[Telecine]

I have recently re tubed an ECP 4000 with known good tubes from a 3000. Everything went well with all aspects of the machine lining up with little trouble. There is one problem left and I am unable to determine if is a fault, or a design flaw.


At this point I should say that the projector is level and square to the screen, the rasters have been centered on the phosphor area, and the yokes have been adjusted to match skew to the red tube (which has its original yoke, and therefore considered a reference.


This is a basic geometry problem.


It seems to be impossible to get geometry correct using KEY bow and Pincushion correction only.


The projected image can only ever be correct on three sides, the other having keystone distortion, as if the correction waveform is acting asymmetrically.


Obviously I can correct for this using green only convergence, but the quantities required produce unacceptable banding on the resultant image. (it seems to me that green only convergence is only of use for removing the smallest of errors before banding becomes a problem, and not really helpful in a "real" situation....a design limitation in itself ??)


Are there any adjustments which affect the symmetry of the keystone waveform so that I may cure this problem ? (it is almost as if I require a separate left/right keystone adjustment) or indeed may this be a problem with the Keystone module, or the waveform module itself ?


I am quite keen to resolve this, as in all other respects it is producing stunning pictures (much brighter and sharper than the sony 4200 it was meant to replace (although some of this can be due to the superior glassware in the ecp...)


Any help offered much appreciated....

 

[jcmccorm]

Telecine, there's a keystone adjustment on the waveform module (the one with all the pots facing the back). I think it's R8. It moves the "kink" or pivot point for the keystone. Adjust it and you'll see. This may not be what you're looking for though. I suggest you try the pin-cushion phase pot

which is also on this board, R78 I think.


Someone more knowledgable may have some better info. Just make sure your yokes are lined up right.


Cary

 

[Jay Wilson]

Also, as recommended by M NEWMAN:

After squaring up as much as possible to screen, equalize any top/bottom keystone error so you have same amount on top as you do on bottom by phisically tilting the PJ or twisting defl yokes. This will also help insure that your H centerline is straight too. The reason for this is that the waveform parabola governing this part of the geometry is usually not quite perfect, causing a "coke bottle" effect. There are actually adjustment pots on the waveform bd (toward the edge connector) that can sometimes help get it dialed in better, but its very difficult, requires a scope, a DMM, and a lot of patience, and you're not gonna get it much better.


Anyway, (now that I've taken a COMPLETE tangent - I'm starting to ramble like Ken) after balancing out the keytone error, use a smidgen of green convergence to turn both areas into pincusion errors (in other words, don't fix the problem with green convergence, just make it an error you can tweak with something stable, like, uh, why PINCUSION! of course...). Now exit out and fix it with top/bottom pincusion and you'll have used MUCH less green convergence and the PJ will be that much more stable.


Of course this should be done after a full reset with zero convergence in memory.

 

[Telecine]

Still no joy !


Even after a hard reset, the projected grids from all three tubes converge almost perfectly in the center indicating that the toe ins and raster centering are pretty good. The errors elsewhere in the image are also about uniform for all three colours, which to me suggests that this may not be a yoke problem as this would almost certainly not be the case. I've tried balancing out the errors to spread it evenly l/r and top/bottom and re adjusting pincushion, but whatever I do, it always involves too much green convergence for the end result to have no banding and be stable.


the situation seems to be that with zero correction, everything is symmetrical error wise, and as soon as any correction is dialled in some problems get fixed, and others created ! I realise that these controls are a little interactive, but this is getting silly !!


Key bow and pincushion seem to do exactly what one would expect, except for one corner only. Green convergence seems to be a last resort, and would rather not go there if at all necessary.


Any moe suggestions guys, and thanks for the input so far !!


I must confess I'm at a bit of a loss with this one, but thanks very much for the help so far.


Any more suggestions guys .... ???

 

[stefuel]

Stupid question, but did you make sure your lens mounts were reset back to 1 inch after replacing the tubes? If the green is not set exactly, in the horizontal plane, it would result in uneven keystone and edge focus as soon as you start.


Chip S.

 

[jcmccorm]

Telecine, can you take a picture? If it's just a bit in the corner I wouldn't worry. Heck, I just setup my 4500 (again) and it needed green convergence in the upper left corner. Nothing I could do about it. Blue and Red lined up to it without any trouble.


Cary

 

[Jay Wilson]

Yep, Chip's gpt a good point. It must be exactly 1" on all corners. An dlike Cary said, a picture may help

 

[Telecine]

I must confess, I didn't actually measure the distance you mention. All I did was set the corner and edge wingnuts to the center of their range, then adjusted the reference bolt for sharp center at this setting, then proceeded to do the lens flapping. This resulted in pin sharp focus right across the frame. I take your point about eneven distances causing a magnification change (maybe just in one corner perhaps) but since the focus is pin sharp throughout the frame, I took this as being close enough.


The distance you refer to I presume is the distance from the back of the lens mounting plate to the tube surface (I don't have the manual....)..should this be set for the reference bolt only then proceed as I did previously ?


I will try to take a picture tonight if my digital camera is up to it, but in the meantime, the geometry I end up with (using key, bow and pin only is like this...




....................

. .

. .

. .

. .

.........................


A bit of an exaggerated drawing but the angles top left and bottom right are perfect 90 degrees, and the problem keystone area is ovviously at the screen rhs.


The strange thing is, that after doing a hard reset, and doing toe ins and raster centering only, all three rasters are of similar shape, and perfect sharpness throughout the frame, which I


.

.

.

.

 

[Telecine]

...oops, sent that one a bit prematurely I think !


As I was saying....I dont think perfect focus could be achieved if the mounting distance was that far off. I shall check though, and I'd be delighted to find out I was wrong about this...


Thanks once again for your input, and sorry about the crazy drawing, but you should be able to get the idea !!

 

[Jay Wilson]

Yes, 1" EXACT at all 3 corners with bolts...this will in turn cause the 4th corner to be exactly 1". As fr as exact, get a ruler with 1/32 increments...does that tell you how exact? Don't do just the reference...do all of them. Then, DON"T TOUCH THE REFERENCE ONE! It is from the tube plate to the back of the lens assembly. You may be right in all you assumptions about focus etc, but this is cheap insurance to be sure, and really a necessity for best picture.

 

[Telecine]

Thanks Jay !!


I think I was being a bit too hasty in getting this thing back together !!


I'll check, and reset these distances, and (hopefully) post a picture of the result !


Still seems a little bizarre that the focus is good across the frame though.


Thanks again for your help, I'll let you know what happens !

 

[stefuel]

It is most important to set the ones that had or have the red caps attached. you could make a 1 inch block as a gauge to insure that they are all the same.


Chip S.

 

[Telecine]

All bolts reset to exactly 1 inch gap (which after measurement seems is exactly 15 turns on each bolt....might be useful to remember this....) then lens flapping etc set up again....with no real improvement to the original problem.....worth a try though, and the sharpness is now slightly improved overall, only a minor improvement, but worthwhile !


I had a bit of a brainwave on the way to work this morning....since the rasters are all of the same dimensions and distortions after a hard reset, with the center lines on all three perfectly horizontal, the problem only creeps in when keystone and pincushion correction are applied. To which yoke are these corrections added? If it is to the D.C. centering or convergence yokes, then it seems to me that maybe these may be slightly out of true with respect to the convergence yoke which may give rise to the problem of keystone etc being asymetrically applied....what do you think ?


I also noticed that green convergence doesn't act in quite the same way as the othe colours, ie. a horizontal shift applied at the remote, will give a slightly diagonal on screen shift. I think this may well be the root of the problem, and shall try this this evening.


As always, any input gratefully received !!

 

[Jay Wilson]

The forward most yoke is the deflection yoke and should be pushed forward as far as possible on the neck of the tube. The rear one is convergence and must be alighned with the deflection. This was a mistake I had one mine recently...there is a small "casting" line where the yoke was in the mold. It is barely visible, but can be easilt felt with your fingernail. Get a doog flashlight and you can make it out. Make certain that these lines on both yokes are in perfect alignment, and pointing straight down (for ceiling mounted) and see if that improves anything.


Also, have you pulled ALL the boards and reseated all the IC chips on them? ECPs are notorious for having loose connections on the ICs. Do it one more time even if you've done it already.

 

[Telecine]

Jay, do you have the circuit diagram, and can confirm whether the keystone and pincushion correction waveforms are added to the deflection, or elsewhere. I would have thought that the deflection yoke was the norm, but if not, then my theory may hold some water.


Cheers.

 

[Jay Wilson]

I do have teh schematics, but not with me. I'll check tonight if no one else jumps in before then.

 

[Telecine]

Thanks Jay !!!

 

[jcmccorm]

Keystone and pincushion act on the deflection circuitry. There are two pots on the board near the connector. My 3100/4100 manual says to act on R51 to make the sides straight when 15% keystone is applied.


Cary

 

[Telecine]

Already tried that ! Not quite what I was looking for, but thanks anyway

 

[Telecine]

Already tried that ! Not quite what I was looking for, but thanks anyway