Stedy's RGB to SCART cable

[Mr B]

Originally Posted by Stedy

Is the cable made to my specification?

You tell me. It's this one. (Shock and horror, they've made it 3.30 Euro cheaper. I will have to demand a refund. Retro stuff is supposed to GAIN value over time, not lose...)

Originally Posted by Stedy

The most common problem, by far is the fact that the CSYNC signal is nearly 5X larger than the TV expects, hence the 330 resistor to reduce it to <1V. The Analog Front End in the TV works off 3.3V, expects 1V and you can feed it a 5V signal, not good.

This, i believe was implemented in the cables when i went shopping. The order was made in January, 2011. And delivered well in time to my B-day, if i remember right.

However, i think the inclusion of the capacitors is a later concept, and as such isn't a part of my cable. I haven't done anything "to it" on my own, so it's still as it was when i posted here a couple of years ago.

[ShK]

Originally Posted by Stedy

They've not failed yet. Did you try the cable without the optional AC coupling capacitors?

Yes, but only with the APTUS 1 RGB to component video convertor (It needs female SCART header).

Originally Posted by Stedy

This is going off-topic now, will write this all up soon in my retro video guide.

Please, do. This is interesting reading!

[ShK]

Does anyone know where I could buy a Tasker TSK1063 cable which is sold cut to length by the metre?

[bladecgn]

Yesterday I took an old Scart cable (2m) with single shielded RGB leads and built a cable according to Ian's 1.0 schematics (thanks a lot btw).

Didn't have a 330 Ohm resistor so for testing I directly connected the lead for now.

Picture quality is better than my old Amiga -> Scart cable. I have a recent Panasonic Plasma TV.

Some questions:

(1) Text has some white ringing. Will this be bettered by adding the resistor?

(2) If I add the three capacitors, will the black level become lower? Black is dark grey at the moment- Did anyone notice an improvement in (general) picture quality by using this option?

(3) Is it in any way better to use the signal coming from the Amiga 1200's composite port as sync? I so, I do NOT need the 330 Ohm resistor and the connection to pin 23 on Amiga side, correct?

(4) On this page (http://www.ntrautanen.fi/computers/h.../amiga_rgb.htm) it is stated at the bottom to not connect pin 13 (GNDRTN) to GND. Ian, what do you make out of this?

[pandy71]

Originally Posted by bladecgn

(4) On this page (http://www.ntrautanen.fi/computers/h.../amiga_rgb.htm) it is stated at the bottom to not connect pin 13 (GNDRTN) to GND. Ian, what do you make out of this?

Signal integrity http://en.wikipedia.org/wiki/Signal_integrity

Pin 13 is only for external clock source - this is different GND than video signal GND.

[ShK]

Originally Posted by pandy71

Pin 13 is only for external clock source - this is different GND than video signal GND.

This information is incorrect. D-connector chassis, pin 13 and pins 16-20 are grounded together on the solder side.



The same applies also at least the A500, A3000T, A4000 and A4000T videomodule.

[pandy71]

Originally Posted by ShK

This information is incorrect. D-connector chassis, pin 13 and pins 16-20 are grounded together on the solder side.

The same applies also at least the A500, A3000T, A4000 and A4000T videomodule.

Perhaps it was valid for A1000 (separate GND planes) - and as a rule is valid and correct - with clock source and clock distribution.
Never analyzed Amiga PCB's on signal integrity but PCB routing technology evolved from last 30 years seriously (and generally all related signal integrity techniques).

[bladecgn]

Originally Posted by ShK

This information is incorrect. D-connector chassis, pin 13 and pins 16-20 are grounded together on the solder side.

Yes, that's a definite proof, thanks very much. Does anyone know the answer to my questions 1-3 in my last post?

[ShK]

Originally Posted by bladecgn

(1) text has some white ringing. Will this be bettered by adding the resistor?

Did you go through the television menus and disabled all of these "enhancement" features, like the Edge Enhancement?

Do you use the 75 Ohm cable?

Amiga video port:
3 RED Analog Red (75 Ohm)
4 GREEN Analog Green (75 Ohm)
5 BLUE Analog Blue (75 Ohm)

What happens if I use 50 ohm cable for video application which needs 75 ohm cable?
If 50 ohm cable sees a 75 ohm load (the receiver), a substantial part of the signal will be reflected back to the transmitter. Since the transmitter is also 75 ohm, this relected signal will be substantially reflected back to the receiver. Because of the delay, it will show up as a nasty ghost in the picture. Multiple ghosts like this look like ringing. Also, the reflections cause partial signal cancellations at various frequencies.

Read more at: http://www.epanorama.net/documents/w...impedance.html

Originally Posted by bladecgn

(3) is it in any way better to use the signal coming from the amiga 1200's composite port as sync? I so, i do not need the 330 ohm resistor and the connection to pin 23 on amiga side, correct?

RGB sync is needed. I don't know is there a better way than take it from the composite signal.

RGB sync:
Amiga: pin10, 5v (47 Ohm)
Scart: pin 20, 1v (75 Ohm)

We need to do voltage level conversion from the Amiga 5v TTL level to RGB sync 1v scart, so the 330 Ohm resistor is required.

Maybe we should also match those two unequal impedances (47 Ohm to 75 Ohm). I find some kind example calculation formula for the impedance matching, but it does not apply in this case:

Impedance matching network usign resistors
The matching network shown below can be used to match two unequal impedances,
provided that Z1 is grater than Z2. 

          ____
     ----|____|---+---------
           R1     |        
                 | |        
   Z1            | | R2     Z2
                 |_|        
                  |         
     -------------+----------

The resistor for this circuit can be calxulated using the following equations: 

  R1 = Z1 - Z2*R2 / (Z2+R2)
  R2 = Z2 * sqrt(Z1) / (Z1-Z2)

The table below will show some precalculated values for some most common
interfacing situations: 

  Z1      Z2      R1      R2     Attenuation
 (ohm)   (ohm)   (ohm)   (ohm)   (dB)

   75      50     42,3    82,5    5,7
  150      50      121    61,9    9,9
  300      50      274    51,1   13,4
  150      75      110     110    7,6
  300      75      243    82.5   11,4

[Stedy]

Hello,

It's approaching that festive time of the year and I have been busy stringing many LEDs together Will now try and address some of the recent comments.

With regard to pin 13/DGND, I covered most if it here:
http://eab.abime.net/showthread.php?...ing#post784306

I doubt it will make much difference how you ground it, but I always separate Analogue ground (Video) from Digital (CSYNC). It's not directly related to Signal Integrity, more EMC and good practice, from my friends at Analog Devices, http://www.analog.com/static/importe...0Grounding.pdf
It may make a difference on the A500/A1000, but not on the other machines. The cable diagram is for all.

@shk
The 270-510 ohms, nominal 330 ohm, is just a potential divider. I want to ensure the signal seen by the TV is <1V. The TV has 75 ohms to ground, The Amiga 47 ohms and the cable 330 ohms, giving 377 ohms from the Amiga, as seen by the TV. Vout = 75/(75+377) x Vin.
in = 4.4V minimum, so vout is 0.73V minimum and 0.82V maximum.

Did adding the AC coupling capacitors improve the picture in any discernible way?

@bladecgn

Try the optional composite video to pin 20 of the SCART socket. This is the actual signal all TVs expect to see, not a sync signal.

Some TVs, 'alpha blend' the composite video with the RGB, to enhance the edges. The user has no control over this. If the composite signal is just that, it will work. The Amiga CSYNC TTL signal, when attenuated, will look like bright white video and could create sparkles/ghosting if blended. This needs some photographs to explain.

It is for that reason, the composite video plug has been shown on the diagram since 2009.

What are your contrast/brightness settings?
Brightness controls black level, contrast controls white level. For a TV, I typically use 50% brightness and 80% contrast, with a Broadcast standard monitor, I would use 0% brightness and 50% contrast.

@MrB
Sorry I missed your earlier post. The only LG TV I have is a 2010 19" widescreen, which is shown in a few photos in the thread.

Are you confident adapting the SCART cable to feed composite video to the SCART cable from the Amiga?

@thread

Part of the next stage of my website development is to write a retro video guide, which will detail some of the lessons learnt adapting various non-standard video signals to normal TV/monitor displays. This is intended to go alongside the often delayed, Retro Video Adapter I have in prototype development. The 240p/288p video issues have taken a bit of planning to work around, efforts are focused on this development stage before updating the website.

If I go quiet for a week or two, it's probably because I'll be cursing at Xilinx ISE/Isim or eating festive food and drink

[Mr B]

Originally Posted by Stedy

@MrB
Sorry I missed your earlier post. The only LG TV I have is a 2010 19" widescreen, which is shown in a few photos in the thread.

Are you confident adapting the SCART cable to feed composite video to the SCART cable from the Amiga?

Sure. I'm worthless at actually getting things done, and i'm more likely to buy a second cable, to modify / upgrade it, then to go in to the one i got with the soldering iron. Since this one works with my mate's TV, i mean. Should be able to pick up a short cable from a local store, and cut it, add a cut down connector, and hook it up.

Hell, i've been meaning to add the capacitors, since they aren't in my cable from AmigaKit, but i haven't even got around to that. Hoping that would fix the ghosting my mates TV has. And just now i cant seam to find the schematic. *shrugs*

[bladecgn]

Originally Posted by ShK

Did you go through the television menus and disabled all of these "enhancement" features, like the Edge Enhancement?

Do you use the 75 Ohm cable?

Yes, I disabled all ehancement features. I think the cable should have 75 Ohms, as it is a single-shielded SCART cable I used.

Originally Posted by Stedy

@bladecgn

Try the optional composite video to pin 20 of the SCART socket. This is the actual signal all TVs expect to see, not a sync signal.

Some TVs, 'alpha blend' the composite video with the RGB, to enhance the edges. The user has no control over this. If the composite signal is just that, it will work. The Amiga CSYNC TTL signal, when attenuated, will look like bright white video and could create sparkles/ghosting if blended. This needs some photographs to explain.

It is for that reason, the composite video plug has been shown on the diagram since 2009.

I'll try that. I won't need the 330 Ohm resistor then as the composite video signal already has the correct voltage level, right?

What are your contrast/brightness settings?
Brightness controls black level, contrast controls white level. For a TV, I typically use 50% brightness and 80% contrast,

That's exactly what I use.

Ian, what possible effect could the three capacitors have on the picture? Will it get darker or sharper in theory or what is the desired effect?

Thanks everyone for their help, very much appreciated!

[bladecgn]

Originally Posted by Stedy

@bladecgn
Try the optional composite video to pin 20 of the SCART socket. This is the actual signal all TVs expect to see, not a sync signal.

Some TVs, 'alpha blend' the composite video with the RGB, to enhance the edges. The user has no control over this. If the composite signal is just that, it will work. The Amiga CSYNC TTL signal, when attenuated, will look like bright white video and could create sparkles/ghosting if blended. This needs some photographs to explain.

It is for that reason, the composite video plug has been shown on the diagram since 2009.

I tried the composite video port signal (connected by the optional plug) as sync but picture is very (=too) dark then. I didn't expect hat, why could this be??

[ShK]

Originally Posted by Stedy

The 270-510 ohms, nominal 330 ohm, is just a potential divider. I want to ensure the signal seen by the TV is <1V. The TV has 75 ohms to ground, The Amiga 47 ohms and the cable 330 ohms, giving 377 ohms from the Amiga, as seen by the TV. Vout = 75/(75+377) x Vin.
in = 4.4V minimum, so vout is 0.73V minimum and 0.82V maximum.

Thank you for clarification!

Does it affect the calculation method, if we use the 'switch transmission' cable or 75 Ohm coaxial cable for RGB sync signal, because of their different impedances?

Originally Posted by Stedy

Did adding the AC coupling capacitors improve the picture in any discernible way?

I should buy more cable, cos it's difficult to compare now, without two similar high quality cable.

I'm still looking for retailers who sells Tasker TSK1063 cable which is sold cut to length by the metre.

[ElectroBlaster]

I am buying in parts now to build a cable and is there any preference over carbon or metal film resistors?

[demolition]

Originally Posted by ElectroBlaster

I am buying in parts now to build a cable and is there any preference over carbon or metal film resistors?

The only place it could matter is the one on the composite pin, but here I think the bandwidth is still too small for it to matter. I'd just use regular carbon resistors.

[Stedy]

I did a couple of tests recently to ascertain what happens in various fault scenarios and how different sync sources work. The test results are below.

If the SCART socket is missing the 1-3V on pin 16 and selects composite video, with the Amiga TTL Csync signal, you will not get a display.
This is most likely due to the analogue front end (video decoder) looking for, but failing to find a black level in the back porch of the video. On my test systems, I had a completely black screen, no ghosting or blurred images.

The synchronisation signal you use can shift the RGB image slightly.


Using TTL Cysnc from pin 10 of the Amiga
The masking tape is the reference point for all horizontal shift measurements.


With a CVBS (composite video) signal, the image shifted to the left by 7-8 pixels.


Feeding composite video only into the SCART port shifted the video to the right. Notice the poorer definition.

The next test, which will take some setting up, will be to replicate the Alpha blending ability of some TVs, this combines the composite video signal, with the RGB, to increase edge sharpness. I believe this will cause issues with the Amiga's TTL sync. Will test it an report the results.
I am trying to explain some of the shadowing/ghosted images people see.

That's all for now.

Ian

[bladecgn]

Originally Posted by Stedy

Try the optional composite video to pin 20 of the SCART socket. This is the actual signal all TVs expect to see, not a sync signal.

With or without the 330 Ohm resistor in between?

[bladecgn]

Originally Posted by Stedy

The next test, which will take some setting up, will be to replicate the Alpha blending ability of some TVs, this combines the composite video signal, with the RGB, to increase edge sharpness. I believe this will cause issues with the Amiga's TTL sync. Will test it an report the results.
I am trying to explain some of the shadowing/ghosted images people see.

Wow Ian, you're incredible!!

Awesome thread.

However, after reading Ian's explanations about the potential issue caused by Csync (and Hsync/Vsync ?) signal amplitude one question still stays on my mind preventing me from sleeping.

Do you remember the official cables commodore provided for connecting the Amiga 500 to a 1084S monitor, for example?
Something like this:




This cable seems to use individually shielded wires for R,G, B and Csync but does not use any capacitors or resistors inside the plugs.

Could this mean the 1084s monitors were built to tolerate signal amplitude beyond recomended values or simply that someone at commodore didn't care for monitor specification/ longevity?

I remember seeing these cables being used in shops and I never heard stories or complaints about them.

Anyway, adding caps and the 330ohm resistor to a custom made cable should not do any harm, am I right?