|
|
|
#32
05-09-2009, 01:38 PM
|
||||
|
||||
|
Joe:
the tint drift depends IMHO on temperature and changing drive voltages. The Porta-color is not a standard PAL, but a simple PAL receiver with very simple circuits and without the PAL delay line. I will check whether I find the tint drift with other simple PAL receivers. - Eckhard 
|
|
#34
07-21-2021, 12:35 PM
|
|||
|
|||
|
This thread is so old, the schematics do not load anymore, so I am attaching the schematic that I found in another thread. If there is a more recent schematic, please re-post it to this thread for others.
 The mod only does DC restoration for the Y luma signal, but the mod does not also add DC restoration to the Y-R, Y-B, Y-G chroma signals that are attached to the CRT's G1 grids. So the described mod would improve luminosity consistency, that is, the black levels and white levels would be consistent, but chrominance consistency will still vary depending on the amount of average red, green, and blue in the scene. Since Y-R, Y-B, Y-G are independent signals that are AC coupled to the CRT's G1 grids, the level of color saturation for each color will depend on the average amount of red, green, and blue in the scene. Has anybody tried DC restoration for the Y-R, Y-G, Y-B chroma signals that the GE Porta Color sends to the CRT's G1 grids? Last edited by LukeSimon; 07-21-2021 at 01:05 PM.
|
|
#35
07-21-2021, 01:30 PM
|
||||
|
||||
|
The grid signals have about 70% DC coupling already. I think this was done to reduce sensitivity to tube drift in the color difference amps.
I think you could get 100% by removing the .0068 cap across the 270K in the diff amp output circuit. This would reduce the AC gain to the same as the DC, and the blue and green brightness pots would still work. I'm guessing the color gain has enough range to make up for the 30% reduction in AC gain. I have thought of doing this fix to my CTC-5, but never got around to trying it. The reduced DC coupling is not noticeable in most pics unless they are close to a solid single color.
|
| Audiokarma |
|
#36
07-21-2021, 01:59 PM
|
|||
|
|||
|
Quote:
 In this screenshot, wouldn't removing C92 cause the G-Y color signal to be attenuated by R103, significantly reducing chroma gain? Also, the output impedance of the color difference amplifier would be very high, significantly reducing the bandwidth of the amplifier. Instead, if C92 was shorted, the result would be DC coupling, so DC restoration is not necessary, but then the green brightness pot no longer works. Ignoring that pot for a moment, why not just use DC coupling from the plate of the amplifier to the CRT G1 grid? I still do not fully understand that.
|
|
#37
07-21-2021, 02:39 PM
|
||||
|
||||
|
Removing C92 would reduce the AC signal by the same ratio as the DC signal, that is to 75% of the signal at the amp plate. I don't know what the CRT grid capacitance is, but a high frequency rolloff could be countered by a small cap in place of C92.
Example if CRT grid capacitance is 10 pf (wild guess): it is across about 800k (R104 and R8). 30 pf or so across R103 should fix it, if I've got my head screwed on straight. Last edited by old_tv_nut; 07-21-2021 at 06:35 PM.
|
|
#38
07-21-2021, 06:24 PM
|
|||
|
|||
|
I have one and will get a you a Pic tomorrow the plastic is very dirty and hard to see clock and I'm afraid of cracking it if I try to remove. Mine works OK and the funny thing the inside looks quality compared to whats made today
|
|
#39
07-21-2021, 06:48 PM
|
||||
|
||||
|
Regarding the use of less than 100% DC coupling in the chroma channels:
As you stated, this does shift the overall color balance towards the opposite of whatever the dominant (average) color is. In TV sets, this reduces the effects of DC drift in the color difference amplifiers. In color photography, this is a simple form of correction for overall color shift in the lighting or the processing of prints from negatives. Kodak's processing machines went through several versions of this. Early ones did the balance individually on each negative and could suffer "subject failure" where, for example, a red sweater covering a large portion of the image would shift everything in the print towards cyan. Later machines averaged all the shots on a roll, thus coming closer to taking out the variations in the particular batch of negative film. At this point, the snapshooter had to use blue flashbulbs to match daylight when taking indoor photos. Later machines measured all the negatives on a roll, but additionally classified them into orangish (taken with clear bulbs or incandescent light) or bluish (taken with daylight or blue bulbs). These technique are called "gray world" color constancy algorithms. https://www.codeproject.com/Articles...orld-Algorithm
|
|
#40
07-22-2021, 01:30 PM
|
|||
|
|||
|
Quote:
Last edited by LukeSimon; 07-22-2021 at 01:34 PM.
|
| Audiokarma |
|
#41
07-22-2021, 04:07 PM
|
||||
|
||||
|
DC coupling (which exists in the circuit) couples any change in the plate voltages to the grids. Changes can happen due to changes in components with temperature and age, or just changing the amplifier tube. DC restoration at the grids would remove such changes due to the preceding stages and would give the most stable result.
Both full (100%) DC coupling and DC restoration give the same result in the picture except that the first one can still drift over time and require readjustment of the gray scale tracking. The reduction in DC coupling by 25% (to 75%) is not really a big improvement in stability, but the designers seem to have thought it worthwhile. It would make sense to me to try 100% coupling first, and see how well it works and whether it drifts noticeably. If it's OK, you're done; if not, you can proceed with DC restoration. Maybe the color difference amps cannot make the necessary 33% gain increase; in that case you will need to do the DC restoration; but it's simple enough to try first. Last edited by old_tv_nut; 07-22-2021 at 04:11 PM.
|
|
#42
08-22-2021, 09:17 PM
|
|||
|
|||
|
Below are details of my implementation of the DC Restoration mod. First, the parts list:
 Two 4.7uF capacitors in parallel to keep the voltage coming off of the brightness pot very stable. The rectifier clamps the luma signal's sync pulses to the voltage coming off of the brightness pot.  Lug terminal is attached via screw to full metal chassis. Two caps are soldered to lug terminal, and hookup wire is run from the lug terminal.  Hookup wire is run through the TV's crawl space.  Rectifier is soldered on the underside of the TV, between two test points as indicated in the schematic.  Oscope readings of the CRT's cathode voltage. Top row is before the mod. Bottom row is after the mod. Left column is a solid white test pattern. Right column is a solid black test pattern. "Max" is peak voltage of the CRT's cathode. Before the mod, max cathode voltage (which is used for CRT electron gun cutoff) varies by at most 33.6 volts between bright scenes and dark scenes, and after the mod, max cathode voltage varies by at most 3.2 volts between bright scenes and dark scenes. That is more than a 10x improvement in brightness consistency between bright scenes and dark scenes. Last edited by LukeSimon; 08-22-2021 at 09:50 PM.
|
|
#43
09-04-2021, 04:51 PM
|
||||
|
||||
|
That is a very nice improvement! Have you watched some video on the set, and confirmed that it works as well as the numbers indicate it should?
__________________
Chris Quote from another forum: "(Antique TV collecting) always seemed to me to be a fringe hobby that only weirdos did."
|
|
#44
09-13-2021, 01:06 PM
|
|||
|
|||
|
I took a bunch of before and after pictures to show the benefits of DC restoration. The video source is a 1992 SNES that outputs a 240p video signal.
Top is before, bottom is after. Picture has high average luminosity, so without DC restoration, the brightness is too low. Top is before, bottom is after. Picture has low average luminosity, so without DC restoration, the brightness is too high. Left is before, right is after. This is an extreme close up of a 240p test pattern of 240 horizontal alternating white and black lines. The ideal picture is for the black lines should to be thicker than the white lines, since this is 240p video where every other line of video is supposed to be blank, and so each black line is displayed next to a blank line in the 240p video that is displayed on a 480i TV. It is a great test pattern for measuring cathode ray spot size (beam diameter), which is useful for tuning the electron gun for a maximally sharp image. DC restoration helps keep the peak cathode voltage very consistent, which is crucial for making the cathode ray as sharp a possible. Here I set peak CRT cathode voltage, which the DC restoration mod clamps to, to be 250 volts (near max). CRT G1 anodes are 175 volts (near min). CRT G2 anodes are 670 volts (max). Focus anode is 270 volts. Not sure why this is not talked about more openly in vintage TV discussions, but cathode ray beam sharpness is highest when cathode to G1 voltage is as high as possible. This means lowering the TV's "brightness" as low as it can go (this increases cathode to G1 voltage), so that the picture is completely black, and then achieving correct black levels or cutoff by increasing the G2 screen anode voltage. Many TVs will hit their max screen voltage, and brightness will be too low. So the "brightness" setting, (which adjusts the cathode voltage to G1 voltage) needs to be slightly increased to achieve correct black levels. Maximizing cathode to G1 voltage squeezes the cathode ray to a sharp point in the beam forming region that is in between the G1 and G2 anodes, but this also means that G3 screen voltage must be maximized to prevent the picture from being too dark. Since this technique relies on setting the peak cathode voltage to a very high value, it only works correctly if DC restoration is in place. Once the beam is sharpened this way, then the focus voltage should be adjusted to further sharpen the beam. The reason why you should only tune focus voltage AFTER tuning K, G1, and G2 voltage is because changing the focus voltage changes the focusing power of the lens... and changing the focusing power of a lens impacts two aspects of the electron gun "optics": (i) focal length, and (ii) magnitude of spherical aberration. Spherical aberration is never a good thing because it causes the electron beam to scatter, even when the lens has been tuned to have the correct focal length to the phosphor CRT face. So when you are tuning the focus voltage, you have to balance focal distance with the amount of spherical aberration to get the sharpest image. The thinner the beam is, when it enters the lens, the less spherical aberration occurs. So sharpening the beam in the beam forming region by maximizing K to G1 voltage means that the focus lens can potentially achieve a better trade off between spherical aberration and focal length Top is before, bottom is after. Here is a comparison of the improved sharpness. Keep in mind that this is a first generation Portacolor that uses a low voltage focus CRT with a gigantic dot pitch. Last edited by LukeSimon; 09-13-2021 at 04:34 PM.
|
|
#45
09-13-2021, 01:27 PM
|
||||
|
||||
|
Very nice thorough demonstration!
|
| Audiokarma |
 |
|
|
Linear Mode
