what would be the philosophy behind the opamp configuration?

[Analog_Fan]

Simulation of the circuit

according to the simulator it works perfect, but what would be the benefit?
input is 10V.p.p and out is 24V p.p.,

it's the output section of this:
Hexmix - Befaco
schematic

 

[abbey road d enfer]

There is no "philosophy" here.
It is a basic inverting stage with Vbe multipliers in the feedback, which is supposed to contain the signal whenever it tries to exceed a certain level.
It's often misappropriatedly called a "limiter", but in fact the proper name is a clipper.
It's a somewhat "soft" clipper because the level containment happens over a relatively high range of level.
This circuit is the basis of many overdrive/distortion pedals. It's also used in applications where excessive peak levels must be avoided, such as AM transmission (CB, walkie-talkie...).
Distortion increases as level approaches the clipping level, so it's not a clean limiter.


BUT: with the actual values, clipping appears at about +100/-120V at the output, which can never be attained wit +/-15V rails.

So, either there is a significant mistake in the printed values, or whoever designed the circuit took it from a cookbook and didn't know how to tune it.
A better operation could be achieved using complementary transistors, for equal positive and negative clipping.

Note that you have transcripted the schemo wrong. The upper transistor has its collector to the right.

Don't neglect the possibility I have missed something and I'm utterly wrong.

 

[moamps]

...Don't neglect the possibility I have missed something and I'm utterly wrong.

I played with the simulator for about 10 minutes and I just wanted to write:
"That circuit isn't working as it should. You need to swap the E and C terminals of the upper transistor, and change the resistor ratios in the NFB."

 

[abbey road d enfer]

"That circuit isn't working as it should. You need to swap the E and C terminals of the upper transistor, and change the resistor ratios in the NFB."

The original schemo is correct in this respect (reversed transistor). However it is a quite complex circuit for replacing a single resistor.

 

[abbey road d enfer]

I had not seen that the 220k is in the middle of two 15k resistors, not 1.5k. With these values, clipping appears at about -12/+10V.

 

[Analog_Fan]

There is no "philosophy" here.
It is a basic inverting stage with Vbe multipliers in the feedback, which is supposed to contain the signal whenever it tries to exceed a certain level.
It's often misappropriatedly called a "limiter", but in fact the proper name is a clipper.
It's a somewhat "soft" clipper because the level containment happens over a relatively high range of level.
This circuit is the basis of many overdrive/distortion pedals. It's also used in applications where excessive peak levels must be avoided, such as AM transmission (CB, walkie-talkie...).
Distortion increases as level approaches the clipping level, so it's not a clean limiter.


BUT: with the actual values, clipping appears at about +100/-120V at the output, which can never be attained wit +/-15V rails.

How did you calculate or envision that?
What values would suggest for a bi-polar 12/15V system?

i updated the schematic with an input opamp with it's correct resistor values.
the output follows the input now, 10V.p.p.
https://tinyurl.com/y6dq4ltr
it seems the 220K can't be changed, making it smaller reduces amplitude.

now i see what it does if you change both resistor next to the 220K to 39K it clips.
here is with 27K
https://tinyurl.com/yymsuael
boosting the gain with the first opamp, it shows what it does, 27K seams to be required value.
Cool, i made another circuit using a Analog Devices rail to rail opamp to do this, but this seams to work quite fine.

So, either there is a significant mistake in the printed values, or whoever designed the circuit took it from a cookbook and didn't know how to tune it.
A better operation could be achieved using complementary transistors, for equal positive and negative clipping.

I'm giving it a shot as well (since it will take some time to get parts and PCB), spend the last fews hours creating, revising footprints for SOT666, SOT457, SOT363 for BCM847, BCM857.

Told ya, many people just copy and paste, but in general "Befaco" leaves a good impression to me as they are pretty much the only ones that "create" circuits no one else tried, not even roland, korg, they have some pretty clever solutions for certain things, like extracting a sawtooth wave from a comparator.

Note that you have transcripted the schemo wrong. The upper transistor has its collector to the right.

Don't neglect the possibility I have missed something and I'm utterly wrong.

Thanks, that escaped me and for the help.

 

[Analog_Fan]

So, either there is a significant mistake in the printed values, or whoever designed the circuit took it from a cookbook and didn't know how to tune it.

the B.O.M. says 15K, as does the schematic, it was caused by because i can't zoom in further (max 97%) and it takes time to render that pdf.

 

[Analog_Fan]

I had not seen that the 220k is in the middle of two 15k resistors, not 1.5k. With these values, clipping appears at about -12/+10V.

Schottky diodes or normal 1n4148?
i got BAT85 and 1N60.

Seams the Schottky is best here.

i did not the last post you made, but the prevous.

 

[JohnRoberts]

View attachment 93422

Simulation of the circuit

according to the simulator it works perfect, but what would be the benefit?
input is 10V.p.p and out is 24V p.p.,

it's the output section of this:
Hexmix - Befaco
schematic

These are sometimes called diode break amplifiers, the diodes in the feedback network conduct at different voltages changing the transfer function.

For small voltages the feedback resistors dominate. When the output voltage is large enough to turn on the bottom NPN the output voltage is clamped (clipped). The steering diode doesn't conduct in this path for negative voltage swings. During negative voltage swings the top NPN reverse connected base-emitter junction zeners at roughly 6.8V.

The "philosophy" is unclear... asymmetrical clippers will generate different sounding distortion from symmetrical clipping. There is a possibility that the top NPN is drawn incorrectly. Reversing it would make a symmetrical level clipper.

JR

 

[Harvey Birdman]

Yes, just a drawing error, the emitter and collector are reversed on the top transistor.

 

[abbey road d enfer]

Yes, just a drawing error, the emitter and collector are reversed on the top transistor.

Even after correcting that mistake, simulation shows that clipping is not symmetrical. That is not a big issue, since the sound quality is already degraded when it happens, but it just shows the designer did not exercise the best of his talent...

 

[JohnRoberts]

That is part of the challenge from working with old hand drawn schematics (guess the actual circuit design).

For another example of a diode break clamp/clipper

There is a lot more going on but the two anti-parallel diodes near the top of the schematic create a soft clipper, to protect the ne572 gain element that would spike if its max current was exceeded.

JR

 

[Analog_Fan]

Even after correcting that mistake, simulation shows that clipping is not symmetrical. That is not a big issue, since the sound quality is already degraded when it happens, but it just shows the designer did not exercise the best of his talent...

this is how it sounds.

(since Befaco employed this circuit in another design before the mixer, that's where i discovered it some 2 years back, i assume they also used it here, on this module the didn't publish the schematic.)

as long as i have chosen the right resistor value and leaves 10V.p.p. clear and untouched accept when exceeding that "design requirement", it should be good.

 

[JohnRoberts]

this is how it sounds.

(since Befaco employed this circuit in another design before the mixer, that's where i discovered it some 2 years back, i assume they also used it here, on this module the didn't publish the schematic.)

as long as i have chosen the right resistor value and leaves 10V.p.p. clear and untouched accept when exceeding that "design requirement", it should be good.

It is good design practice to make clippers symmetrical. Clipping in one polarity before the other introduces a DC component into the signal... Some circuits can ignore varrying DC content, others don't. YMMV

JR

 

[Analog_Fan]

These are sometimes called diode break amplifiers, the diodes in the feedback network conduct at different voltages changing the transfer function.

For small voltages the feedback resistors dominate. When the output voltage is large enough to turn on the bottom NPN the output voltage is clamped (clipped). The steering diode doesn't conduct in this path for negative voltage swings. During negative voltage swings the top NPN reverse connected base-emitter junction zeners at roughly 6.8V.

The "philosophy" is unclear... asymmetrical clippers will generate different sounding distortion from symmetrical clipping. There is a possibility that the top NPN is drawn incorrectly. Reversing it would make a symmetrical level clipper.

JR

it's still a weird but cool looking circuit.
yeah, i figured that the resistor still obey normal feedback path.
Thank for explaining.

this is what i considered diode clippers so far.
This circuit I'm working on is now 368 days old (i have now 4 revisions).



A and C where my ideas, B came from a user on Stack Exchange.

C is expensive and my previous supplier couldn't deliver that, since he was going in pension in few he didn't want to put the effort.
the Opamp is strict and sharp.

the Befaco Circuit seams to add some charrater to the audio while A and B might not sound good, hate to breadboard.

 

[Analog_Fan]

That is part of the challenge from working with old hand drawn schematics (guess the actual circuit design).

For another example of a diode break clamp/clipper

There is a lot more going on but the two anti-parallel diodes near the top of the schematic create a soft clipper, to protect the ne572 gain element that would spike if its max current was exceeded.

JR

the circuit around Q1 (2N3904) is interesting.
i would never come up with C9 and R12.
: )

seems like a compressor, but the normal feedback is 44.4 k and the seemingly optional path is 45,8K (if that's a vactrol).
wouldn't than become the paralegal resistor in action? ... and would would still result in a value between 44.4k and 45.8K, right?
not sure what it is.

 

[JohnRoberts]

A is a simple voltage clamp,
B is less simple (and less linear) where output follows input until diodes run out of current and pull up/down resistors dominate.
C looks like one of those hand drawn schematics that doesn't make sense.

JR

 

[Analog_Fan]

B is less simple (and less linear) where output follows input until diodes run out of current and pull up/down resistors dominate.

don't own a spectrum analyzer yet, to expensive.

C looks like one of those hand drawn schematics that doesn't make sense.

C is a Analog Devices precision rail to rail opamp.

 

[JohnRoberts]

the circuit around Q1 (2N3904) is interesting.
i would never come up with C9 and R12.
: )

don't hurt you head... that is a very complicated circuit using features inside the NE572 in ways other than the IC maker intended. To understand that circuit you need to understand the innards of the NE572 (a cheap compander IC from last century).

seems like a compressor, but the normal feedback is 44.4 k and the seemingly optional path is 45,8K (if that's a vactrol).
not sure what it is.

Its a compander for tape noise reduction... compress the input then expand the output (similar to dbx NR. The goal is end up with a transparent recording reducing the tape noise floor.

That circuit is a rather non-linear compressor/expander side chain that does multiple tricks to respond quickly to short transients , then upon playback restores as much as possible. In an ideal world you will not hear this working, only quieter recordings.

JR

 

[abbey road d enfer]

don't own a spectrum analyzer yet, to expensive.

If you have a computer and a soundcard, it takes only loading one of the free softwares and you're good to go. REW, ARTA, Audacity...