RPCRT Overscan Reduction - SHIMMING - AVS Forum

AVS Forum > Display Devices > Rear Projection Units > RPCRT Overscan Reduction - SHIMMING

Rear Projection Units

superleo's Avatar superleo
01:43 PM Liked: 231
post #1 of 179
03-15-2009 | Posts: 2,171
Joined: Nov 2006
This subject has been very well covered and documented by now. The idea is to bring the projected image closer to the mirror thus reducing the size of the image. By reducing the size of the image we accomplish two main objectives; the first one is the reduction of the inherited RPCRT overscan and the second one is the more concentrated light which will give a sharper better depth perception image.

The creator of this mod goes to Owen. Below is a quote directly from him and his original thread:

"There is a link to my original post on the shimming concept. At the time it was pretty much ignored, I suspect that the idea of modifying a then new or near new HDTV was just too much to contemplate so know one wanted to go there. Almost 5 years on things have changed and many more people are willing to have a go, probably because the financial risk is now minimal. Better late then never.


http://archive2.avsforum.com/avs-vb/...frame+overscan

Here is a link to a more recent post with pictures of my Hitachi.

http://www.avsforum.com/avs-vb/showt...2#post14085492

I posted some mods for the SXRD 18 months ago and got a similar lackluster response from owners obviously too scared to hack into a new TV. Some people even went so far as to suggest my mods where of no value without ever having seen a modified set. I suppose its human nature to dismiss ideas that you are unwilling to pursue, but it’s a disappointing attitude IMHO.

I modify TV’s for my own benefit not for accolades on this forum; I simply post the results of my work so that others may benefit.

Good luck with the shimming mod, an effective lens shield is worth exploring as well."

Also credit needs to be given to Mr. Bob for spreading the word on this mod and making it possible for the rest of us. Here is a link to one of his many shims... http://www.avsforum.com/avs-vb/showt...5#post15843565
superleo's Avatar superleo
01:55 PM Liked: 231
post #2 of 179
03-15-2009 | Posts: 2,171
Joined: Nov 2006
And now to the good stuff. This mod was done to a Hitachi 57F59, so any specifics apply to this model.

This model has the gun array attached to a single tray.



this tray is sandwiched by a board on each side held by screws and by 3 very tricky screws on the lower part of the housing of the array.





The way I remove the tricky screws was, as previously suggested by others, by removing the speakers, the side handles and using a small wrench.







However, after having done the mod, my suggestion would be to take the top boards first and remove the screws from the top instead of going through the speakers. Going through the speakers is quite cramped, but doable. going through the top is alot easier.


superleo's Avatar superleo
01:58 PM Liked: 231
post #3 of 179
03-15-2009 | Posts: 2,171
Joined: Nov 2006
Here are some pics on the array the way is prior to shimming.




superleo's Avatar superleo
02:06 PM Liked: 231
post #4 of 179
03-15-2009 | Posts: 2,171
Joined: Nov 2006
Now to the shimming!!

I ended up using a mixture of 2x4s and paint sticks, taken from LastButNotLeast shimming (great idea). The array tray is placed on top of the sets board, the wood and the paint stick that would gave me approx. 2 3/8". Sorry for the out of focus picture.





Then the boards are put in place. The array needs to be center as much as possible at this time.




superleo's Avatar superleo
02:11 PM Liked: 231
post #5 of 179
03-15-2009 | Posts: 2,171
Joined: Nov 2006
By shimming the array, pretty much the whole picture goes. The main issues are covergence, geometry and mechanical focus. Also, by refocusing your grayscale changes and ofcourse then color changes.

Here is how the set looks just after moving the gun array.


superleo's Avatar superleo
02:13 PM Liked: 231
post #6 of 179
03-15-2009 | Posts: 2,171
Joined: Nov 2006
Since I already had the set open, did deep optic cleaning of the lenses and mirror.




superleo's Avatar superleo
02:17 PM Liked: 231
post #7 of 179
03-15-2009 | Posts: 2,171
Joined: Nov 2006
Then a quick recovergence to at least make the set watchable before geometry and full calibration.

This how It came out after the first pass. Notice that major geo work is required.




superleo's Avatar superleo
02:26 PM Liked: 231
post #8 of 179
03-15-2009 | Posts: 2,171
Joined: Nov 2006
Preparing the set to correct geometry and correct covergence as much as possible to finalize the overscan... LastButNotLeast's string technique was used after trying to wing it with no real positive outcome. Using the string technique is tedious but very effective.

Here is the 57F59 pattern. I adjusted it by 2.5% first but it was too much, making the edges to hard to adjust. Also, you know how the saying goes... "measure twice cut once", well...



This pattern is just so the string can be visible to the camera.



Doing green firs is a must.


superleo's Avatar superleo
02:30 PM Liked: 231
post #9 of 179
03-15-2009 | Posts: 2,171
Joined: Nov 2006
My opinion on this mod is that is well worth it... and worth all the trouble and time. It was alot of fun.

Still have grayscale and color to finish but so far excellent results.







And blue color decoding ... Perfect ... Where are the flashing squares?


superleo's Avatar superleo
02:33 PM Liked: 231
post #10 of 179
03-15-2009 | Posts: 2,171
Joined: Nov 2006
And the final touch up for the inside ....

cover the inside with black cloth





And cover the magic focus wiring with tape to avoid reflection from it.


superleo's Avatar superleo
02:35 PM Liked: 231
post #11 of 179
03-15-2009 | Posts: 2,171
Joined: Nov 2006

superleo's Avatar superleo
08:01 AM Liked: 231
post #12 of 179
03-16-2009 | Posts: 2,171
Joined: Nov 2006
Opinions and comments ... Welcome.
gpounders's Avatar gpounders
10:49 AM Liked: 10
post #13 of 179
03-16-2009 | Posts: 106
Joined: Feb 2002
Quote:
Originally Posted by superleo View Post

Opinions and comments ... Welcome.

Thanks for sharing. I just did this mod to my pioneer and have some question. I believe I have some geometery problems just outside the boundries of visable space but I believe they existed before I did this. There is a lot of geometery work to do, more than I thought but I'm getting the hang of it now. I have some colors that have really bent lines at the outer edges. At first I decided to just resize the image larger to avoid the problems but that appears counter to my objectives and I can't get convergence at the edges.

I was wondering, would it be wise to bring in the size enough to see the bent lines on the outer bounds in attempt to get them as straight as possible first (doing each color independent), then increasing the size to about 3% overscan, then doing the color convergence again?

Any other ideas, I will post some picture when I can

Thanks
jwebb1970's Avatar jwebb1970
01:47 PM Liked: 35
post #14 of 179
03-16-2009 | Posts: 8,352
Joined: Aug 2006
Nice thread, superleo!

Have a "bachellor's week" coming up (wife heading out of town for a work conference). Might be the best time to tackle the shimming process, as well as any other internal modding I might shoot for.

If I can nail it along the lines you appear to have done (2.5 % o'scan), I will be stoked.
gpounders's Avatar gpounders
01:50 PM Liked: 10
post #15 of 179
03-16-2009 | Posts: 106
Joined: Feb 2002
I the circle hatch and the screen shot I see a bit of green. This is a similar problem I'm having but it's red and I think it's caused by the red convergence on this outer edge.
jwebb1970's Avatar jwebb1970
02:08 PM Liked: 35
post #16 of 179
03-16-2009 | Posts: 8,352
Joined: Aug 2006
Quote:
Originally Posted by gpounders View Post

In the circle hatch and the screen shot I see a bit of green. This is a similar problem I'm having but it's red and I think it's caused by the red convergence on this outer edge.

Not to speak for leo, but I do think that his shots were of the "quick & dirty" variety - taken before final touchups.

However, still want to see how it turns out in the end & if leo does address the green stuff you mention. I have seen similar "green edges" after o'scan redux (not the shim variety) before.
superleo's Avatar superleo
02:09 PM Liked: 231
post #17 of 179
03-16-2009 | Posts: 2,171
Joined: Nov 2006
A more realistic number, if you want to nailed it perfectly, and can't put up with any imperfections, is 3%.

I did the mode mainly to reduce the overscan, but I think that the real benefit is the gain in picture quality.
jrcorwin's Avatar jrcorwin
02:13 PM Liked: 10
post #18 of 179
03-16-2009 | Posts: 3,188
Joined: May 2006
Quote:
Originally Posted by superleo View Post

A more realistic number, if you want to nailed it perfectly, and can't put up with any imperfections, is 3%.

I did the mode mainly to reduce the overscan, but I think that the real benefit is the gain in picture quality.

How much of a difference do you think the black fabric made?
superleo's Avatar superleo
02:21 PM Liked: 231
post #19 of 179
03-16-2009 | Posts: 2,171
Joined: Nov 2006
Quote:
Originally Posted by jwebb1970 View Post

Not to speak for leo, but I do think that his shots were of the "quick & dirty" variety - taken before final touchups.

However, still want to see how it turns out in the end & if leo does address the green stuff you mention. I have seen similar "green edges" after o'scan redux (not the shim variety) before.

You are correct, that is not the final pass. That is why I'm saying that the end result is around 3%. If you can live with minor details you might even be able to go less than 2.5%, but if you are like me, you know is there, may be that no one else would notice ... but you know where the little things are.

I haven't finish with the calibration yet. I think a weekend with no distraction will do.
superleo's Avatar superleo
02:24 PM Liked: 231
post #20 of 179
03-16-2009 | Posts: 2,171
Joined: Nov 2006
Quote:
Originally Posted by jrcorwin View Post

How much of a difference do you think the black fabric made?

It doesn't have to be fabric, as long as there is nothing that might cause a reflection. Fabric is the easiest to do and it might also be the most light absorbent (?!)
jrcorwin's Avatar jrcorwin
02:36 PM Liked: 10
post #21 of 179
03-16-2009 | Posts: 3,188
Joined: May 2006
Quote:
Originally Posted by superleo View Post

It doesn't have to be fabric, as long as there is nothing that might cause a reflection. Fabric is the easiest to do and it might also be the most light absorbent (?!)

Well, the inside of my case is black plastic. I'm not sure how light absorbent it is however.
Mr Bob's Avatar Mr Bob
04:50 PM Liked: 63
post #22 of 179
03-16-2009 | Posts: 9,470
Joined: Jan 2000
Awesome expose, Superleo!

I just donned my mad scientist cap again yesterday - it said, "I want you, boy!" and I had to comply...

So I performed all that was necessary to remedy the fact that my red and blue guns had been absolutey MISERABLY aimed by Mit at the design stage. They were both hitting the screen slightly cross-eyed, meaning that most of the pic was at one edge, with a whole bunch of space at the other. I learned years ago directly from Jim Doolittle that you never let your image go thru your lenses at an angle, they need to shoot straight thru or you get aberration of your image.

I know that this makes the same 9" gun array fit both the 73" and 65" Mit versions, and can't wait to see a 65" version again, to see if it's still THAT much off as what I found on my 73".

Be that as it may, the full potential of my CRT faces was NOT being utilized! There was space there that the image could be expanded out to. Since my CRT faces have no visible wear on them yet - I am VERY sensitive to ANY kind of screenburn, and watch for that like a hawk - including the natural amount that comes from 16x9 images being shown on 4x3 guns, I knew I could expand my pic to take in that previously unused face area and get away with it, maximizing the potential of my display. Which is what the shimming op is all about. Reducing the o'scan can be done in the sm, but doesn't increase the amount of the CRT phosphor face being utilized. The shimming op does, which is what gives us the added spaciousness in our images, like the difference between 7" guns and 9" guns aimed at the same size screen. I had optimized it for the most part, but after examining it last week and finding the above results, I realized that it had not been maximized yet.

So I loosened the inner screws on each of the outer guns and released just the outsides of my red and blue guns enough to let them fall back a little bit, angling them correctly, then increased the size on all 3 of the CRT faces, since a whole bunch was now available, that had not been available before.

This also involved adding another thickness of shim to my already triple-pack, for a 4'ple pack. My shim distance is now a full 3", up from the former 2.25", of multiple 3/4" thicknesses. If I had known then what I know now, I woulda used 2 thicknesses of 2x4 on each corner and been done with it.

I will send in pix soon.


For CRT RPTV, unless you can actually do the shimming method, I would not recommend even trying to go below 4%. You might squeak by with 3.5%, but the conv feathering at the edges on your hor lines gets pretty bad after that. With the shimming I have done on mine I can now go to just under 3%, but since you have to have your hor and vert be ID, the entire size is limited by the shortest one.

I have also found that the HD DVE o'scan pattern is slight off. With all edges at exactly the same measure, if you actually take a yardstick to your image on the screen, you'll find that the circle is actually a vertical oval, by about 3/8" on a 73". I therefore have decided to go for 3% at my sides, and 2.5% at t/b, which makes the circle go back to being a true circle.

Perhaps my Mit is deranged, but this is also from the rectangles in my set's internal grid, so I don't think the display is the issue. I think it's the test pattern. We all know his o'scan pattern on the original SD version of VE was worthless for 16x9 geometry, so it's really no big surprise for this version to be slightly this way also. The SD version of VE was completely worthless. The HD version of DVE is almost perfect, with just this slight offness.

If you try to go any lower on CRT RPTV tech, you begin exposing the fact that on one side some of the lines go up while on the other side the same lines go down, as they leave the screen. In this event, the feathering can only be averaged between them, not remedied. If both sides go up and you have a Mit or Pio, you can use the invisible side columns to get it right. But if one goes up while the other goes down, averaging is the only answer.

If that makes 2.35:1 movies with black t/b bars show conv error at the extreme sides, you can mitigate that at those areas by moving the error in towards the middle a bit. This makes the area closer to the center a bit mulchier for 3"-4" in from the side edges, but does not overshow the aberration at the far edges as much.

This kind of thing is why they did overscan in the first place - to hide it.


More later -


Mr Bob
superleo's Avatar superleo
04:58 PM Liked: 231
post #23 of 179
03-16-2009 | Posts: 2,171
Joined: Nov 2006
Quote:
Originally Posted by Mr Bob View Post

Awesome expose, Superleo!

I just donned my mad scientist cap again yesterday - it said, "I want you, boy!" and I had to comply...

So I performed all that was necessary to remedy the fact that my red and blue guns had been absolutey MISERABLY aimed by Mit at the design stage. They were both hitting the screen slightly cross-eyed, meaning that most of the pic was at one edge, with a whole bunch of space at the other. I learned years ago directly from Jim Doolittle that you never let your image go thru your lenses at an angle, they need to shoot straight thru or you get aberration of your image.

I know that this makes the same 9" gun array fit both the 73" and 65" Mit versions, and can't wait to see a 65" version again, to see if it's still THAT much off as what I found on my 73".

Be that as it may, the full potential of my CRT faces was NOT being utilized! There was space there that the image could be expanded out to. Since my CRT faces have no visible wear on them yet - I am VERY sensitive to ANY kind of screenburn, and watch for that like a hawk - including the natural amount that comes from 16x9 images being shown on 4x3 guns, I knew I could expand my pic to take in that previously unused face area and get away with it, maximizing the potential of my display. I had optimized it for the most part, but after examining it last week and finding the above results, I realized that it had not been maximized.


So I released just the outsides of my red and blue guns enough to let them fall back a little bit, angling them correctly, then increased the size on all 3 of the CRT faces, since a whole bunch was now available, that was not available before.

I will send in pix soon.

For CRT RPTV, unless you can actually do the shimming method, I would not recommend even trying to go below 4%. You might squeak by with 3.5%, but the conv feathering at the edges on you hor lines gets pretty bad after that. With the shimming I have done on mine I can now go to just under 3%, but since you have to have your hor and vert be ID, the entire size is limited by the shortest one.

I have also found that the HD DVE o'scan pattern is slight off. If you actually take a yardstick to your image on the screen, you'll find that the circle is actually a vertical oval, but about 3/8" on a 73". I therefore have decided to go for 3% at my sides, and 2.5% at t/b, which makes the circle go back to being a true circle.

Perhaps my Mit is deranged, but this is also from the rectangles in my set's internal grid, so I don't think the display is the issue. I think it's the test pattern. We all know his o'scan pattern on the original SD version of VE was worthless for 16x9 geometry, so it's really no big surprise for this version to be slightly this way also. The SD version of VE was completely worthless. The HD version of DVE is almost perfect, with just this slight offness.

More later -


Mr Bob

Coool!!!!! more stuff to do to the set. Let us know.
juancmjr's Avatar juancmjr
05:22 PM Liked: 10
post #24 of 179
03-16-2009 | Posts: 893
Joined: Dec 2006
I'll have to do this to my Panasonic one day, including an optics cleaning and maybe even a full calibration by a certain Mr. Don't dump your CRT RPTV. I'll start collecting every penny I find...

Doing a shimming isn't that hard is it?
Angelo M's Avatar Angelo M
05:26 PM Liked: 10
post #25 of 179
03-16-2009 | Posts: 365
Joined: Feb 2003
Superleo, very nicely done report!!!, and thanks for sharing.


Mr Bob, it just amazes me how you are able to refine these tweaks the way you do, and look for improvements! This is just like the programs monster cars/ trucks, and Pimp my Ride!


I have a question about shimming again, but still havent the nerve to try it.
Can you run out of optical focus with these lens with the shimming? Or you would be beyond 0% overscan before that happens?
Mr Bob's Avatar Mr Bob
06:20 PM Liked: 63
post #26 of 179
03-16-2009 | Posts: 9,470
Joined: Jan 2000
Quote:
Originally Posted by Angelo M View Post

Superleo, very nicely done report!!!, and thanks for sharing.


Mr Bob, it just amazes me how you are able to refine these tweaks the way you do, and look for improvements! This is just like the programs monster cars/ trucks, and Pimp my Ride!


I have a question about shimming again, but still havent the nerve to try it.
Can you run out of optical focus with these lens with the shimming? Or you would be beyond 0% overscan before that happens?

Thanks! Come on over and watch my pimped out display sometime! Love to have you observe it and write in about it.




On my 73" Mit my shimming difference is only 3", on a folded throw distance of maybe 5'. It's not going to upset ANY optical focusing, you won't come anywhere near to running out of room to move on your lens barrels. I changed my lenses when refocusing by maybe a MAX of half an inch, edge to edge, barrel to turret, on the turning of the lenses, which probably has at least 2 inches of total turning capacity.

I have never come close to running out.


b
Mr Bob's Avatar Mr Bob
06:27 PM Liked: 63
post #27 of 179
03-16-2009 | Posts: 9,470
Joined: Jan 2000
Quote:
Originally Posted by juancmjr View Post

I'll have to do this to my Panasonic one day, including an optics cleaning and maybe even a full calibration by a certain Mr. Don't dump your CRT RPTV. I'll start collecting every penny I find...

Doing a shimming isn't that hard is it?

You have to be ready for anything. On one I did the whole thing only to find that the aging footprint was too great and the difference in centering between the 2 sources too great, resulting in too much showing in too many places. It resulted in having to dismantle the whole thing and put it right back where I started.

It's not rocket science, but it does take a WHOLE lot of tenacity!


b
superleo's Avatar superleo
08:14 PM Liked: 231
post #28 of 179
03-16-2009 | Posts: 2,171
Joined: Nov 2006
Quote:
Originally Posted by Angelo M View Post

Superleo, very nicely done report!!!, and thanks for sharing.

My pleasure... I hope this serves others to do it. And if you have any major trouble or need help, Master Mr. Bob is available.
gpounders's Avatar gpounders
09:28 PM Liked: 10
post #29 of 179
03-16-2009 | Posts: 106
Joined: Feb 2002
Quote:
Originally Posted by Mr Bob View Post


For CRT RPTV, unless you can actually do the shimming method, I would not recommend even trying to go below 4%. You might squeak by with 3.5%, but the conv feathering at the edges on your hor lines gets pretty bad after that. With the shimming I have done on mine I can now go to just under 3%, but since you have to have your hor and vert be ID, the entire size is limited by the shortest one.

I have also found that the HD DVE o'scan pattern is slight off. With all edges at exactly the same measure, if you actually take a yardstick to your image on the screen, you'll find that the circle is actually a vertical oval, by about 3/8" on a 73". I therefore have decided to go for 3% at my sides, and 2.5% at t/b, which makes the circle go back to being a true circle.

Perhaps my Mit is deranged, but this is also from the rectangles in my set's internal grid, so I don't think the display is the issue. I think it's the test pattern. We all know his o'scan pattern on the original SD version of VE was worthless for 16x9 geometry, so it's really no big surprise for this version to be slightly this way also. The SD version of VE was completely worthless. The HD version of DVE is almost perfect, with just this slight offness.

If you try to go any lower on CRT RPTV tech, you begin exposing the fact that on one side some of the lines go up while on the other side the same lines go down, as they leave the screen. In this event, the feathering can only be averaged between them, not remedied. If both sides go up and you have a Mit or Pio, you can use the invisible side columns to get it right. But if one goes up while the other goes down, averaging is the only answer.

If that makes 2.35:1 movies with black t/b bars show conv error at the extreme sides, you can mitigate that at those areas by moving the error in towards the middle a bit. This makes the area closer to the center a bit mulchier for 3"-4" in from the side edges, but does not overshow the aberration at the far edges as much.

This kind of thing is why they did overscan in the first place - to hide it.


More later -


Mr Bob

Feathering, I think that is exactly the problem I'm seeing with my Pioneer. I have invisable columns on the top and right side but not the bottom and left. Also, I may have to live with 6%. After the shimming mod, assuming the geometery was correct befor that when I change the horizontal size I need to change the vertical size proportionately? Can I just calculate the percentages and work out the fine convergence afterwards to get better convergence?
Mr Bob's Avatar Mr Bob
09:31 PM Liked: 63
post #30 of 179
03-16-2009 | Posts: 9,470
Joined: Jan 2000
Quote:
Originally Posted by gpounders View Post

Feathering, I think that is exactly the problem I'm seeing with my Pioneer. I have invisable columns on the top and right side but not the bottom and left. Also, I may have to live with 6%.

The right side affects both the right and left. But in a crazy kind of way. I have a system that I discovered that allows it to be dialed in just right, but it took me years to figure out...


b

Reply Rear Projection Units

Subscribe to this Thread

Powered by vBadvanced CMPS v3.2.3