Although Sorgelig's idea using an RGB->YPbPr cable was brilliant, and he made the necessary FPGA video module as well, only a few developers (actually only Gehstock) utilized his work, and limited number of cores contains it (Menu, Commodore 64, ZX Spectrum, BK0011M, SAM Coupe, Apogee, Specialist, Vector 06C, NES and Amiga/minimig).
I built recently a more or less universal RGB->component converter, as a cheap DIY solution. I would like to share it's wiring diagram, perhaps others will find it useful also (see attachment at the end of the post).
I took idea of the converter from the SEGA-16 forum below:http://www.sega-16.com/forum/showthread ... e-BA7230LS
The picture quality is not perfect (it's a bit brighter and the picture is 'washed' compared to native RGB out, especially in cases of progressive signal), but usable. Did not test all cores so far, but as long as video output meets the PAL or NTSC TV-standard's vertical resolutions and refresh rates (240p/480i/480p@60Hz or 576i/576p@50Hz), you have a great chance that the converter will work with the given core.
Non-TV standard resolutions (800x600 - i.e. GameBoy, TRS-80, some Arcade cores; 1024x768 - Apple MacIntosh) or refresh rates (56Hz - Atari ST PAL, Videoton TVC VGA; 70Hz or more - Atari ST B&W, Amstrad's 73Hz VGA mode) will likely not work.
I tested it with my Samsung LE40B650 LCD TV, which is very sensitive and picky in aspect of non-standard video signals. In most cases it works flawless, in some cases the picture is too bright (i.e. Atari 2600 core), but connected to the Gonbes chinese scaler clone it's fine.
Tested only a few arcade and console with this scaler, they work fine, althought tried only 15kHz NTSC so far:
In other cases it works with some output mode of the core, and doesn't work with another resolutions or TV standard. I.e. the Atari ST core is OK in NTSC 240p, but 'unsupported mode' if set to 480p. The C64 core is fine using PAL standard in either 576i or 576p, but 'no signal' in case of NTSC setting.
* In some of my early experiments (and screenshots) appeared pale green horizontal bars on the screen, especially noticable over dark background or borders. Meanwhile I solved the problem: replaced the Huawei phone charger which feeded the circuit with another, designed for iPhone-4 (althought both have 5V 1000mA specification, the second one provides more stable voltage), and the green bands totally disappeared!
* This circuit may be a good solution for people living in NTSC countries (like U.S.) to connect their TC64 to retro displays. Here in Europe it's much better to use the minimig VGA to SCART cable - but not all cores support it, and only in interlaced (scandoubler_disabled=1) mode.
* The only exotic component of the circuit is the BA7230LS IC, I ordered it on EBay from Singapore. Althought it's designed for NTSC TV standard, works well with PAL resolutions/refresh rates as well. The color balance is acceptable also, despite the theoretically different color space. Attached some screenshots for reference, compared to the SCART and VGA image.
* In case if progressive signals (VGA out) you need to merge the separated horizontal and vertical sync using a BC547 or 548 transistor and two resistors (680ohm & 1Kohm as drawn with blue color on the schematic), as the IC accepts only composite sync ('sync on green'), and it does not have separated H/V sync input. The picture quality will be a bit weaker in this case. I built a short VGA extension cord for this purpose, which contains the necessary components in one of the connector's socket, and can be added if necessary or removed if not. If you have interlaced signal and composite sync on VGA pin 13 (this is the recommended mode if the core supports it as this results better image quality), this is unnecessary (see the dark red part of the schematics: all scandoubler_disabled=1 compatible cores have this option).
* You may connect the composite (or merged horizontal + vertical) sync signal to the IC's green input, or after the last 75ohm [R2] resistor. This does not make big difference in picture quality, and it's marked with green on the schematics.
* In some cases you need to left out the 270ohm [R1] resistor: especially if you connect a progressive signal - merged H & W sync. I used DIP switches for this purpose, but you may install a 250ohm potentiometer (they have mostly a bit bigger maximal resistance, somewhere between 260-270ohms) instead of fix resistor. In this case sometimes the value of [R2] needs to be reduced to 39ohm as well. Again I recommed DIP switches or a 100ohm potentiometer.
* I powered the circuit using a good quality (iPhone 4), standard 1000mA 5V USB phone charger.
* I could not acquire C945 transistors as the original scematic suggested, used BC182B instead.
* The IC's datasheet recommends to ground it's unused input pins (#3, #4, #5 & #20) through a 1uF capacitor, this was missing from the original circuit.
* The outer contacts of the component RCA connectors need to be grounded as well of course.
Feedback is welcome, if you have quastion about a specific core's compatibility, I'll try to answer it ASAP. I intend to test all the cores with and without the Gonbes scaler as well, but this may took a while, perhaps I'll complete it only after Christmas.
You do not have the required permissions to view the files attached to this post.