That is a good start, but color #6 is dark yellow by that schematic when it should be brown. If you want to further, the attached schematic should give you a proper brown.

I just studied the CGA/RGBi dark yellow/brown problem. And i have learned this so far.
- On PC's with CGA the actual conversion from dark yellow (DY) to Brown (B) is done inside the monitor.
- The NTSC composite output from the early CGA was DY - NOT Brown.
- Most games made in the early days focused on the NTSC Composite output, and made the games for DY. Reference for this is CGA article on Wikipedia who gets this information from Byte Magazine https://archive.org/stream/byte-magazine-1982 … ge/n37/mode/2up
- The most common palette used for NON composite games does no use the DY/B. (my belief after looking trough many games.

- On Commodore 128 which also has RGBi output states in the C128 Programmer's Reference Manual, on page 35 that Brown is not listed as a colour in 80 column mode. Dark Yellow is.
- The Commodore 1902 monitor, which was produced for the C128, doesn't change dark yellow to brown, so the correct colour is dark yellow. Using a PC CGA monitor gives the wrong colour.

Is the DY/B discussion all wrong? Should not a correct RGBi to SCART adapter produce Dark Yellow - at least for the C128? On the PC the answer should be that some games are made for DY and some (newer) are made for Brown?

Jinxter wrote:

- The NTSC composite output from the early CGA was DY - NOT Brown.

The same goes for late CGA.
The 16 colours are slightly different, but the brown is still dark-yellow.

Jinxter wrote:

Is the DY/B discussion all wrong? Should not a correct RGBi to SCART adapter produce Dark Yellow - at least for the C128? On the PC the answer should be that some games are made for DY and some (newer) are made for Brown?

On PC I don't think it's a question at all: an RGBI to SCART adapter *must* produce brown, since that is what RGBI does.
NTSC composite is an entirely different mode. RGBI is 320x200, 4-colours. You can't get 320x200 resolution on composite, only 160x200. So what composite does on CGA is completely independent of RGBI. The RGBI modes don't work correctly on composite anyway, and games generally have separate composite modes, which use all 16 colours in 160x200 resolution.

Jinxter wrote:

Is the DY/B discussion all wrong?

RGBI monitors hooked up to an RGBI output produced brown on the display, so any conversion device must do the same.

What the composite output port did with colors is not relevant to the topic of converting RGBI output to something else.

Jinxter wrote:

- On Commodore 128 which also has RGBi output states in the C128 Programmer's Reference Manual, on page 35 that Brown is not listed as a colour in 80 column mode. Dark Yellow is.

This lists brown:
http://rvbelzen.tripod.com/c128sg/sect-06a.htm#6.2.1.3

Jinxter wrote:

- The Commodore 1902 monitor, which was produced for the C128, doesn't change dark yellow to brown, so the correct colour is dark yellow. Using a PC CGA monitor gives the wrong colour.

Are you sure?
Screenshot here shows brown: http://www.commodore128.net/index.php?topic=2209.0

The C128 manual says Dark Yellow, but the manual could be wrong. - It probably is.

Regarding the 1902 monitor i heard that from another person saying this on a Facebook forum for C128 Enthusiasts.

I think the discussion is dead. The CGA cards and the C128 deliveres Dark Yellow, but the screens convert it to brown. - As MobyGamer writes above.

Jinxter wrote:

The CGA cards and the C128 deliveres Dark Yellow, but the screens convert it to brown. - As MobyGamer writes above.

Technically they don't really 'deliver' anything. It's a 4-bit digital signal, which is interpreted by the monitor.
The only thing is that they have assigned Red, Green, Blue and Intensity to the 4 bits.
There's no 'yellow', let alone 'light' or 'dark' yellow.
It's not like an analog signal such as composite where the colours are already encoded into an NTSC color carrier, and basically 'set in stone'.

The CT1902 is a composite and RGB monitor, so you must make sure you are seeing the RGB output, not the composite output. Also some Commodore 1084s supported analog and digital RGB.

Great Hierophant wrote:

The CT1902 is a composite and RGB monitor, so you must make sure you are seeing the RGB output, not the composite output.

The C128 doesn't have composite for its 80 column mode though.
It's a weird setup, which is sorta 'dual monitor', kinda like having an MDA and CGA card in a single system, and selecting which one to use with 'MODE MONO' and 'MODE CO80'.
The VIC-II outputs the composite signal, the 80-column VDP chip outputs the RGBI signal.

bjt wrote:

Holy thread resurrect Batman! […]
Show full quote

Holy thread resurrect Batman!

First of all though, before I put this away last year I was still having intermittent problems with the battery light flickering on during disk access, and read failures to boot. This continued even when I replaced the floppy with the HxC emulator, so my next suspect was the power supply. Sure enough, some of the supply voltages were well out of spec.

The innards of the PSU. I can't see anything obviously wrong but I'm not about to poke around with that massive cap in there, so time to look for a replacement. Thankfully a kind soul posted a pinout for the power connector at http://www.actsirius1.co.uk/pages/docs/v86p_power.pdf. Note that I found Pin 1 to have voltage on it when I probed the original PSU, so I left it disconnected for the replacement.

The original PSU is 8.5V with fair whack of current, more than most cheap generic PSUs can supply.

I've also got a V86P, except mine's missing the PSU totally. I'm just in the process of gathering as much information about it as possible, then I plan to document it as well as I can and put it all online. So far, I've identified that the v86p_power.pdf pinout is indeed incorrect - pin 4 isn't just "part of the battery circuit" it's wired directly to the battery +ve terminal, which means if you have a battery pack in there and you blindly throw 8.5V from a PSU that can push up to 4A, you run the risk of overcharging the battery pack.

Pin 1 is also a bit of a mystery. The PSU connector on the system is attached to a little power input PCB, which I've documented here:

Do you still have the guts of the original PSU? If so, some hi-res photos of the board showing the components and traces might help (assuming it's just a single or double-sided PCB, and not multi-layer). Given pin 4 goes straight to the battery, I'm also trying to figure out if the charge logic is in the PSU, or on the motherboard.

I've also got a working hard disk in mine, so if I can find someone locally with a logic analyser, maybe I can actually confirm whether it's essentially MFM signalling or not.

Made a quick video comparing two CGA to SCART converters. One with Brown fix, and the other without.
https://www.youtube.com/watch?v=dASDx-oVFbQ

not that it is needed but here is a picture of a 1084S monitor with TTL input suitable for IBM CGA and APPLE computers, straight connections from cga->1084 :

The monitor does the brown-fix internally. The same happens with the later 1084S-P1 - that was also used with C128.

keropi wrote:

The monitor does the brown-fix internally. The same happens with the later 1084S-P1 - that was also used with C128.

Yes, and the same goes for the Philips CM8833, which is the 'original'. The Commodore 1084 is a rebadged Philips.

exactly, these Philips monitors got rebadged a bazillion times: from commodore to Acorn to who knows what other system... all of them are really good monitors so no complains here 😀

as a sidenote, after quite some searching I managed to get a NTSC 1084S locally (really really really lucky coincidence!) and now I have my "dream" cga monitor with color in composite mode 🤣 🤣 🤣

ciernioo wrote:

Hi Bjt! Thank you very much for your work on the CGA -> SCART adapter. I`ve built it according to your schematic, and it worked perfectly for my CGA + TV:
This is sooo much fun!

Bjt, thanks from me too! I also decided to build your version, because I wasn't sure whether the others would work on a CRT TV. I am glad to report that it works for me on both my LCD and CRT TV in all CGA modes, including 640x200.

I have designed the breadboard circuit layout in Blackboard (freeware), the attached zip contains two versions of my .bb file:

CGAAdapter_Variant3_Breadboard_80x80mm.bb: if any of you want to modify my layout, then better start with this version. This one is easier to understand, as each CGA or SCART pin number corresponds to the respective pin row number. I'm still not sure whether audio ground should be connected to all the other SCART ground (I'm no expert in electronics) - if so, then add a connection between SCART pin rows 4 and 5.

CGAAdapter_Variant3_Breadboard_80x80mm_finallyBuilt.bb: this is the version that I finally built. I rearranged some of the SCART pins to have them closer together, and I had to make some space for the 12V plug (for automatic 4:3 switching, however I later found out that I don't need that for my TVs). Pin number equals row number unless noted otherwise. I also connected audio GND to the other SCART ground since I had no more free wire left in the SCART cable.

Note that this was the first time ever that I have designed such a circuit layout, so there could be some potential for improvement. However, I can say that at least my layout seems to be error-free.
Although I used a stripboard to build the circuit and Blackboard also supports stripboard layouts, the Blackboard file shows a breadboard because IMHO this looks much clearer, it is much easier to see where you need to make breaks.

P.S. The only problem ist that on my small CRT TV set the picture is a little off-centered to the left, but this is a general problem of that TV, the picture of my MIST board looks the same on this TV. I will soon test it on another 72 cm CRT TV and on this one the picture will most likely be centered correctly (just like the MIST's picture on that TV).
Does maybe any of you know a DOS tool that allows centering the screen "by software"? Many old games from the early 1980s allowed horizontal screen adjustment in their startup menus, like for example Galaxian, Moon Patrol, Decathlon and Jungle Hunt. However, all these were (initially) self-booters that had their own "operating system" - maybe this allowed more control over the hardware.

Nice work.
Also: I like your EuroPC. I am about to make some more videos on the Schneider in a day or to.
https://www.youtube.com/channel/UCSUf7nCpwTgFbjGmcTYGZEQ

EuroPC does not support composit output, but I am currently thinking about an COM-Switch with a MM12 (Hercules) and a CGA Monitor.
I own a 1084S and it can be switched to green, so there might be 4 green tones possible? I don't know, I will have to test this.
I also have an old IBM "Multicolor" Monitor. It can display EGA and might display CGA as well. This monitor has the ability to switch
to all kinds of colors. (green, white, amber)...

Wow! I also have a very old XT/CGA Laptop...would be a fine challange to get a color picture out of it. Nice!

dr.zeissler wrote:

EuroPC does not support composit output.

I know. But I wonder, does anybody know if it would be possible to get at least a b/w VBS signal out of it ? With a little bit of hacking ?
Some consoles, like the french NES, also had RGB (via built-in transcoder) but used the the composite signal to provide "sync" to the SCART connector.
That's why I wonder-- Does this CGA chip provide a pure sync signal or a mixed one (sync+video) ?

Edit: Never mind. To err is human. I forgot that CGA uses dedicated H/V-sync.
So it is unlikely that this chip has a combined ouput also.

If there's anyone else out there with a Victor V86P that has a dead (or missing) PSU and wants to use an existing 8.5V PSU (like the one for the Playstation2 suggested earlier), I've reverse engineered the battery charging circuitry that's in the original PSU, and made a little board to implement it:

I'm still waiting on getting the board back, and I need to rebuild the battery pack (it's a stack of Sub-C NiCd cells in a cardboard tube with a couple of thermal cutouts) - so I don't actually know if the board works yet, but the idea is that you have a barrel jack input that'll take the output from a generic 8.5V brick, and it'll give you the necessary power and signalling for the V86P's charge circuitry.