Yes - according to my knowledge he first published this invention in his Stanford thesis "A new type resistance-capacitance oscillator" (1939).

Thank you so much Lutz. Nice to hear from you here on "Research Facebook" ;-)

Here comes a reference to W.R. Hewlett's patent !!

This question gives rise to place another question regarding the WIEN oscillator.

The WIEN oscillator is one of the classical oscillator circuits and mentioned in (nearly) all textbooks dealing with „linear and harmonic“ feedback oscillators. And - of course - this circuit and its advantages are also explained in a textbook published by the authors

U. Tietze/Ch. Schenk: Electronic circuits: Handbook for design and application (Springer, 2008).

In German: Edition 12 ( !).

However - the book contains some statements which I cannot agree upon - and I like to ask you experts if I am right or wrong.

It is well known that the symmetrical WIEN band pass (two equal time constants) has a „gain“ of 1/3 and a rather poor selectivity and, thus, a phase slope, which is rather flat. Therefore, to build an oscillator we need an amplifier having a gain of +3.

Now - the referenced book states that because of the poor passive phase slope any amplifier phase error would cause a relatively large frequency displacement. Therefore, it would be advantageous to complete the WIEN path by using two resistors - thus creating a matched bridge (the known WIEN-ROBINSON bridge) having a very steep phase response.. As a consequence, now we need an error amplifier with again approaching infinity - but the frequency error would be smaller.

My position is as follows: I do not see any „advantage“ using this bridge together with an (ideal) opamp because this alternative is nothing else than another interpretation of the first circuit (band pass and a gain of 3 amplifier).

The frequency error will be the same in both cases:

Alternative 1: Bad (small) passive phase slope and a rather small amplifier phase error (due to negative feedback for realizing a gain of 3);

Alternative 2: Very steep passive phase slope and a relatively large amplifier phase error (no negative feedback for open-loop gain).

Am I right? Do I overlook some important arguments? I am interested in your opinion.

(Of course, I have assumed that in both cases the same opamp type is used) .

Wien bridge oscillator is named after its inventor.

From Wikipedia, the free encyclopedia

Max Wien has invented a specific RC bandpass but not a bridge and certainly not an oscillator.

I could not find the original paper from M Wien "Messung der Inductionsconstanten mit dem „optischen Telephon” (without paying an unknown amount of money to Wiley)

Referring again to WIkipedia and reading your statement make me guess that M Wien proposed a measurement circuit (the Wien bridge) that , many years later Hewlett used to build an oscillator. patented : US 2268872, Hewlett, William R., "Variable Frequency Oscillation Generator", published July 11, 1939, issued January 6, 1942.

I probably have made a mistake, my first sentence should be "Wien bridge oscillator is named after the inventor of a bridge that lately was used to build an oscillator"

To be sure, we should verify reading a copy of the original work of M Wien.

As a matter of interest, in your extension to the original question, does "frequency error" mean a difference between the frequency calculated using the formula and the practical frequency of the output electrical signal of the oscillator? In todays world of digital frequency measurement, is it important to get a perfect match between theory and reality ? Of course, on an academical point of view, errors are a recognized subject of interest. Further digging of your question could lead to a scientific explanation why Wien bridge oscillators are not used at high frequencies.

I have just realized that WIen could not have built a Wien bridge oscillator because he would have missed the triode amplifier. Wien bridge paper: 1891. Triode usage : early 1900's.

Hello Monsieur Bergogne ,

regarding "frequency error" - yes, you are right. I mean the frequency displacement of the real circuit if compared with the formula based on ideal opamp properties. Neglecting all errors caused by passive tolerances this frequency error is caused by the phase error introduced by the frequency dependent opamp gain. Because the phase of the closed-loop is always zero, this phase error is compensated by a corresponding phase shift of the passive feedback network leading to the mentioned frequency shift.

What is the difference between WIEN bridge and WIEN-ROBINSON bridge ?

Hello Erik,

I am not sure - but as far as I know M. WIEN has introduced an RC bandpass (RC series and RC parallel combination) which was completed by Mr. Robinson to form a bridge (using two resistors). Perhaps I am wrong?

To Monsieur Bergogne:

A revised and updated version of the document requested by you is available here in RG as

"A novel harmonic Oscillator: GIC Resonator"

thank-you I have downloaded the paper. Great.

Popularly, William Hewlitt is the acknowledged inventor. He was in fact responsible for commercializing an AF version of it, which is still considered to have excellent performance even by today's standards. See US patent 2,268,872 as noted earlier.

Max Wien invented the Wien RC filter as noted in earlier answers. When gain stages (triode vacuum tubes) became available, Frederick Terman used the Wien Bridge in a feedback oscillator, but the oscillation amplitude and distortion were very difficult to stabilize.So I guess that was the first "wien bridge oscillator". Willian Hewlitt knew Terman from school (they both attended Stanford in the 1930's). Hewlitt's invention was to use a non-linear gain element in Terman's circuit so that the oscillation would achieve a stable amplitude with extremely low distortion as explained in the patent. The model HP 200CD Audio Oscillator is pretty much identical to the original design and is still a serviceable instrument and a much sought-after collector's item among electronics buffs.

Thank you Mr. M. Rodighiero for this valuable information.

I must confess that - up to now - I never heard about F. Terman.

In context with the subject discussed here (WIEN oscillator) I have another question to all "oscillator experts":

There is another 4-element passive bandpass (RC-CR ladder topology) which has the same transfer function and the same properties as the WIEN bandpass (Q=1/3 and gain=1/3).

In every textbook the WIEN oscillator is mentioned - however, it is hard to find a remark about the above mentioned alternative.

WHY? Because of historical reasons only? Are there some (hidden) disadvantages of the RC-CR topology?

According to my knowledge it has absolutely the same properties as the WIEN type.

Concerning the discussion of the Wien Bridge oscillator above:

Only thing to me is that the "gain" of the amplifier is not constant equal to 3.

The gain vary with time as it should do. The eigenvalues (poles) of the

linearized Jacobian (the small signal model) move around in the complex frequency plane.

The open loop circuit is just an active filter. When you close the loop

you may change the load of the power source. A steady state oscillator is

a nonlinear circuit with a time varying bias point.

Instead of a nonlinear lamp resistor a diode circuit is more appropriate.

By means of cut and try you may reduce phase noise and harmonics.

PS.

Fred Terman is mentioned in w-r-hewlett-05669988.pdf

Please see link above.

Hi Erik,

may I ask you why do you think that diode stabilization is "more appropriate" than a temperature dependent non-linear resistor (lamp or thermistor)?

Ohh -may a add a second question?

I cannot see how a diode network (which is active during amplitude peaks only) can cause a "varying bias point". Please, can you explain?

Hi Lutz,

(1) The lamp mechanism is more complicated. It depends on

temperature. It is difficult to set-up a proper SPICE model ;-)

(2) It is not the diodes as such which give rise to the

time varying bias point. The open loop circuit is just an active filter

for which you may refer to a dc bias point. The amplitude of the

input signal may influence the current in the power source.

When you close the loop you may load the power source in quite another way.

All the nonlinear elements give rise to a time varying bias point i.e. a time varying

small signal model. The diodes do not operate as ON/OFF switches.

Erik, thank you for replying.

@1: Yes - a thermistor (NTC or PTC) depends on temperature. No doubt about it.

Nevertheless, I have achieved extremely good results (in terms of THD) with NTC-thermistors (thermal time constant at least 10...50 times the oscillation period).

The same applies to a well-designed FET control mechanism.

But I agree - diode stabilization is the most simple method.

@2: OK - now I see what you mean by saying:

"A steady state oscillator is a nonlinear circuit with a time varying bias point".

Of course, the current drawn from the power source is not a constant one but changes with the momentary signal level.. No doubt about it.

However - I think this also applies to all amplifiers and, in particular, to transistor-based amplifiers? The drawn current always depends on the applied/delivered signal.

Is this such an outstanding property, which - in case of an oscillator - must be explicitely mentioned (underlined)? As you already knows, I am not completely with you regarding this point. Or did I misunderstand something?

Hi Lutz,

Thank you for your remarks above. Autonomous oscillators depend on the constant power source applied. To me it is important to study how energy is transferred to the load of the power source in order to understand how and why the circuit oscillates ;-)

I have a feeling that it is impossible to find a kind of criterion which is sufficient and guarantee oscillations. We start with Barkhausen and then we continue with hit and try ;-)

Erik,

Yes - it is really an interesting question "how and why". Sometimes I simply tend to say: It is just an amplifier that creates his own input signal from its output - but this is nothing else than another formulation of Barkhausen`s rule. And it cannot explain the starting phase. For my opinion, the start-up phase is the most challenging part of the story.

The second aspect mentioned by you (sufficient criterion) surely deserves an extra chapter within this forum.

Lutz,

Thank you for your comments above. I agree. We must discuss the starting phase of the hit and try design procedure !!!

The closed loop circuit is an autonomous circuit excited by a dc voltage power source. The bias point of the circuit vary with time so it is important to study the power source current in order to understand the mechanism behind the behavior. The start-up phase begins with switching on the power source. We may "stop" time and calculate a DC bias point and derive the small signal model. If the poles are in RHP the signals will increase and either steady state oscillations or a new DC bias point will occur. If the poles are in LHP we may change some elements so the poles come to RHP and we have the situation with the unstable small signal model.

We have examples where the poles of the initial small signal model is in LHP so it becomes necessary to apply an impulse somewhere in order to start-up steady state oscillations. The size of this impulse is crucial for the behavior. A nonlinear circuit may have several stable DC bias points but some of these points may be potentially unstable.

l

Mark Rodighiero,

Thank you so much for your information above. Sorry for my delayed response. Fred Terman published his famous book in 1943. Do you have access to lecture notes by Terman from before 1938 indicating that he was the first to introduce the Wien bridge ?

Dear All,

Thank you for your comments above.

Concerning the Wien bridge oscillator family I want to point out the following:

If you remove the capacitor in the RC parallel circuit of the Wien bridge you have a common multi-vibrator based on the capacitor in the RC series circuit.

,

.

Yes - so we have changed the order of the system from "two" to "one".

This assertion assumes an ideal opamp.

However, if we assume a real opamp with frequency-dependent gain it is possible (for large frequencies) to produce again sinusoidal signals at the output - unless the slew rate disturbs this effect.

YES - I agree 100 pct  with you as usual  ;-)

Hermon Hosmer Scott, in 1937.

US patent 2173427A

https://patents.google.com/patent/US2173427

Hello Steve Byan

Thanks a lot for the link to Scotts wien bridge oscillator patent

Best regards

ERIK LINDBERG