GBS-8200

A place to discuss everything Dragon related that doesn't fall into the other categories.
Post Reply
pser1
Posts: 1655
Joined: Sun Mar 25, 2012 7:32 pm
Location: Barcelona (SPAIN)

GBS-8200

Post by pser1 »

Hello,
I bought one of these YUV2VGA converters.
It has inputs for signals as RGB and YPbPr too.
Looking at the MC6847 Datasheet, there are three pins with that later kind of signal:
Pin 10 has Fib = R-Y or Pb
Pin 11 has Fia = R-Y or Pr
Pin 28 has Y (luma)
I was thinking of connecting the signal from these pins to the input of the GBS-8200 but maybe it will need a current amplifier
For instance an emitter follower with one Transistor (which one will suit it well?) or with a Darlington pair ... I don't know
Should it be tied with or without condensators (ac or dc link)?
It would be nice to improve the stability of the Dragon output video.

Any advice will be very appreciate!
thanks a lot

cheers
pere
sorchard
Posts: 530
Joined: Sat Jun 07, 2014 9:43 pm
Location: Norwich UK

Re: GBS-8200

Post by sorchard »

That's a handy looking board you've found there.

Emitter followers would be of benefit and would help protect the 6847 from the effects of driving cables. General purpose NPN transistors should be fine for experimentation (like 2N3904, 2N4401)

The problem might be that the Y signal from the 6847 is inverted and you may also need to pick up the sync signals from other places on the Dragon main board. It's worth getting this working as the 6847 output is really clean.

Try it and see what happens...
Stew
pser1
Posts: 1655
Joined: Sun Mar 25, 2012 7:32 pm
Location: Barcelona (SPAIN)

Re: GBS-8200

Post by pser1 »

Hi Stew,
thanks for your advice
Just a question, how do you tell that signal Y is inverted in the 6847?
It would mean to add a common emitter transistor to invert it (0<->5V), I hope.
According to the 6847 datasheet, Y contains: Voltage sync at 1V and all other levels below that value ... don't know what should be the right 'non inverted' values.
I am attaching the very poor GBS-8200 documentation for if it helps.
You will see that there are only three inputs for the YUV signal, so if there is need for the syncs to be included, they should be added to the Y signal in the
transistors buffer board. But which one(s) ... FS, HS or the inverted ones /FS, /HS?

cheers
pere
Attachments
GBS-8200 Manual.zip
(190.26 KiB) Downloaded 166 times
sorchard
Posts: 530
Joined: Sat Jun 07, 2014 9:43 pm
Location: Norwich UK

Re: GBS-8200

Post by sorchard »

Nice manual. "Don't Demolition" :-)

We can tell Y is inverted because the syncs should be the lowest voltage on the signal, below the black level. The 6847 outputs a high voltage instead. I remember looking at this signal on a monitor and it did look like inverse video.

You are right, you would need a common emitter stage to invert the signal and then an emitter follower to buffer and level shift. If you have a look at the Siemens schematic page 3 you will see two transistors doing exactly that.

You might as well make life easy and use the video signal at the main board connector (or even the monitor socket) to drive the Y input. This will already have the correct sync signals. You will also need to disable the Dragon's colour using your favourite method. (e.g. remove IC9 or stick a cap across the colour crystal)

That should get you a mono display, with just the UV inputs to figure out.
Stew
pser1
Posts: 1655
Joined: Sun Mar 25, 2012 7:32 pm
Location: Barcelona (SPAIN)

Re: GBS-8200

Post by pser1 »

Hi Stew,

I have been looking at the Dragon 64 schematics and are similar to the D32 ones ...
The IC is number 17 and there is a link from Chroma signal to the base of the common emitter transistor at Resistor num 44 (num 65 in Siemens D32)
This wire is labelled as "Disconnect for B/W" so we are looking at the same point.
The ICs are not socketed, so no way to remove it.
I could try to cut the resistor terminal but for sure the best will be you idea of adding a capacitor in parallel to the crystal XL2.
Will 100nF be enough to prevent it working?

Then for the emitter followers (Pb and Pr), I would use a 1K2 resistor to the Base and a 330 Ohm for the emitter
I hope this values will not reduce the output level too much.
The chroma signals will be taken right from the board but corresponding to the 6847 pins 10 and 11 (Fib and Fia)
Thanks a lot for your advice

cheers
pere
sorchard
Posts: 530
Joined: Sat Jun 07, 2014 9:43 pm
Location: Norwich UK

Re: GBS-8200

Post by sorchard »

Hi Pere,

100n across XL2 should stop it working and disable the colour.

The resistor values you've chosen are a good starting point. The 330R emitter resistor in the Dragon connects to -5V so that the composite signal doesn't get clipped, but a 0V connection may be sufficient for the chroma signals. It's worth trying both ways.

If luck is on your side, and the right colours appear, then it will be worth looking for faster transistors (ft >= 500MHz) to see if you can get a better image. The 2N2369A used in the Dragon is still available, but surface mount transistors should be lower cost.

I would be interested to know how well this works.
Stew
pser1
Posts: 1655
Joined: Sun Mar 25, 2012 7:32 pm
Location: Barcelona (SPAIN)

Re: GBS-8200

Post by pser1 »

Hello Stew,
thanks one more time.
On the Dragon 64 the emitters are tied to -5V too. I did not pay much attention at that point :(
I think that i saw somewhere a picture of the chroma signal and its level seemed to me to be always above 0V, so now I think
that maybe the -5V wouldn't be mandatory for the two signals Pb, Pr, but it is a good idea to try both possibilities ...
I will search among the materials I have in my electronic components drawer and will choose the most similar to the 2N2369A.
I will not attempt to use SMD by now ... I have never used them
I will let you know how this evolves ...

cheers
pere
sorchard
Posts: 530
Joined: Sat Jun 07, 2014 9:43 pm
Location: Norwich UK

Re: GBS-8200

Post by sorchard »

The thing to remember is an emitter follower outputs a voltage approx 0.7V lower than the input. This kind of helps here because it brings the 6847 output of 0.8 - 2V somewhere closer to typical video signal levels. The -5V connection will have the benefit of reducing the variation in emitter current, and so give a more constant gain, but the downside will be a larger current on average and therefore more base current and bigger voltage drop across the base resistor.

As always a bit of experimentation will be necessary :-)
Stew
pser1
Posts: 1655
Joined: Sun Mar 25, 2012 7:32 pm
Location: Barcelona (SPAIN)

Re: GBS-8200

Post by pser1 »

Hi Stew,

great explanation, thanks for those reminders.
I had worked quite a lot in the electronics domain but that was in the 80's, then I went into software development ... so almost everything has been forgotten!
This weekend I am going to have very little free time, that's good too ;-)
Hopefully next week I could be performing the first tests with the GBS-8200 board
I will share the results here, it would be really nice to improve the quality of the colored image ... so that we could work again mixing images of different color palettes
It would be great to get 16 colors in PMODE3 with still images, I mean without flickering.

cheers
pere
Vato
Posts: 114
Joined: Fri Apr 03, 2009 3:00 pm
Location: Mariestad, Sweden & Tromsø, Norway

Re: GBS-8200

Post by Vato »

Looking forward to see how this works!
Tom
Post Reply