Building an oscillator matching HP200A specs

[PurpleAmaranth]

I am hoping to build an oscillator with the same specs as the very first HP product: the HP 200A (can be found here: http://www.hparchive.com/Manuals/HP-200A-Manual-1951.pdf). I expect it to take a while, but I figure I'll learn a lot in the process. I'd like to do everything analog, but don't really have any other design requirements. I've done some research and figured I'd use a Wien Bridge oscillator just like the 200A did, and found the very informative discussion here: https://www.eevblog.com/forum/projects/wien-bridge-project/ .

This led me to Linear Technology's Application Note 43 (https://www.analog.com/media/en/technical-documentation/application-notes/an43f.pdf), which begins its discussion of Wien Bridge oscillators on Page 29. From there, I found an LTSpice schematic of a similar circuit on the LT1037 page.

I figured my first step was to really understand how the most basic model of the circuit, such as the one found on Wikipedia or in Application Note 43 worked. I figured I could try to derive the Transfer Function myself. But that quickly led me to my first question. How is it possible that there is no input to the circuit? Looking at the LTSpice simulation, the only voltage sources present are both op amp supplies. My understanding from university classes is that op amps provide differential amplification of whatever is fed in to the +/- terminals, but it looks like in the oscillator, nothing is fed in. How is it possible at all that the output is non-zero?

I figure I'm missing something basic in my understanding of op amps, but haven't found much online.

[Kim Christensen]

but it looks like in the oscillator, nothing is fed in. How is it possible at all that the output is non-zero?

Oscillators are basically amplifiers with positive feedback. There is always a little bit of noise or turn-on transients to get it started.
 

[jonpaul]

Purple: BRAVO,

Jim W at LT did extensive research on this topic and rebuilt several HP200.

See the BSTJ Mecham paper

http://www.alcatel-lucent.com/bstj/vol17-1938/articles/bstj17-4-574.pdf
 1939 Hewlett   patent US2268872: A  Variable frequency oscillation generator 

http://www.hp.com/hpinfo/abouthp/histnfacts/museum/earlyinstruments/0002/0002history.html

HP 200 used tubes, but   same in theory and simulation.

try it using the original tubes (valves)

"History: Just one damn thing after another!"   
Winston Churchill

Jon

[PurpleAmaranth]

Kim: That makes a lot of sense. Thank you!

Jonpaul: I'm slowly making my way through the Meacham paper (that link is down by the way, but it can also be found here: https://ia801900.us.archive.org/20/items/bstj11-1-67/bstj11-1-67.pdf). Might order some crystals to build the Bridge-stabilized oscillator he presents and gain some more intuition about it. Will then move on to the HP patent, and hopefully really understand how the Wien Bridge oscillator worked. Thanks for the HP history page as well. I would try it with tubes, if only by curiosity, but I'm not comfortable working with the voltages I see in the service manual (300V), so will stick to op-amps for now.

I'm in no particular hurry, just hoping to learn as much as possible. I'll report back at some point.

[jonpaul]

PurpleAmaranth bonjour are you in India?

The history of oscillator design is interesting.

Some valves  can be run at low plate V and work fine.

The late and great Jim Williams at Linear Technology did a lot of work restoring and studying and improving the HP200.

Simulation is not easy as nonlinear thermal response of an incandescent lamp is not well characterized.

Finally  see HP204C and 204D:  Modern solid state HP descendants using transistors and  FET or optocoupler for the gain control

https://bama.edebris.com/download/hp/204c/HP%20204C.pdf

Enjoy,

Jon

[chuckb]

Here is some work done to model an incandescent lamp by INTUSOFT.