Triple monochrome monitors from an Amiga a500

[Golden Child]

I had an idea about being able to use 3 monitors (in monochrome b/w) by redirecting each of the R G B lines to an individual monitor (along with the sync signals).

 

Using 1 bitplane for red, 1 bitplane for green and 1 bitplane for blue, you could do three 640x200 b/w monitors for a total mono resolution of 640x600 pixels.

 

Another possibility is that you could do a single b/w monitor with a second monitor that had 4 levels of grey.

 

That would be an interesting experiment.  I remember how the PC had the ability to do a mda as well as a cga simultaneously.

 

It might be a little bit easier now that you have cheap VGA 4 port splitters and cheap monitors.

 

Did anyone ever try this?

[Brek Martin]

On 8/11/2024 at 10:01 AM, Golden Child said:

I had an idea about being able to use 3 monitors (in monochrome b/w) by redirecting each of the R G B lines to an individual monitor (along with the sync signals).

 

Using 1 bitplane for red, 1 bitplane for green and 1 bitplane for blue, you could do three 640x200 b/w monitors for a total mono resolution of 640x600 pixels.

 

Another possibility is that you could do a single b/w monitor with a second monitor that had 4 levels of grey.

 

That would be an interesting experiment.  I remember how the PC had the ability to do a mda as well as a cga simultaneously.

 

It might be a little bit easier now that you have cheap VGA 4 port splitters and cheap monitors.

 

Did anyone ever try this?

It's harder to find a neat A500 schematic online than it was years ago. The vector schema page I used to use appears to be unresponsive.

I just wondered if the HSYNC, VSYNC, or CSYNC signals are internally buffered, or if the port pins are direct from Agnus.

The reason I bring it up is because I wonder if they intended the fanout of those signals like that.

On the CD32 debug board I had, the sync signal from the CD32 expansion connector was buffered with some stock logic on the debug board,

and that was just because the debug board provided an output for one single RGB monitor that the CD32 doesn't have.

 

You could tie one input of an AND gate high, send the sync signal to the other input, and there's your buffer.

 

Sounds like a cool idea, I just don't see a practical use when it will only work with your own custom software.

There could be practical use that I don't see of course.

 

 

 

Edited September 10 by Brek Martin

[Daedalus2097]

Yeah, the AmigaWiki page seems to have lapsed. There are a couple of people trying to get the various content and host it elsewhere, so I'm sure it'll be available again in some form or another. But to answer the question: the H- and V-sync lines aren't buffered and are driven directly from Agnus / Alice, with the intention being that they can also be driven externally by a genlock. Loading these signals too heavily will cause the machine to fail to boot because it sits waiting for an external clock signal that never arrives. But simple logic would work fine on them - after all, the standard buffered VGA adaptors simply use a 74 chip to buffer both sync lines.

 

C-sync is buffered before being fed into the video hybrid. I don't know how much load the hybrid is designed for, but putting too much load on that line won't disrupt the operation of the machine.

[Brek Martin]

4 hours ago, Daedalus2097 said:

Yeah,...

That should sort him out then, so long as he can make a small board that doesn't directly short the 5V supply,

and blow the current limiting resistor  

 

I don't know the Amiga enough from the software side to know that it's bitplanes are directly associated with

the hardware RGB colour channels, but you'd have pointed it out otherwise I guess,

or an image could be drawn with a palette that includes pure R,G,B components to achieve the same thing.

 

The whole idea has a slight "out of the box-ness" that I like.

 

 

 

 

 

[Golden Child]

My idea was to use a buffered VGA adapter on an a500 and use a 4 port VGA splitter which creates 4 vga outputs, duplicating all of the necessary sync signals.

 

I've used a 4 port VGA splitter which works fine with 15khz signals and will work with VGA monitors with 15khz support.

 

I bought some VGA terminal breakout boards from ebay so I could rewire them to make a custom pinout.

 

Another alternative is just to get a regular VGA cable and remove the pins you don't want to connect, so one monitor will only get the red signal, one will only get blue, and one will only get green.

 

Viewsonic makes some monitors that support 15khz and will also allow you to set it into a monochrome color mode.

 

 

[Daedalus2097]

On 9/10/2024 at 7:04 PM, Brek Martin said:

I don't know the Amiga enough from the software side to know that it's bitplanes are directly associated with

the hardware RGB colour channels, but you'd have pointed it out otherwise I guess,

or an image could be drawn with a palette that includes pure R,G,B components to achieve the same thing.

Well, the bitpanes don't have any alignment whatsoever with the colour signals. I was just assuming the OP would be using a specifically crafted palette to have it work that way. Setting pens 1, 2 and 4 to red, green and blue should have the desired effect.

[phoenixdownita]

1 hour ago, Daedalus2097 said:

Well, the bitpanes don't have any alignment whatsoever with the colour signals. I was just assuming the OP would be using a specifically crafted palette to have it work that way. Setting pens 1, 2 and 4 to red, green and blue should have the desired effect.

Well he needs to set all 8 so he can set a pixel on more than one screen at a time. Obviously each of the 8 colors in the palette will always carry max for r g b, but he does need to set all of them for example pen 8 (0x7 = 0b111) to 4095 (all 12 bits to 1) as that will turn on the same pixel on all 3 screens as needed (for example if he happens to set to 1 the same bit/pixel on all 3 bitplanes), pen 3 needs to max out (4 1s) the 2 respective colors  (whatever he sets on 1 and 2 … he can literally binary OR those 2) and leave zero in the rest etc…

Edited Thursday at 05:31 PM by phoenixdownita

[Brek Martin]

1 hour ago, phoenixdownita said:

Well he needs to set all 8 so he can set a pixel on more than one screen at a time. Obviously each of the 8 colors in the palette will always carry max for r g b, but he does need to set all of them for example pen 8 (0x7 = 0b111) to 4095 (all 12 bits to 1) as that will turn on the same pixel on all 3 screens as needed (for example if he happens to set to 1 the same bit/pixel on all 3 bitplanes), pen 3 needs to max out (4 1s) the 2 respective colors  (whatever he sets on 1 and 2 … he can literally binary OR those 2) and leave zero in the rest etc…

I was assuming the content on all three monitors is supposed to be unique, otherwise the method described to use three monitors is over-complicated.

 

"you could do three 640x200 b/w monitors for a total mono resolution of 640x600 pixels."

 

 

Edited Thursday at 06:26 PM by Brek Martin

[Brek Martin]

1 hour ago, Daedalus2097 said:

Well, the bitpanes don't have any alignment whatsoever with the colour signals. 

Of course they don't! A few more synaptic connections were made since my last post.

Addressing four monochrome bitplanes in parallel would provide sixteen indexes to address palette tables of wider values that can appear on-screen simultaneously.

 

 

 

[Daedalus2097]

Yup  I was assuming different content on each of the monitors too. You have 6 bitplanes at your disposal, so you could have 2 bitplanes for each display for 4 possible pixel values (e.g. Gameboy display). Unfortunately, there are only palette entries available for 5 bitplanes, so one of the displays would be just a single brightness level for the entire display.