Problem Solved, and it was easier than it could be.
When the trace was intermitently shorten (squished), I tested XY, and
it was solid steady, so I was relieved to know that everything from
U760 (the sweep horizontal control chip) up to the CRT was okay, since
XY is a secondary input to this chip, along the A Sweep Input. If the
A Sweep input was shaking the screen, and the XY was not, the problem
was before that chip, and at the A Sweep circuitry.
I followed the sawtooth waveform from the horizontal "Miller Sweep"
generator, from page 4 and it was perfect. The output of the
generator being okay eliminated a bunch of problems, all the rotary
switches, etc.
Scoping probe point [31] (entry of the U760 chop) the image was
deformed. Scoping probe point [30] (generator output) it was okay.
In between both points there is R740, a 1k trimpot.
This trimpot can be seen on page 6 of the foldouts (schematics),
grid F4, it is physically located at the edge of the timing board
(A4), it controls the "gain" of the horizontal trace. The "gain" is
the adjustment of the trace width on screen, and obviously, the width
of the image. Changing the adjustment of this pot will mess with the
accuracy between the trace (image) and the acrylic marks, so you might
read a wrong period of time from a waveform.
Before the R740 the sawtooth was neat, after the trimpot, it was
deformed and shaking... hmmm... using a small screwdriver I just
rotated the trimpot to one side and another, several times, as "to
clean the trimpot contacts over the carbon film", it solved the
problem for few seconds, but always returned, until the trace
collapsed completely. Suddenly the trimpot was showing more than 2M
Ohms. As the trimpot is at the edge of the board, I was able to
replace it without disassembling anything. Used a different multiturn
1k trimpot. Now it is working nicelly.
Of course this trimpot should be adjusted with a very precise waveform
and coincide the waveform with the acrylic marks (graticule). I
adjusted using a crystal 10MHz squarewave (verified with a HP
frequency counter).
Later I opened the tiny trimpot and found that one of the end-of-
course lead was making bad contact with the carbon film metallic
termination. Unfortunately, if you grab one trimpot lead with a
pliers and turn few degrees it will loose the good contact with the
termination. Looking at the trimpot, it seems the Tektronix assembly
required the trimpot lead to be rearranjed with a pliers, to fit the
pcb holes, it then reveal the problem was "planted" at the production
line. Probably the pin broke solid connection with the termination,
years later and temperature changes increase oxidation in the bad
contact. So, rotating the trimpot adjustment back and forth will not
reveal the problem, since the bad-contact was not between the carbon
film and moving contact. It was not super easy to pinpoint the
problem, but evident when the trimpot presented 2M Ohms of
resistance.
On the same page 6, grid F6, there is another exactly the same
trimpot, R730, 1k, for the B Sweep Gain. B sweep is when you expand
part of the actual trace. This R730 adjust the width of the B
sweep. Testing and comparing, R730 is actually also presenting some
problems, just touching it with a pen makes the image shakes and
squishes. As it is somehow stable, I let it be quiet.
Will try to find two better brand trimpots that might fit well in the
space and replace both later. In real it doesn't need to be a high
precision trimpot, it only need to be small and top adjusted. Even a
plastic body will do it. Found some replacements at Jameco for less
than $2 each.
The trimpot Tektronix p/n is 311-0635-00, the manufacturer p/n is
82P-6-2-102k, manufacturer code is 73138. It is a NonWir 1k 10%, 0.5W
round, metalic top, very small trimpot (0.25" diameter). It seems to
manufactured by Bourns, even that part/number doesn't match:
http://search.digikey.com/us/en/prod...2LF-ND/1088227
I wonder, there are several other similar trimpots all over the scope
boards... hmmm... new problems soon?
Used the opportunity to clean the interior of the scope with
compressed air. It is a nice scope returning working to the bench.
Thank you for the help.
Wagner
Orlando Florida
On 1/18/2012 10:54 AM, Lip* wrote:
You are correct, Jeff.
If the Horizontal Position adjust (R726) can move the "squishing"
trace to the right or left, beyond the dead zone, the amplifier is
working and has enough voltage to show the trace in that region.
Perfect debugging thought. *Will do it tonight.
Look at a signal.
Does the trace stay the same when it squishes? *Or does the trace
get truncated? *That should tell you whether it's a sweep timing issue
or a gain issue.
If the XY works all over the screen, for example, a 60Hz sine
generating an almost perfect ring on screen, it eliminates voltage,
amplifier, plates, etc. *Will be around the trace sweep generator, or
something like that.
I already identified all the scope points on the diagram and their
position on the board printout, so it will be easier to follow this
thing.
The only place in the diagram where I can see different circuits
(bipolar) for left or right side of the screen (half trace) is the
output of the U760, the guy who does the sweep controls, and it is
called "Horizontal Pre-amp". *It has +SWP and -SWP outputs that drive
directly the output transistors. *The input of U760 is not separated,
it comes from one linear ramp from Q742 that represents the whole
screen trace, feeding the U760 "A SWP IN".
So, the fact that the right half of the screen is more "damaged" and
"squishing", lead me to think about connections between those boards,
or something at the output transistors driving the right plate. *As
far as I remember, the right side failing was not steady, but as if
there is a lot of noise in the right side sweep ramp.
I need to find it out... it seems easy. *The first test was to make
sure +100Vdc is steady and Okay. *Then the Horizontal Position knob to
make sure the "sqhished" trace can be moved to the left or to the
right of the screen. * Next, the XY test, to eliminate forward or
backward circuits in the path. *There is also a -5V regulator (U750)
specifically feeding the U760, and I guess this negative rail is only
used to generate the negative sweep ramp to control the output
transistors for the right plate.
It will be nice to have this2235back to life, even with the Tek2246
available, I always choose the2235, it is smaller, lighter, faster to
operate and I can transport it anywhere I need it, without pulling a
muscle in my back.
Thanks, will report.
Wagner.
On Jan 18, 12:54 pm, Jeff *wrote:
If the position control will move what trace you do have to the right,
that might eliminate the output stage. Look for it on the print, in
some scopes it is a dual pot wired in llike an "X" formation and may
feed the CRT plates "directly", but I don't think Tektronix does this.
(they have good reason on an engineering level)
I would definitely check it in XY the very next thing. If it works
properly in XY you've eliminated the entire amplifier section and can
proceed directly to the oscillator. However if whatever you try to
display compresses on the right, you know it's that amp, and whether
the position control moves the flattened portion or not is the next
test. If the whole squished trace moves, it is before the position
control, if the trace squishes (love those technical terms) without
reaching the right side then it is after the position control.
If the problem is one of the output stages to the plates, the non
linear portion of the trace whould exhibit defocussing or astigmatism,
I think. It may not with certain CRT designs. To go any farther I'll
have to get the print, but this should be a good wham bam eliminate a
bunch of stuff. In other words, quichly find out what ituisn't/u..
JURB
Another good tip Mike, will look at it.
Thanks a bunch.
Wagner