

Have you tried adjusting the line length + or - 1, 2 or 3
(If the screen all goes noisy instead of clean on one of those settings that is probably the right value but you then have to run the calibration)

Also where have you connected the RGBtoHDMI inputs?

[dadecoza]

I have tried playing with the line length but it did not help much. Not giving up though will mess around some more,

As for how it is connected ...

Osborne 1	RGBtoHDMI
VSync	VSync
HSync	HSync
Video Out	Grn3
Gnd	Gnd

[IanSB]

@dadecoza
The video output is really an analog signal which is at least one reason why the signal is noisy as you are feeding that into a digital input. Looking at the schematic:

The video is fed from a contrast control which will vary the voltage level and also the low intensity setting will reduce the voltage level too when active so characters with that attribute may appear noisy or blank.
Normally I would suggest using an analog interface addon but that can only use composite sync and in this case you can't easily combine the H and V sync signals with diodes to make composite sync because the Hsync signal is a very non-standard square wave.

It should be possible to modify an analog interface to inject separate syncs if you want to go that route but to get this working properly with the digital board you would have to connect internally to the inputs of those four open collector (O.C.) drivers shown above (Hsync, and Vsync with Grn3 and Grn2 connected to the video and intensity signals)

Turning up the contrast to max might help but the low intensity setting is still going to reduce the voltage level anyway.

[dadecoza]

Hi Ian,

Thank you so much for your patience and detailed explanation.

As you probably gathered my knowledge of how video output work is very limited. At least I now have an explanation for the noise, and it really is not that bad,

Thanks again and feel free to close this ticket,

[IanSB]

@dadecoza

Can you post your current profile.
Will either be in /Saved_profiles if you made any changes after creating your custom profile or if not then it will be in /Profiles

[dadecoza]

Custom_Profile_0_.txt

as requested...

[IanSB]

@dadecoza
There were some settings in your profile that needed tweaking:

As the hsync edge is so close to the active video that can cause problems so I switched from Leading to Leading+Delay (that also meant changing to 3bpp capture). Plus I set the pixel aspect ratio and the palette to mono.
It should still be possible to get this working with intensity although the connections might have to be different with a custom palette.

BTW the line length of 512 looks correct as it is a nice round binary number so all the noise is probably due to the analog signal.

Can you check that this profile is still OK and if so I will add it to future releases as it at least gets the timing correct so would be a good starting point for a fully correct profile:

Osborne_1_(No_Intensity).txt

Put it in this folder on the sd card:

\Profiles\6-12_BIT_RGB\_Other

reboot and select it from 'Other' in the profile menu

[dadecoza]

Looking good!

[dadecoza]

Seems the intensity bit is more important than I thought ... In the above example the CP/M logo has a lower intensity and is completely absent from the RGBtoHDMI screen capture.

[IanSB]

@dadecoza

Seems the intensity bit is more important than I thought ... In the above example the CP/M logo has a lower intensity and is completely absent from the RGBtoHDMI screen capture.

There are several options to implement this:

1. Internal TTL connection
2. Use an RGBS analog addon board (some mods will be required to get the separate vsync in)
3. Use a mono/lumacode board. This is a board that contains the CPLD and part of the analog board, enough for multi level mono signals and it is mainly used with c0pperdragon's lumacode adapters to get RGBtoHDMI working with systems like the C64 / C128, Atari 800 & 2600, TI99 etc but it can be configured by using solder jumpers to input analog video and separate H and V syncs on a 3.5mm TRRS connector and that option was included for this kind of situation. (I have tried that on my Commodore PET.)
4. Use a couple of analog comparators on a breadboard between the Osborne 1 socket and the RGBtoHDMI to extract the two analog video levels. (This is effectively what the analog or mono/luma board does)

[dadecoza]

Great thanks. :)

Our local vintage computing group is planning an "Osborne repair day" and I will then take the opportunity to try some of your suggestions. The repair day is also the reason I wanted to get the RGBtoHDMI working so that we can bench test/debug the main boards outside of the case.

[dadecoza]

But man it still looks great!

[IanSB]

@dadecoza
Here is a speculative profile for intensity:
Osborne_1_(With_Intensity).txt

I had to revert to Leading rather than Leading with delay as that only works at 3bpp and intensity support requires 6bpp
Can you check this still works. The image will be dimmer as there is no intensity bit until you connect that to Grn2

[dadecoza]

Amazing