Analog Scope about to fail?

[fstd]

Hi everybody,

I've got this old analog Tektronix scope (2235 or AN/USM-488 which I think is the mil spec version of that) and it's worked fine for me so far.

It does something that make me suspect there's something wrong, though; the trace occasionally drops down slightly, then jumps back in place. 
This seems independent of vertical scale, happens on both channels, signal present or not, even when grounded.  It's also independent of the horizontal scale.



What could that be?

[tggzzz]

Check all the PSU voltages and ripple.

See if any of the PSU voltages change when the trace position changes.

See if any switch contacts need cleaning, probably with IPA on a piece of smooth paper.

Wiggle connectors to dislodge any oxide buildup.

If only one trace moves then the problem is in one channel. If both move then it is either after the two channels are multiplexed onto the single beam, or in the global PSU.

[Yansi]

Vertical output transistor breakdown.

[fstd]

Thanks -- this might be a dumb question but can I use the scope to probe its own power rails?  Because otherwise I'd need another scope to look at the power rails

[fstd]

Vertical output transistor breakdown.

Would that be likely a failure mode of the transistor itself, or of whatever is driving the transistor?

[tggzzz]

Thanks -- this might be a dumb question but can I use the scope to probe its own power rails?  Because otherwise I'd need another scope to look at the power rails

Use a voltmeter, AC setting for the ripple.

Most scope's vertical accuracy specification is 2%.

Get the scope's manual and follow the faultfinding procedures. The tekwiki q140.com probably has it, if not google is your friend!

[Yansi]

Vertical output transistor breakdown.

Would that be likely a failure mode of the transistor itself, or of whatever is driving the transistor?

Failure mode of the vertical transistor. I've seen this failure many times in old CROs. I admit, that not Tektronix, but mostly the vertical transistors are heavily stressed. High voltage (that may have become even higher due to many reasons), heat dissipation, etc.
And due to manufacturing defects or due to the old high voltage transistor technology not being perfect, the breakdown voltage may decrease with time,  until it fails completely.

Problem is, that checking such transistor is not easy, as the failure mode is not easily detectable, unless testing in circuit. So get the relevant schematics of that scope, and get another scope to probe the input signals to the vertical output stage and probe the very output of that stage.

[akimpowerscr]

The first thing I would do would be to check whether or not this fault can be caused by vibration.

To do this, you must lightly tap in working all the printed circuits with an isolated object (the handle of a screwdriver for example)

This makes it possible to detect bad contacts and cold solders and their approximate location.

[med6753]

Vertical output transistor breakdown.

Would that be likely a failure mode of the transistor itself, or of whatever is driving the transistor?

I wouldn't go there just yet. The 2235 PSU is a notoriously terrible design and failure prone. As been suggested monitor not only ripple but voltage level. It may be hard to spot with a DMM because of refresh rate but an analog meter, such as a VTVM or VOM, will spot any quick voltage sag immediately. 

[2N3055]

Check contacts, including wires going to the CRT deflection pins.