Topics

7904 - Very Dim Trace

Jim McIntyre
 

Guys, this is my first post here. I've been tinkering at Tek scopes for along time - since I first bought a 531A in the late 70's. I'm looking for some tips on a 7904.


I picked this one up at an auction, and at first, it looked pretty good. Good enough that I've been scrounging plug-ins for it, and generally cleaning it up, and adjusting it. I switched it off one night, and the next day, I had no trace. Zilch. Beam finder, maxing the intensity controls - no help. Even in a darkened room, no light at all.


The filament is lit.


All LV supplies check out OK.


I don't have a HV meter to allow me to measure the anode supply. I know there's decently high voltage there, judging by the length of the spark when I discharge the electrostatic charge, however.


The 2950V measures high, about 3250V. Yet, curiously, the cathode supply regulator (U1635) isn't railed - in fact, based on my DC measurements in this area, I think the regulator is actually regulating - just at too high of a voltage. Weird. I wish I knew what the cathode supply measured before the issue popped up...


On a whim, I tweaked the CRT grid bias (R1674) and can now get a faint trace. It's well focused, just dim. Pushing Beam Find does not h elp - it just gets blurry, reduces the sweep span, but it's still dim.


Any tips?


Thanks,


Jim

 

On 07 Feb 2016 19:15:56 -0800, you wrote:

Guys, this is my first post here. I've been tinkering at Tek scopes for along time - since I first bought a 531A in the late 70's. I'm looking for some tips on a 7904.

I picked this one up at an auction, and at first, it looked pretty good. Good enough that I've been scrounging plug-ins for it, and generally cleaning it up, and adjusting it. I switched it off one night, and the next day, I had no trace. Zilch. Beam finder, maxing the intensity controls - no help. Even in a darkened room, no light at all.

The filament is lit.

All LV supplies check out OK.

I don't have a HV meter to allow me to measure the anode supply. I know there's decently high voltage there, judging by the length of the spark when I discharge the electrostatic charge, however.

The 2950V measures high, about 3250V. Yet, curiously, the cathode supply regulator (U1635) isn't railed - in fact, based on my DC measurements in this area, I think the regulator is actually regulating - just at too high of a voltage. Weird. I wish I knew what the cathode supply measured before the issue popped up...
I assume you measured the output of U1635 and found it to be within a
couple volts of ground.

A high cathode voltage could be produced by leakage through CR1639
which protects the non-inverting input of U1635. U1635 could also be
bad.

On a whim, I tweaked the CRT grid bias (R1674) and can now get a faint trace. It's well focused, just dim. Pushing Beam Find does not h elp - it just gets blurry, reduces the sweep span, but it's still dim.
If the cathode voltage is high then the horizontal and vertical
deflection sensitivity should be proportionally lower. Is that the
case? You can check the calibration using the probe calibrator output
signal.

I would defocus the beam before conducting further tests in case you
fix the problem so that the CRT phosphor does not get burnt.

Marian B
 

*If* the X / Y deflection works as before, so calibrated time base steps
are still correct and calibrated vert amp steps, too, then you are
likely looking at a Z axis issue.

I'd recommend to check if the Z axis amplifier output is as expected. If
it is not, check the amplifier, if the amp is okay, check the Z axis logic.
If the amp output is okay, replace all HV caps and all diodes in the Z
axis DC restorer circuit.

Cheers, Marian

 

It has to be more than a z-axis amplifier problem because Jim measured
the 2950 volt cathode voltage at 3250 volts. I wonder if that by
itself would be enough to cause the problem.

Alternatively, could the measurement be wrong? Jim, can you verify
proper operation of the probe you used to make that measurement?

On Mon, 8 Feb 2016 10:24:59 +0100, you wrote:

*If* the X / Y deflection works as before, so calibrated time base steps
are still correct and calibrated vert amp steps, too, then you are
likely looking at a Z axis issue.

I'd recommend to check if the Z axis amplifier output is as expected. If
it is not, check the amplifier, if the amp is okay, check the Z axis logic.
If the amp output is okay, replace all HV caps and all diodes in the Z
axis DC restorer circuit.

Cheers, Marian

Jim McIntyre
 

Thanks all for your ideas and responses,

To David - My HV probe is a bit suspect. I didn't want to apply 3kV to my Fluke, so I ginned up a voltage divider. It's spot on at 100V, but maybe the carbon resistors misbehave at higher voltages. Thanks for bringing this up. I'll poke around and see if I can find a decent HV probe.


To Marian and David - XY deflection is OK - certainly not off by more than a few %.


To David - Yes, U1635's output measures 0.6 - 0.9V.


Please keep noodling on this - I'll get back in here with a better measurement of the cathode voltage...




Thanks again,


Jim

Jim McIntyre
 

One other minor clue -

the 'A' sweep, 'B' sweep, and on-screen readout are all affected in the same way...

 

On 08 Feb 2016 18:25:32 -0800, you wrote:

To David - My HV probe is a bit suspect. I didn't want to apply 3kV to my Fluke, so I ginned up a voltage divider. It's spot on at 100V, but maybe the carbon resistors misbehave at higher voltages. Thanks for bringing this up. I'll poke around and see if I can find a decent HV probe.
To David - Yes, U1635's output measures 0.6 - 0.9V.
My guess then is that the cathode voltage is correct and it is the
probe which is a problem. Carbon resistors have a poor temperature
coefficient so self heating will be a problem and at high voltages,
their voltage coefficient will contribute a significant error as well.

To Marian and David - XY deflection is OK - certainly not off by more than a few %.
I thought you posted that the trace was "compressed".

There must be more than one problem if the trace is dim, compressed,
*and* the beam finder does not work.

Jim McIntyre
 

No, I didn't say the trace was compressed.

RE: beamfinder. This has 3 functions, right? It blurs the trace, limits the deflection,and intensifies the beam. Mine limits the deflection and blurs the trace. It does not make it brighter.


Only one symptom - dim trace.

 

On 08 Feb 2016 22:06:38 -0800, you wrote:

No, I didn't say the trace was compressed.
That was my mistake. I got the 7904 and 7904A threads confused.

RE: beamfinder. This has 3 functions, right? It blurs the trace, limits the deflection,and intensifies the beam. Mine limits the deflection and blurs the trace. It does not make it brighter.
That is right . . . maybe and sort of.

I tried tracing the schematic and there are a bunch of drafting errors
around the beam finder switch in the schematic so I powered up my 7904
and made these observations:

1. Traces are compressed horizontally and vertically but the readout
is only compressed horizontally. Checking the CRT vertical amplifier
schematic shows that this behavior with the readout is deliberate.

2. Intensity was *not* changed with the beam finder activated.

3. Only traces were blurred; the readout focus was unchanged.

My manual says the beam finder switch can be locked in the outward
position but my 7904 does not do this - weird.

Only one symptom - dim trace.
I agree with Marian. Check the output of the z-axis amplifier for
proper operation and if it is good, replace the high voltage
capacitors and diodes in the DC restorer circuit.

For what it is worth, correct deflection means that the cathode
voltage is correct.

Jim McIntyre
 

Thanks for detailing the beamfinder operation.

At this point, I'm pretty sure I made a big mistake using carbon resistors in my homebrew HV probe. I had no idea how bad their voltage coefficient was.


So, I agree - the cathode voltage is probably OK, and I'll start going through the Z-axis amp. I also ordered a Z-axis board off eBay for a parts source, if needed.

Carl Hallberg
 

If you look up the breakdown voltage of carbon composition resistors, you'll see for individual units it is between 350v to 500v. Also depends on manufacturer, how many in series, how protected from arcing. Look inside a High Voltage probe and you will find a spiral wound film resistor.
Carl
--------------------------------------------

On Mon, 2/8/16, David @DWH [TekScopes] <TekScopes@...> wrote:

Subject: Re: [TekScopes] Re: 7904 - Very Dim Trace
To: TekScopes@...
Date: Monday, February 8, 2016, 11:25 PM


 









On 08 Feb 2016 18:25:32 -0800, you wrote:



> To David - My HV probe is a bit suspect. I didn't
want to apply 3kV to my Fluke, so I ginned up a voltage
divider. It's spot on at 100V, but maybe the carbon
resistors misbehave at higher voltages. Thanks for bringing
this up. I'll poke around and see if I can find a
decent HV probe.



> To David - Yes, U1635's output measures 0.6 -
0.9V.



My guess then is that the cathode voltage is correct and it
is the

probe which is a problem. Carbon resistors have a poor
temperature

coefficient so self heating will be a problem and at high
voltages,

their voltage coefficient will contribute a significant
error as well.



> To Marian and David - XY deflection is OK - certainly
not off by more than a few %.



I thought you posted that the trace was
"compressed".



There must be more than one problem if the trace is dim,
compressed,

*and* the beam finder does not work.











#yiv2496582887 #yiv2496582887 --
#yiv2496582887ygrp-mkp {
border:1px solid #d8d8d8;font-family:Arial;margin:10px
0;padding:0 10px;}

#yiv2496582887 #yiv2496582887ygrp-mkp hr {
border:1px solid #d8d8d8;}

#yiv2496582887 #yiv2496582887ygrp-mkp #yiv2496582887hd {
color:#628c2a;font-size:85%;font-weight:700;line-height:122%;margin:10px
0;}

#yiv2496582887 #yiv2496582887ygrp-mkp #yiv2496582887ads {
margin-bottom:10px;}

#yiv2496582887 #yiv2496582887ygrp-mkp .yiv2496582887ad {
padding:0 0;}

#yiv2496582887 #yiv2496582887ygrp-mkp .yiv2496582887ad p {
margin:0;}

#yiv2496582887 #yiv2496582887ygrp-mkp .yiv2496582887ad a {
color:#0000ff;text-decoration:none;}
#yiv2496582887 #yiv2496582887ygrp-sponsor
#yiv2496582887ygrp-lc {
font-family:Arial;}

#yiv2496582887 #yiv2496582887ygrp-sponsor
#yiv2496582887ygrp-lc #yiv2496582887hd {
margin:10px
0px;font-weight:700;font-size:78%;line-height:122%;}

#yiv2496582887 #yiv2496582887ygrp-sponsor
#yiv2496582887ygrp-lc .yiv2496582887ad {
margin-bottom:10px;padding:0 0;}

#yiv2496582887 #yiv2496582887actions {
font-family:Verdana;font-size:11px;padding:10px 0;}

#yiv2496582887 #yiv2496582887activity {
background-color:#e0ecee;float:left;font-family:Verdana;font-size:10px;padding:10px;}

#yiv2496582887 #yiv2496582887activity span {
font-weight:700;}

#yiv2496582887 #yiv2496582887activity span:first-child {
text-transform:uppercase;}

#yiv2496582887 #yiv2496582887activity span a {
color:#5085b6;text-decoration:none;}

#yiv2496582887 #yiv2496582887activity span span {
color:#ff7900;}

#yiv2496582887 #yiv2496582887activity span
.yiv2496582887underline {
text-decoration:underline;}

#yiv2496582887 .yiv2496582887attach {
clear:both;display:table;font-family:Arial;font-size:12px;padding:10px
0;width:400px;}

#yiv2496582887 .yiv2496582887attach div a {
text-decoration:none;}

#yiv2496582887 .yiv2496582887attach img {
border:none;padding-right:5px;}

#yiv2496582887 .yiv2496582887attach label {
display:block;margin-bottom:5px;}

#yiv2496582887 .yiv2496582887attach label a {
text-decoration:none;}

#yiv2496582887 blockquote {
margin:0 0 0 4px;}

#yiv2496582887 .yiv2496582887bold {
font-family:Arial;font-size:13px;font-weight:700;}

#yiv2496582887 .yiv2496582887bold a {
text-decoration:none;}

#yiv2496582887 dd.yiv2496582887last p a {
font-family:Verdana;font-weight:700;}

#yiv2496582887 dd.yiv2496582887last p span {
margin-right:10px;font-family:Verdana;font-weight:700;}

#yiv2496582887 dd.yiv2496582887last p
span.yiv2496582887yshortcuts {
margin-right:0;}

#yiv2496582887 div.yiv2496582887attach-table div div a {
text-decoration:none;}

#yiv2496582887 div.yiv2496582887attach-table {
width:400px;}

#yiv2496582887 div.yiv2496582887file-title a, #yiv2496582887
div.yiv2496582887file-title a:active, #yiv2496582887
div.yiv2496582887file-title a:hover, #yiv2496582887
div.yiv2496582887file-title a:visited {
text-decoration:none;}

#yiv2496582887 div.yiv2496582887photo-title a,
#yiv2496582887 div.yiv2496582887photo-title a:active,
#yiv2496582887 div.yiv2496582887photo-title a:hover,
#yiv2496582887 div.yiv2496582887photo-title a:visited {
text-decoration:none;}

#yiv2496582887 div#yiv2496582887ygrp-mlmsg
#yiv2496582887ygrp-msg p a span.yiv2496582887yshortcuts {
font-family:Verdana;font-size:10px;font-weight:normal;}

#yiv2496582887 .yiv2496582887green {
color:#628c2a;}

#yiv2496582887 .yiv2496582887MsoNormal {
margin:0 0 0 0;}

#yiv2496582887 o {
font-size:0;}

#yiv2496582887 #yiv2496582887photos div {
float:left;width:72px;}

#yiv2496582887 #yiv2496582887photos div div {
border:1px solid
#666666;height:62px;overflow:hidden;width:62px;}

#yiv2496582887 #yiv2496582887photos div label {
color:#666666;font-size:10px;overflow:hidden;text-align:center;white-space:nowrap;width:64px;}

#yiv2496582887 #yiv2496582887reco-category {
font-size:77%;}

#yiv2496582887 #yiv2496582887reco-desc {
font-size:77%;}

#yiv2496582887 .yiv2496582887replbq {
margin:4px;}

#yiv2496582887 #yiv2496582887ygrp-actbar div a:first-child {
margin-right:2px;padding-right:5px;}

#yiv2496582887 #yiv2496582887ygrp-mlmsg {
font-size:13px;font-family:Arial, helvetica, clean,
sans-serif;}

#yiv2496582887 #yiv2496582887ygrp-mlmsg table {
font-size:inherit;font:100%;}

#yiv2496582887 #yiv2496582887ygrp-mlmsg select,
#yiv2496582887 input, #yiv2496582887 textarea {
font:99% Arial, Helvetica, clean, sans-serif;}

#yiv2496582887 #yiv2496582887ygrp-mlmsg pre, #yiv2496582887
code {
font:115% monospace;}

#yiv2496582887 #yiv2496582887ygrp-mlmsg * {
line-height:1.22em;}

#yiv2496582887 #yiv2496582887ygrp-mlmsg #yiv2496582887logo {
padding-bottom:10px;}


#yiv2496582887 #yiv2496582887ygrp-msg p a {
font-family:Verdana;}

#yiv2496582887 #yiv2496582887ygrp-msg
p#yiv2496582887attach-count span {
color:#1E66AE;font-weight:700;}

#yiv2496582887 #yiv2496582887ygrp-reco
#yiv2496582887reco-head {
color:#ff7900;font-weight:700;}

#yiv2496582887 #yiv2496582887ygrp-reco {
margin-bottom:20px;padding:0px;}

#yiv2496582887 #yiv2496582887ygrp-sponsor #yiv2496582887ov
li a {
font-size:130%;text-decoration:none;}

#yiv2496582887 #yiv2496582887ygrp-sponsor #yiv2496582887ov
li {
font-size:77%;list-style-type:square;padding:6px 0;}

#yiv2496582887 #yiv2496582887ygrp-sponsor #yiv2496582887ov
ul {
margin:0;padding:0 0 0 8px;}

#yiv2496582887 #yiv2496582887ygrp-text {
font-family:Georgia;}

#yiv2496582887 #yiv2496582887ygrp-text p {
margin:0 0 1em 0;}

#yiv2496582887 #yiv2496582887ygrp-text tt {
font-size:120%;}

#yiv2496582887 #yiv2496582887ygrp-vital ul li:last-child {
border-right:none !important;
}
#yiv2496582887

Siggi
 

On Tue, 9 Feb 2016 at 12:00 jimmcintyre22@... [TekScopes] <
TekScopes@...> wrote:

At this point, I'm pretty sure I made a big mistake using carbon resistors
in my homebrew HV probe. I had no idea how bad their voltage coefficient
was.
I wonder if you can't use the existing HV divider in the scope to get a
better reading. If you measure the resistance of the resistors in the
divider with the scope off, and then measure the voltage over the low
element of the divider, wouldn't that give you the voltage (assuming the
scope's divider is sane)?

 

On Tue, 09 Feb 2016 18:03:45 +0000, you wrote:

On Tue, 9 Feb 2016 at 12:00 jimmcintyre22@... [TekScopes] <
TekScopes@...> wrote:

At this point, I'm pretty sure I made a big mistake using carbon resistors
in my homebrew HV probe. I had no idea how bad their voltage coefficient
was.
I wonder if you can't use the existing HV divider in the scope to get a
better reading. If you measure the resistance of the resistors in the
divider with the scope off, and then measure the voltage over the low
element of the divider, wouldn't that give you the voltage (assuming the
scope's divider is sane)?
That could work but measuring the actual divider resistance will not
be trivial and we do not know its voltage coefficient of resistance
without testing it with a calibrated high voltage source. All we can
really determine is if the cathode voltage is in regulation by
measuring the output of U1635. If we assume the voltage divider is
exactly as marked, then that makes the cathode voltage -50V * 29.65M /
500k = -2965V. The error from the input bias current is cancelled by
R1637 but the input offset voltage and offset bias from the
operational amplifier add about another +/-10 volts to that. There
are some direct substitutions for U1635 which would lower that error
by an order of magnitude but what is important is that the cathode
voltage remain stable over time and temperature; its absoluate value
is not critical.

You could make a simple high voltage probe with the appropriate
resistor(s) for less than $10 with a total error within 1.5%. To do
better than that would cost as much as buying a used high voltage
probe like a Fluke 80k-6 off of Ebay and they are only good to 1%.

 

Do you know what kind of phosphor the CRT has ?Some phosphors are optimized for photographic applications instead of the human eye.Make sure the 10X horiz mag is not turned on; it can dim the trace significantly if the trigger rep rate is low.If the timebase has variable holdoff, that can affect it too. HankC, Boston WA1HOS

[Non-text portions of this message have been removed]

Jim McIntyre
 

It's P31. Sweep mag is off. (I may not be the brightest bulb in the string, but I'm not overlooking something like this...)

I'll have a HV probe tomorrow to check the cathode and PDA voltages and will report back. Let's just assume, for the moment that these are OK.



There's something fundamental here I'm missing...



Assume we have a static display (no blanking because all chop modes are off, and we're in XY mode)


The cathode is at -3 kV. Doesn't the control grid (Pin 3 of the CRT) need to be slightly more negative? ie -3.050 kV or something of that order? How does this happen?

Ed Breya
 

There are typically two ways to get DC-coupled Z-axis control on scope CRTs. One is to generate a high voltage, similar to that used for the cathode supply, and stack it onto the ground-referenced Z-axis signal. This is what's applied to the control grid. The other way is to use a DC-restorer circuit, which is basically an AC-coupled charge pump, with the AC swing proportional to the ground-referenced Z-axis signal. It's coupled capacitively and rectified, and stacked onto the cathode supply, then applied to the control grid.

Either way, the control grid is indeed usually negative with respect to the cathode, by maybe up to 100V or so, and it controls the beam current. There are pluses and minuses to each approach.

Ed

Jim McIntyre
 

OK - reading p 3-65 (Control Grid DC restorer) from the 7904 manual...

Starting to make sense...

 

Diagnostics on DC restorers are not much fun. Measurements require a
high voltage probe and component tests require a high voltage power
supply.

The most common failure is leakage in the high voltage capacitors and
diodes but since testing would require their removal anyway, it is
usually easiest to just replace them if a fault is suspected. New
parts are not expensive so if I had to order them, I would rebuild
both the z-axis and focus DC restorers.

The grid bias control and the output of the z-axis amplifier together
clip the peak-to-peak value of the high voltage drive from the
transformer. The clipped peak-to-peak voltage drives the charge pump
to produce the high negative DC output voltage. High frequency AC
from the z-axis amplifier it capacitively coupled to the high voltage
negative output.

Before messing with the DC restorer, I would verify that the output
from the z-axis amplifier is correct. This can be done with just a
voltmeter if necessary.

On 09 Feb 2016 21:23:08 -0800, you wrote:

OK - reading p 3-65 (Control Grid DC restorer) from the 7904 manual...

Starting to make sense...

Jim McIntyre
 

I looked at TP1876 with a DVM and saw +20 to +80-ish volts, as I varied the intensity control.

What's the recommended way to check the Z-axis amp?

David DiGiacomo
 

On Wed, Feb 10, 2016 at 12:52 AM, David <@DWH> wrote:
Diagnostics on DC restorers are not much fun. Measurements require a
high voltage probe and component tests require a high voltage power
supply.
In the ones I have repaired, the diode leakage has been pretty obvious
when tested with a 50V power supply. Of course that might not always
be the case. As you say, the parts are cheap.