21-CT-55 CTC2B on eBay

[Pete Deksnis]

Here's what I think. Until I see an engineering drawing (vintage, of course) with phosphor types specifically spelled out, all I know for sure is that the 15GP22 had it.

Rumor has it that the 21AXP22 had it but that the 21AXP22A didn't. Who knows for sure?

Purist? Maybe.

Pete

[Steve D.]

Originally Posted by yagosaga

Hi Steve (D.):
on your web page I saw you own a CT-100 with a 15GP22 and a CTC-5 with a 21AXP22A. We know that the 15GP22 has the early type phosphors according to NTSC norm. What is with the 21AXP22A? If you compare the colors, is there a difference, and when in which color?

Eckhard

Hello Eckhard,

My CTC-5 has a replacement 21CYP22A tube. I can say that the color image on the CT-100 is more vivid and true to the primary color spectrum.
Most appearent is the natural deep reds. There is a higher brightness level on the 21CY and overall the color reproduction, is a bit more muted and similar between the 21CY and 21AXP22A's I have owned. My CTC-5 is the Deluxe chassis with X & Z color demodulation and additional color amp. Let me say that both a well set up CT-100 and CTC-5 produce far more natural, IMO, color pictures then later model color sets with sulphide tubes and "improved" phosphors.

-Steve D.

[old_tv_nut]

If you have a 15GP22 and a 21-inch side by side, you can easily see if the green phosphors are the same. Put on color bars and turn up the color level (and maybe turn down the brightness) so there is absolutely no red or blue in the green. Or, disable the red and blue guns Try to get both about the same brightness to make it easier. The non-P1 (sulfide) green phosphor will look yellower than the original P1 green.

Another clue (but harder to see) is that the P1 green persistence is longer than the sulfide green, and if you know how to run your eyes quickly over a white object on a dark background, you will see a green smear on the original P1, more than on a sulfide tube. Another way to see this is to stand far away (like 40 feet) from the two sets while they show a black and white picture and run your eyes rapidly left and right. Inthis case, the whole screen is the "white object". Another way to see a difference in persistence is to spread your fingers apart and wave your hand rapidly in front of an all-white screen, looking for color fringes.

[old_tv_nut]

By the way, the differences in color between chassis have a lot to do with the color demod gains and angles as well as the phosphors. The "deep" reds in a properly adjusted CT-100 are as much due to the use of a standard NTSC chroma matrix as anything else. In later sets with sulfide tubes, the green is yellower, which means that turning on the green gun is also adding some red to the color. This would mess up flesh tones, so an adjustment is made by increasing the R-Y gain, thus turning off some of the red whenever the green is turned on. However, this has the opposite effect on red colors, turning the red on excessively and making reds overly bright. This defect would not occur if picture tubes were linear devices. It is the picture tube voltage-current characteristic ("gamma") that makes the gun turn on more strongly than expected and turn off less strongly than expected. This brightness distortion in different colors is the trade-off in order to make hues roughly correct with the new green phosphor (especially to maintain the natural variations in flesh tones). There is an opposite effect on saturated cyan colors, which are reproduced too darkly.

PAL solved this problem by setting the camera matrix to match the newer phosphors. Thus, PAL receivers cannot reproduce the saturated "NTSC green" either, but at least the saturated colors are all the correct brightness and hue. Same goes for HDTV - the camera matrix has been correctly specified for the modern phosphors.

[yagosaga]

old_tv_nut:

thank you very much for this detailed explanation. The difference in the green color is the most significant difference between the 21AXP22A and PAL color tv sets. I have wondered why our tv broadcasts look so strange when displaying green on the CTC-5. Now I know that this is due to different camera matrix. But on the other hand, NTSC videos looks much richer in colors on the CTC-5 than on a PAL color tv set.

Eckhard

[yagosaga]

P.S.: what is the difference between sulphide phosphors and P1 phosphors? Which CRTs used sulphide phosphors and which P1 phosphors?

[Steve D.]

Originally Posted by yagosaga

P.S.: what is the difference between sulphide phosphors and P1 phosphors? Which CRTs used sulphide phosphors and which P1 phosphors?

I believe the sulphide tubes have a higher brighness level. The red phosphor more closely matches the green and blue in light output. The sulphide tubes were introduced in the 1962 model year (ie: RCA CTC-11) with the 21FBp22 & 21FJp22 picture tubes.

-Steve D.

[Pete Deksnis]

Originally Posted by old_tv_nut

PAL solved this problem by setting the camera matrix to match the newer phosphors. ...Same goes for HDTV - the camera matrix has been correctly specified for the modern phosphors.

I echo Steve D's characterization of 'deep' red on the CT-100. It is often a computer-generated graphic including text but also occurs in video from a camera.

Since the phosphors in a '53 display device (15GP22) controlled the colorimetry of the camera of the day (TK-40) and the same thing happened today with PAL and HDTV, and in between SMPTE in 1982 slewed the '53 standard for a more bright picture, what the dickens is feeding my CT-100 today? Is that 'deep' red we see something that would have happened with a TK-40 generating the video?

[Pete Deksnis]

This thread seems to have become aligned with a quest of mine: trying to discover what 1953 NTSC color actually looks (looked) like. To that end, this morning I did another side-by-side comparison between the CT-100 and a modern HDTV operating in NTSC mode, with both sets driven by the same OTA rooftop antenna.

As inspired by old-tv-nut earlier in the thread, I balanced the sets in terms of brightness, contrast, color, and tint. This meant pushing the CT-100 to a point just below where, on super-bright screens, the raster begins to bloom when the load is so great the 6BD4 H-V regulator can no longer maintain adequate ultor voltage (which occur quite often but usually only in commercials where the screen is driven very white with only a minor amount of pastel text). Conversely, the HDTV set was throttled back to match the brightness and contrast and color of its elder.


In a nutshell:

CT-100 displays better white than the modern set, which tends to reproduce more 'eggshell' in hue than the CT-100's white.

CT-100 displays more saturated blue than the modern set, which seems to reproduce a less 'bright' blue of possibly even a different hue.

CT-100 displays a less bright red than the modern set, which tends to reproduce the 'red' with a clear shift in hue towards orange.

CT-100 seems to display green with greater saturation in general and hue seems to be simply more 'green' than the modern set.

Finally, it's important to remember that these observations were made with sets that were adjusted to produce pictures that look identical most of the time. The 'greens' matched each other. Ditto the blue and red. Only occasionally would a given hue be different between the two sets.

The above observations were made watching PBS kid's shows on channel 13 from NYC.
--------------------------------
A further observation was made this morning watching GMA on channel 7 from NYC. Here the cameras are HDTV where the camera matrix, if I interpret old-tv-net correctly, mates the camera and modern HDTV-spec phosphors.

In a nutshell:

Much of the GMA studio has colors that do not translate from HDTV to 1953 NTSC very well. Areas of the GMA set swing very bright on the CT-100, away from the subdued orange-like hue, and on almost to white. This is the only OTA video I've seen that shows a striking difference between the two television standards.


[I wanted to switch the modern set between NTSC and ATSC modes to confirm that there was essentially no visual difference between the two, but the ABC digital transmitter (ch. 45 in NYC) was either off the air or another channel 45 from somewhere skipped in and swamped my ATSC front end. [the ATSC-equivalent of FM capture ratio must suck!]

[ha1156w]

I worked on a 15" IBM PS/2-era monitor back in the mid-90's for a friend of mine. It needed a new VGA cable grafted on to it due to a short in the connector of the original. When testing the finished repair, I was BLOWN AWAY by the color deepness in the red and blue, compared to the other monitors I had on hand. The Windows 95 logo on bootup had a significantly more blood-red cast instead of the orange-ish color that I was used to seeing. There wasn't anything wrong with the brightness or contrast adjustments on the monitor, as everything else was "normal". I haven't seen another monitor with such intense, clear coloration since. Could it be that the IBM monitor also used a true-color phosphor in their tube, at least for that model?

[andy]

---

[old_tv_nut]

Pete D - thanks for the heroic attempts at a comparison. I say heroic, because trying to get various factors (like white point color) to be the same in different sets so you can compare the phosphors and matrices is VERY difficult. You cannot even do the white point and gray scale matching with a colorimeter (a device having multiple filters and sensors) unless the colorimeter has been calibrated for the particular phosphors. [Then how DO you calibrate it, you ask.] The only accurate electronic instrument to measure phosphor color is a spectroradiometer, which should be calibrated to a standard source traceable to NIST. However, spectroradiometers are much more expensive, and used to be much slower reading, than colorimeters. At Zenith, our usual technique was to use a colorimeter calibrated to match a spectroradiometer, with a different calibration setting for each new phosphor set.

The other way to guarantee correct white point and grayscale tracking is by means of an optical comparator that presents a D65 field and an immediately adjacent view of the screen to the person making the adjustment. We had one of these for a while at Zenith, but I haven't seen one in a long time. The optical comparitor gave very precise results for one particular observer - so precise that the next observer would get noticeably different results (looking in the comparator) depending on normal slight variatons in color vision from one person to the next. For example, one of our older engineers always required more blue current to match the comparator than I did. I think this was a combination of his brown eyes and age (indicating more pigmentation in the eye itself) vs. my blue eyes. I'd also guess that I require more blue than I used to as my eyes age. Color experts who have had cataract surgery on one eye have written papers on the greater blue-violet sensitivity when the lense is removed. Gas stove flames, in particular, may go from dull to vivid violet/blue. We have one engineer at work who had lens implants in both eyes - but the doctors put a UV blocking lens in one eye and a clear one in the other, and he reports he can see such differences between his two eyes.

A similar thing: if you wear glasses that turn dark in the sun (as I do), put them down on a white surface and see if they aren't slightly amber rather than fully clear indoors - this will affect your sensitivity to the blue phosphor, possibly to a greater extent on the newer tubes where the blue phosphor is more towards violet.

[Pete Deksnis]

You bring up an interesting point about problems with the use of the human eyeball as an objective instrument for measurement. I've been curious about it too. Last February during an eye exam I was able to prod the doctor into a color test. A nurse administered just a quick test. (Nothing too advanced when they're working with Medicare payment rates.) There is a slight color degradation, something I first learned in 1960 when I was twenty. No big deal, as far as I can see...

So, knock on wood, today it's no glaucoma or cataracts and an unaided vision of 20/30 in the 'instrument' I've been using to make these subjective measurement/observations.

I figure that if this measuring instrument 'sees' the same colors on the modern and vintage sets that were balanced using the 'instrument', the whole idea is a wash, because the sets were designed to be different. But in actuality, I have for example seen pale yellow 'look' the same on both sets. Then, in the same program, more saturated yellow areas are observed in different hues. So, could it be that, when I see the pallet of color in 'eyeball' agreement between the two sets, it becomes a 'reference' that confirms other differences in hue valid? I guess I'm going under that premise. Without a bevy of pricey instrumentation though, who knows if it holds water...?

I have fun messing with it.

'deep' 1953 red forever

Pete

[old_tv_nut]

Pete, I presume the quick test was done with pseudo-isochromatic plates (where you try to see the numbers hidden in a bunch of polka dots). This is a very rough test when given in the GPs office, good for finding gross problems only.

There actually is a fairly good test that can be done on your computer, which will identify color discrimination problems. The software is not expensive, but not free, however, so you may not want to bother.

http://www.visualmill.com/shop/catalog/index.html

Even this will not tell you if you have color mixing curves within the normal range, just if you have color confusion on certain hue axes; complete characterization of color vision takes something called an anomaloscope:

http://4colorvision.com/files/newnagel.htm

The basic principle of this is to present an actual yellow wavelength of light, and for comparison, a mixture of red and green that can be adjusted by the subject until the mixture appears to match the "real" yellow. The ratio of R/G and the total amount then define several types of long-wavelength or mid-wavelength color defiiciency. For example, if the subject in repeated trials selects a wide range of match ratios, it may indicate a lack of discrimination between red and green hues.

As the web page explains, recent versions have also been developed to measure short-wavelength deficiencies by matching a cyan wavelength with a mixture of blue and green.

You are actually doing an anomaloscope type of experiment on yourself if you try to adjust the R, G and B drives on a monitor to match a given external color (which has a different spectrum from the mixture of phosphor spectra that reproduces it).

Visually matching the screen white point to natural daylight has some of these aspects to it, but I note two things:

1) the R/G vs. yellow and G/B vs. cyan measurements involve only two variables each, so are much easier to do
2) sensitivity to hue is much greater at higher saturations, so the variance in measurements of your vision would probably be huge for data derived from white point settings.

[Pete Deksnis]

Wayne, you are right about the test being pseudo-isochromatic, the same test that was administered back when I was twenty, basically to see if I could read the color code.

I'm going to bite the bullet and spring for the visualmill test.