RGB switching

[protek]

I got an RGB to Scart cable from Retro Computer Shack recently. It seems that this cable isn't providing the blanking signal to pin 16 in Scart as I only get black and white picture. The TV I'm using doesn't have an option to force the RGB mode on so it has to be done by the source signal.

Having looked at the CPC monitor connector pinout, it seems that pin 4 provides sync and is used for RGB switching and I saw some schematics that have the pin 4 from CPC and pin 16 from Scart connected. I haven't tried with multimeter but I suspect the sync is connected to the composite sync pin in Scart in the cable that I have.

What  voltage is the pin 4 supplying? Could I just switch the wire from one pin to another in the Scart cable or is it possible to supply both pins by putting a jump wire?

Thanks!

[Badstarr]

Is your TV set definitely RGB capable? I know some TV's and VCRs with a SCART connection basically take the RGB and use it to produce a composite signal, for example like shorting the RGB signals together. Have you tried a different TV and only got a black and white picture? The SCART cable I use produced a black and white picture on one old TV I have, I tried connecting a DVB box via the SCART and switched its output to RGB and the picture went black and white indicating that it cant deal with a RGB signal.

[Bryce]

Hi Protek,
      to get around this problem I created an alternative solution which you can find here: TV SCART cable - CPCWiki - The Ultimate Amstrad CPC Community & Encyclopedia!
As you can see, I used pin 4 to bias pin 16 of the SCART and this usually works fine, however, some modern TVs permanently probe this signal at a very fast rate, causing either colour flickering or in some cases the OSD (on screen channel display) to keep appearing, so the capacitor was added to hold the voltage permanently at the correct threshold. Give this a go, it might solve your issue.

Bryce.

Edit: Just to answer the technical part of your question: Pin 4 should be oscillating between 0V and 2V peak, but the average voltage (if measured with a multimetre instead of an oscilloscope) will probably read somewhere around 1 - 1.2V. By adding the capacitor, this signal charges the capacitor and makes sure that the voltage is always around 1.8V which is enough to switch the TV to RGB input mode constantly.

[protek]

Thanks, Bryce and Badstarr.
Yes, the TV is RGB capable. It works with my Amiga, where I got the blanking directed to pin 16 via a resistor to get the provided 5 Volts within the treshold. It also works with my +3, in which I have a 3 Volt coin battery attached to Scart plug. Unfortunately, it draws the battery empty within a couple of hours and the picture disappears.
I did have a look at your alternative solution. Looks simple enough to do. I just have to check the cable I have to see how it's connected. Do you know hiw many Volts pin 4 is pushing out? I can check that with my multimeter too, though.

[Bryce]

I added the technical info to the bottom of my last post.

Bryce.

[protek]

Thanks! I'll check the continuities in my cable and report back.

[ralferoo]

As you can see, I used pin 4 to bias pin 16 of the SCART and this usually works fine, however, some modern TVs permanently probe this signal at a very fast rate, causing either colour flickering or in some cases the OSD (on screen channel display) to keep appearing, so the capacitor was added to hold the voltage permanently at the correct threshold.

Technically, they're doing the correct thing as this is the "fast blanking pin" and is designed to allow RGB to over-ride the composite for specific portions of the picture only. That said, it's a real PITA if you don't have a steady voltage handy to apply to this pin.

This is definitely something I want to do some further investigation into though. I've got a feeling for some sets this needs to be low during the sync pulses - certainly I'm having interesting results with one of my TVs with the current output from my current SCART breakout board for my FPGA stuff where it refuses to display anything, even when there's an RGB voltage applied permanently to this pin, yet this same TV was working fine with an earlier SCART breakout board where I was toggling the fast blanking.

[protek]

I tested the continuity in my Scart cable and pin 4 is indeed connected with pin 16. It seems that I'll need to put that capacitor in there.

[protek]

I was wondering, if a similar solution would work in Spectrum +3? In +3 the sync is connected the composite pin of Scart as there's no luminance but I was thinking that if I put a jump wire with diode from composite pin to blanking pin and then add the 100 uF capacitor between the blanking and blanking ground, would I get a steady voltage within RGB threshold?

[protek]

I put the 100 uF capacitor between pins 16 and 18 in the Scart. I keep still getting a black and white picture. When I measure the voltage with a multimeter I get practically no voltage.

[Bryce]

A Multimeter (depending on the type and whether it's an expensive one or not) might not show you a proper value because the signal is not stable. If your TV is still showing B&W then you will probably have to add a battery to hold the voltage at 3V and if this doesn't work, then your TV needs to be manually switched to RGB.

Bryce.

[ralferoo]

If your TV is still showing B&W then you will probably have to add a battery to hold the voltage at 3V and if this doesn't work, then your TV needs to be manually switched to RGB.

Or you've plugged it into a socket that only does composite or s/video... You've probably got at least 2 SCART sockets on the back of your TV - try them all before changing the cable again!

[protek]

I am using an LG Flatron 173ST Monitor TV. It has a single Scart socket. It supports RGB but it isn't manually selectable so it depends upon the blanking signal. I've got my Amigas, my old ST and my SNES to work in RGB. I even got my +3 to work with a coin battery soldered into the Scart plug. Problem is that the LG drains the battery in a couple of hours after which the screen goes dark.

I actually tried the same method to the +3 cable as was in the Amstrad Cable. In +3 the composite sync goes to pin 20 as there's no separate luminance, so I put a diode between pin 20 and pin 16 to enable one way sync signal for blanking. I also soldered a 100 uF capacitor between pins 16 and 18 to even out the signal. My multimeter read slightly over +2V on pin 16 but for some reason it still didn't switch to RGB. I've got a UK telephone cable coming to me so I may just rip a proper 12V from the serial port or lightgun port and provide the blanking with that. Or maybe I'll just put my Commodore 1081 together so I don't have to worry about blanking.

[protek]

I ended up taking the easy way around and bought a 14" Salora CRT Color TV for 5€. It has Scart and the input mode can be selected to RGB so the Amstrad and my +3 display a picture without any hassle.  The image quality is really nice. It has also RCA jacks on the front so I'm able to connect my Commodore machines too.

[MacDeath]

the blue and yellow is nice, the grey attributes is less...

Anyway the Speccy+3 is clearly the best of the bunch of speccies... just like good old CPC6128.

[Bryce]

The speccy looks like it has a slight Moiré issue propogating from the left side, but that may be just from the camera used to take the picture. Is this visible in real or just in the picture?

Bryce.

[MacDeath]

would be nice to compare the size of the screen (border) and the pixels generated by both machines (CPC and Speccy+3) on the same monitor...

CPC is supposed to generated smaller pixels in Mode1... but I would like to know about the differences with the border size (in centimeters/milimeters perhaps)


concerning the moiré...
first the speccy picture has more angle.
also isn't grey supposed to produce more moiré than "monocolour"  RGB colours (R or G or B only...)
hence the Blue produce less moiré than Grey (dark white actually).

[arnoldemu]

would be nice to compare the size of the screen (border) and the pixels generated by both machines (CPC and Speccy+3) on the same monitor...

CPC is supposed to generated smaller pixels in Mode1... but I would like to know about the differences with the border size (in centimeters/milimeters perhaps)


concerning the moiré...
first the speccy picture has more angle.
also isn't grey supposed to produce more moiré than "monocolour"  RGB colours (R or G or B only...)
hence the Blue produce less moiré than Grey (dark white actually).

to do that, you could try simulating a spectrum sized screen on the cpc, then measure them on the same television. should give a closer approximation. I would measure the visible area not the border. Visible area has known size.

[ralferoo]

Spectrums have a wider pixel than the Amstrads - they use a slower pixel clock (IIRC it's 7MHz rather than 8MHz for CPC mode 1). So, the spectrum screen should be about 36.6 CPC characters wide. This is why the "spectrum port" looks so much worse on the CPC - as well as having less colours, the border is comparatively wider than the Spectrum's even though there are the same number of pixels.

[MacDeath]

to do that, you could try simulating a spectrum sized screen on the cpc, then measure them on the same television. should give a closer approximation. I would measure the visible area not the border. Visible area has known size.

that's it...
I though about having the camera on a tripod so it is stable and take the exact same picture... so try with both machines...

change the border's colour for each computer so we can get a better approximation from the picture, would be cool to see with some sort of equivalent colours...

This is why the "spectrum port" looks so much worse on the CPC

sort of... having smaller pixels is not actually worse... ditherings are better looking I guess.
the "worse" part is that the "playfield" is then ridiculously smaller off course, hence the CPC bad rep as having less pixels than the speccy (less surface with same resolution actually)...

[Bryce]

You could always just connect an oscilloscope to the output signal and then measure and compare the exact times of the colour signals.

Bryce.

[protek]

The speccy looks like it has a slight Moiré issue propogating from the left side, but that may be just from the camera used to take the picture. Is this visible in real or just in the picture?

Bryce.

It's because of the camera. No moire with plain eye.

[MacDeath]

You could always just connect an oscilloscope to the output signal and then measure and compare the exact times of the colour signals.

why use a crude and boring meter while you could use a particle accelerator connected to a nuclear plant and a neutrinos charge.

[TFM]

Indeed!

[Bryce]

why use a crude and boring meter while you could use a particle accelerator connected to a nuclear plant and a neutrinos charge.

Unfortunately I came home after a few too many beers and accidently knocked over my bucket[nb]Before you ask, the bucket uses super-conductor magnets at -273.15°C to hold the neutrinos in place (until you knock it over).[/nb] of neutrinos sending them hurtling through the earth in pico-seconds (Though they probably arrived in Italy slightly earlier the other places )

Bryce.