ddwyer wrote:


A multimode delay line oscillator can be
achieved by introducing a peaked gain response at the required overtone.

How? A tuned filter? There's a hole in the bucket... (circular
implementation)

Alternatively with more than 360 degrees phase shift the oscillation
which cannot change frequency instantaneously can be induced by
continuously increasing phase and jumping from 360 to 0 degrees to sweep
over multimodes.

Wha? Could you expand on that?

Thanks

--
Scott

**********************************

DIY Piezo-Gyro, PCB Drill Bot & More Soon!

http://home.comcast.net/~scottxs/

**********************************

In article vLWrc.13935$JC5.1310262@attbi_s54, Scott Stephens
writes
ddwyer wrote:


A multimode delay line oscillator can be
achieved by introducing a peaked gain response at the required overtone.

How? A tuned filter? There's a hole in the bucket... (circular
implementation)
Not quite.
The delay line can be multioctave and the phase slope will be
proportional the the delay. It may go through many 360 deg phase
rotations over its bandwidth.
An amplifier with much shallower phase slope can still have sufficient
selectivity to determine which particular 360 deg rotation is
oscillated. This has been done successfully with PAL bulk acoustic delay
lines.
The following is more contentious but I think it will work

Alternatively with more than 360 degrees phase shift the oscillation
which cannot change frequency instantaneously can be induced by
continuously increasing phase and jumping from 360 to 0 degrees to sweep
over multimodes.

Wha? Could you expand on that?
Not a lot. progressively changing phase with varicaps cannot carry on
but the circuit does not know the difference between 360deg and 0 deg.
Goniometers? used to do this for direction finding.
Can also be done with mixers but that is more of a circular argument.


Thanks


--
ddwyer

ddwyer wrote:

In article vLWrc.13935$JC5.1310262@attbi_s54, Scott Stephens

ddwyer wrote:

A multimode delay line oscillator can be
achieved by introducing a peaked gain response at the required overtone.

The delay line can be multioctave and the phase slope will be
proportional the the delay. It may go through many 360 deg phase
rotations over its bandwidth.

The phase slope is proportional to the delay. When I think of phase
slope, I think of inductance or capacitance, reactance based phase shift.

An amplifier with much shallower phase slope

An amplifier with broad bandwidth, having the real, resistive impedance
greater than the reactive impedance?

can still have sufficient
selectivity to determine which particular 360 deg rotation is
oscillated. This has been done successfully with PAL bulk acoustic delay
lines.

I've played around with ring oscillators made from 3 (or more)
transistors, 3 FETs and my favorite, 3 CMOS unbuffered inverter gates. I
find I can tune these sine-wave ring oscillators between 8 and 20 : 1
range, by changing the bias gate voltage and source current. Perhaps
more, after I try a few tricks. 20:1 isn't bad, but I don't see why I
shouldn't be able to get 100,000:1 out of a transconductance-tuned phase
shift oscillator. I guess you'd say these have a very shallow "phase
slope".

I was hoping to use them as a wide-band tuned circuit, by setting the
gain just under the oscillation point. But I found they are very
sensitive to harmonics.

I can bet a delay line oscillator is going to have this same problem then?

I've tried several ways to make a clean wide-range sine-wave oscillator
and regenerative variable tuned circuit.

--
Scott

**********************************

DIY Piezo-Gyro, PCB Drill Bot & More Soon!

http://home.comcast.net/~scottxs/

**********************************

ddwyer wrote:

In article vLWrc.13935$JC5.1310262@attbi_s54, Scott Stephens

ddwyer wrote:

A multimode delay line oscillator can be
achieved by introducing a peaked gain response at the required overtone.

The delay line can be multioctave and the phase slope will be
proportional the the delay. It may go through many 360 deg phase
rotations over its bandwidth.

The phase slope is proportional to the delay. When I think of phase
slope, I think of inductance or capacitance, reactance based phase shift.

An amplifier with much shallower phase slope

An amplifier with broad bandwidth, having the real, resistive impedance
greater than the reactive impedance?

can still have sufficient
selectivity to determine which particular 360 deg rotation is
oscillated. This has been done successfully with PAL bulk acoustic delay
lines.

I've played around with ring oscillators made from 3 (or more)
transistors, 3 FETs and my favorite, 3 CMOS unbuffered inverter gates. I
find I can tune these sine-wave ring oscillators between 8 and 20 : 1
range, by changing the bias gate voltage and source current. Perhaps
more, after I try a few tricks. 20:1 isn't bad, but I don't see why I
shouldn't be able to get 100,000:1 out of a transconductance-tuned phase
shift oscillator. I guess you'd say these have a very shallow "phase
slope".

I was hoping to use them as a wide-band tuned circuit, by setting the
gain just under the oscillation point. But I found they are very
sensitive to harmonics.

I can bet a delay line oscillator is going to have this same problem then?

I've tried several ways to make a clean wide-range sine-wave oscillator
and regenerative variable tuned circuit.

--
Scott

**********************************

DIY Piezo-Gyro, PCB Drill Bot & More Soon!

http://home.comcast.net/~scottxs/

**********************************

In article vLWrc.13935$JC5.1310262@attbi_s54, Scott Stephens
writes
ddwyer wrote:


A multimode delay line oscillator can be
achieved by introducing a peaked gain response at the required overtone.

How? A tuned filter? There's a hole in the bucket... (circular
implementation)
Not quite.
The delay line can be multioctave and the phase slope will be
proportional the the delay. It may go through many 360 deg phase
rotations over its bandwidth.
An amplifier with much shallower phase slope can still have sufficient
selectivity to determine which particular 360 deg rotation is
oscillated. This has been done successfully with PAL bulk acoustic delay
lines.
The following is more contentious but I think it will work

Alternatively with more than 360 degrees phase shift the oscillation
which cannot change frequency instantaneously can be induced by
continuously increasing phase and jumping from 360 to 0 degrees to sweep
over multimodes.

Wha? Could you expand on that?
Not a lot. progressively changing phase with varicaps cannot carry on
but the circuit does not know the difference between 360deg and 0 deg.
Goniometers? used to do this for direction finding.
Can also be done with mixers but that is more of a circular argument.


Thanks


--
ddwyer