From: "Boris Pinson"
Subject: Re: spurious oscillations
Date: 21 Sep 2002 13:26:53 GMT
Organization: The Salivation Army of Alibama
X-Newsreader: Forte Agent 1.8/32.548
On Sat, 21 Sep 2002 07:31:53 GMT, Robert Baer wrote:
> Gabriel Horner wrote:
> > Hi,
> > I'm trying to repair a pulse generator which for now only oscillates at
> > 10 MHz. I've verified that the spikes which trigger the pulses are correctly
> > made when I vary the frequency knob but I've also discovered a 500 mV
> > peak to peak 10 MHz oscillation. As I trace the signal down further, the
> > spikes die away and the generator ends up triggering on these
> > oscillations. The whole problematic section is composed of transistors and
> > includes a Schmitt trigger. Any ideas on how to find the troublemaker if
> > my diagnosis is right?
> > Thanks,
> > Gabriel
> Nominally speaking, a pulse generator does not oscillate.
Sure they do (or can). An astable multivibrator is nothing but an
oscillator, and a Schmitt trigger is a variation on the monostable
multivibrator design. In all cases, very high gain is employed in
order to drive the transistors into saturation -- to make them go to
one and only one state, rather than to amplify their input current.
When one of the circuit components drifts, or when one of the
transistors starts to fail, it is common for the high gain to be lost,
and for oscillations to set in via the feedback mechanism which is
already present. There is no longer complete saturation of one of the
stages, and that allows some conventional amplification to take place.
Usually, the designer adds additional circuitry intended to suppress
such oscillations, but it may not be present in simpler equipment. In
any case, the old tried and true methods will apply.
Check out the passive components (resistors, inductors, capacitors)
first, and see if one of them has drifted or failed with time. If you
have a voltage chart (as in a Sam's Photofact schematic), start with a
check for correct operating voltages at the various solder points.
You can also try "freezing" various components by means of a special
spray can you can purchase. Sometimes, lowering the temperature of a
resistor or a capacitor will reveal the problem part.
In these days of printed circuit boards and VLSI, it is common to swap
out the board rather than to test components, but it looks as though
the old methods will be needed here instead. I'll bet that a passive
component has changed value, or that a solder joint is bad. (I'd
suspect the transistors last.)