The originals were 600-volt rated paper caps, which I replaced during the first recapping pass. C64 is .035 and C65 is .05, which look like pretty standard values for those caps in my old TV books.

I replaced them with modern film caps (the brown chiclet style) rated for 630V. I paralleled two caps to make up the .035 cap.

I guess it's possible that one of the replacement caps is bad or that I (shudder!) installed a wrong value. I can disconnect and test them.

Phil Nelson
Phil's Old Radios
https://antiqueradio.org/index.html

I have less confidence that they are the problem, since you replaced them, but it doesn't hurt to double check. There seem to be two things going on: left and right not the same width (usual sort of linearity problem), and sides much wider than the center. I don't understand the circuit well enough to point to a possible cause or causes.

Did you replace C-76 and C-77?

C76/C77 are 1000-mfd/15V electrolytics, which I replaced with new 25V-rated electrolytics. All paper & electrolytic caps have been replaced.

Phil Nelson

On other sets of the day that were based on the RCA 630 those caps on the linearity coil were 1000V IIRC. Are you sure the 600v ones you found in it were original?

They are specified as 600V in both the Sams and Riders parts lists, for what that’s worth.

Phil Nelson

OK, they are just huge filters to bypass all AC. I was guessing that, but wasn't sure if they could be smaller values that could shape the waveform. Never mind.

In the damper circuit, there is a 25-watt wirewound filter resistor (R86), which turned out to be open!

https://antiqueradio.org/art/Emerson...rSchematic.jpg

It's one of those guys that stands upright inside the high voltage cage.

https://antiqueradio.org/art/Emerson...esistorBad.jpg

https://antiqueradio.org/art/Emerson...esistorNew.jpg

Replacing it cured the worst horizontal geometry problems, but then the width was insufficient. Earlier in the project, I had removed a .25 paper cap that someone tacked in parallel with the width coil (L8).

https://antiqueradio.org/art/Emerson...AddedCap02.jpg

I experimented with a decade box and found that .25 was indeed about the right value to bring the width back into range.

https://antiqueradio.org/art/Emerson...Experiment.jpg

I guess the guy who added that cap was onto something :)

There's more tweaking to do -- there is still a very slight crushing of width on both far edges -- but I have a watchable picture for the first time in this project.

https://antiqueradio.org/art/Emerson...orReplaced.jpg

Yay!

Phil Nelson
Phil's Old Radios
https://antiqueradio.org/index.html

Hi Phil,
Your picture looks really good now.:thmbsp:
Ed

I've been watching this thread with great interest. Very well done. Just goes to show "never give up". Working rear projection sets would have to be as rare as rocking horse poo.

Very nice work. Interesting to see how these sets performed. They must've really impressed people back then with their, for the time, huge picture.

:D:banana::banana:

Phil you da man.

Wow, lookin good! :thmbsp::thmbsp::thmbsp:

Do you intend to experiment with other values of R-86 to address the slight crushing that you observe?

jr

Always great when black magic turns into problem solved!
Looking very watchable!

Gotta 'member that one. :lmao:

Unfortunately, I don't have an assortment of 25-watt resistors on hand. I had to mail-order the replacement that I used.

The symmetrical edge crushing is slight, but noticeable with a test pattern. In a crosshatch pattern, the edge-most two or three columns of squares are narrower than the rest.

Any guess as to what range of values might be reasonable to experiment with? The stock resistor is 7.5K. One local store sells 25-watt resistors in values as high as 5.6K. For an experiment, I can wire resistors in series to make other values, of course.

Phil Nelson
Phil's Old Radios
https://antiqueradio.org/index.html

Does the .25 cap affect both overall size and the edge squeeze? If I follow what's happening, will you need a joint adjustment of the cap and the resistor for best overall result?

I didn't keep a library of screen photos while testing, but I believe changing the .25 cap value did affect the squeeze somewhat, to a lesser extent than the width.

Experimenting with both the cap and resistor values is a possibility, if I get a little assortment of 25-watt resistors to play with.

Phil Nelson

For a quick test you could bridge the 7.5k with a much lower wattage resistor, perhaps starting with something like 20 to 50k and observe the change (if any). Also a lower wattage resistor, something like 500 to 1000 ohms could be added in series for a quick test. :scratch2:

jr

You can't do it with the high wattage components, but if it were lower power, simultaneous adjustment with both a capacitor decade box and a resistor decade box could make short work of a problem like this.

This has been deprecated in the past as "dec-box engineering" because there is no pre-calculation of the correct values. But it need not be looked down on if two things are true:
1) You already have a reasonable starting point
2) You are observant enough to understand which aspect of performance is changing as you make the adjustment.

In a hypothetical example, you might note that a small resistor change modifies overall width (and is it larger or smaller) without much change in edge compression, while the cap affects both width and compression, and adjust first the cap in a small step to get better edges, then the resistor in a compensating small step to get the right width, and so on.

NOTE: you also have to be aware whether the transient condition of an open circuit while switching the dec box could cause a destructive fault. In your case, you already know nothing is being destroyed even with the components open-circuited.

Phil, it might be worth trying the old "seven seconds off & back on" test. Does the raster come back at full width, or is it slow filling out? If unusually slow, are there any peculiarities with the lin during fill-out?

This is interesting to me. Does slow recovery point to weak tubes or other things as well?

Exactly. It's an old 'quick & dirty' field test for borderline-weak tubes like hor osc and 3.58 osc. (slow achieving sync), hor. out (slow fill-out), vert. out (slow fill-out), RF amp (slow clearing snow), etc.

Same principle as the Life Test on a tube tester which drops the heater voltage down.

This was followed by the 'tap test' of each tube to show up any arcing (like the damper), intermittent shorts etc.

Made service calls more effective with less chance of call-backs.

OK, I'll try the seven-second test later today and observe the results.

I spent some time yesterday playing with different values for R86, using a handful of 25-watt resistors that I picked up locally. It is a cumbersome and very slow process, to clip a resistor (or series-ed resistors) into place, slide the chassis back into position, turn the set back on, observe the result, and try to compare the pattern to what you saw before. I tried taking photos with the camera on a tripod, and noting the resistor values for later comparison.

Subbing the capacitor values was so much easier with a decade box, where I could keep my eyes on the screen and flip between different values, back and forth, looking at different parts of the screen, etc.

Anyway, yesterday's results were not very informative. When I put a crosshatch pattern back on, I realized that the linearity wasn't quite as good as I remembered.

I also realized that I had begun with the TV in a somewhat random state -- that is, with the various controls set to produce an image with best horizontal linearity & centering. I had also left in the .22 cap across the width coil. What kind of baseline is that? Perhaps it would make more sense to remove that .22 cap and center ALL of the pertinent controls, then repeat the rigamarole of subbing the cap and resistor values, looking for changes that make sense.

There are lots of variables in this equation. In addition to the horiz width and linearity adjusters, you have the horiz centering and horiz drive adjusters, all of which affect width and linearity to some extent. Not to mention the mechanical centering adjusters on the optical box.

Does anyone have a clear idea what changes you would expect to see on a perfectly-linear set when changing the value of this damper resistor (R86)?

Phil Nelson

On a perfectly linear set no. I did mess with that and parallel capacitance on a Hoffman that had width and linearity issues a year or two ago.

Some sets that used that big damper resistor (RCA 630?) had taps on the resistor to rough in one of the horizontal adjustments (forget if it was the width or lin) so the control of that name would be in the correct range to work.

Yes, now I recall seeing a tapped damper resistor in one of my old TV books.

The 630TS factory manual says this is a coarse linearity adjustment:

"If repeated adjustments of drive, width, and linearity fail to give proper linearity, it may be necessary to move the tap on R209, which is located in the high voltage compartment. Adjustments of drive, width, and linearity must then be repeated."

The parts list describes R209 as:

"Wire wound resistor, 5300 ohms, 20 watt; 500 ohms, 2 watt; and 500 ohms, 2 watt."

Not sure what to make of that description. Why only 2 watts for the 500-ohm segments?

In the schematic, that resistor is drawn with three taps, and the default connection made at the center tap. I interpreted this to mean it's a 5300-ohm resistor, which you can adjust to either 5800 or 4800 ohms by moving the tap. Am I guessing wrong about the resistor construction?

Phil Nelson

Assuming a constant current through the resistor, a 500 ohm section will have roughly 1/10 the voltage drop across it as the 5300 ohm section, hence 1/10 the wattage (actually it'll be slightly less than 1/10).

(If the 5300 were 5000, the figures would be exact.)

[QUOTE=
The parts list describes R209 as:

"Wire wound resistor, 5300 ohms, 20 watt; 500 ohms, 2 watt; and 500 ohms, 2 watt."

Not sure what to make of that description. Why only 2 watts for the 500-ohm segments?
[/QUOTE]

Power in each segment uses the formula P=(I^2)R. Since the resistor section are in series, the same current would go through all resistor sections. The 500 ohm sections are approximately 1/10 the value of the 5300 ohm section; hence, their power handling capacity would need to be only 1/10 the wattage of the 5300 ohm section.

I think it means the first section is 5300, in series with a 500 and another 500, so total is 6300. Power is I squared times R, so if the current is approximately the same magnitude for all settings, the smaller resistors in the series string need proportionally smaller wattage. So, one tenth the R --> one tenth the watts.

Ah, another forehead-slap moment. Lower resistance means lower wattage, duh.

I reviewed the photos that I took while experimenting, and it's clear that increasing the resistance stretches the left half of the image, while decreasing resistance squeezes the left half.

Changing the resistance also changes the overall width somewhat. And it does not really cure the non-linearity I'm seeing. The far left edge is still squeezed, etc. And there is an additional variable -- the value of the .22 cap added to increase the overall width.

Finding the best setting looks like a very iterative process:

-- Try new resistor value
-- Adjust the width, horizontal linearity, and horizontal drive controls for best linearity
-- Try another resistor value
-- Adjust all the etc.
-- Repeat, repeat (maybe also adjusting the value of the .22 width-coil cap to see how that effects everything)

That would be simpler with a direct-view TV, which you could plop onto the workbench and easily view while adjusting. Not so much fun when you are crouching on the floor to make adjustments and peering around to view the results in a little serviceman's mirror on the other side of the cabinet.

OK, enough whining already! I hope to make some progress this weekend.

Phil Nelson

P.S. I did try the seven-second test suggested by old_coot88. The raster comes back at full width. You can hear that the HV spools up slightly after everything else, so there's a brief interval with no raster. But when it reappears, the image is at its full width.

My favorite way to think of it is if you want to replace a 5000 ohm 10W with a series pair of 2500 ohm units those need to be 5W a piece, if you want to replace with 4 1250 ohm 2.5W a piece, 8 625 ohm units at 1.25W each and on down*...If your changing the main resistance only slightly things should be safe doing the calculations based on the present current/voltage going through it, but if you change the resistance a lot the current flow may change enough to require recalculation of wattage.
*My college only stocked 1/2W resistors so when the maths for a circuit design called for a 10W+ resistor that needed to last 30min of various measurements and tests without burning up we had to get creative with series-parallel combos to get a safe wattage...And the professors and parts room folks always wondered why there were so many off tolerance resistors in stock. :screwy:

What kind of connectors does the main chassis use to hook into the cabinet projection system?... If it is just a set of octal and or Fat-pin tube base/socket connectors (and maybe a CRT socket/HV lead) why not make 5'-12' extender cables so you can keep the cabinet next to the bench and have the chassis on the bench connected to the cabinet for easy adjustment and testing....I did just that when I was working on my 21CT55 to make things easier for me.

Extender cables would be handy, although not super-easy.

Two of the three cables would be do-able. The yoke connector is 9-pin, roughly the size of an octal socket with a ninth pin near the center. A second cable goes to the HV power chassis with a 4-pin connector. I don't have such connectors on hand, but I suppose you could scrounge them somewhere.

The CRT cable looks tricky. It uses a funky 5-pin connector of a type I haven't seen before. The cup has terminals on the inside and it slides over the butt end of the CRT, which has little metal tabs around its edge.

https://antiqueradio.org/art/Emerson...TConnector.jpg

Weird,wild stuff, as Johnny Carson used to say.

You wouldn't need an extender for the HV cable to the CRT second anode, since that doesn't come from the main chassis. It only goes between the two boxes of the Protelgram unit (HV power chassis and optical box holding the CRT).

Phil Nelson
Phil's Old Radios
https://antiqueradio.org/index.html

CRT base appears to be a euro tube connector...I once was at a swapmeet where someone had some really old Brittish or German tubes with that style of connector...Some had a different number of 'pins'. I had the same impression when I saw them there for the first time...

In your shoes if I did not much like the set, badly wanted/needed service convenience, or had a situation where the wiring to the interconnects was damaged or otherwise unoriginal I'd be tempted to cut the harnesses and insert some other connector that I could easily make (or use an existent) extender for...In any case it's your set, your back/knees, your call.

I think probably patented in the Netherlands by Philips. I have one protelgram projection unit made by Mullard, much more robust and heavy duty. The optics are claimed to be f 0.62. Now THAT's FAST !!!

Hello Phil,
Great to see your set working that well:banana:
I'm 99% sure that the CRT socket is an Au74 like it's used on an AB2 tube for example. In German language this socket is called Außenkontaktsockel.

Thanks, that does look like the tube base. The Phillips 3NP4 data sheet is at:

https://frank.pocnet.net/sheets/201/3/3NP4.pdf

In an effort to eliminate more variables, I tried injecting video and audio directly at the amplifiers, bypassing the tuner and IF stages. The results were good, so I wired in cables and made a little junction box where I can plug in A/V cables from a DVD player, etc.

https://antiqueradio.org/art/Emerson...tionTrial1.jpg

The picture looks better than before, and the audio is outstanding:

https://antiqueradio.org/art/Emerson...tionResult.jpg

More details in my article at:

https://antiqueradio.org/Emerson609P....htm#Injection

The screen geometry is approaching normal. The picture is still somewhat off-center horizontally, a condition I hope to cure by raising the optical box using its bottom "tilt" adjusters.

Phil Nelson
Phil's Old Radios
https://antiqueradio.org/index.html

There is something unusually appropriate about how a B&W movie looks on a period projection...TV. :D :thmbsp:

I agree Tom. I think the key word is "projection". What we see is an intensification of the brightness in the center if the screen and less in the outer regions. I would think in both the movie theater and the projection TV you have this effect to some degree due to the use of projection lenses taking a small image and projecting it magnified onto a screen.

In any event I love the way it looks.

OK, I'm declaring this project finished. The picture is very watchable and I got around to touching up the cabinet:

https://antiqueradio.org/art/Emerson...onFinished.jpg

This article provides many details:

https://antiqueradio.org/art/Emerson...onFinished.jpg

One question in parting: why did Emerson use a pattern of tiny concentric grooves in the face of the display screen? In this extreme close-up, I heightened the contrast to exaggerate the circular pattern, which is imperceptible at a normal viewing distance:

https://antiqueradio.org/art/Emerson...ernCloseup.jpg

I have a couple of guesses (see article above), but optics is not my strong suit, and I haven't restored any other projection sets, so if you know the answer, kindly clue me in!

Regards,

Phil Nelson
Phil's Old Radios
https://antiqueradio.org/index.html

Fresnel lens. Used to magnify the picture.

Awesome! For the first time since I started collecting these sets you've changed my mind about projection TV's. I had no idea the picture could be that nice. Often you see them not restored as well as yours and the picture is rather muddy.