Tip by Mark Rejhon: (8/30/99) Convergence
HOWTO - NEC / Runco projectors
[MARCH 2000 ADDENDUM: For HTPC users, it is preferable to use equivalent test patterns
using DisplayMate computer software or using bitmap test patterns instead of using the
projector's built-in test patterns.]
"Yes, a howto would be nice. But in the meaintime, remember this mantra when doing
convergence: CENTRE first, EDGES next, CORNERS last! And avoid zone/point convergence
until the very last step (even though it's tempting to do it sooner). Before adjusting,
leave your projector for half an hour to warm up before you begin to do the mechanical
convergence/focus.
Without further ado, here's my personally recommended steps for NEC and Runco
projectors. I'm not an ISF technician, but I do consider myself to be more
technology-literate than the average Home Theater user:
Make sure electronic static convergence is set to 0 for all CRT's before mechanical
convergence. First, I do the mechanical convergence and mechanical focussing, as well as
the electronic focus/astig. I won't bother going into detail about this, but if you have
zone focus/astig, do the center first, edges the next, and corners the last. Do mechanical
focus before you do electronic. Do electronic focus before you do astig. I normally start
with the green CRT first, then red CRT, then blue CRT. Also, sometimes the astig test
pattern is no good - use a DisplayMate computer test pattern (or as a crude measure, the
computer mouse arrow pointer as a reference of focussing/astig) ... for more proper astig.
I just try to adjust astig left/right until I get the sharpest image in that part of the
screen, then I adjust astig up/down until I get an even sharper image in that part of the
screen. (Sometimes I repeat these steps twice, to make sure I did not overshoot). After
proper adjusting, I can now just about barely see line structure at 1024x768 on my NEC
XG135 (8" CRT) when I'm staring few inches from the screen - that's more or less how
sharp I've managed to get it.
Secondly, I do the green CRT geometry - it's gonna be your reference CRT when doing
red/blue convergence later on. Make sure the green CRT is approximately positioned and
sized (use image size and position - don't use static convergence to adjust position of
green CRT image at this point). I Use tilt/bow/amplitude/linear/keystone/pincushion to do
the best you can, and then use RGB point convergence to make green CRT look corect. Repeat
the steps for any further small changes and adjustments you need to do. At this point you
DO NOT adjust the green CRT further anymore - you're finished with the green CRT even
before you touch the red/blue CRT. So you have to try to get the green CRT visually as
geometry perfect as you can at this point, especially with the edges lining the boundaries
of the screen properly.
Thirdly, I do the convergence of the red CRT onto the green CRT. (red/green CRTs
enabled, blue CRT disabled at this point) I use the fine grid pattern. First, I do static
convergence to make the exact middle look correct. Next, I look at the middle of the 4
edges and I do it roughly in this order: static/tilt/amplitude/linearity/bow ... First
horizontally, then vertically. Sometimes I have to do "key balance",
"linear balance", "linear distortion" to compensate for minor
imperfections, you should learn how these adjustments work by experimenting with a spare
memory slot in your projector. While doing all these adjustments, I completely ignore the
corners for now, while I focus on making the middles of the 4 edges as perfect as I can. I
try to make the middle horizontal line (of the grid pattern - the line that divides the
screen into a top and bottom half) and the middle vertical line (of the grid pattern - the
line that divides the screen into a left and right half) as perfectly converged as I can,
converged both horizontally and vertically, by using the grid crosslines as my point of
reference during adjustment. (Don't worry about the corners for now) Next, I focus on the
corners - I only do edge-specific keystone and edge-specific pincushion adjustments (to
tilt and curve the 4 edges of the image individually, one edge at a time). Edge-specific
keystone/pincushion doesn't move the middle of the edge, but affects the corner
convergence by tilting the edge or curving the edge. No further general geometry
adjustments are usually needed at this point when you are converging the corners. Finally,
when I am done with the red CRT, I repeat all of these above steps with the blue CRT (red
CRT disabled, but with the green/blue CRTs enabled).
Finally, I do point convergence. I start with the red CRT (red/green CRT's enabled,
blue CRT disabled). If you've done the geometric adjustments very well like I have, you
DON'T need to do zone convergence - only point convergence. (Zone convergence = move many
points of the grid at once, Point convergence = move one point of the grid at once). With
NEC XG135 CRT projectors, use medium point convergence (coarse normally not necessary) to
converge the red CRT as well as you can onto the green CRT. I work convergence in a spiral
pattern going outwards from the middle, going from one grid square to the next, converging
each one by one - always start in the middle. Ideally, you should do the center first,
edges next, and corners last, but going in a spiral pattern pretty closely approximates
that and is easier to do. Don't bother doing "fine" convergence (208 point
convergence for NEC XG135) since it is a lot of work, unless absolutely necessary -
sometimes you come across spots that can only be accessed by "fine" convergence
though usually "medium" is usually all you need and will properly converge the
whole image most of the time. Next, I repeat these steps and do the blue CRT convergence
onto green CRT (with the red CRT disabled). As a general rule of thumb, try not to create
more than 15-20 minutes of red/blue point convergence work - spend more time on the
general geometry-related convergence adjustments sooner.
REMEMBER: Avoid point convergence until the very last step, don't be lazy - you will
have less convergence drift problems in the long run if you use the general geometry
adjustments first to the best of your ability before hitting the point convergence.
REMEMBER: Save your settings once every few moments while converging. Don't rely on the
auto-store feature to back up your settings. Also, when you're finally done, make a copy
of your memory slot as a backup. And also, if this is your first real convergence attempt,
you should save a copy of your settings into the master memory slot (the
"default") so it's quicker to converge the next video mode, by using the default
settings that will sometimes do 98% accuracy, saving you a lot of work - even a whole
weekend of work!
REMEMBER: You have to repeat all these steps for each different video mode or different
computer resolution/refresh rate. It's a big hassle, so keep your resolution selection to
a limited subset. You don't need to repeat the mechanical focus/convergence, but sometimes
you may need to repeat the focus/astig steps. Often, if you save your settings to the
master/default memory slot, you usually don't have to re-focus/re-astig but you should
doublecheck everytime you create a new memory slot for a new video mode that's going to be
used frequently.
I may have missed a few steps, but I think all of the above basically sums up what you
need to do. You have my permission to integrate the above into whatever FAQ comes up -
just make sure you give me a chance to proofread it, and give me credit.
And if there's any errors, please do not hesistate to let me know! Also, be noted that
this HOWTO does NOT cover greyscale calibration. Just DIY project convergence.
