Re: 7904 Mainframe damaged

Chuck Harris

Hi John,

I am not like that. I have done all of these dumb
things myself (and more). I am quite happy to call
myself dumb, but you will have to decide how harsh
you are willing to be with yourself.

If there is no zap mark on the ground clip, there was
very little current involved. I don't believe that it
caused your main problem.

I have burned ground wires clean off during probe
misadventures on my 7000 series scopes without any
untoward result... Phone switch -48V supply, I'm
looking right at you!... (I was even warned, sigh...)

Anecdotal evidence, I know, but I am very certain that
tektronix tested these scopes for sensitivity to that
accident during their design process.... And way, way
worse! And a heck of a lot of us have repeated the,
uhmmm, testing.

The most I can say that has ever happened when I did
such stunts, was the scope triggered from the RF
generated by the zap.

What I think we have here is two events that although
coincident'ish in time, were totally unrelated.

The favorite time for a 7904 to eat a tantalum is
within a second, or so, of turn on. Indeed, that
is a very common time for any component failure.

I have never had a tantalum fail during operation...
they seem to self repair during operation. The truly
explosive failures happen when they are employed on
high current supplies, without any current limitation.

I have had more than a few 6.3V 100uf tantalums on 5V
logic power busses scare the Bejezus out of me, on
power application... but I digress.

The 7904 power supply is very sensitive to over current.
It will go into tick mode for some combinations of
plugins... for instance, mine cannot power a 7L18.

The 7904 is an early 7000 series scope, and was made
long before any of the current hog plugins were made.

It is fine with any combination of 7A and 7B plugins
that I have tried.

The 50V supplies are favorites for knocking a 7904
supply down into kick mode. They are low current, and
a little 4.7 ohm 1/4 watt resistor can do the job on
power up.... though it would burn up if the short
happened after power is up.

You say it wouldn't respond to any controls. Really?

-Chuck Harris

JJ wrote:


There are no burn marks on the scope probe, or ground wire, or probe
alligator clip. The probe works fine. No burn marks on the 7A26. I tested
it everything out on my 7834 storage scope - works fine. All plugins work
fine, There are no burn marks inside the plugin area. I have inspected all
boards that are visible with a magnifying glass - no signs of component
failure. I can't inspect the main interface board until I remove it.

I was measuring a voltage on a leg of a bridge rectifier (110v) when there
was a snap - and I realized what I had done. The scope itself didn't go off
- the CRT was still active with a noise signal on the CRT - but the scope
wasn't responding to any actions on the controls. I was hoping against all
hope that if I powered it off and then back on again that I would get
lucky. Or, maybe a fuse when it didn't come on - all fuses were fine.

Please don't tell me how dumb I was - tell me something that I don't
already know. :).


On Wed, Apr 11, 2018 at 1:35 PM, Chuck Harris <> wrote:

The problem is, the current from the probe's ground
lead is directly connected to the chassis at the BNC
connector on the plugin's front panel, and at the
finger stock around the front panel, and at the side
rails along the bottom and top of the plugin slot...

And then the chassis of the scope is directly connected
to the ground pin on the power cord... Which was
connected to the ground pin on the isolation tranny,
which was connected to the building's ground, and at
the service panel to the mains ground lead (neutral)...

So, how much current were we talking here?

Enough to vaporize the ground wire? Not from what we've

Did it even burn the plating off of the ground clip?

Enough to blow a fuse somewhere? Not from what we've
heard.. AFAICR.

And, exactly what was shorted by the ground wire?

Mains? A filter cap on a switching supply? An 800V
power supply on a monster tube amp?

I'm having more than a little trouble believing the
presumed cause fits the symptoms. Something else is
going on.

A very popular failure in the 7904's 50V power supplies
is a 50V tantalum capacitor on the Horizontal amplifier
board. It doesn't seem to show up on the schematic in
the obvious place, and I don't remember if it is on the
+50, or the -50V supply, but it is there. Usually, when
it shorts, it incinerated a little 10 ohm resistor placed
there as a fuse, and eliminates the horizontal amplifier
from the circuit. If the resistor is not killed, the
resistor takes down the current overload, and shuts down
the supply.

I'd look there.

-Chuck Harris

Robert Hay wrote:
Just to re-cap, your initial message indicated your probe ground lead
contacted a
high voltage point on a device under test with a common ground. This
would raise the
scope vertical input ground and cause high current to find its way thru
your scope.
This apparently caused a failure in the scope.

Your resistance tests don't point to a short on any of the LV supplies.
The author
of the old article used a VOM so I would tend to compare VOM readings to
those given
in the article. If you tested using both Fluke and VOM did you find
much differences
in readings?

The minus 51 V load resistance is much higher than the article
measurement. The plus
5 is lower resistance. So, those might be the places to look. Did you
check the
diodes and caps for the -51 volt supply? If they are good, you may be
looking for
something failed open on the mainframe side since your measurement of
31.6k with the
LVPS disconnected is much higher than the 2k given in the article.

Usually, the tic-tic mode is the power supply letting you know there is
a short on
one of the supplies and saving itself from further damage. As far as
testing it
without load to see if you still get the tic-tic or not it would be good
if one of
the more experienced members with testing the 7904 supply jumps in here.


On 4/10/2018 4:37 PM, JJ wrote:
In reading that doc, I'm concerned that the problem is in that inverter
control chip. That's probably impossible to find. I found the schematic
what's inside it:

Z-axis Board voltages PSU in Mainframe
+130 8.3k
+50 3.7k
+15 90.1 ohms
-15 205 ohms
-50 4.2k
+5 47 ohms

+5V lamp 2 ohms

Z-axis Board voltages PSU out of Mainframe
+130 9,9k
+50 4.6k
+15 92.3 ohms
-15 216 ohms
-50 31.6k
+5 48 ohms

+5V lamp 23k

On Tue, Apr 10, 2018 at 6:55 PM, Robert Hay <> wrote:

Ok, you said the resistances were very high, how much higher than those
suggested by the article?

+130 volts 6k
+50 2k
+15 90 Ohms
-15 100 Ohms
-50 250 Ohms
+5V lamp 800 Ohms


On 4/10/2018 1:28 PM, JJ wrote:

Bob, I found that the diodes seem OK when I disconnected one lead and
measured both resistance and diode forward and reverse voltage. I
that the short across those two diodes was due to a shorted capacitor
the rectifier board - I was actually measuring the winding tap
through a weird path. I reconnected the PSU back into the mainframe
changing the cap. There was no glory - a very low tick coming from the

I'll need to continue debug by following the procedure in the that
document. I measured the resistances in z-axis board as the procedure
indicates while the PSU was out of the mainframe - they are pretty
much in
line - a couple of resistances were much higher than the table - none
lower . I'll measure them now while the power supply is in the
Also, I get different results using a VOM and a DVM. Maybe there's
something wrong with my VOM - it's pretty old. Is a VOM required to
proper measurements as indicated in the table?


On Tue, Apr 10, 2018 at 2:41 PM, Robert Hay <> wrote:

They say in the article that a low resistance indicates a problem in
mainframe which is true if you are looking at something shorting to
ground. But, in your case I would not discount that high current
open circuit while trying to get back to the source. I suppose you
looked for any signs of burning or smoked components, leads, traces,


On 4/10/2018 7:44 AM, JJ wrote:

Yes the resistances were very high. According to that doc, the issue
be on the PSU. I'm going to put the PSU back into the mainframe,
the cables, and make those voltage measurements on the LV regulator
I didn't check for voltages before taking the PSU out because I
at that time that the PSU needs to be under minimum load.

I found a shorted cap C1360 on the rectifier board (+54v filter cap
output of Pi filter). I'm going to lift one lead to ensure it's the
problem. I'm also going to validate that the two power diodes that I
are shorted - by disconnecting the wires going to those diodes and
measuring. If the diodes are bad, I'll replace the bad 10A diodes
5A diodes in parallel temporarily - hoping that's OK, I'll then
hook up
mainframe's cables to the PS and check the voltages. I think the 3
are long enough for the PSU to hang out the back. If not, I plan on
extender cables.


On Tue, Apr 10, 2018 at 10:16 AM, Robert Hay <> wrote:

Have you checked resistances as suggested in the Service Scope

And then voltages with the supply outside the scope but still


On 4/10/2018 6:38 AM, JJ wrote:

Hi Tony, So, based on your findings, you have confirmed that the

needs to be under load in order to be functional. I will check
all the
on those 3 boards as well. Thank you for the info - it's very,
I hope the experts on this forum can help you out with the
your 7904's issues.


On Tue, Apr 10, 2018 at 9:29 AM, Yiu On Tony C via Groups.Io <> wrote:

Hi JJ :

I am from Hong Kong ! I just complete the repair of PSU of
TEK7904 , I
lucky , I find a SHORT E-Cap. once I replace it and turn on PSU
alone ,
still sound shut-down ! once I connect to main unit , it come
I found few E-Cap short on A9 H amp board, A2 Main interface
A12 Rectifier board
For the power diode , I did check each by de-solder the to wire
My 7904 still have other issue , H ok but the read-out still
units , Y are not function -- position knob no response .

RegardTony CheungAPR 10 2018

From: JJ <>
Sent: Tuesday, April 10, 2018 7:16 PM
Subject: Re: [TekScopes] 7904 Mainframe damaged

Dave, I went through hundreds of files in the files section of
and wasn't able to identify the schematic describing the minimum
requirements for power supplies. Tried searching, browsing, ad
glory. Are you sure that Jerry uploaded it?


On Mon, Apr 9, 2018 at 4:54 PM, Dave Daniel <>

On some of the 7xxx supplies (SMPS supplies, I believe), one
needs to

a minimum load on the power supply output in order for the
supply to
I don't remember if that is true for the 7904 power supply.

Jerry Massengale built one of these loads. I think he uploaded a
the schematic to the files section back in maybe2015.


On 4/9/2018 12:22 PM, JJ wrote:

I removed the power supply from the mainframe - it's on my

I'll remove the wire from the diode terminal and check to see
if the
is actually shorted as others have suggested. I was thinking of
the 10A diodes with two 5A diodes in parallel that I have
parts bin temporarily to see if I get all the other voltages
back -
now there are no voltages at the test points of the rectifier

low voltage regulator. I wouldn't think there would be that
the supply removed? That way I can continue to debug. Thoughts?

On Mon, Apr 9, 2018 at 2:06 PM, Jeff Urban <

Incidentally, that stuff about the mini EMP is not very likely.

does it actually have to happen at a strength to do something,

needs to be oriented the right way to do something. Also I erred
vaporized foil on the board, it was about ½ cm., not inch.

Anyway, if you are sure about the diodes I guess you know
what to

possible. It might be rough to gt everything hooked back up
for a

without actually assembling it. Changing the diodes right
away ?
do that. And of course watch for mounting screws that also

ground, that has tripped me up a couple of times.
Since you have one + and one - diode bad, assuming you don't
have a
to ground, I would check see if there is a short between the +

of that supply. At this time I have no idea what those sources

could be a series arranged push pull output to something and
the output devices could be shorted, the load isn't low enough

or maybe even electrostatic, to read a short to ground. It

Join to automatically receive all group messages.