Nice question, Glen!

I like Fig. 3 as it is clear from details. I think it gives a good answer to another RG question:

https://www.researchgate.net/post/what_if_feedback_resistor_of_opamp_is_replaced_with_diode

Cyril, 

I am still laughing out loud 

at your presentation of the Two OpAmp Circuits : 

""" In non-inverting circuits, the op-amp feels a "good envy" """

""" In inverting circuits, the op-amp feels a "bad envy" """

Yes, Fig. 3 is a "Result" ,

to me,  a nearly non-sensical example ,

just something  that is detached from a useful application. 

It is so that a student must think about the circuit, 

and the interactions of currents, 

but not think about his application. 

I use much wider ranging gain settings, -50 to +50 roughly, 

and R(diode) = 3000 Ohms , 

in my application.   

In my application, the driving signal V(inA) is variable, 

and my gain on that signal V(inB)  is variable, 

thus I can select different balances and different limiting roof values.  

This is an option which I am still playing with 

after several years of using this approach. 

As soon as I get back , then I will apply Barrie's posting 

and calculate the V(forward) for Fig.3.    

I have refused to simply let V(forward) = V(out) - V(Rdiode2)

...  that is no fun at all.  

Perhaps I am getting old,  but I need peace-and-quiet 

and then three runs through the calculations 

to make sure I have not been led astray by some errant thought. 

Something along the lines of 

V(forward) = 0.6V + XXX mV  *  log (I(forward) / YYY mA) 

based on an aproximation of V(forward) standard room temp

and shifting away to a different I(forward) measured).

So, Til then, Buena Suerte!. 

Glen

This type of limiter is often used for "cheap" oscillator.

"Starting gain" is great - at the desired amplitude compression occurs.

Josef,

I must investigate that idea on my protoboard. .

Well, yes, now that I think about it ... just need positive feedback.

I could use fixed gain and positive feedback.  

Thanks.  

Glen,

I just do not know what will "feel" the opamp in the case of a positive feedback:)

Cyril,

I think the word has lost good meaning in translation. 

"FEEL" ?  

An oscilloscope is used for to "feel" an OPA.

Did you mean "Experience" ? 

I do recall soldering  eight  LM741 , one on top of another, 

to "experience" the effect. 

So, when I placed my finger on top of the stack, 

it was not warm.  

f(max) was about 1 KHz,   and   I(out) was about 60 mA. 

On the other hand, this student did put an OPA in backwards, 

and burned the chip

and melted it into the white plastic Proto-Board !

I had the "feel" of embarrassment, 

so I told the instructor it was an "experiment". 

Try a different word than "feel".  laugh

Oh!  Now I recall the thread 

... "What_feedback_convention_do_you_prefer" 

""" In non-inverting circuits, the op-amp feels a "good envy" """ 

""" In inverting circuits, the op-amp feels a "bad envy" """

Perhaps the OPA circuit "feels" that it should hop-along-faster !  

to "feel" the magical state of " good resonance "

for which it was designed.

I am suggesting that

there are always more than two sides to a story.   

That is something that my mentor explained to me, way back. 

So, there are more than simply "good" and "bad" 

there is also "resonance",  and surely other states. 

the apprentice, Glen

Glen

some very basic "ideas" we can see in "guide 12". Very inspiring is AN-263.

Josef

Josef,

Thank you for the pertinent tutorials. 

While I am away for the weekend, 

I will read through both.  

They look like well prepared tutorials.

Already, I see the Weinbridge Osc with the light bulb , 

which I recall using, far back.  These will be good to read again.

Now, looking around my lab shelves, 

I find a single 6x6 perf-board with several oscillators 

which I used years back.  Also,. my old digital Hex-Decoder-Readout, 

and a multiplexer which I designed to display eight channels of digital onto a single channel Oscope.  

I kept all these circuit things, along with 10 Proto-Boards and four power supplies in a Suit-Case, along with my NLS-Mini-Oscope for school labs.   I did school work at home, went to the college lab and just opened the suit-case and demonstrated my working circuits "Pronto!".  PIctures of these things are on my website.  Now-a-days, I use this set-up most every day.  It is very convenient to just close the lid and protect all the wiring.  From that I transfer modules to the soldered-up Perf-Boards with bus-pad-per-hole patterns.   Has been very convenient way to keep my work "straight and true" during construction.     

Now, I find  my old  Weinbridge Osc section,  with Germanium diodes,

and the much more used cousin ,

which produced Square-Waves and Triangles,   

Now-a-days, I pull Radio Signals from the Ham Bands to test filters. 

Thanks again, Josef,

very appropriate texts to review for this week-end. 

Good things to read

as I think about anayzing the Varient OPA Diode Limiter. 

the apprentice, Glen

Folks,

I surely did not think that I had "invented" something "new" 

when I assembled the ideas for this Variant Diode Limiter, 

but ( as has been written by BG ) 

"I was infused  with a sense of joy and youthfulness." 

Reading Josef's very well done chapter on Wein-Bridge variations, 

I see that I am sharing old and good-working ideas.  

Thank you, Josef, for the good reading. 

Now, among many other duties that are mine,

I am back to the drawing board ,

In the previous post, I have edited and attached 

some little JPGs of the "Suit Case" lab.  

Very impressive photos!

Cyril,

( The old things were certainly more impressive

than my math skills today. )

Attached is a working Spice Schem of this Variant Diode Limiter,

with notes taken from the Transient plot. 

It may well be the FINAL schem of this sub-project, 

after much prodding and encouragement from this forum, and BG. 

Perhaps calculating forward into a dynamic diode's performance 

is part of the "fools errand" Barrie mentioned relative to temperature.  

Perhaps it needs a younger mind !   or just more humility in heart. 

At any rate, this circuit has a  'working'  Spice analysis, 

and discussion / critique can be placed on it  

to move it forward if anyone desires. 

Next, I need to do one for the actual circuit, and include that in the dox. 

Should include a Transient for each of the currents, 

since several overlap and become covered. 

just the apprentice  playing with the 'broom' in the master's shop, 

Glen

Cyril,

I have made edits to the above posts, 

correcting  "pencil errors" and clarying some terms. 

With great thanks to several expert electrical engineers ! 

I have uploaded a fresher  scholastic schematic

and resulting transient plots . 

I Concede to using  calculations,  made via Spice , 

to  show both Currents  and  Voltages.  

Others may differ,

but I am coming to see that a "technician"/apprentice

will do well to make use of Spice, 

& should not spend his years re-inventing the wonderful wheel. 

(*) I note

that real-time measurements of the "head-room" of the limiter ,

the  diode-roll-over effect [V(forward)]

is consistently in the -9 dB range.  

The Head-room has a proportional ( log ) response 

to the given Gain-Ratio .  

ie,

(1) Given a V(out) that shows the start of Diode Limiting action,

then ,

(2) when the diodes are removed  the V(out) signal rises +9 dB . 

(3) Same readings occur at Low Gain-Ratio  and High Gain-Ratio. 

(4) This approximates the expectations of a limiting action

that is proportional to the the given gain-ratio effect.  

(*) I note 

that given the V(in) 1 V signal, 

(1) that by changing the R(Rd) from 100 Ohms to 1500 Ohms

the V(out) rises from aprox. 333 mV to aprox. 556 mV.

(2) that by changing the R(Rd) from 100 Ohms to 5500 Ohms

the V(out) rises from aprox. 333 mV to aprox. 770 mV.

(3) that by changing the R(Rd) from 100 Ohms to 15000 Ohms 

the V(out)  rises from aprox. 333 mV to aprox. 850 mV. 

...

This will be Spiced, then proto-typed and measured. 

(*) I note

that the V(null) shifts by a very small voltage, 100 uV , 

related to the time of the V(diode) change-over. 

Plans are to measure the delays of the OPA model 

during the transit time of the negative-feedback currents. 

With respect to the varying V(null) (+/- 100 uV) : 

In my opinion,

this whole circuit is a "servo loop",

which  runs  continously, 

producing a "running-average response" V(out) ,

with respect to the entire circuit's continuous interaction. 

still the apprentice,  only tinkering  in a long shadow,  

Glen

Cyril,

I am adding the data from the latest run in Spice

on my Current Circuit, running right now,

as I am listening to CW radio signals on the 14 MHz band.

My active project uses a higher limiter "head-room"

than the scholastic example as presented above.  

This higher R(RD) of 33000 Ohms reduces the background noise

detected by both ear and Oscope.  

Recall, that less diode limiter action will produce less IMD. 

This higher R(RD) makes this stage function in a more linear fashion. 

Regarding the "Head-Room" with/without the Diode-Limiter, 

I ran a set of real-time experiments using R(RD) =< 33K Ohms. 

(1) With a signal of -27 dBm, installing the Diodes V(out) = -28 dBm.

(2) With a signal of -14 dBm, installing the Diodes V(out) = -8 dBm.

In practice, this means

(1) that a weaker signal will pass with far less limiting,

(2) that a stronger signal will pass with more limiting.

Since this circuit

(1) should produce minimal distortion on weak signals 

(2) should limit more on strong signals ( lightning static )

...

thus, this 33K Ohm R(RD) seems suitable.  

Since my application is not referenced to a rigorous lab standard, 

but rather to the human effects found in the ear/brain combination

as the user is active in receiving / detecting information in the signal,

the system function of circuit and user

might be called "soft science"    

( ie, very relative to the user ).  

Similar considerations must be made in the values selected 

for the main Active Log-Limiter of this project .  

The Active Log Limiter stage uses a set of resistances

summing > 3.3 Ohms.  

Lower Signal levels receive more 'linear' control .

Higher Signal levels receive more 'logrithmic' control. 

It is not obvious when viewing this single Pre-Amp stage 

but, there are two gain controls in the larger project

which can be balanced 

(1) Gain control using R(in) in the Pre-Amp with limiter

(2) Gain control using R(audio) in the Audio Pre-Amp final.

Thus, the user can balance between

(1) more Pre-Amp to push the Active Limiter stage into action,

(2) less Pre-Amp and more Audio-Amp, for more linear signal.  

I have wandered afar from the original "question" 

about a Variant Diode Limiter, for student consideration. 

------------------------------------------------------------------------------------------------- 

Last Schematic includes two Diode Limiter stages 

with more current (i) markings.    All paths add up properly. 

-------------------------------------------------------------------------------------------------

Now it is time for some Active Walking on my Creek-Side Trail, 

to push my clock a little further along the Path of Life. 

"Probably Ordinary Time"  compared to a friend of mine, 

but it is mysteriously engaging to me. 

The First Dance was such a joy,

but the Last Dance is all about the company

which I am allowed to keep. 

Glen

Note from lab measurements of running circuits 

indicate one of the errors that Spice can slip in, unawares. 

Spice assumes that a standard OpAmp is IDEAL, 

so current flow must be analyzied carefully.

I replaced one of the above basic schematics,

so, now, all the input currents add up to the feedback currents !!! 

Old eyes picked the wrong schematic. 

Thank you all for the comments. 

180116. Have just reviewed this interesting little question/project.

Thank you , excellent engineers,

for your comments and encouragements.

www.GeoCities.WS/glene77is

Glen

Fellows,

I have never discussed the Inter-Modulation-Distortion

nor the effects of energy displacement coming from the Limiter Action.

.

In my radio operations,

When I am working a very weak signal ( S/N = 1 )

these effects are very noticable.

.

Question :

To "Where" do you think the Limited Energies are transferred ?

To "Where" do you think the Filtered Energies are transferred ?

...

This has been observed, and some preliminary documentation done.

.

Energy is not Destroyed, so to "where" does it go ??

Maybe, just up in smoke ???

.

Glen