So I decided that it would be a good idea to completely replicate the DVE geometry pattern on my screen using black thread. I then converged the green over the crosshatch pattern I made using lots of balance controls and some point. It was fun indeed. I still need to touch up a few spots with point and get a better angle for the screenshot, but I think it looks pretty good. Yea, I'm kind of anal about geometry :D.

http://www.avsforum.com/avs-vb/attachment.php?attachmentid=95105&d=1195961927

http://www.avsforum.com/avs-vb/attachment.php?attachmentid=95106&d=1195961927

Man, you are one square dude

Interesting way of doing a setup. An alternative is to use laser levels.

Another alternative is to use a digital to run test patterns. Digitals may have lousy black level but they're usually pretty tight on geometry. Negative is that if you want to get really slam on you'd need a 1080p digital to do the patterns - but if you were anal enough you could manage it with XGA or so. I've done it a couple of times for rough setups, and it makes life a lot easier to not have to carefully do physical center lines and squares.

It also makes you feel warm and fuzzy to wave your hand in front of the digital but not the CRT and see the black levels come back... :D

pretty darn good dude ;) I used the good old fashioned string method with a level and tape measure, here is my 1352 (camera distortion ruins the shot some on the perimeter):

http://home.bigsandybb.com/gmurrell/theater/screenshotgeometery.jpg

I honestly don't know how anyone would get as perfect a geometery without the controls available only on the XG projectors, no way it could be done without them, maybe 99% perfect but not 100% :p

-Gary

Crikey.

Mine's crap compared to that Jeff.

Screen shots that serve a useful purpose. I like the idea of a 1080p digital test pattern projector. Its mariages like that that will eventually eliminate projector typism in America.

Why bother?

It's impossible to tell if geometry is off, not unless it's off by a mile.

Deron.

Why bother? WHY BOTHER?

Its true, perfecrt geometry really accomplishesnothing. Do some research about CRT set ups. In the old days, geometry was the least important of any set up.


BUT. Every projector has a built in grid generator and anybody who looks at it can see geometric errors and can access the controls to fix them. It gives them something to do and something to accomplish. Its harmless except if banding is introduced. Keeps you off the streets. :)

pretty darn good dude ;) I used the good old fashioned string method with a level and tape measure, here is my 1352 (camera distortion ruins the shot some on the perimeter):

I honestly don't know how anyone would get as perfect a geometery without the controls available only on the XG projectors, no way it could be done without them, maybe 99% perfect but not 100% :p

-Gary


Yep. All the balance controls were very very useful. Even with them I couldn't get the right setting between linearity and linearity balance and I had to use point on a lot of spots.

Why bother?

It's impossible to tell if geometry is off, not unless it's off by a mile.

Deron.

I'm anal about it. If I'm gonna spend around an hour setting the geometry by eyes, I might as well spend two hours and get it perfect. But yea, it's not like I'm going to be watching TV and think "Wow look at that geometry!" :D.

Interesting way of doing a setup. An alternative is to use laser levels.

Agreed - a laser level with horizontal and vertical reference lines works great.

Delete point. It will cause banding. Doug

WTF? :(

anyone try to watch film credits on a PJ with bad geometery? our eyes expect perfect geometery and when it isn't near it things look screwed up, pans in films are easiest to spot it on

-Gary

http://home.new.rr.com/arseneau/HT%20images/seagull.jpg

I have a bit of seagull on my geometry that I can't get rid of. Is there a secret to this? Am I trying to run by raster too close to the edge?

I currently run at 1280 x 720 @ 60Hz.
XG 135 AC
124" distance from 92" wide screen (image area is 92")

Looks like my G70. I haven't tried real hard to fix it but I understand it's normal/common on G70's. Not on XG's though.

http://home.new.rr.com/arseneau/HT%20images/seagull.jpg

I have a bit of seagull on my geometry that I can't get rid of. Is there a secret to this? Am I trying to run by raster too close to the edge?

I currently run at 1280 x 720 @ 60Hz.
XG 135 AC
124" distance from 92" wide screen (image area is 92")
Me too.
If there are any professionals(Doug, of course) who could post some advice about getting rid of seagull, it would be appreciated.
I know the XG has a control to address it, but personally, I've never seen it adequately deal with it in a real world setup...

WTF? :(

anyone try to watch film credits on a PJ with bad geometery?

-Gary

I like good geometry, but could care less what the credits look like.

For a lot of commercial applications "perfect geometry" is a must. For HT, the best requirement should be the 5 circles (one each corner and one in the center). That and the four borders being straight, there's really no need to waist time going after something that you'll not have a problem with in most of what we watch... unless you're also into the credits.

Me too.
If there are any professionals(Doug, of course) who could post some advice about getting rid of seagull, it would be appreciated.
I know the XG has a control to address it, but personally, I've never seen it adequately deal with it in a real world setup...

Yea, I believe the "line distortion" control on the NEC takes care of this. I still have some banding on the left edge that's always there no matter what - even on the last screen/physical setup. I can move the raster around and change image shift and size and it's still there. Probably something with my source then?

I like good geometry, but could care less what the credits look like.

For a lot of commercial applications "perfect geometry" is a must. For HT, the best requirement should be the 5 circles (one each corner and one in the center). That and the four borders being straight, there's really no need to waist time going after something that you'll not have a problem with in most of what we watch... unless you're also into the credits.

I need to work on my corners/edges - geometry and focus. It's hard when you go close with the NEC because the line to the left of the right edge will be too far to the right, yet the edge is too far to the left. Same situation with the other edge. This is why perfect linearity was impossible to achieve without significant point, which caused banding as master Doug predicted :p. I guess the Marquees don't band with point because Mike uses it a lot, but then again they band with the bow control :D.

Interesting way of doing a setup. An alternative is to use laser levels.

I'll try this idea. Very cool tip! :)

Jeff,
Banding:
What you are talking about is on the side of the image. I know its been called banding but lets just say for now its timing 'harmonics'. This is different then the banding I'm talking about that will run across the image usually in the top portion of the image and be a band you can see best on a all white test pattern (appear darker) or in sky shots of the actual image. This is usually from convergence and point being over used. Point uses memory up much faster and is the leading cause.

When using point to correct for seagull in the top and bottom of the image it can bring on banding quite fast. Jerry called it right this in 'old-school' was called seagulling.

Seagull:
You can correct line distortion center and edge but you also have to start at the center and make sure bow is straight and your not fighting pincushion or pin balance. Many things to consider as well as the internal test patterns being 10% larger and over-scanned. Line distortion at best will get seagull balanced but this will usually still have a rise at center and the outside edges. Considering the yokes and tube necks are round it will have some error making a rectangle image. It does seem to show a bit more on the G70 and XGs but they both use the same tube.

It takes a bit of patience using a external test pattern or Avia then find a medium that works best. If using 16:9 run a 2:35 format movie checking for straightness across the image. Its hard to explain but easier for me to correct the actual image working from the center out without using point. That and I'm able to use my signal generator w/fine cross-hatch not upconverted using the exact freq.

None74, yes real world I understand. Its always a bit different between chassis / raster / sources and signal chain. Some magnetics going on also, has to be when you have metal, earth core even just a little bit.

Last have you ever put a cross hatch test pattern on a regular 4:3 TV before or took a look at how the 16:9 image looked on it? Most people don't think about it or see it but it bugs the heck out of me. Doug

Jeff/Doug

The dark/light bands running vertically on the LH edge are typically called "Ringing" here at least.

But yep, caused by the beam changing speed as it settles after retrace - ringing in the yokes. TSE explained it really well recently. Beam slows down, more energy hits phosphur = brighter, and vice versa.

Bigger porches on the LH side can fix it, or at least help. It gets worse if you expand the raster...one of the tradeoffs..

Mark,
Yes thanks, way back in the Guy Kuo days we worked on this and finally off set the raster and changed porches. 1080I was just coming into play. TSE explained it very well. Doug

Sounds like a big headache :). It doesn't bother me that much... tweaking phase and some convergence settings toned it down a bit.

On another topic, how can I fix hotspotting on the blue tube? It's because the center is defocused, but the edges remain very focused, almost as if the blue focus tracking was only applied to the center. This makes the center whites bluish and the whites on the edges very yellowish.

Jeff,
You will have to redo blue starting with B defocus set to off.. this would be from the start of mechanical on blue including wave board pot. Its either that or you blew the blue F drive channel. When you get it done then turn on B defocus. Might also check your connection at the F yoke. Doug

Jeff,
You will have to redo blue starting with B defocus set to off.. this would be from the start of mechanical on blue including wave board pot. Its either that or you blew the blue F drive channel. When you get it done then turn on B defocus. Might also check your connection at the F yoke. Doug

I just did a complete mechanical setup and convergence this weekend with B defocus set to off. Wave board pot? I tried to defocus the blue with the edge and corner controls, but there was no effect when b defocus was on.

There are three focus pots on the wave board in the card cage at the back Jeff - labelled for colour.

You set the menu driven focus to zero, and set for best focus using the pots. The menu focus is just a fine control on top of the coarse focus pot.

Jeff,
Set blue defocus to off. Move all center, edge and corner electronic blue focus to mid curser on screen from the remote. Redo the two lenses for optimum. With center focus at mid point adjust the wave board pot marked blue (B) on the metal cover of the card cage top. Make this optimum focus at the center of the screen only.

Recheck lenses then go through electronic focus again center, edge and corner. Also go over electronic astig.

With that completed for best focus turn on blue defocus and see what you have. Doug

Ah okay. I just always stayed away from the pots in the NEC after reading countless times that the pots should not be touched. But I guess its okay for the focus pots.

Nice work Jeff... how long did it take you to tape up the string?

I use non-marking tape, and run it around the edges and centres, with lines drawn on the tape for the column/row widths

If I see seagull errors, I fix it with the "SEAGULL CORRECTION" menu item.

[808s gloat over] :D

If I see seagull errors, I fix it with the "SEAGULL CORRECTION" menu item.

[808s gloat over] :D

Look at Jeff's geometry! I guess there's just no pleasing some people...

Jeff,

your geometry is stunning, especially your left to right linearity is very, very good.

If you want these perfect on the XG you need to use less raster as you do and results will be very good, with more raster usage you will need a lot more point to get where you are.

Oliver

Look at Jeff's geometry! I guess there's just no pleasing some people...

Ain't that the truth. From what can be seen in the screenshot the geometry is perfect, you cannot improve upon that.

Jeff,

your geometry is stunning, especially your left to right linearity is very, very good.

If you want these perfect on the XG you need to use less raster as you do and results will be very good, with more raster usage you will need a lot more point to get where you are.

Oliver

You are very correct about the raster usage. The problem is that when you use more raster, the far edges tend to be stretched too far out while everything in between the center and far edges tend to be squished, as if the edges need less amplitude and in between the edge and center needs more amplitude. You can correct a little bit with linearity and linearity balance, but it's impossible to get perfect without using enough point to cause banding.

Screen shots that serve a useful purpose. I like the idea of a 1080p digital test pattern projector. Its mariages like that that will eventually eliminate projector typism in America.

I asked about this a while back and everyone thought I was crazy. I used the string method last week and got it looking better than my first few attempts but nothing like what Jeff has done here. I can't imagine taping all of those strings up on the screen and then trying to see them all...

No question about thinking. Don't fell singled out, most everyone one here is certifiable.

Nice work Jeff... how long did it take you to tape up the string?

I use non-marking tape, and run it around the edges and centres, with lines drawn on the tape for the column/row widths

Thanks Gino. It took around 1.5-2 hours to figure out the dimensions and tape up all the string (thank god for millimeters...if I had to do this in inches, oh man). Coincidentally my dimensions were 80mm between each space, and I have a million 80mm diameter fans laying around, so I could double check the dimensions using the fan.

I asked about this a while back and everyone thought I was crazy. I used the string method last week and got it looking better than my first few attempts but nothing like what Jeff has done here. I can't imagine taping all of those strings up on the screen and then trying to see them all...

You should try it with your XG. I did the horizontal lines first and tried to do the horizontal linearity without the vertical lines, and i was way off. Even if you dont get it perfect, the lines help a lot. You could even do every other line and it still helps significantly.

Jeff, if you ever make a trip down to the Carolinas I would love to put you up for a night or two so you can tinker with my XG110 a bit. By the way, I solved my "rotated" blue problem with another hour of convergence work. I guess I was being lazy and was just ready to give up. The family I had visiting over the weekend was impressed despite the obvious convergence errors and my amateur setup...

You are very correct about the raster usage. The problem is that when you use more raster, the far edges tend to be stretched too far out while everything in between the center and far edges tend to be squished, as if the edges need less amplitude and in between the edge and center needs more amplitude. You can correct a little bit with linearity and linearity balance, but it's impossible to get perfect without using enough point to cause banding.


What you describe is exactly what I observed - you cannot get the horizontal squares even with maximized raster and without using lots of point adjustment and I thought it fair to point (pun intended ;)) this out to those who max the hell out of their XG's :)

(thank god for millimeters...if I had to do this in inches, oh man).



Maybe there is hope for the US afterall.

:)

Maybe there is hope for the US afterall.

:)

Haha :D. I think of you every time I use millimeters or celcius... honestly. My favourite sports are hockey and soccer, my favourite cars are european, I prefer using the metric system, and I spelled favorite as "favourite" just for fun. I do not belong in America :o.


Jeff, if you ever make a trip down to the Carolinas I would love to put you up for a night or two so you can tinker with my XG110 a bit. By the way, I solved my "rotated" blue problem with another hour of convergence work. I guess I was being lazy and was just ready to give up. The family I had visiting over the weekend was impressed despite the obvious convergence errors and my amateur setup...

Heh... I don't know how much help I'd be. The real person you want to tinker with your XG is Mark or Doug, although they both live far away. The XG is hard to figure out, but you just need to take your time (as well as asking Mark and Doug 400 questions :p).

Here are some threads from the archives discussing the use of laser levels:
First Thread (
http://archive2.avsforum.com/avs-vb/showthread.php?t=611632&highlight=laser)
Second thread (
http://archive2.avsforum.com/avs-vb/showthread.php?t=710812&highlight=laser)
Third thread (http://archive2.avsforum.com/avs-vb/showthread.php?t=626698)

Jeff, is there any way you could get a shot of your rasters so we can see how much you are using? I would be curious to know if I am pushing mine too much compared to what you have. I think I have plenty of margin between the raster and tube edges but perhaps not.

Jeff, is there any way you could get a shot of your rasters so we can see how much you are using? I would be curious to know if I am pushing mine too much compared to what you have. I think I have plenty of margin between the raster and tube edges but perhaps not.

I'm ~109" away from an 84" wide screen, which gives a throw ratio of 1.3. You should be about the same throw ratio away.

Okay, I adjusted the blue focus pot and it's uniformly defocused... thanks for the tips. And what exactly does focus balance do, and what are the symptoms if it's not set properly? I mean, I'm leaving for college in around 9 months. I don't think it's wise to spend 600$+ to have it calibrated only to leave in 9 months. I could probably get it 80-90% of the way there with an eye-one colormeter and some time for focus, convergence, astig, etc. But if focus balance is really critical, I might as well hire a tech before I spend another 20-30 hours tinkering only to need a tech in the end. So if someone could clear me up on focus balance, it would be much appreciated :).

edit: and nevermind about the green focus. I had contrast at 70% so I could see the blue and I guess the green starts to bloom around there.

edit 2: did some searching and found this thread
http://archive2.avsforum.com/avs-vb/showthread.php?t=653800


Geese thanks guys,
Yonexsp,
Depending on the exact model you have in the 'method of adjustment' section it explains the focus balance adjustment for the OSC board to align the input waveform which is global for the particular model kHz range ie: The 1350 uses a 135 kHz signal for adjustment. This can and will change H position and H width as well as a lot of other things in the projector that a newbie wont be able to correct or do. This is just the start point but why I dont like to comment on the procedure. The OSC board adjustments are critical and if the board was replaced it needs to be done or you will have oddities that focus balance cant correct using this function from the remote. In other words you can make things worse real fast because its hard to notice changes in a defocused state or if your not familiar with what you are trying to correct. What you change is global and will effect the whole freq range and at different freqs. The band-aid would be adjusting this without first checking the OSC board.

There are very few new out of the box virgin projectors but many that have been rebuilt or repaired by end users with help from the Forum. This one adjustment is best left to the rebuilders that have the test gear and knowledge to do it.

Id like to suggest that you get the service manual for your exact model and go over the circuit descriptions and also the method of adjustment so you have a better understanding of what things do back in the card cage and how it can affect the projector as a whole.

Most people understand there are things that we dont like to comment on for a reason and that isnt arrogance but more of a respect we have had to learn when posting on a open Forum that techs and newbies use. Doug

Doug, can you elaborate on the "a lot of other things in the projector" part? We can talk about this in PM if you prefer. I just want to know how important it is to have the focus balance set properly so I can decide whether to have someone calibrate things.

the majority of geometry errors are because of incorrect use of some of the electronic controls.

Trying to use one adjustment in an attempt to correct a problem created with another adjustment.

Half the battle is using the correct adjustments to correct the right issues.

By the time you have used layers upon layers of correction trying to get it right. You have confused yourself so much nothing will help it. :)

You are very correct about the raster usage. The problem is that when you use more raster, the far edges tend to be stretched too far out while everything in between the center and far edges tend to be squished, as if the edges need less amplitude and in between the edge and center needs more amplitude. You can correct a little bit with linearity and linearity balance, but it's impossible to get perfect without using enough point to cause banding.

I had started a thread on how it would be nice to have convergence software for your computer, being able to adjust RG&B before it even reaches the PJ. Not sure how hard it would be to do with software, but it could possibly fix a lot of problems.

Deron.

That would be nice... could even hook a camera up to the computer for an ACON type system :eek:. I'm sure the internals would enjoy having little to no convergence adjustments.

Actually Nec gets horizontal and their white (brighter). One or two 1' wide across.
Nec is sensitive to vertical point more than horizontal.
If it shows up while setting up you just reinitialize (one single command) the points to its default values and start over again. Not a biggy.
It is tough but doable to do a no point convergence and geometry with Nec but it won't be perfect.

the majority of geometry errors are because of incorrect use of some of the electronic controls.

Trying to use one adjustment in an attempt to correct a problem created with another adjustment.

Half the battle is using the correct adjustments to correct the right issues.

By the time you have used layers upon layers of correction trying to get it right. You have confused yourself so much nothing will help it. :)

I agree 110%. I just watched your 1292 set-up dvd again. Your logical approach is great. I went thru my 1272 again following your steps and what a difference it made!

My only trouble was being able to get the astig really good as my pj does not focus to the fine detail as well as the 1292.

Either way, thanks again for putting the dvd together.

wallace

the majority of geometry errors are because of incorrect use of some of the electronic controls.

Trying to use one adjustment in an attempt to correct a problem created with another adjustment.

Half the battle is using the correct adjustments to correct the right issues.

By the time you have used layers upon layers of correction trying to get it right. You have confused yourself so much nothing will help it. :)

Graham

What I have found is if you go through the convergence adjustments a few times you will find yourself adjusting some settings more towards zero. It's kinda like a learning process, the more you go through the convergence the more you find which settings that are best for getting correct convergence. This in turn allows you to find which adjustments you are using too much of and which ones do the most good.

You get your convergence perfect, then you decide that you can get better focus, or you want to try a new screen, or you want to use more phosphor, new lenses... and it starts all over again, we should call this a hobbie, CRT PJ Hobbie:D

Deron

Graham

You get your convergence perfect, then you decide that you can get better focus, or you want to try a new screen, or you want to use more phosphor, new lenses... and it starts all over again, we should call this a hobbie, CRT PJ Hobbie:D

Deron

Man you got that right. I was "experimenting" with running digital output to my Barco using a HDFury. Doing a little "adjusting" and the next thing I know, not only do I have vertical banding on the left side of the image, but the right edge of the raster has actually folded over on top itself. :eek: If I moved the image far enough to the right with H Phase I could actually see the image folding under the image and moving back to the left. :eek::eek: What's that all about?

After about a half and hour of head scratching and getting no where I just said screw it......I deleted all the mem blocks and reset everthing to midpoint.....basically starting over. All the banding and folding is now gone..but the geometry and convergence isn't as good as I originally had it. Sigh...it never ends.

Regards,
John

I'm looking into getting a 1352. What's this banding about? Where you get shadows when viewing grass or looking at the sky?

If this banding occurs when you use too much point convergence, how do you go back and start from scratch and start over? Is there a "reset all point convergence" in the service menu. Or do you set all point banding options in the menu to their original value?

The sky on the right and the grass on the left have banding in this picture. I HATE banding.

http://webpages.charter.net/coreyg/banding.jpg

What I think you see in that image is 4:3 AR tube wear. The banding I know about is horizontal dark shading caused by using too much point correction. Yes you can "normalize" your settings. It’s really the first step of a setup.

Mike

Look at Jeff's geometry! I guess there's just no pleasing some people...

Uhh, I wasn't dissing his geometry, I was jokingly bragging because my projector happens to have seagull correction when someone else was saying theirs didn't. I'm not sure how you extracted an insult from that... argh.