RPCRT Overscan Reduction - SHIMMING - Page 4

Ah, I think I understand now! So, I should forget about if the focus changes with the screen in its normal position. Once I get it close, I should use the screen-in vs screen-out to determine what changed (no memory needed), and for the final result. Correct?

They may not be perfectly so, as your lenses may not be the best available, but the edges should all wind up at least equivalent to each other when refocused tightly again.


b

Right.


b

This is another great thread you have superleo. I also enjoy watching the screen shots in your other thread "RPCRT Reference Material - Color Decoding". Waiting on more fabulous RPTV screen shots, keep them coming.

I have a question about the string technique. For those that have done it, how did you get the string so straight horizontally? What kind of string are you using. Sounds like another long night for me again.

Mr. Bob. I have a question for you. I have the service menu for my TV but it doesnt mention what the dimensions for the horizontal and vertical spacing. Its a Mits WS-A65. Have you worked on one of these sets? I believe it is similar to a WS-65311.

Sorry about the bad attachment.

I don't use screen templates. I make sure the boxes are all the same size, whatever size that is. It differs when you change the overscan. And that the rows are all the same thickness of box, and same for all the columns.

Of those few who make templates for their sets, NO manufacturer makes o'scan reduced templates. But they used to have some homespun ones for the Mit's over at the SPot. The sizes should have remained the same, no Mit HDready models have changed over the years in size except for the 46" and 48" models. The 55, 65 and 73s have not changed dimensions over the years, since day one.

And I make sure that the circles are all circles.

That's it.


The green takes red and blue with it, tho, and this happens ONLY on a Mit. Get that right and the blue and red have followed, and you're pretty close already.


b

Sounds like a good idea, but doesn’t the horizontal strings sag a bit? Looking at your pics in the beginning of this thread, they looked perfectly straight.

I knew about the red and blue following green when making adjustments. Thanks


That is what I was thinking of using last night after I posted my question.

I measured the boxes and they were all pretty consistent. I then drew it up on AutoCAD. I have access to a plotter at work and was thinking of using Vellum paper to plot out a full size templete. Also with autocad you can change the width of the lines so they dont plot out too big.

Last week I used some left over blank Vellum paper to adjust my focus. It work great!!. Used five shims and clamped them where I measured the screen to be at. Two on each side and one on the top. Then held the paper with tape. Had to be carefull breathing though, the bottom of the paper would move slightly.

Greetings,

How thick of washers are you using? The first time I tried Scheimpflug I used (more or less) standard thickness washers found in my toolbox. That was a mistake. I never ended up with a decent focus pattern using these washers.

The second time I tried I went to our local hardware store and bought the thinnest machine washers they sold. Actually they're more like a C-Clamp. I used those to adjust the Red Scheimpflug and finally got a decent result. Even afterward I didn't get an exact lens angle with equal defocus on either Rt/Lft far edge. In fact before Scheimpflug adjust the worst Red defocus was on the left edge of the screen and after Scheimpflug adjust the worse defocus was on the right edge of the screen, which tells me even the thinnest washers I could find were still too thick... But it was better than not adjusted at all so I took "yes" for an answer and kept the washers in.

Regards

[/url]

[/url]

Large version. Hit F11 to see it all

[/url]

Use the Cantilever Technique and none of that extra work is necessary.


b

I just wanted to say that what Mr Bob told me here was huge for me. After removing the doorstops (no extra piece of wooden paint stirrer needed either), and sliding the array up (higher)--this eliminated all of my red and green scheimpflug issues, which were caused by tilting the array with the doorstops (and leaving them there). Using washers on blue cleared up the scheimpflug issue that was origianally there before the shimming. I did have to use the centering magnets for all 3 (RGB) to lower the picture after removing the doorstops because using VSTA in course green (which had to be cranked way up) caused some weird looking thin lines at the top edge. I checked to see if I needed to adjust astigmatism, but, luckily, it didn't need it. So, once again, Mr Bob saved the day! Thanks so much--I really appreciate all of your help and patience (I'm pretty thickheaded sometimes)!

Also, for anyone using a Mits. template--my understanding is that using the Avia crosshatch pattern with the template is for SD (480) only, and that for HD (1080i) you have to rely on the internal Mits. pattern without using the Avia pattern. Well, I have a hard enough time as it is using both the template and the Avia pattern (I have no idea how Mr Bob can do it by eye), so without the Avia pattern, I was in for a big struggle. But, after trying many things (and failing), I managed to get the new version of PowerDVD to strech the Avia crosshatch pattern perfectly using the "linear stretch" option. EDIT: In case it wasn't clear, this "fix" was for HD (1080i).

I used #10 washers from Home Depot. They actually fit perfectly into the little "hole", which held at least two in position without them going anywhere. I bought a whole bunch, but ended up only needing 3 washers for blue (none stacked--3 different screws), and none for red and green.

hehe. Thanks

I ended up using the template I created in the attachment. But I suggest you not print this one out since converting it to a PDF causes the scale to change.

My lines are now straight .

Mr Bob is just the best.

My HD was fine, it was the 480 that was off for some reason. Not much though. Now its back to where it should be. I tried it by eye since it was only a small section but ended up making it worse. At that point I put up my template and matched up the lines that were correct and fixed the ones that werent.

 

Mit Template.pdf 10.0693359375k . file

Keeping the horizontal plane exactly how it is OOB has caused me no sch issues. The doorstops are ONLY for sliding the array forward and back on the existing plane, just as it is. I don't change any angles.

Nor do I change any VSTA or the centering mags.

VSTA is known to cause problems when you alter it to near its extremes, and with the doorstops you only need to alter it a tiny bit afterwards, not more than 50-100 numbers away from where it was OOB, preferably keeping it to 50 changed numbers or less.

I don't change the centering mags because doing this can change the astig magnet influence and cause you to have to redo that whole astig alignment. This is one thing the manufacturers always get right as rain, if your set is equipped with those astig magnet sets in the first place, and as such I like to leave it totally alone.

So I kept mine as they are and used HSTA to change the positioning of the image on my CRTs for recentering the image on the r and b guns. This forces the HSTA value to around plus and minus 230 respectively - started around plus/minus 50 - but that does not seem to insert any non-linearities between points, so I am leaving that alone. On the Pios you gotta worry when you put things too far out from OOB settings in both Coarse and Fine/point, but evidently that's not a worry on the Mit's. So I am keeping those regs there, even tho the ideal thing to do would be to alter them with HSTA to zero and realign the astig. Just not into doing unneeded alignments these days, too much going on that needs my attention.

In your case, the vertical positioning can be done by sliding your array forward and back until it's just right according to the HD DVE overscan grid, then inserting the doorstops to keep the array there until it can be screwed down tightly, whereupon you can remove them. That's what I did. You get it where you want it with the OOB settings and screw it down, then if you need to fine tweak that vertical positioning later you're so close it won't screw anything up in the sm because it's not going anywhere near any extremes.


Using the internal grid on any of these CRT RPTVs is usually the worst thing you can do for the final supertightening. It's OK for geometry and for getting the point sys close, but their grids are usually too thick and too hot - too super bright - to accurately match up to real world video light level content. As such getting it perfect for the internal grid makes it really messy on real world video.

So I use the Accupel for my supertightening. Its super thin grid is nailed right at 50IRE, where you have the AVERAGE light level of all video material that is being shown at videophile settings of contrast 50 or below.

Both grids show up, but on a Mit you go with the dimmer grid beneath the super monster internal grid whenever you have the choice, which is at each point. The internal grid will show to be off, but who cares. It's the grid I have sent in that counts.

If the only other HD grid I had was the DVE circlehatch, I would dim the contrast down as much as necessary to give me thin, dim lines to use. This can be done on a Mit by setting the User contrast down way low and THEN going into the sm section, where that setting is carried into the sm. All user settings are carried into sm, in the conv sm section.

For other brands I would do whatever it takes to have a nice thin NOT superbright grid to use. If that took lowering my contrast in user, like in the Mits's, fine. On the Pios I'd temporarily alter the sm Contrast reg value. On Hit I'd lower the DCUCONT and DCUBRT to half their hex factory value, in their sm. This lowers the light level of their internal grids, which would allow me to use THEIR internal grid for the supertightening. The internal grids would then show up only half as bright - and thus half as thick - as their OOB sm settings have them show up. Hitachi's user conv grid is excellent when not too bright, and allows for use even in User point alignment only - no need to even go into DCAM! It stays where put even after leaving sm, unlike Mit and Pio's user grids, where any corrections you make in User vaporize upon leaving the sm. Hitachi's user corrections stay.


b

Time is coming to do the shim on my 51F59A....

AS I mentioned before here, my 1st job is to return the F59 to factory oscan - H/V sizing first. Will either use the screen jig I have for that.

Once back to factory, then it's on to shims. I see in Bob's case, 3" did the trick.

Anyone here do this on a 51" Hitachi? Am thinking that I'll go with several smaller pieces of wood that will allow me to slowly build up to the 3" area, checking where it's at before adding more/taking out any shims.

And the "quick'" internal blackening of some black fabric will come last (or just before a final optics cleaning).

My observation of the HD DVE overscan pattern is that I have to do the opposite. At 2.5" at the t/b and 3% at the sides, the circle looks like a true circle and measures out that way via a yardstick. At 3% all around, it looks like an oval that is higher than it is wide.


b

There's no way to really tell without trying. 3" worked for my 73", but only after recapturing the remainder of the CRT face that was not showing when I did it at 2.25". Yours is only 2/3 the size of mine. So in a perfect world, something under 2" should work for you.

But Hit may be differently set up from Mit, on the OOB sizing, so then you could be back to the drawing board...

Add to that if you can recapture the unused portion of your CRT face like I did, that could change everything.

First thing I would do is put up the DVE o'scan pattern and use the h and w trimpots - mark them first of course - to expand out that white pic and see if you can get away with recapturing lost areas of screen face. If your aging footprint is too dark, revealing virgin phosphors out at the edges, then you can't and must go back to the factory o'scan size before doing anything.

If you CAN, if no aging footprint shows up when you expand the DVE pattern out beyond the OOB parameters, then resize it to its maximum potential before you do anything else!


b

Considering the amount of paranoid care I have taken in terms of torch mode & any phosphor burn ready images on the set, I'm assuming going back to "factory" - which in terms of DVE patten o'scan is around 5.5% or so - won't expose anything unwanted.

I'd be resizing the image via Hit's service manual instructions - involves specific green dcam patterns that must be measured out to certain amounts H & V after putting dcam in an uncorrected state (but I imagine you have seen this before ) to be done prior to popping up the screen jig.

I did this service op before after my 1st failed attempt at o'scan reduction - 1st hand knowledge of the mess such a job can make of your image. Used the jig to get it back to OOB geometry (better than, I actually imagine). Entire op took under an hour last time I did it. I figure this will be the job on night #1.

Then, it's off to experiment with shims. Getting the CRT tray loose looks to be the biggest PITA. But others here have done it without much complaint, so we'll see. Those "cones" - forgot about them. They will limit shim size, unless they could be sawed off....but let's not get that crazy.


Solid geometry & convergence, the above mentioned o'scan level & the jump in PQ that getting just more of everything on screen results in is all I seek. And I do expect to deal with at least some focus issues (not worried much based on earlier reports here) as well as the potential need to do/redo some color/grayscale tweaking, get the inside blacked out & clean those optics & mirror.

Should go down at some point next week or 2. Will try & post visual evidence of results.

I'd still encourage you to try to regain the lost areas of pic unused in there if you can, maximizing the usable area of each screen's phosphor face. Meaning if there's no visible aging footprint by now. There's no better time to do this than when you're going to be installing the shims anyway -


b

Meaning increasing o'scan via sm/ H&V pots before shimming?

My set sits @ around 4% now, down a bit from it's factory setting. Did this the "old fashioned" way. I figured I was doing what you mention by getting it back to factory settings beofre shimming. Yes?

There's 2 options, I've done them both.

First is to use the original sizing already on your set from the factory. Shim it and that can stay and your overscan will still be reduced.

The other one is to INcrease the overscan to fill your CRT faces as much as possible, increasing the used area of your phosphors on each gun. THEN shimming.

2 different things, but when joined together rendering a specacular result, better than either one individually. And each one better than reducing the o'scan ONLY by way of the sm/trimpot method, which actually keeps the amount of phosphor area used to display your images the same, and is what you are currently contemplating. Increasing resolution in terms of viewed pixel count, but not achieving nearly the net POTENTIAL resolution increase possible by taking advantage of currently unused phosphor face area. Shimming PLUS size increase on your CRT faces is essentially giving you a bigger set of guns to display with, than you had before.

That's why my set looks so much better now than before.


b

OK - got it.

MY past o'scan reduction was the trimpot/sm method. That's how I got to the current 4% (just a hair under actually). But thi did lead to the slightest amount of line feathering on the right side (mainly). Not super noticable, but I know it's there.

Going back to factory oscan - by increasing size via trimpots - was to be my 1st step. Get that dialed in & tight, then shim.

Now, I wonder if I should take it back to factory o'scan, get it all dialed in, then further increase the size of the image before shimming.

I do know that we are tweaking different sets, but roughly how far out did you go when increasing the size?

edit: Went back to your post #55 here - the shots from thru the lenses. I see what you mean about exapanding the usable image & what to look for/shoot for. That'll be the 1st thing I check out when I get my set open. And I know DVE will be getting a workout.

Check it for the aging footprint first. If you have that and it is noticeble when you expand your DVE o'scan pattern out beyond the factory setting - if the outside portion at the edges is noticeably brighter than the regular section, you gotta observe that by looking down INTO the lenses - then you can't expand your image out beyond where the factory had it initially.

It would show.

I had no such worries. So I went out to where the absolute edge of my raster - which was not going to be visible - was touching up against the face edge. This allowed the portion I wanted to actually be visible on the viewscreen to be what was visible - in my case I was able to go to just under 3% sides, viewed at the viewscreen - and it would stay completely coherent except for the inherent feathering.

Also 2.5% t/b because of the non-linearity of the DVE circle, but again we have plenty of room to move there, which is not the case with the sides.

Meaning that the part where it butted up against the CRT face side edge would NOT show, not even at its most opportunistic. Which you will see when you move your eyes from one extreme of the s/s to the other, which is the only way you can see the whole pic when looking thru the lens. From one end you'll see all of the far side but not the close side, and it will be the opposite on the other, etc.

The edge of the CRT glass curls a bit at the extreme edge. That was my limit. At its most opportunistic, the part that butts up against that face edge will not be incoherent nor show that curling, even tho it's not going to be visible on the viewscreen anyway.


b

Thanks as always Bob.

Should I run into any problems once the mod starts....well, I may have to reactivate the old PayPal account & give you a ring (scheduled in advance, of course!)

OK




b

superleo,

How do you do a screengrab like that?