Michael A. Terrell wrote:
David Farber wrote:

Michael A. Terrell wrote:
David Farber wrote:

"Michael A. Terrell" wrote in message
...

David Farber wrote:

One of my ~30 year old Sound Technology distortion/power
analyzers has a problem. It's been sitting around for a number
of years because I had a spare. The symptom is that when you are
measuring distortion and move the rotary selector switch one
step from the 1% range to the .3% range, the meter goes from a
near zero reading to full deflection and then some. If I feed
the signal output to my other analyzer, the distortion is very
low so I know the oscillator is ok.

Here is a copy of the schematic:
http://members.dslextreme.com/users/.../st-1700b.html

At the output of U202, pin 6, the signal goes from zero (meter is
working properly) to a nice sine wave (meter pegs) when the
switch is rotated.to the .3% range and below. The signal is too
low to measure at the input of U202 no matter where the switch
is. There is a very detailed circuit description in the owner's
manual. However I have a general sense that there's an open
circuit somewhere causing the gain to go full blast. I cleaned
the switches but it wasn't of any help. Anyone have any clever
ideas as to how to pinpoint the trouble?

Troubleshoot it. Is the switch part of an attenuator, or does
it switch in more gain for the last range? Look to see if the op
amp is oscillating. Look for bad electrolytics on the supply
rails. It isn't rocket science. A distortion analyzer is a
tunable notch filter and attenuator, followed by an AC voltmeter.


The first thing I did was to check all the caps. They're ok.

You can see from the schematic that the switch is part of the
attenuator and that U202 is before the attenuator switch. So my
question is why does U202 suddenly have a wild signal swing when
switched to the next lower step?

Thanks for your reply.


You have a lot of DC coupled stages, and some are not very good op
amps. A dc offset can make a string like that unstable. Did you
look at pin 6 of U202 with a low capacitance scope probe? How does
the DC voltage there compare to the working unit?

Are the op amps marked 2605 Harris HA2605? If so, they have been
out of production for some time. Metal cased op amps, (and other
metal cased ICs) started disapearing 10 years ago.

You might luck out and fix it with a couple .1 caps to ground of
the supply pins of U202 if it's oscillating. Another thing to
check is all the mounting screws for the PC board and any shields.
Loosen the screws and tighten them up to remove any oxide.
Generally op amps oscillate from defective bypass caps, or signals
being coupled from another circuit. Also, did you test the
resistance for the contacts in that mode after you cleaned the
switches? An open or high resistance contact cn upset the circuit.

One question. Is the sine wave close to the frequency the filter
is tuned to?

Hi Michael,

The oscillating is at the same frequency of the signal output (1kHz)
so I assume that is not a parasitic oscillation. I make a mistake in
identifying the correct IC, I should have said U203 is where the
wild voltage swing takes place.

Here are some readings I made with my low capacity 10:1 probe:

U202, pin 6 output, selector set to 1% range, 4.8 Vp-p.
U203, pin 6 output, selector set to 1% range, 0.0 Vp-p.

U202, pin 6 output, selector set to .3% range, 3.8 Vp-p.
U203, pin 6 output, selector set to .3% range, 2.8 Vp-p.

I was about to end my message at this point when I went back and
re-read your message to make sure I didn't leave anything out. I
noticed the part where you mentioned comparing the good analyzer to
the bad one. So I disassembled the good one and checked the dc
voltages first. They were ok as were the dc voltages on the bad one.
On the AC side, of course the voltages were zero at the output of
U203 on the good one. Here comes the fun part. While probing around
the bad one again, I could hear the meter pegging as I was touching
different pins trying to hold my hand steady. All of a sudden, the
signal was gone and the meter stopped pegging. It seems to be
working correctly now. BUT if I move the frequency select switches
from X100 on the top row, X10 on the next row to X10 on the top row,
X100 on the next row (still maintaining a 1kHz signal), the
distortion is much higher, over 3%. If I gently tap the X100 switch
on row 2, the meter jumps around like a bad tape monitor switch but
never drops below 3%. There is a yellow sticker on the shield that
covers the frequency selector switches that says, "WARNING DO NOT
SPRAY SWITCHES. PUSHBUTTON SWITCHES ARE PERMANENTLY LUBRICATED."
etc. etc. Perhaps the switches are causing all of these symptoms?


Pushbutton switches are a pain in the ass. From the age of the
equipment I would guess that they were made by ALPS? I have had a lot
of bad interlocking pushbutton switches over the years. The silver
plating is low grade, and I've even seen it flake off. The contact
lube used in some switches is a low grade grease that washes away with
contact cleaner, exposing the contacts to contaminants. Mouser
carried some push button switches the last time I looked. I always
used some GC Tunerlube to replace the lube in pushbutton switches
when I had to clean them. Check out the http://www.mouser.com site
to see if you can find new switches, or turn the unit of and press
the buttons 50 to 100 times each to wear away any oxide, then
relubricate the switches.

If you can't find what you need, post some pictures of the switches
and I'll see if I have any in my collection.


Switchcraft made my favorites, but they are quite expensive.

Hi Michael,

Since I now can get good readings on the 100 x 10 range, I am not going to
worry about the 10 x 100 range. The switches are buried under some pc
boards. Not worth the trouble to fix.

Thanks for your reply.
--
David Farber
David Farber's Service Center
L.A., CA