"Paul Keinanen" wrote in message
...
On Sun, 06 Jun 2004 17:59:00 -0500, Ed Bailen wrote:

The last issue of QST had a short article on the CircuitMaker Student
edition circuit design package. I downloaded it and spent several
hours trying various oscillator designs (including the phase shift
oscillator discussed in the same issue of QST). I couldn't get any of
the designs to oscillate! I was mainly trying Colpitts oscillators.

Typically, an oscillator is just a noise amplifier with some frequency
selective feedback. In a real oscillator, there is always some wide
band thermal noise present, which is amplified, some selected
frequencies are circulated back to the input and reamplified and so
on. Finally, there is going to be a narrow spectral line and a wide
band noise floor due to the thermal noise, assuming of course that the
phase, amplification and feedback loss conditions are correct.

If the simulator does not simulate the thermal noise inherent in all
resistors and semiconductors in temperatures above absolute zero, it
is not very likely that the design would oscillate. Inject a very low
level wide band noise into the amplifying stage and the oscillation
should start. However, I have never seen the CircuitMaker or how to
generate a wide band noise signal with it.

Injecting a pulse usually works with SPICE, I've found.

Leon

"Leon Heller" wrote in message
...
"Paul Keinanen" wrote in message
...
On Sun, 06 Jun 2004 17:59:00 -0500, Ed Bailen wrote:

The last issue of QST had a short article on the CircuitMaker Student
edition circuit design package. I downloaded it and spent several
hours trying various oscillator designs (including the phase shift
oscillator discussed in the same issue of QST). I couldn't get any of
the designs to oscillate! I was mainly trying Colpitts oscillators.

Typically, an oscillator is just a noise amplifier with some frequency
selective feedback. ...

If the simulator does not simulate the thermal noise

I do not believe this is necessary. The normal charge-up of any coupling
caps will get you "off center" at the start. On PSpice ver 9 one of my
students "couldn't get an oscillator to go". I lenghtend the simulation
time and found that it took longer than he had set for the transient
analysis.
Out of the box on PSpice I got both a 555 and Colpitts to fly. I can't
speak to CircuitMaker, but there may be analysis set-up parameters such as
the aforementioned initial condition which may be the key. PSpice also has
something about "skipping" an initial transient state/solution which I
haven't experimented with to comment further, but could be a consideration.
Watch the options...

--
Steve N, K,9;d, c. i My email has no u's.

On Mon, 7 Jun 2004 17:21:03 -0500, "Steve Nosko"
wrote:

Out of the box on PSpice I got both a 555 and Colpitts to fly.

The 555 should start unconditionally, since it is based of the RC time
constant and two comparator levels. Thus, it should not matter, what
the initial charge in C is, since sooner or later one of the
comparators will change state.

A startup transient in the Colpitts may very well start the
oscillation, since the transient is rich of harmonics and if one of
the harmonics is connected back with sufficient gain through the
frequency selective feedback, the oscillation should start.

Paul OH3LWR

On Mon, 7 Jun 2004 17:21:03 -0500, "Steve Nosko"
wrote:

Out of the box on PSpice I got both a 555 and Colpitts to fly.

The 555 should start unconditionally, since it is based of the RC time
constant and two comparator levels. Thus, it should not matter, what
the initial charge in C is, since sooner or later one of the
comparators will change state.

A startup transient in the Colpitts may very well start the
oscillation, since the transient is rich of harmonics and if one of
the harmonics is connected back with sufficient gain through the
frequency selective feedback, the oscillation should start.

Paul OH3LWR

"Leon Heller" wrote in message
...
"Paul Keinanen" wrote in message
...
On Sun, 06 Jun 2004 17:59:00 -0500, Ed Bailen wrote:

The last issue of QST had a short article on the CircuitMaker Student
edition circuit design package. I downloaded it and spent several
hours trying various oscillator designs (including the phase shift
oscillator discussed in the same issue of QST). I couldn't get any of
the designs to oscillate! I was mainly trying Colpitts oscillators.

Typically, an oscillator is just a noise amplifier with some frequency
selective feedback. ...

If the simulator does not simulate the thermal noise

I do not believe this is necessary. The normal charge-up of any coupling
caps will get you "off center" at the start. On PSpice ver 9 one of my
students "couldn't get an oscillator to go". I lenghtend the simulation
time and found that it took longer than he had set for the transient
analysis.
Out of the box on PSpice I got both a 555 and Colpitts to fly. I can't
speak to CircuitMaker, but there may be analysis set-up parameters such as
the aforementioned initial condition which may be the key. PSpice also has
something about "skipping" an initial transient state/solution which I
haven't experimented with to comment further, but could be a consideration.
Watch the options...

--
Steve N, K,9;d, c. i My email has no u's.