[unfolds] #01 - Franklin oscillator

[RoGeorge]

Sometimes, a quick lookup for an unknown term happens to unfold a story more interesting than the main reading.  Such was when I've looked up for the term 'Franklin oscillator'.

The Franklin Oscillator, by the patent from 1930, is this:


Source:  Franklin oscillator patent - US2028596 - page 1

An oscillator made out of two identical common-cathode inverter stages (can be made as well with common-emitter transistors), chained one after another as a non-inverting amplifier, and with a resonant parallel LC tank to ground, that is very loosely coupled between the output and the input of the amplifier.

The US2028596 patent for the Franklin oscillator also claims a specific mechanically thermocompensated mechanism and arrangement for a variable LC tank, as in the Fig.1 and Fig.2 drawings.

The legend says Franklin oscillators can make very stable VFOs.

Another example, with BJT:

Source:  http://www.arthropodsystems.com/FranklinOscillator/FranklinOscillator.html


To stay brief:
- very stable as a VFO (when powered from a well stabilised voltage)
- well behaved and easy to build, the 2 inverting stages can be anything: tubes, BJTs, FETs, CMOS gates, etc.
- L desn't require taps
- the LC parallel tank has one side grounded (advantageous for shielding a variable C or L, or for adding a varicap diode to ground, in series with C, so to turn it into a VCO)
- the LC is very loosely coupled to the amplifier, so thermal and other variations in the the amplifier have little to no influence over the LC resonance
- being very loosely coupled, the quality factor of the LC tank remains unaltered
- typical coupling to the LC tank is made with very small capacitors, a few pF when the oscillator is made with tubes, a few tenth of pF when with BJT (i.e. a 2pF for tubes can be made by twisting isolated wires - sometimes called a gimmick capacitor - depending on the number of twists it can give 1pF/inch, or 0.4 pF/cm - the capacitance can be reduced in circuit, by clipping away the length of the twisted wires)
- voltage feedback, if the negative impedance of the amplifier is >= R_losses of the LC tank then the circuit will oscillate, or apply Barkhausen criterion to determine the oscillation condition
- the point with maximum amplitude oscillations and lowest distortions/harmonics is at the LC tank (the output signal can be taken from the LC, if a loosely coupled buffer with a high enough input impedance is added to the schematic)

Franklin oscillator seems to be one of the most obvious types of oscillators:  a non-inverting amplifier with a resonant LC to ground in its positive reaction.  I wonder why other types of oscillators (e.g. Collpits or Hartley) are taught in every RF101, while the Franklin Oscillator is never mentioned.


It's tempting to give it a try, and see how it behaves.  Don't need any oscillator right now, but it has a certain elegance and simplicity, not sure why this oscillator seems so appealing.

There are plenty of schematic examples online, though I'm still not sure what are the must have of a Franklin oscillator.  A positive feedback network, with a loosely coupled parallel LC to ground seems to be a must have, but does the amplifier has to have two inverting stages?

Would it still be a Franklin oscillator if some other non-inverting amplifier is used instead?

[RFDx]

This type of oscillator is quite a rare bird. The last time I saw one in a practical application was as a VCO in a synthesizer for a legacy, premium HF-receiver from SRT (Standard Radio & Telefon). The amplifier was implemented with four noninverting instead of the more common, inverting, two stages variant.

[sv3ora]

You asked for it. Here you go
http://qrp.gr/supervfo/
http://qrp.gr/minivfo/

Note, that my designs are different than anything you will find. They take the output directly from the resonator, very lightly coupled. This achieves extremely low distortion. Then driving with it a high gain ALC amplifier (again my design), to achieve low distortion and relatively stable amplitude amplified signal.

And it is exactly like you mentioned in the advantages.

Under my temptation microwave1 tried. You judge yourself



I won't built any other type of HF VFO from now on.
Whenever a stable and good tone VFO is needed, I go for the Franklin VFO.

One thing I have not managed to do yet is to achieve the same good tone and stability performance, using a varicap.

[wofritz]

I built a 3.5 MHz transceiver some 45+ years ago (which still exists today) that uses a Franklin oscillator. It is indeed extremely stable but I think that's more due to very stable mechanical design and a good portion of pure luck. 

[TristanLaguz]

Franklin oscillator seems to be one of the most obvious types of oscillators

I can attest to that, considering that I came up with this version of it myself:


 A few days ago, I asked on StackExchange (see https://electronics.stackexchange.com/questions/728401/what-is-this-swinging-circuit-made-up-of-lc-circuit-and-ac-booster-called) what this circuit is called, as I was almost certain that such a simple swinger had already been invented. I was right. Shortly thereafter, I got my answer: that it's the Franklin oscillator. So I ecosied and googled that term and thereby found this post and also learned that my design seems to be suboptimal, as C2 and C4 are so big that they load the tank too much. Nevertheless, even my design works fine and gives out a nice sine (rhyming intended).

it has a certain elegance and simplicity

Indeed.

not sure why this oscillator seems so appealing.

I am: Because it's so simple and easy to understand, and in particular because it needs no inductor taps or capacitive voltage splitting.

Franklin oscillator seems to be one of the most obvious types of oscillators:  a non-inverting amplifier with a resonant LC to ground in its positive reaction.  I wonder why other types of oscillators (e.g. Collpits or Hartley) are taught in every RF101, while the Franklin Oscillator is never mentioned.

For me, the Franklin swinger is the most obvious kind of oscillator; it's just a simple LC tank hooked up with the input and output of an AC booster. My Electronics book wonderfully explained tuned circuits and pointed out that they have to be pushed like swings to keep them going. It then presented Colpitt's oscillator as one such solution. That was all nice and well, but since that I had just read about simple resonant circuits, I would've liked to see how some of these could be kept going. However, neither the book nor Wikipedia gave me a solution, so I worked it out myself. Only when I had learned the term "Franklin oscillator" did I find an article about it on German Wikipedia. However, that Wiki article isn't available in other tongues. So I have the same question as you: Why is the Franklin so seldom mentioned?

a loosely coupled parallel LC to ground seems to be a must have

If the loose coupling be a must-have, then my swinging circuit be no true Franklin swinger after all.

And thanks for the picture of the patent!

P.S.: How can I make the image scale with the browser window?

[RoGeorge]

The pictures in the first post were inserted with fixed width.  They should enlarge if you click on them.  Same for the thumbnails at the end of the post, they should enlarge when clicked.