Transistor oscillator.

The LC pi-network between collector and base. The author set it at 10MHz resonance. Frequency is L in parallel with the C in series or 1/L=(2pix10e6)^2 x 500p=> L=0.5uH, use 0.47uH stock part.

Isn't one of the common pathologies of bipolar emitter followers that they often end up as Colpitts oscillators, due to the parasitic diffusion capacitance between the base and emitter?

So set up the tranny in a pathological emitter follower config and see what happens?

Relevant reference from Pease:

One fairly simple method is to put small resistors of equal value from the base and collector to ground, stick an appropriately biased sinusoidal current into the emitter, and use a 2_channel scope to look for the base and collector waveforms to be equal in amplitude. You might start with 10 ohms.

Cheers

Phil Hobbs

Yabbut...

What diode would you use? Schottky?

OP said germanium transistors, so, have fun with that... :-)

(Did they even make Ge "hot carrier" diodes? Want to say, even in the

Tim

Just set up a pulse test. Measure t_r or t_f or the delay (t_d or t_stg). It'll be close enough to representative of fT, at least within a gross factor.

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website:

Is there a rock bottom oscillator that will free run at the frequency limit of whatever transistor is put in it?

I was thinking of the common base job with a capacitor strapped from collector to emitter, but am unsure what to use as a collector load.

Basically for testing how good are a collection of vintage germanium transistors.

Thanks for any help.

As long as its the transistor Ft that's the limiting factor - I doubt that saturation is going to be a problem.

Having anything other than the transistor itself set/limit the frequency, defeats the object of the excercise.

The common base I was thinking of is just the textbook automatic bias common emitter with the decoupling capacitor moved from the emitter to the base.

The capacitor from collector to emitter is basically the feedback capacitor - I was thinking a gimmic capacitor would probably be enough.

Usually these oscillators have an LC tank in the collector circuit, but sometimes just an RFC. What would happen if I just gave it a collector load resistor?

On Friday, January 15, 2016 at 12:04:27 PM UTC-8, Ian Field wrote: ....

This circuit works by intentionally running at a constant oscillating frequ ency somewhat below Ft.

The series base capacitor is adjusted until oscillation just starts as indi cated by the meter. The lower the capacitor value, the higher the gain of the transistor. Assuming the transistor has a single pole with the gain dr opping off at 6db/octave the chosen capacitor value then indicates the Ft o f the transistor.

kevin

On Friday, January 15, 2016 at 12:10:38 PM UTC-8, Ian Field wrote: ....

There is no current gain from emitter to collector so it won't oscillate.

The normal Colpitts oscillator relies upon the resonance of the inductor with a capacitor tap give the current multiplication. In many circuits the base emitter capacitance forms the bottom half of the tapped capacitor.

kevin

If you put a resistor in series with the emitter, the intrinsic base resistance becomes negative. If the extrinsic resistance plus the is too small, it'll sing like a bird.

Early POR circuits were frequently just an emitter follower hung on an RC delay network. Smart people stopped doing that for a variety of reasons, one of which was that it often oscillates like crazy.

Cascading PNP and NPN followers to eliminate the V_BE offset is another good way to make oscillators.

Cheers

Phil Hobbs

Plus the what?

So, I admit that my powers of circuit design aren't great enough to see this negative intrinsic resistance thing inside my pointy little head. I had heard that this was the case with an inductive load on the emitter -- but a resistor?

Can you point to a web page, or suggest a starting point for modeling this myself? Just the good ol' pi model with an emitter resistor?

IIRC it's because there are two lags: high frequency beta has a pi/2 phase shift, and so does C_BE. You can find it in Hollister, for instance.

Cheers

Phil Hobbs

You're sounding a bit too imperious for my liking, so I'll leave you with your fantasies.

MHz

y,

quency somewhat below Ft.

dicated by the meter. The lower the capacitor value, the higher the gain o f the transistor. Assuming the transistor has a single pole with the gain dropping off at 6db/octave the chosen capacitor value then indicates the Ft of the transistor.

He's oblivious to the concept of current gain bandwidth product, with that being his common base Ft he's after. The circuit is deliberately run common emitter to reduce the operating frequency to the more manageable Ft/beta x

I think the situation should become apparent if one analyzes the emitter folllwer topology using the hybrid-pi model, and the high frequency equation for transistor beta - beta is a complex (in the sense of complex numbers) function of frequency.

And include the parasitic capacitance in the model, of course.

Another circuit I have in mind is the LO used in some AM transistor radios, they have 3 windings, one each in the collector and emitter circuits - the third is tuned by the oscillator gang.

Maybe remove the tuning cup and any capacitors from an oscillator can and see what it spins at.

There's a CFL board sitting on the desk with a 3 winding toroid that might work in that type of oscillator - won't be fast enough for what I want to try with the draw full of old germanium transistors, but might be amusing to play with for a while.

It might even be simpler to do it properly - a varicap tuned oscillator fed by a ramp generator.

The problem is; my best scope is down to 0.707 at 100MHz and I'm hoping a few of the transistors can do better than that.

Also I'd have to be able to switch to manual voltage control to set up markers.