Oscillator stops when probed.

[David C. Partridge]

I've an old Tektronix 184 with an 10 MHz OCXO hooked up to a nuvistor
tetrode (valve). The crystal is connected between the cathode and the
first grid. The cathode end of the crystal connects to earth through a
variable 6 to 10 micro Henry inductor.

2nd grid should be at about 50V (I think) as there's a 75V zener between
there and the 125V anode supply, but in fact it measures at around 70V DC on
my DMM???

The anode (plate) is connected to a tank tuned to the 5th harmonic.

Without a 10x scope probe at the junction between the inductor, cathode and
crystal, the system produces an output, so the oscillator is clearly running
when I don't try to probe it.

If I put the scope probe very near the junction, it acts as a aerial, and
picks up the signal at a very low level. The moment I touch the connection
with the probe tip the oscillator stops (sometimes I see the correct output
of about 70V peak to peak as shown on the schematics for a fraction of a
second).

I suspect something is wrong with the circuit, but can't work out what it
might be ...

Why is this likely to be happening?

 

[Michael Black]

I've an old Tektronix 184 with an 10 MHz OCXO hooked up to a nuvistor
tetrode (valve). The crystal is connected between the cathode and the
first grid. The cathode end of the crystal connects to earth through a
variable 6 to 10 micro Henry inductor.

2nd grid should be at about 50V (I think) as there's a 75V zener between
there and the 125V anode supply, but in fact it measures at around 70V DC on
my DMM???

The anode (plate) is connected to a tank tuned to the 5th harmonic.

Without a 10x scope probe at the junction between the inductor, cathode and
crystal, the system produces an output, so the oscillator is clearly running
when I don't try to probe it.

No divider probe? That's likely the problem.

An oscilliscope may have 1meg of resistance across it's input, but when
you start connecting a cable to that input you won't be seeing 1meg impedance
at the other end. The capacitance of the cable will load down the circuit
under test. The higher in frequency, the more significant the cable will
load down the circuit.

The divider probe ensures that the circuit under test will not see
the load of the cable. It will see a relatively small load. This is
exactly what the probe is for, the fact that it divides the input level
by 10 is only a byproduct of this.

Michael

 

[[email protected]]

Even a 10x probe can sometimes load a crystal oscillator sufficiently
to kill it. Is there a particular reason you need to probe the circuit
at the crystal?

 

[David C. Partridge]

OH! I didn't realise that an 10 meg loading along with the capacitance of
the probe could quite normally shut a crystal oscillator down.

I was using a x10 probe, so the load shouldn't have been much.

I was just trying to check everything out according to the service manual
and waveforms throughout the circuit as this piece of kit is new to me (but
very old!). A through check through seemed in order.

The service manual clearly shows a (drawn) scope shot of the output on the
xtal circuit, so I expected to be able to probe it. Maybe a 100x probe
might do the trick, but I don't have one of those.

David

 

[Asimov]

"David C. Partridge" bravely wrote to "All" (19 Oct 05 16:30:45)
--- on the heady topic of "Oscillator stops when probed."

DCP> Reply-To: "David C. Partridge" <[email protected]>
DCP> Xref: core-easynews sci.electronics.repair:345698


DCP> I've an old Tektronix 184 with an 10 MHz OCXO hooked up to a nuvistor
DCP> tetrode (valve). The crystal is connected between the cathode and
DCP> the first grid. The cathode end of the crystal connects to earth
DCP> through a variable 6 to 10 micro Henry inductor.

DCP> 2nd grid should be at about 50V (I think) as there's a 75V zener
DCP> between there and the 125V anode supply, but in fact it measures at
DCP> around 70V DC on my DMM???

DCP> The anode (plate) is connected to a tank tuned to the 5th harmonic.

DCP> Without a 10x scope probe at the junction between the inductor,
DCP> cathode and crystal, the system produces an output, so the oscillator
DCP> is clearly running when I don't try to probe it.

DCP> If I put the scope probe very near the junction, it acts as a aerial,
DCP> and picks up the signal at a very low level. The moment I touch the
DCP> connection with the probe tip the oscillator stops (sometimes I see
DCP> the correct output of about 70V peak to peak as shown on the
DCP> schematics for a fraction of a second).

DCP> I suspect something is wrong with the circuit, but can't work out what
DCP> it might be ...

DCP> Why is this likely to be happening?


If this is with the direct probe, then the stray and cable capacitance
is surely damping the oscillation. If it is stalling with the 10x
probe then the tube gain might be marginal. If you know your probe's
input capacitance then use a series cap 10 or 100 times smaller in
value to increase the impedance further but you will need to set the
gain correspondingly higher by 10 or 100. A voltage divider works
wonders here. BTW, what kind of trouble do you suspect having?

A*s*i*m*o*v

.... This message transmitted on 100% recycled photons.

 

[Jim Adney]

OH! I didn't realise that an 10 meg loading along with the capacitance of
the probe could quite normally shut a crystal oscillator down.

Not all 10x probes are alike. I believe 10x probes exist with tip
capacitances anywhere between about 7pF and 40pF, maybe higher. The
low capacitance ones tend to be pricier, but it makes a siginificant
difference once you get up into that frequency range.

A random 100x probe won't necessarly have lower C than a random 10x
probe.

-

 

[David C. Partridge]

What kind of trouble - well, the output waveform after the pickup
transformer is quite distorted, so I though I'd check things closer to the
source so to speak.

Yes, the gain on the tetrode could well be marginal, if it's the original
(likely) it is now almost 40 years old.

Looks like I may need to get a job lot of 7587s (almost but not quite pure
unobtanium).

"David C. Partridge" bravely wrote to "All" (19 Oct 05 16:30:45)
--- on the heady topic of "Oscillator stops when probed."

DCP> Reply-To: "David C. Partridge" <[email protected]>
DCP> Xref: core-easynews sci.electronics.repair:345698


DCP> I've an old Tektronix 184 with an 10 MHz OCXO hooked up to a nuvistor
DCP> tetrode (valve). The crystal is connected between the cathode and
DCP> the first grid. The cathode end of the crystal connects to earth
DCP> through a variable 6 to 10 micro Henry inductor.

DCP> 2nd grid should be at about 50V (I think) as there's a 75V zener
DCP> between there and the 125V anode supply, but in fact it measures at
DCP> around 70V DC on my DMM???

DCP> The anode (plate) is connected to a tank tuned to the 5th harmonic.

DCP> Without a 10x scope probe at the junction between the inductor,
DCP> cathode and crystal, the system produces an output, so the oscillator
DCP> is clearly running when I don't try to probe it.

DCP> If I put the scope probe very near the junction, it acts as a aerial,
DCP> and picks up the signal at a very low level. The moment I touch the
DCP> connection with the probe tip the oscillator stops (sometimes I see
DCP> the correct output of about 70V peak to peak as shown on the
DCP> schematics for a fraction of a second).

DCP> I suspect something is wrong with the circuit, but can't work out what
DCP> it might be ...

DCP> Why is this likely to be happening?


If this is with the direct probe, then the stray and cable capacitance
is surely damping the oscillation. If it is stalling with the 10x
probe then the tube gain might be marginal. If you know your probe's
input capacitance then use a series cap 10 or 100 times smaller in
value to increase the impedance further but you will need to set the
gain correspondingly higher by 10 or 100. A voltage divider works
wonders here. BTW, what kind of trouble do you suspect having?

A*s*i*m*o*v

.... This message transmitted on 100% recycled photons.