|
|
Post by huggy111 on May 30, 2016 6:56:43 GMT -5
Guys, I bought a digital pj for perfect geometry and convergence. I reset convergence to midpoint and realised I've stuffed something up and cant figure out what I;ve done wrong. Problem is this.
I have projector aligned center line to screen center ( both of them ) rasters are perfectly centered and image width reflects that on screen however, even though image is at perfect width, the center crosshatch is off to one side one whole geometry block width. Using the digital as alignment I can only bring it in with phase control but that only fixes the middle not the sides, so when I fix the middle, one side is off the screen one grid width and the other side is too far in the screen one grid width. I've obviously done something wrong but what? Any ideas?
Dave
|
|
|
|
Post by gregstv on May 30, 2016 7:39:36 GMT -5
Do have a picture? Horizontal linearity maybe?
|
|
|
|
Post by huggy111 on May 30, 2016 8:00:59 GMT -5
Hi Greg Cant take a pic but the arrow on this pic will show you how far my green center is off.  |
|
|
|
Post by mastertech on May 30, 2016 10:00:46 GMT -5
Correct me if I am wrong, but on the 909, even if you select "genlock" in random access mode, doesn't it still use an internally generated pattern and it is only based off the input frequency?
|
|
|
|
Post by gregstv on May 30, 2016 15:46:19 GMT -5
When you did the reset did it turn the COG off? I would think you would need this on to line it up.
|
|
|
|
Post by huggy111 on May 30, 2016 16:17:15 GMT -5
Greg
I can line the grids up but I have to use extreme convergence in corners, I was hoping to avoid that due to drift on the stack.
|
|
|
|
Post by gregstv on May 31, 2016 4:13:13 GMT -5
So it wasn't like this before, Or is just that the new grid is different? Are both projectors having the same issue trying to line up against the digital?
I had a play doing side by side stack with two 808s and needed to use just about all the convergence range to get them to line up on the edges.
|
|
|
|
Post by huggy111 on May 31, 2016 4:41:16 GMT -5
Hi Greg
Yes it was like this before as well. I was wondering if it was normal but from what youre saying, it is.
Being at convergence extremes cant be good for drift. Perhaps an over/under scenario is better suited for a stack? I could pull them down and make a custom housing if it means I dont have to converge the bastards every few weeks lol
|
|
|
|
Post by Casethecorvetteman on May 31, 2016 5:00:19 GMT -5
Correct me if I am wrong, but on the 909, even if you select "genlock" in random access mode, doesn't it still use an internally generated pattern and it is only based off the input frequency? Correct yes. Selection of 'selected source' will allow use of whatever you like. |
|
|
|
Post by Casethecorvetteman on May 31, 2016 5:26:15 GMT -5
Hi Greg Yes it was like this before as well. I was wondering if it was normal but from what youre saying, it is. Being at convergence extremes cant be good for drift. Perhaps an over/under scenario is better suited for a stack? I could pull them down and make a custom housing if it means I dont have to converge the bastards every few weeks lol There should be no reason you cant get this correct with those sets. Because youre shooting on an angle, the middle of the green raster is NOT the middle of the green image. You need Green Convergence turned ON. There is never any good reason to turn it off. Phase should be set to where the image is slightly to the right inside the raster, and the image is what you centre on your tubeface, not the raster. Get the image in the correct spot on the raster, then use the green coarse raster shift to make it centre of the tube face. Fine raster shift should be set to 49 or 50 in both directions. From there, you get pin and keystone pretty close for your screen, check image is still centre of tube face, more it using coarse raster shift. Move the projector until the correctly set up green image lines up with both edges of your screen. Linearity on green will be required to get the middle of the picture to the middle of the screen, and from there you will need to adjust image size and use green raster shift to align again with your edges, start with coarse, the move to fine. It will be a juggling act to get this right. After this is done, you should have the green image lined up with the middle of the screen as well as the edges, and it should be still centre of tube face. Now you set the red and blue, using their coarse raster shift you move the image so it is centre of tube face, angle the tubes so that both sides line up at the edges of the screen. Adjust the width coils so the size is the same, but DO NOT use a metal screwdriver to adjust these with the set turned on. I use porcelain tools for this, you can get them at Jaycar. There are pots on the V board to adjust vertical size and position. Now you should have a perfect green, and red/blue should be pretty near perfectly aligned at the middle of the sides. Use coarse raster shift on red and blue to align the middle of the screen, finish it off with fine raster shift, then use pots on the V board to get the vertical position set right, the fine raster shift should be at 49 or 50 for vertical on red and blue. Now do red convergence to the green starting with coarse adjust, followed with fine adjust. When red is done, the red image should be dead centre of the tube face. Blue is next, either line it up to red or green, i prefer lining up with red. This is about the best way to get it perfect. |
|
|
|
Post by Casethecorvetteman on May 31, 2016 5:28:16 GMT -5
So it wasn't like this before, Or is just that the new grid is different? Are both projectors having the same issue trying to line up against the digital? I had a play doing side by side stack with two 808s and needed to use just about all the convergence range to get them to line up on the edges. Considering the step sizes for linearity id certainly go for side by side. |
|
|
|
Post by Casethecorvetteman on May 31, 2016 5:43:02 GMT -5
Basically, you need to start from scratch with everything at midpoint geometry wise and do it all properly, following the methods ive briefly described above.
Adjusting linearity will have an effect on image size and position. It will be a balancing act to get it correct, but it all must start with proper projector placement for both sets so green can line up with the sides and be centre of tube face. Then you work on the middle of the image. Get it as perfect as you can, then use green only settings to get the last bit of perfection.
Lining up the image sides with the screen sides is critical for an image to be centre of tube face, the middle of the image is totally irrelevant at that point of the setup.
The same again applies to red and blue, centre the image on the tube faces, line up the edges with the screen, then you know that when you converge the sides of the image it will once again be centre of tube face, and there will be no guess work involved at all.
You should never need extreme adjustments of any convergence setting if you have the width coils set correctly and the vertical adjustment pots are set perfectly. The middle of the image should align pretty close without even doing and zone convergence, and if all size/position pots and coils are set right, you wont need much adjustment to get it aligned.
|
|
|
|
Post by huggy111 on May 31, 2016 7:50:39 GMT -5
Hi Greg Yes it was like this before as well. I was wondering if it was normal but from what youre saying, it is. Being at convergence extremes cant be good for drift. Perhaps an over/under scenario is better suited for a stack? I could pull them down and make a custom housing if it means I dont have to converge the bastards every few weeks lol There should be no reason you cant get this correct with those sets. Because youre shooting on an angle, the middle of the green raster is NOT the middle of the green image. You need Green Convergence turned ON. There is never any good reason to turn it off. Phase should be set to where the image is slightly to the right inside the raster, and the image is what you centre on your tubeface, not the raster. Get the image in the correct spot on the raster, then use the green coarse raster shift to make it centre of the tube face. Fine raster shift should be set to 49 or 50 in both directions. Case , I dont understand the above. Youre saying to centre image on tube face not raster, then once image is centred on tube face make raster centre of tube? What am i missing here? |
|
|
|
Post by Casethecorvetteman on Jun 1, 2016 0:04:00 GMT -5
Forget about the raster, its not really that important.
The image you display must be centre of tube face, the image does not have to be centre of raster.
H-PHASE moves the image within the raster, you want that image positioned within the raster so it is as even and straight as it gets, which is generally off to the right of the raster, because the left side is where the line starts and at this point the raster is not as stable as it is further across to the right.
You can do this without even looking into the tube face, and the raster will be blanked up to the image edges anyway.
On my Cine 9 the left of the raster has some distorted lines, so ive got the image almost all the way to the right inside the raster to compensate for that.
|
|
