PDA

View Full Version : CGA MCG2502 composite components



archeocomp
December 25th, 2013, 10:57 AM
Hello, I have a CGA RAM MCG2502 card, that has some components missing.


16492

Obviosly they would be there for composite out connected to pin 7 of the 9-DSUB connector. I can not use the card without composite out. So I searched for some pictures hopin that I could find photo wit fully populated card. Unfortunately I only found one russion place, where someone apparently tried to do the same thing. So my question is, what transistor would fit in and can it work without the rest of resitstors, and if not what value should I use ? Maybe someone here has a fully populated card ?
16493

http://vgamuseum.ru/gpu/ram/ram-mcg2502/

reenigne
December 25th, 2013, 07:41 PM
I haven't seen a card like this before, but the composite output circuitry does look a lot like the one from IBM's CGA. The transistor there is a 2N3904 - probably any NPN with reasonably similar characteristics should work.

The resistors are more of a puzzle since there are several variations on the resistor values in the IBM cards and the given values don't look much like any of them. I'd suggest starting by just populating the resistors whose values are given and seeing what happens. The other resistors might be for changing the brightness of the various direct colours (as is done on the newer IBM CGA cards). It might help to use a multimeter's connectivity tester to see how the pins of the 74LS244, the resistor pads and the pins of the RGBI connector are connected up. If I'm right the RGBI outputs will be connected to inputs of the 74LS244, the corresponding outputs will each be connected to one side of one of the resistors and the other side of the resistors will go to the base of the transistor.

The designers of this circuit might have set it up this way because they were aware of the different composite output circuits of the different IBM CGA cards and wanted this card to be able to act like any of them, which would be really neat!

archeocomp
December 26th, 2013, 09:28 AM
Thank you. I populated it with known resistors (4k7 instead of 5k1) and 2N3904 as well as BC548, but it does not work with none of them. Voltage on base of transistor is 0.19V. On the output (emitter and pin7 on DSUB connector) it is 0V. And there is no image as well. This seeems to be more complicated issue than i would have expected. Those resistors seem to be too big maybe?

Those unknown value resistors R5, R6, R7 are connected to GND, the other side goes to pins 3, 5, 7. R8 is connected to VCC the other side to pin 9 of 74LS244. They are connected to the double row header next to DSUB, but not to the DSUB connector pins. Maybe it was for light pen?

reenigne
December 26th, 2013, 05:09 PM
I don't think any analogue stuff should be needed for a light pen - at least on the IBM CGA card the light pen stuff is all done digitally (the analogue parts are done inside the light pen itself). Also on the IBM CGA the light pen connector is a 1x6 rather than a 2x5 (two data inputs from the light pen, ground, +5V, +12V and one not connected). So that makes me suspect that the 2x5 pins are not a light pen connector at all. My guess is that they are jumpers that are normally connected but can be removed to disconnect pins of the DSUB connector - it'd be interesting to see if you can verify that with a multimeter. It might just as simple as just a missing jumper!

Some more things you can try:
* I noticed in the picture from vgamuseum.ru that the second two pins of P1 are connected - have you done that with yours?
* If you have an oscilloscope, check what the signals on each of the 74LS244 pins look like. With any luck you'll be able to recognize composite sync, color burst and intensity. Those signals should come out of the 74LS244, through resistors and all end up together at the base of the transistor, which amplifies the signal and sends it out to composite.
* Otherwise, see if you can use the multimeter to reverse-engineer a schematic for that whole output section, and I'll compare it to the IBM CGA one and see if I can spot anything obviously missing.

archeocomp
December 27th, 2013, 09:09 AM
Thanks. I connected the two pins of P1, but as I said there is 0V already at the emitter of transistor. I looked with the oscilloscope only at the output, access to internals of the cards is really bad. I have put this little "project" on hold, because there are only THREE resistors going to the base of the transistor, and that is IMO not enough. I have some more CGA cards with composite out to use. Thank you again.