S-Video from the VIC

[eslapion]

For reasons that are beyond my understanding, mistermsk recommends a 330 ohms resistor on the chroma when connecting to a non-commodore monitor.

It seems the chroma signal from a VIC or 64 is actually stronger than a standard S-Video signal coming from a DVD or digital TV receiver, for example.

Will have to investigate with an oscilloscope.

[e5frog]

In series I guess?

How will this affect the picture (if you don't have it) picture seems just fine with my TV.

[e5frog]

I know I could buy a cable to solve this, but I'm a bit short on cash @ the moment & seeing that I do have some soldering experience (I used to build guitar fx pedals & I built myself a tube amp) & a LOT of extra RCA cables laying around, I'd be willing to make my own cable.

It's fairly easy:



Pin 4 is used for video on C64 so if you use that it will also work for one of those. I don't think I've ever tried the pin 5 for video.

So pin 3 audio, pin 4 video and pin 2 is gnd for both signals.

[a4000bear]

I have recently obtained two more VICs. One is a CR model, the other is the earlier version with the two pin plug. This gives me the opportunity to also modify and document S-Video for the early model VIC. I have already done the 'before' screenshots. I had forgotten how terrible the video quality out of an unmodified VIC really is!

There may also be a minor addition to the existing mod, which may provide a further slight improvement in quality.

[Pedro Lambrini]

Bearing in mind that, in general, the output of the two prong is better quality than the CR Vic, does this hold true for the S-Video mod? To be clear, is the S-Video mod better on the two prong vs the CR?

[rga24]

I have recently obtained two more VICs. One is a CR model, the other is the earlier version with the two pin plug. This gives me the opportunity to also modify and document S-Video for the early model VIC. I have already done the 'before' screenshots. I had forgotten how terrible the video quality out of an unmodified VIC really is!

There may also be a minor addition to the existing mod, which may provide a further slight improvement in quality.

Hi,

Some thoughts on the S-Video mod.

In the existing S-Video mod for VIC-20 CR computers, the luminance signal is buffered by Q2 but the chrominance signal has no buffering. That's why the chrominance output is lower after the mod has been done. Some have suggested adjusting the setting on R10 so that the display looks right again. This works by reducing the luminance output to the buffer transistor.

I would suggest duplicating the transistor circuit for the chrominance signal instead. C13 can be disconnected from C17 and the wiper of R10, and instead connected to a copy of the full transistor buffer circuit, i.e. C17, R11, R12, Q2 and R13. It may also be useful to duplicate the loading on the chrominance output provided by R9 and R10.

The circuit diagram on Bo Zimmerman's archive disagrees with the diagram in the earlier thread on Denial as to where in the circuit the chrominance output is applied. Commodore's schematic has C13 connected to the junction of R9 and R10, the circuit in the Denial thread has it connected to the wiper of R10.

http://www.zimmers.net/anonftp/pub/cbm/ ... 1027r1.gif

The biggest difference between the Revision E video circuit and the Revision N video circuit is the use of transistor Q6 to provide the current for the luminance output of the VIC chip. Pin 3, the luminance output, is open drain and has to have a source of current in order to function correctly.

http://www.zimmers.net/anonftp/pub/cbm/ ... 4001_1.gif

The Revision N circuit in the VIC-20 CR uses a 510 ohm resistor R9 for the pullup function, whereas the original VIC-20 uses R29, C23, Q6 and R19.

This may produce differences in the linearity of luminance steps between the two types of VIC-20.

This circuit does not need duplicating for the chrominance output, since pin 2 of the VIC chip is open source and can simply be connected to the base of a biased transistor buffer via a capacitor.

[e5frog]

... or you can adjust the color on your TV.

[rga24]

... or you can adjust the color on your TV.

This mod will allow the S-Video equipped VIC-20 to work at the same chrominance level as unmodified VIC-20s, so you don't have to change the colour control when you switch input signals.

The circuit in a4000bear's diagram is the correct one; it seems Commodore changed their mind about where to apply the chrominance signal but didn't update their schematic diagram. The effect of the change is to increase the amplitude of the chrominance signal in the combined composite video output.

[a4000bear]

Bearing in mind that, in general, the output of the two prong is better quality than the CR Vic, does this hold true for the S-Video mod? To be clear, is the S-Video mod better on the two prong vs the CR?

There will be no difference between the two after the mod, except the E-version (two pin) VIC will retain the ability to adjust the sync level in the luminance signal, using the existing trimpot.

I am still working on the mod (after an interruption caused by the new year)

It should be ready soon.

There will also be some minor improvements that will further reduce banding in the picture. They are not part of the S-video mod, and can be applied separately if desired.

[a4000bear]

In regards to the concerns about the colour circuit in the S-video mod, I have considered the following:

The primary concern for me is to get the highest possible picture quality with the minimum modification to the VIC. Especially in regards to cutting too many tracks and adding components that do not already have a place for them to fit on the PCB. In other words, I want the VIC to at least look as though it has not been hacked unless a close examination was carried out. I also wanted the mod to be as minimalist as possible, bearing in mind the wide variation out there of people with electronics skills.

I certainly did think long and hard about adding a buffer stage to the colour output (and still thinking about it). I agree the colour output is a bit low, but not out of the bounds one normally experiences with the variations found in all the typical video sources out there.

Its worth noting whenever I have performed the modification, there was only a mild reduction in apparent colour saturation as it appears on the screen after the mod. The monitor I was using still has plenty of room to turn up the colour more. Are there people out there that still have low colour with the colour turned up to maximum?

I have seen the additional pics posted on the wiki page from someone else who has done the mod. I agree the colour is a lot lower looking compared to what he had before he did the mod. However his picture also shows a severe case of 'Venetian Blinds' (thin alternating horizontal lines in the coloured areas pf the picture). This is an indication there are serious phase/frequency errors somewhere in his system. It could be the VIC, or it could be the monitor. The poster may have played with C48 (the trimmer capacitor on the CR board) and caused gross errors in the PAL encoder in the 6561 VIC chip. Note that this only applies to PAL. Phase errors in NTSC cause major changes to the colours rather than a drop in colour level, ie: it behaves just like the 'tint' control on an NTSC TV. Note if the trimmer is adjusted too far off, the colour will disappear totally.

I would strongly advise that C48 be not touched, as it needs a frequency counter to be set correctly.

Hope this clarifies my thinking on the modification.

Are there any people out there that have done the mod and want to share their experiences?

[e5frog]

I didn't cut a single trace when I made the mod on my two prong PAL VIC, just solder loose the components in that end instead, no need for any violence. I also put a switch in so I could swap between the mod and the original configuration - I thought that was a neat solution.

I like to be able to restore my machines to it's original state again (apart from the same solder tin) so I usually try to use the least invasive method.

I thought the original method was fairly easy but there's improvements that can be done, someone mentioned some kind of adapter for the VIC for even better results which would be very handy for people with less electronics interest, just plug in and it's done.


Does anyone have equipment to check the resulting output signals, are they up to S-video standard or does something really need to be done?

[rga24]

I did the S-Video mod last night with the buffer transistor. I made two track cuts in total, and one was to break the connection between pins 4 and 5 of the video connector.

I retained the ferrite bead FB7 and 220pF capacitor C13 of the original circuit and added as many components of the new buffer circuit inside the metal shield as I could. This worked by using solder pads from unfitted components where possible.

C13 is connected to a new pulldown resistor 330R and to the negative side of a new electrolytic capacitor 47uF. The positive side of this capacitor is soldered to an unused throughhole location in the PCB. Bias network resistors 1K8 and 2K7 are also connected to this point, by soldering to the lead of the 47uF capacitor. There are unused component pads for C52 which provide the correct voltages for the bias network.

Thus four of the new components are mounted inside the metal shield, only two need to be mounted on the solder side of the VIC-20 PCB. These are the buffer transistor itself and a 270R emitter resistor. I used a ZTX300 for the transistor because of its compact size (and because I had a few in my parts drawer), but many other small signal NPN transistors will do.

The mod worked first time and gives an amazing degree of sharpness to the VIC-20's picture. Most of this comes from not mixing the chrominance signal in with the luminance channel; the luminance signal is pretty sharp by itself without all the checkerboard chrominance dots.

The colour is the same intensity as the composite signal, there was no large drop in colour level. There was still a small amount of smearing between colour edges, e.g. from red to cyan, from magenta to green or from blue to yellow.

This is because the 220pF capacitor and the 330R pulldown resistor together form a 2.19MHz high pass filter on the chrominance output. This filtering is useful in the original composite circuit because it reduces the amount of chrominance interference in the luminance signal, but here there is no need for it.

[a4000bear]

Thank you e5frog and rga24 for your experiences.

I modified those two VICs I recently received yesterday, and have come to the conclusion that the low colour levels are caused by faulty VIC chips. Both new machines had low colour. My two original modified VICs have plenty of colour. I confirmed this by swapping the VIC chips between the good and bad machines.

This may explain why I thought the video quality of the two new computers were especially bad before I did the modification. The chips were already faulty.

I have an oscilloscope, and was able to examine the chroma signal. With the faulty chips, it was low as expected. The colour burst was also badly distorted (which would be the real cause of the low colour saturation as seen on the monitor) There was also a spurious DC level on the outputs, varying on alternating lines, probably due to a failure the VIC chip that is allowing the PAL switch signal to corrupt the output. There was also a lot of noise on the colour signal.

I also tried a buffer transistor, and while it does increase the colour signal level, it made very little difference to the colour level as seen on the monitor, due to the distorted colour burst as mentioned above. (Its the ratio between the level of the colour signal and the level of the colour burst that determines how much colour appears on the monitor)

From this, I have found that:

Low colour is caused by a bad VIC chip.
Low colour cannot be corrected with a buffer transistor.
Not having a buffer transistor will not harm the VIC chip (I have two VIC 20s that have had this mod for years, without the buffer.

One thing I'll be doing today is to see what effect the buffer has when a good VIC chip is installed.

I will also compare a bad chip and a good chip in an unmodified VIC to see what effects there are. Hopefully it will allow us to spot a bad chip beforehand.

[rga24]

Hi a4000bear,

I've noticed the colour level in the composite signal is different between original VIC-20s and VIC-20 CRs. Here in the UK, our original two prong VIC-20s have a relatively low chrominance signal in the composite video output, which produces quite a clean video signal on a monitor with not very much chrominance noise interference but relatively unsaturated colours.

Our VIC-20 CRs with 7 pin DIN power connectors have a much higher chrominance signal in the composite output, and having studied the circuit diagrams I think this is because Commodore moved where in the video circuit the chrominance signal is applied, from before the 1K preset R10 as drawn on Commodore's schematic, to after it as seen in actual manufactured units.

Other factors may be the use of 6561-101 VIC chips instead of 6561E, the 240R series resistor on the luminance output (pin 3) rather than the 180R resistor in the original, and the 510R current source resistor instead of the Q6 circuit.

The effect on the screen is that the colour saturation in VIC-20 CR displays is much higher than in original VIC-20 displays, but the chrominance noise is much more noticeable as well.

[a4000bear]

Hi a4000bear,

I've noticed the colour level in the composite signal is different between original VIC-20s and VIC-20 CRs. Here in the UK, our original two prong VIC-20s have a relatively low chrominance signal in the composite video output, which produces quite a clean video signal on a monitor with not very much chrominance noise interference but relatively unsaturated colours.

Our VIC-20 CRs with 7 pin DIN power connectors have a much higher chrominance signal in the composite output, and having studied the circuit diagrams I think this is because Commodore moved where in the video circuit the chrominance signal is applied, from before the 1K preset R10 as drawn on Commodore's schematic, to after it as seen in actual manufactured units.

Other factors may be the use of 6561-101 VIC chips instead of 6561E, the 240R series resistor on the luminance output (pin 3) rather than the 180R resistor in the original, and the 510R current source resistor instead of the Q6 circuit.

The effect on the screen is that the colour saturation in VIC-20 CR displays is much higher than in original VIC-20 displays, but the chrominance noise is much more noticeable as well.

I did notice that too. I recently got a CR VIC and a 2 prong VIC and compared them before I did the mod yesterday. The CR was a lot sharper, and had a higher chroma level (to the point it made some characters look chopped up if displayed in certain colours) It also had a lot more noise and banding.

Unfortunately at the time I did not know both machines have faulty VIC chips. I'm going to modify one of my CR VICs back to composite, and compare it with good and bad VIC chips to see what effect there is.

By the way, in your CR VIC, do you have a 180 ohm resistor in the FB8 position on the board?