Wow, Wow, Wow, Wow, Wow! Absolutely brilliant. I learn't about tv waveforms at college back in the late 80s and sadly have forgotten much of the detail since - amazing how a little prompting brought it back. Thankyou for such a superb explanation. Now to build the circuit and try it out on my newly-acquired ZX81 !!
You make such superb videos. Everything you show is so very well thought out and explained. It must take you ages to get all the footage and edit it together, but thank goodness that you do. A big thank you for another great video. Cheers!!
Finally got around to building the circuit (only changed the 22uF cap to 10uF since I was lazy and didn't have any 22uF ones). Works a treat, picture on my flatscreen is very good, on my sony PVM broadcast monitor quality is on par with emulation. I'd post a before-after-picture but for some reason youtube won't let me paste the URL. Thanks a lot for this.
I built this circuit for my Timex Sinclair 1000 and had a few difficulties. It works great now that I've figured it out. Just so any other poor souls try this on a Timex Sinclair 1000, the main thing you must know is you MUST attach the video input to pin 16 of the ULA. The USA builds do some magic between the ULA and the RF box and you can't take the signal from the wire that enters the RF box. Other notes: There are 2(3?) different ULA chips depending on your model. ULA 2C184E needs this circuit or some other video circuit to add back porch. ULA 2C210E does not need this circuit, but might need a simple transistor amplifier. www.classic-computers.org.nz/blog/2016-01-03-composite-video-for-zx81.htm I hope this helps!
It is nice to look at a video made by somebody who knows what he is talking about. Improving the video-signal with so little components is almost magic. Somewhere in the far distance I see two colours coming to the ZX81, in stead of black and white. If you are able to find the back-porch you can be skilful enough to transform the black and white signal into primary colours. Only two colours full screen of course, I know the limits of the ZX81. I am not asking for black, white and a form of grey, that would be impossible.
555 circuit was tested on a kit built ZX80 which suffered from a poor and fuzzy picture even with the brightness at 100% due to no back porch output. Image quality on the 80 is now very good, so much so that the brightness and contrast levels on the flatscreen test television are down in the 50% levels.
Excellent - I dont think this could have been explained any more perfectly. I'm currently building a ZX81 clone using an old ULA (without the back porch) and havent been able to understand whats going on with other approaches so thankyou :)
My ZX81 no longer produces a signal at the mic socket when saving, and I'm wondering if this is because of the composite mod, since isn't the video pin on the ULA also feeding the mic socket? Anyone else experienced this?
I had to use a 100 Ohms resistor load in the video out. If not done the new TV sets do not detect the video signal. My ZX81 uses the new ULA with back porch. Thanks.
The circuit as presented does require the display's input to provide a path of approx. 75 ohms to ground. Most do, even modern ones, but obviously there are exceptions. Adding an extra resistor as you've done is the right thing to do for those that don't. Depending on the design of the display's input circuit, it might be possible and appropriate for the load resistance to be quite a bit higher than 100 ohms; I suggest experimenting to find out.
Brilliant, another masterful improvement to a classic system! I might need to do this mod to my ZX81 - need to get it out and check which revision it is. Do you know how to identify which ZX81s this mod is not required for - ie. what revision board / ULA?
No prob! I just didn't think of the obvious ways to check there. I love this channel - very educational! Good time of day to you!!! < We need a T- Spring campaign with that slogan!!!
1:35 _"Strictly, this is a luminance video signal, not a composite one."_ I am pretty sure that technically that's a luma signal, not luminance. And is the term "composite" so so precisely defined? The signal you show, after after all, is not just the luma information but is the composition of the luma and two sync signals. Combining the horizontal and vertical sync signals together (which is not correctly done with a simple XOR, though that usually works ok) is usually called "composite sync." I deal a lot with split and composite sync signals independent of the video signal. To try to reduce confusion with baseband single-channel signals that include composite sync, I've taken to using the terms CVBS (widely used, with many different ways of expanding that acronym) and MVBS to distinguish those with and without a colour signal mixed in. (This is independent of scan format; e.g. I use MVBS for both 60 Hz (old B/W NTSC) and 59.94 (newer colour NTSC) scan rates, so long as a color signal isn't present). Such quibbles aside, this video is well done and very informative.
That's weird as the Memotech MTX512 actually had a power pack that had a -5v supply as well as -12v, I think that the -5v supply is to make the pc produce the standard video signal that all televisions expect.
Pictures as promised. Top: original modulator output, bottom: composite with 555 based back porch generator. Both taken on my Sony LCD, on the PVM broadcast monitor the picture look even better. imgur.com/aBfK6GI.jpg
Hmm, this circuit (the one using the 555) doesn't seem to work for me. My ZX81 doesn't have a back porch, and the line synchronization pulse is about 2.3 V, the rest is about 4V. To trigger the 555, we'd need to go below 1/3*5V (~1.667V). If anyone has a nice solution, I'm all ears. Otherwise I'll reply to myself when I have figured out something that works.
Got it to work by adding a resistor divider in front (ULA pin 16 → 270 ohm resistor → path 1) 1k resistor → GND; path 2) rest of circuit. (Took some experimentation to get working values ― just calculating normally would drop the voltage more than expected.) Also tried adding fast diodes in front at first (driven directly by the ULA), but that wouldn't work, the signal would end up garbled. (Note: I haven't checked yet how much current the ULA has to push with the added parts.)
I was recently gifted my first ZX81, but for some reason the black to white transistions "bleed" quite bad (white to black is fine) on both of my TVs (Sony PVM tube and KLV LCD). Would this be related to the back porch issue or is it a seperate issue? The board has been modified to output composite (but has no additional circuitry as found in various instructions online).
+Hans “Shockwav3” Meier I wouldn't have thought it would be, but a missing back porch can cause all sorts of odd symptoms. How is the contrast of the picture? Do you have to turn the TV's brightness or contrast control up high to see the picture? Other causes could include noise on the signal that coincides with black to white transitions. Another possibility is that the signal rise time at these transitions is too long. One cause of this could be excessive capacitance or inductance in the interconnect. It'd be worth trying a different (and as short as possible) cable to connect the Spectrum to the TV. Is the video output taken directly from the ULA or has any interposing circuitry been added, even just a capacitor? I think further diagnosis would require viewing the composite video signal on an oscilloscope, with and without the TV connected.
+JoulesperCoulomb Turns out the modification I found was wired up a bit odd. This is the intended circuit on the left and what I found on the right: i.imgur.com/lPqbgMI.gif Threw it out all together now and going to add your circuit at a later point. Above circuit did not fix the backporch issue - which my LCD fixed automatically, but the tube monitor interpreted it incorrectly as both should have.
+JoulesperCoulomb Oh ... one more question ... would a 2N3904 or BC548 work as well in your circuit? I don't have any 2222 lying around - but more than enough of the types mentioned.
Great video, hope you don't mind, I've used a still from this video showing the video waveform in my blog post on improving the video signal on my ZX80 clone. blog.tynemouthsoftware.co.uk/2017/01/minstrel-zx80-clone-video-output.html
Bit late to the party here, but I've just built the 555 circuit and it works perfectly. Come back JoulesperCoulomb we miss your excellent videos.
Thank you for this circuit and the explanation. I built the one based on the 555 and got excellent results. My ZX81 is alive again!
Wow, Wow, Wow, Wow, Wow! Absolutely brilliant. I learn't about tv waveforms at college back in the late 80s and sadly have forgotten much of the detail since - amazing how a little prompting brought it back. Thankyou for such a superb explanation. Now to build the circuit and try it out on my newly-acquired ZX81 !!
You're videos are excellent, to the point, informative and without silly background music or advertisements. Thank you!
You make such superb videos. Everything you show is so very well thought out and explained. It must take you ages to get all the footage and edit it together, but thank goodness that you do. A big thank you for another great video. Cheers!!
Finally got around to building the circuit (only changed the 22uF cap to 10uF since I was lazy and didn't have any 22uF ones).
Works a treat, picture on my flatscreen is very good, on my sony PVM broadcast monitor quality is on par with emulation.
I'd post a before-after-picture but for some reason youtube won't let me paste the URL. Thanks a lot for this.
I built this circuit for my Timex Sinclair 1000 and had a few difficulties.
It works great now that I've figured it out.
Just so any other poor souls try this on a Timex Sinclair 1000, the main thing you must know is you MUST attach the video input to pin 16 of the ULA. The USA builds do some magic between the ULA and the RF box and you can't take the signal from the wire that enters the RF box.
Other notes:
There are 2(3?) different ULA chips depending on your model.
ULA 2C184E needs this circuit or some other video circuit to add back porch.
ULA 2C210E does not need this circuit, but might need a simple transistor amplifier. www.classic-computers.org.nz/blog/2016-01-03-composite-video-for-zx81.htm
I hope this helps!
It does! I am going to be doing this with my recently acquired TS1000. Also planning to do the internal 16k mod.
It is nice to look at a video made by somebody who knows what he is talking about. Improving the video-signal with so little components is almost magic. Somewhere in the far distance I see two colours coming to the ZX81, in stead of black and white. If you are able to find the back-porch you can be skilful enough to transform the black and white signal into primary colours. Only two colours full screen of course, I know the limits of the ZX81. I am not asking for black, white and a form of grey, that would be impossible.
absolutely fascinating. I don't have a ZX81, yet, but this succeeded in explaining video signal generation to me
555 circuit was tested on a kit built ZX80 which suffered from a poor and fuzzy picture even with the brightness at 100% due to no back porch output. Image quality on the 80 is now very good, so much so that the brightness and contrast levels on the flatscreen test television are down in the 50% levels.
Your Videos read like a thriller and I cannot stop watching until the very end. Very educational stuff !!
Excellent - I dont think this could have been explained any more perfectly.
I'm currently building a ZX81 clone using an old ULA (without the back porch) and havent been able to understand whats going on with other approaches so thankyou :)
Everything's allright Sir? You haven't posted a video for a long time.
Built the 555 circuit for my ZX81. Works great :)
Sorry, for commenting on such an old video, but quick question: should this circuit also work on a ZX80?
Exelent video, thank you.
My ZX81 no longer produces a signal at the mic socket when saving, and I'm wondering if this is because of the composite mod, since isn't the video pin on the ULA also feeding the mic socket? Anyone else experienced this?
I had to use a 100 Ohms resistor load in the video out. If not done the new TV sets do not detect the video signal. My ZX81 uses the new ULA with back porch. Thanks.
The circuit as presented does require the display's input to provide a path of approx. 75 ohms to ground. Most do, even modern ones, but obviously there are exceptions. Adding an extra resistor as you've done is the right thing to do for those that don't. Depending on the design of the display's input circuit, it might be possible and appropriate for the load resistance to be quite a bit higher than 100 ohms; I suggest experimenting to find out.
Brilliant, another masterful improvement to a classic system! I might need to do this mod to my ZX81 - need to get it out and check which revision it is. Do you know how to identify which ZX81s this mod is not required for - ie. what revision board / ULA?
GadgetUK164 If your TV produces a picture even on low brightness settings then you only need to do the first circuit I suppose
I guess I could also measure the output on my scope!
GadgetUK164 sure, if you have one. I assumed you didn't.
No prob! I just didn't think of the obvious ways to check there. I love this channel - very educational! Good time of day to you!!! < We need a T- Spring campaign with that slogan!!!
1:35 _"Strictly, this is a luminance video signal, not a composite one."_
I am pretty sure that technically that's a luma signal, not luminance. And is the term "composite" so so precisely defined? The signal you show, after after all, is not just the luma information but is the composition of the luma and two sync signals. Combining the horizontal and vertical sync signals together (which is not correctly done with a simple XOR, though that usually works ok) is usually called "composite sync."
I deal a lot with split and composite sync signals independent of the video signal. To try to reduce confusion with baseband single-channel signals that include composite sync, I've taken to using the terms CVBS (widely used, with many different ways of expanding that acronym) and MVBS to distinguish those with and without a colour signal mixed in. (This is independent of scan format; e.g. I use MVBS for both 60 Hz (old B/W NTSC) and 59.94 (newer colour NTSC) scan rates, so long as a color signal isn't present).
Such quibbles aside, this video is well done and very informative.
I just got a US TS1500. Would this apply? ULA is different? Not much of an electronics boffin, but would really like to connect this to a LCD monitor.
That's weird as the Memotech MTX512 actually had a power pack that had a -5v supply as well as -12v, I think that the -5v supply is to make the pc produce the standard video signal that all televisions expect.
What's that got to do with the ZX81?
Pictures as promised.
Top: original modulator output, bottom: composite with 555 based back porch generator.
Both taken on my Sony LCD, on the PVM broadcast monitor the picture look even better.
imgur.com/aBfK6GI.jpg
Great video :-)
Well done!
I expected better picture without rf modulator, clearly video signal band is to low but i don't know the reason.
Hmm, this circuit (the one using the 555) doesn't seem to work for me. My ZX81 doesn't have a back porch, and the line synchronization pulse is about 2.3 V, the rest is about 4V. To trigger the 555, we'd need to go below 1/3*5V (~1.667V). If anyone has a nice solution, I'm all ears. Otherwise I'll reply to myself when I have figured out something that works.
Got it to work by adding a resistor divider in front (ULA pin 16 → 270 ohm resistor → path 1) 1k resistor → GND; path 2) rest of circuit. (Took some experimentation to get working values ― just calculating normally would drop the voltage more than expected.) Also tried adding fast diodes in front at first (driven directly by the ULA), but that wouldn't work, the signal would end up garbled.
(Note: I haven't checked yet how much current the ULA has to push with the added parts.)
I was recently gifted my first ZX81, but for some reason the black to white transistions "bleed" quite bad (white to black is fine) on both of my TVs (Sony PVM tube and KLV LCD). Would this be related to the back porch issue or is it a seperate issue? The board has been modified to output composite (but has no additional circuitry as found in various instructions online).
+Hans “Shockwav3” Meier I wouldn't have thought it would be, but a missing back porch can cause all sorts of odd symptoms. How is the contrast of the picture? Do you have to turn the TV's brightness or contrast control up high to see the picture?
Other causes could include noise on the signal that coincides with black to white transitions. Another possibility is that the signal rise time at these transitions is too long. One cause of this could be excessive capacitance or inductance in the interconnect. It'd be worth trying a different (and as short as possible) cable to connect the Spectrum to the TV. Is the video output taken directly from the ULA or has any interposing circuitry been added, even just a capacitor?
I think further diagnosis would require viewing the composite video signal on an oscilloscope, with and without the TV connected.
+JoulesperCoulomb Turns out the modification I found was wired up a bit odd.
This is the intended circuit on the left and what I found on the right: i.imgur.com/lPqbgMI.gif
Threw it out all together now and going to add your circuit at a later point.
Above circuit did not fix the backporch issue - which my LCD fixed automatically, but the tube monitor interpreted it incorrectly as both should have.
+JoulesperCoulomb Oh ... one more question ... would a 2N3904 or BC548 work as well in your circuit? I don't have any 2222 lying around - but more than enough of the types mentioned.
+Hans “Shockwav3” Meier I would think both of the alternatives you suggest should work; give them a try.
From www.kolumbus.fi/pami1/video/pal_ntsc.html you can find all timing information of both PAL and NTSC video signals
Great video, hope you don't mind, I've used a still from this video showing the video waveform in my blog post on improving the video signal on my ZX80 clone. blog.tynemouthsoftware.co.uk/2017/01/minstrel-zx80-clone-video-output.html