A Mitsu owner was commenting that "in 4x3 mode there is a noticeable widening of
the picture when graphics are displayed. Also, I can't seem to get the picture lined up
perfectly from top to bottom of the screen (vertically). "
"Mits's have always used a non-servo regulated power supply, with capacitive
timing instead of servo-lock sensing in their HV circuits, like the Sony's do. I've
serviced many Mit's, and I'm assuming that nothing has changed in these newer units, from
what I'm hearing. They're all probably still in warranty, so I haven't been called upon to
service any yet, and so haven't seen the insides of one yet.
But from what I'm hearing, I'd figure nothing has changed lately, in this regard. This
means that, among other things, the HV electrostatic focussing is affected by whatever's
on the screen. It's impossible to accurately focus a Mit electrostatically simply by
having a weak pattern on your screen. With a strong pattern, it'll all change. The best
elec. focussing is done with a medium-level pattern on the screen, or basically with
whatever type of picture you view most of.
This also affects how the vertical lines are drawn, which shows up mightily in the 4x3
mode. Many times, the image within the gray bars "tows" the sides in just a bit
at the vertical center, forming a slightly pincusioned set of vertical edges. From Doug's
report here, it looks like a heavier-duty picture, one with the demanding nature of the
instantaneous turn-on and turn-off of graphics, pulls down the HV just a bit, which widens
up the image just a bit. Then there's the edge of the 4x3 graybars to consider, since the
scanning beam goes from a steady, static-drawing scan line -- the graybars -- to a
demanding and always-changing scan line -- the picture being drawn. Doug and I both
noticed that certain images and pictures kept the sides correctly vertical, others tow
them off -- or in this case in--a bit, or widen them or narrow them just a bit. It all
depends on the light content--see the needle pulse pattern in VE and AVIA, and watch the
needle bend with increasing light content. This is nature of the beast on a unit without
servo-controlled and HV sensor-guided power supply, whose design senses changes in demand
on a microsecond to microsecond basis, and tells the unit to "step up the
volume" whenever the unit requires more current/voltage to stay stable, then takes it
back and maybe even goes down in voltage a bit in the next microsecond, depending on the
demands of picture content.
In this case the current/voltage is fixed by a voltage reference -- the timing cap --
even tho the picture would have stayed stable if the current had gone up to compensate for
the increased demand on the power supply. And so the picture content affects the drawing
of the picture, which wouldn't have happened with a HV-sensor guided servo-controlled
Sometimes that timing cap gets overheated by some hot-running raised resistor placed
close to it in the design of the circuit board -- and sometimes it just drifts off value
because it has become defective -- and winds up overvoltaging the set, causing shutdown. I
have had to replace a number of those timing caps in Mit units, DVs inluded, after which
the 130 volts comes back to EXACTLY 130 -- or within 1.5 volts of it, any more than that
will trip the protection circuits--where it stays from then on, just like it was designed
to, no matter what the picture content. Some brands use exactly 130, some use exactly 140,
etc. Whatever goes into the HV circuit determines what comes out, which makes
stabilization of this HV circuit extremely critical, just from a safety standpoint. But
also from a picture-shaping standpoint.
When I set up the geometry on Doug's unit, I used the entire screen--it was that or
make special exceptions where the vertical lines will bend or expand a bit during 4x3.
Since Mit's geometry/convergence settings are global among the formats, it was either get
one right or the other right. Getting both right is not possible on a global format like
that. Which is one way the Tosh's have it over the Mits's, in that in the Tosh
designs, each format/mode is geometried/converged independently of the others, allowing
for each to be optimized -- the 4x3 on the one hand and the
16x9 formats on the other.
Of course, it costs more to have that done, because most calibrators, me included,
charge extra for each additional format/mode. On the Mit that works in your favor over the
Tosh costwise -- less costly -- but it also has the built-in limitation of being able to
get one format perfect while the other has to be slightly "out", or vice
Note: If the service menu was reset intentionally or not, the HV Regulation
adjustment needs to checked/adjusted per the service manual... it says this must be done
if "E2RESET" was activated.