View Full Version : X/Z color demodulation.


benman94
04-22-2017, 08:40 PM
Can anyone tell me what the X and Z axes of demodulation are that RCA started using in the CTC-7 chassis? Are they 90 degrees apart and just shifted from I and Q? The 7 chassis appears to be RCA's first step away from difference demod since the CTC-2 chassis. Were these just arbitrarily selected?

Electronic M
04-22-2017, 10:18 PM
Pretty sure they just are shifting the demod phase so the two outputs are R and B that can be easily matrixed with Y to get G to drive the CRT with the RGB in the simplest way achievable.

benman94
04-22-2017, 10:34 PM
That's just difference demod: R-Y, B-Y, derive G-Y, mix luma in the gun. I'm pretty positive that X/Z has some difference, but Pete's site and Ed's old site don't say what exactly is different.

old_tv_nut
04-22-2017, 11:47 PM
That's just difference demod: R-Y, B-Y, derive G-Y, mix luma in the gun. I'm pretty positive that X/Z has some difference, but Pete's site and Ed's old site don't say what exactly is different.

As you know, the X-Z demod and matrix circuit that RCA invented and used for many years has color difference output tubes that are connected at the cathodes to matrix R-Y, G-Y and B-Y from the X and Z inputs. Because of the connected cathode circuits, most of the X input shows up in the R-Y output and most of the Z shows up in the B-Y, but some negative Z shows up in the R-Y output and some negative X shows up in the R-Y output. To compensate this, the X and Z angles are adjusted so that the X demod output contains mostly R-Y plus a little B-Y, that is the demodulator angle is between normal R-Y and B-Y angles, but closer to R-Y. Similarly for Z, its angle is moved a bit towards R-Y from B-Y.

The nominal angles to get out pure R-Y and B-Y after the matrix are:
X: 74.8 degrees (instead of 90);
Z: 15.5 degrees (instead of 0).

Pure R-Y and B-Y outputs are actually not optimum for CRT's with sulfide green, and I believe RCA eventually changed the angles and the matrix at some time in later years.

There are more things to know about this circuit:
It is known as low-level demodulation because of the amplification in the matrix section. Sometimes this term is confused with I/Q non-equiband demodulation, which also was always low level. However, X-Z low-level demodulation is equiband, but it does not suffer from the worse bandwidth limitations that are likely to occur with high level (single stage) demodulation (without separate amplification).

A clever thing about the matrix circuit that is never explained in texts:
The matrix circuit has the X and Z signal inputs capacitively coupled. This by itself would be a minor disaster, since AC-coupled color difference signals drift to the opposite background color of whatever is prominent in the picture: e, g, performer enters in red shirt, and the whole picture drifts toward cyan. However, the matrix has negative horizontal blanking pulses coupled into the cathodes. This draws grid current in the three matrix tubes during retrace, making each of them a DC restorer. The circuit is stable not only against scene content, but also against any DC drifts in the X and Z demodulators.

benman94
04-23-2017, 09:46 AM
Wayne,
What exactly would the advtange be to X/Z demod as opposed to demodulating R-Y/B-Y directly. Is it just a cost saving technique? Motorola had their low level diode demodulators set up for R-Y/B-Y; I'm sure other manufacturers used similar arrangements. Would there be a distinct performance advtange over the Motorola method? I feel like I'm missing something fairly obvious here...

Tom9589
04-23-2017, 11:45 AM
Could it be a situation of Motorola not wanting to pay RCA to use an RCA patented circuit of demodulation scheme? Money and pride are strong motivators.

benman94
04-23-2017, 11:51 AM
I was just giving that as am example of a set that used R-Y/B-Y. Westinghouse did with their 15 incher. I think GE used R-Y/B-Y.

I just don't see the reasoning behind the X/Z when R-Y/B-Y worked.

old_tv_nut
04-23-2017, 11:52 AM
For two different low level schemes (both with two stages, demod and matrix/amplifier), with similar gains per stage, the bandwidth advantage will be about the same. It's a matter of overcoming stray capacitances.
However, the RCA circuit has the DC restoration advantage.

Direct high level demodulation is cheaper than having two stages. But any imbalance gets carried right through to the CRT grids, possibly messing up grayscale tracking. in some cases you see balance controls, either a service adjustment, or in some cases a third customer control that can turn the grayscale tracking towards either blue or sepia. If these adjustments are omitted for further cost savings, it may mean that replacing the demod tubes will require a tracking adjustment.

The CTC-5 "Super" chassis has reduced DC coupling in the color difference outputs, which reduces the effect of demodulator offsets, but that means there is also some background shift due to scene content. When everything is set for good general pictures and you put on a test pattern of a full screen red, for example, it gets reproduced with reduced saturation because
1) the demodulator DC balance shifts a bit toward cyan
2) the reduced DC coupling in luma makes the luma output shift upward from the correct 30%

old_tv_nut
04-23-2017, 11:56 AM
Could it be a situation of Motorola not wanting to pay RCA to use an RCA patented circuit of demodulation scheme? Money and pride are strong motivators.

There absolutely were strong efforts to work around the RCA patents, coupled with court fights to make RCA unbundle their patent package - it did no good to find a work-around if you still had to pay for all the patents anyway. Zenith did not start producing color sets until the industry got more favorable license terms.

miniman82
04-24-2017, 02:33 AM
I was just giving that as an example of a set that used R-Y/B-Y. Westinghouse did with their 15 incher. I think GE used R-Y/B-Y. I just don't see the reasoning behind the X/Z when R-Y/B-Y worked.

Keep in mind that just because that's what they chose to call it, that doesn't mean that's what it is strictly speaking. Most of the time the demodulator is angularly close to the actual axes, but not exactly. I think I read somewhere the slight offset reduced crosstalk or noise or something, and wayne is correct that certain things had to change to accommodate the CRT being driven. The sulfide tube mentioned is the best example, since the phosphors used weren't exactly....right for direct R/B demod. Everything aside from the first I/Q chassis was designed with cost savings in mind, which is why we got the tube saving CTC-4 with high level demod.

Most of the RCA books mention that in actual operation viewers couldn't tell the difference between I/Q and other methods anyway, so the cost wasn't justified. Calling later schemes X and Z seems more or less arbitrary, but it was probably that way on purpose. X and Z are not defined in the chroma chart, so the actual demod angles can be whatever you want them to be in practice without being wed to a defined number. Sometimes the angles are called out in literature, sometimes not. Only way to really know for sure is to measure.

old_tv_nut
04-24-2017, 11:39 AM
The offset does nothing to reduce noise or crosstalk, but it does affect color reproduction.

Because the nominal X and Z axes are less than 90 degrees apart, the phase shift can be obtained with a simple RC circuit. In chassis that use an RC instead of a coil quadrature adjustment, no adjustment is needed, and you can actually calculate the phase offset from the schematic values. With no adjustment, there is some variation with parts tolerances, but such an untuned two-passive-component circuit has the least sensitivity to component values of any possible arrangement.

Regarding the effect on color reproduction: spreading the angles to greater than 90 degrees and shifting the over-all phase properly results in a suppression of green and magenta, and is the simplest form of automatic tint control, which was used much later when "one-button color adjustment" was all the rage. It reduced variation in skin-tone tint, but tended to result in the "tan cowboy on a brown horse on the beige range riding into the orange sunset" kind of color reproduction. It is mathematically equivalent to reducing the Q channel gain in an I/Q demodulator.

wa2ise
05-02-2017, 03:02 PM
There absolutely were strong efforts to work around the RCA patents, coupled with court fights to make RCA unbundle their patent package - it did no good to find a work-around if you still had to pay for all the patents anyway. Zenith did not start producing color sets until the industry got more favorable license terms.

What they did do was invent their own patents that improved or made cheaper color TV sets, and after a while RCA, Zenith and others would cross license. Imagine something like a poker game where RCA would lay on the table a patent, Zenith would counter with one of their patents, and so on. "I'll license you if you license me..."

old_tv_nut
05-02-2017, 08:24 PM
What they did do was invent their own patents that improved or made cheaper color TV sets, and after a while RCA, Zenith and others would cross license. Imagine something like a poker game where RCA would lay on the table a patent, Zenith would counter with one of their patents, and so on. "I'll license you if you license me..."

Exactly - two companies would go to court, each with a stack of patents, and they would first argue which were irrelevant and would be thrown out of the contest, and then compare stacks to see who ended up with something that the other company had nothing to trade for.