TEA2000 mod for better CVBS quality?
Has anyone tried this before? :confused:
Anyone knows who was the original author of this modification? I would like to know if it is possible to improve this circuit or build a better one. :cool:
I have been experimenting with various values for the coil and the ceramic capacitor and I *think* that a 15~18 uH coil and a 47pF ceramic capacitor gives the best compromise between picture fuzziness (but still very sharp) and color fringing. :)
Anyone knows who was the original author of this modification? I would like to know if it is possible to improve this circuit or build a better one. :cool:
I have been experimenting with various values for the coil and the ceramic capacitor and I *think* that a 15~18 uH coil and a 47pF ceramic capacitor gives the best compromise between picture fuzziness (but still very sharp) and color fringing. :)
Post edited by Renegade on
Comments
I haven't tried it but I think Ingo has. I've done an s-video mod on a grey +2 and this avoids the issue of chrominance and luminance separation altogether because they are sent separately. No interference whatsoever.
Also, screenshots of the tv screen before and after the mod would be nice to compare.
Bytes:Chuntey - Spectrum tech blog.
https://www.youtube.com/watch?v=GMKGIP5Yuqs
Before and after shots of the modification I did with different values here (sadly done on a +2 BEFORE lifting C31 and twisting the 2N3904 transistors):
http://i.imgur.com/FIzV4zf.jpg
I am the author of that mod and I soldered this into a lot of +2 computers as for myself and for some of my friends. It removes the yellow and blue shaddows from vertcal lines.
The aim is to avoid interferrence of high frequency parts of the video signal with the colour carrier of 4.433619MHz. So this is a series resonant circuit that shorts 4.433 MHz signal parts in the video signal. The parts you used (15 ... 18 ?H / 47 pF) have a series frequency a bit more than 4.433 MHz. But there are tolerances and circuit capacities. So I think you took the right parts. Best is to use a trimmer capacitor to adjust for best picture.
The frequency of the circuit is calculated by: fres=1/(1*PI*SQRT(L*C))
BTW: It might be better to use a lower capacity and a higher inductance to avoid clipping high frequencies in the video signal (leaving a sharper picture) for example 68?H and 18pF - same resonant frequency but lower capacitance.
Greets
Ingo.
@Arjun: Sorry, this is the best my camera does. When it comes up and close it goes out of focus. Hope you can make out what I did in these 2 photos: coil and capacitor are soldered together on one leg. One end goes where C31 was disconnected while the other end is soldered to ground pin on TEA itself.
@rga24: That sounds very interesting. Hoping to see your posts on that subject on your other topic (even though I currently find the making of best possible image, CVBS or S-video, out of a 48K, a greater challenge right now!) :cool:
@Shockwav3: Yes, pretty much what happens to me too. :rolleyes:
@Ingo: Congratulations are in order then! :cool: Thanks for the explanation and the formula. I suppose you are right, although my eyes tell me that the picture is "nice" it may be mathematically incorrect. I must find a trimmer capacitor for better tunning.
BTW, is your mod suitable to an early +2 which has motherboard revision issue 1?
Bytes:Chuntey - Spectrum tech blog.
After a careful tweak this is the best I get from my trusty 14'' tv set:
If it is via CVBS / composite input, did you check that TR4 (video driver) at the +2 is soldered in the right orientation. If it is a 2N3904 then at most boards it has to be soldered vise versa to the orientation printed on the board. Otherwise emitter and collector are interchanged.
Ingo.
Yes, I know about TR4 and I already put it the correct way.
My camera is really bad at taking pictures from a CRT screen :-x but in reality it is not as bad as it looks. :lol:
Anyway, I think you are right, I got a sharper picture from using a 68 uH coil followed by a trimmer capacitor which ranges 9~50 pF.
@BCH: I suppose we can expect this mod to work the same way on the +2A/+3, provided that the composite video out circuit is completed and similar to +2!
Sintech sells a small pcb and components but you could make it youself:
http://sintech-shop.co.uk/sinclair/videomodulator--2a--3/a-6095/
On the other hand, there seems to be a better alternative (according to the author) but I have not tried that myself so I can't speak on how much better it can get. Still, I would like to try it someday.
http://www.secarica.ro/html/plus3_hardware.html#compositevideo
http://www.secarica.ro/zx/plus3_composite_video.pdf
Since that filter works by attenuating frequencies in the luminance signal near 4.43 MHz, it will reduce the magnitude of the black / white / black / white dot patterns in Chromatrons (and any 50% grey pattern). The colours will still be there but reduced in saturation, proportional to the reduction in amplitude.