Need help with Engineer's Thumb compressor build

[Eb7+9]

Measurements (please note that I don't have that 4k7 resistor on wiper), plus probe is at op amp output pin 6 and minus probe at 4.5V Vref:

With NE5534: -434mV with Ratio at minimum (0 Ohm) and -195mV with Ratio at max (1M).
With TL071: 0mV with Ratio at minimum (0 Ohm) and -6mV with Ratio at max (1M).

Quite a difference.

I have also measured voltage difference, with 4k7 resistor on wiper - plus probe at op amp output pin 6 and minus probe at 4.5V Vref:

With NE5532: -300mV with Ratio at minimum (0 Ohm) and -195mV with Ratio at max (1M).

So it is similar to NE5534 (of course), but it is working?

finally, some real numbers // well done Rankot ... completely as expected
anybody with any real knowledge of electronics get where this is all headed ??!

like I suggested three times already, head over to a 570/571 datasheet or app note and look carefully at any AGC/compressor schematic found there, which is where the THUMBS is heavily borrowed from, and see what's missing here for achieving 100% repeatable behavior (what engineering aims for) ... for example, see figure 14 in the following :

https://www.onsemi.com/pub/Collateral/AND8159-D.PDF

notice in particular how the 570/571 IC has a plain old bipolar op-amp in it ... in fact, the discrete three-chip combo that one guy I know makes and has been installing in guitars since the 70's (like Angus Young's for example) uses NE5532 op-amps instead of the internal 570/571 one - but only because the internal implementation is on the crappy/noisy side ...

intrinsically, the problem has nothing to do with the op-amp type used ... the correct version of the circuit will handle any op-amp, however crappy

Until Ranknot tells us what circuit he thinks he has built, this thread is just the blind leading the blind.

um, wrong thread dude ... init? 

[merlinb]

which is where the THUMBS is heavily borrowed from,

...no it isn't

and see what's missing here for achieving 100% repeatable behavior (what engineering aims for) ... for example, see figure 14 in the following :

...a wholly different topological block diagram, which he tells us is secret knowledge but which he won't actually explain to us

that one guy I know makes and has been installing in guitars since the 70's (like Angus Young's for example)

...name drop

... the correct version of the circuit will handle any op-amp

...demonstrates lack of knowledge of opamps and their properties

um, wrong thread dude ... init? 

... mic drop

dude you sound like that kid at school who tells you he is 10th dan black belt in karate but he can't actually show you any moves because reasons

[puretube]

So, is Eb7+9 talking about extracting the CV-envelope from the output instead of from the input?
Or about AC-coupling the rectifier-section to its signal-source? (instead of DC-wise).
(Be it taken from in- or output ...)

[ElectricDruid]

So, is Eb7+9 talking about extracting the CV-envelope from the output instead of from the input?
Or about AC-coupling the rectifier-section to its signal-source? (instead of DC-wise).
(Be it taken from in- or output ...)

If he is, that's the basic "feed forward" versus "feed back" distinction amongst compressors, isn't it? I've never been very clear on the pros and cons of each design (I'm sure each has its strong points) but I've never researched compressors in depth.

[puretube]

Yes, my Druid, it is indeed - I just wanted to avoid these two expressions

[ElectricDruid]

Yes, my Druid, it is indeed - I just wanted to avoid these two expressions

Well then, I'm sorry to have brought them up!

[puretube]

Don`t worry - just don`t want to fuel up that distinction-discussion ...
The "OR"-part of my question is what I am more eager to get to know.
(I`m on the edge of starting to breadboard an E.T. just out of interest, to check what differences show up between hooking up the rectifier to in-/out-put, between doing so AC-/DC-coupled, and maybe even use 3 single opamps with distinct/decoupled Vref, maybe even AC-coupled, to avoid any accumulating offsets ...).

What I do know about the difference between FFW vs. FB control stems back from my 1971 invention of the "Auto-Wah" (Envelope-Filter), which in its very first instance derived the CV from the output of the filter. This has never been done in any commercial product, except for the EHX "Tube Zipper", where the "Trill/Tron"-switch lets you choose the FFW or FB-option.

Compressors until now never occured to me as "sound effects", that`s why I never delved into them too much. (Except for using commercial units for studio-use).

[PRR]

...I've never been very clear on the pros and cons of each design...

Audio Dynamic Range Compression For Minimum Perceived Distortion
BARRY A. BLESSER, Student Member, IEEE
IEEE Transactions On Audio And Electroacoustics Vol. Au-17, No. 1 March 1969

[ElectricDruid]

...I've never been very clear on the pros and cons of each design...

Audio Dynamic Range Compression For Minimum Perceived Distortion
BARRY A. BLESSER, Student Member, IEEE
IEEE Transactions On Audio And Electroacoustics Vol. Au-17, No. 1 March 1969

Thanks Paul, I'll check that out.

[Eb7+9]

So, is Eb7+9 talking about extracting the CV-envelope from the output instead of from the input?

nope, he isn't ...

referring to the SE570/NE570 datasheets for what clearly is a stab at establishing proper nomenclature
we see the FFW topology is called ALC and FBK topology is called Compressor
(I've referred to THUMBS as an ALC type loop all along for this reason)

the blunder applies to both cases

[puretube]

Thank you, JC!
I`ll keep searching for the riddle that is left.

[PRR]

... referring to the SE570/NE570 datasheets for what clearly is a stab at establishing proper nomenclature
we see the FFW topology is called ALC and FBK topology is called Compressor

https://www.onsemi.com/pdf/datasheet/ne570-d.pdf

I do not see the term "ALC" anywhere in this datasheet. Is there maybe an AppNote which uses "ALC"?

[DrAlx]

... referring to the SE570/NE570 datasheets for what clearly is a stab at establishing proper nomenclature
we see the FFW topology is called ALC and FBK topology is called Compressor

https://www.onsemi.com/pdf/datasheet/ne570-d.pdf

I do not see the term "ALC" anywhere in this datasheet. Is there maybe an AppNote which uses "ALC"?

https://www.onsemi.com/pub/Collateral/AND8159-D.PDF    page 8

[puretube]

"ALC" is the new "AGC" (like we called it backindadayz...)
or at least it seems so ...

[PRR]

.....AND8159-D.PDF    page 8

Thank you.

> "ALC" is the new "AGC"

Both terms (and "AVC"!) are old, and there is some slight difference in concept, if not in the store/workbench.

Limiters rather imply a threshold. Sustain may or may not (threshold may 'just happen'). The '570 is intended to be "wide range" (over 100dB) which means you need to do some thinking.

> the FFW topology is called ALC

As Blesser's full paper points-out: it is real awkward to detect very-small signals at the output. Or in our terms: noise gates are nearly-always input-sensing, feed-forward. (If it actually gated, the output would be zero forever.)

Feedforward limiters require simple predictable detector and gain control laws. Feed-back limiters can use all sorts of crappy gain-cells and even lame detectors, because they tend to reduce their own errors.

I still do not know why a bunch of Thumbs are distorting this week. But my 2 cents is on mistakes or, this year, bogus rare parts.

[Eb7+9]

.....AND8159-D.PDF    page 8

Thank you.

> "ALC" is the new "AGC"

Both terms (and "AVC"!) are old, and there is some slight difference in concept, if not in the store/workbench.

Limiters rather imply a threshold ...

I still do not know why a bunch of Thumbs are distorting this week. But my 2 cents is on mistakes or, this year, bogus rare parts.

AGC is the umbrella term to mean a host of self-inflicted Gain alteration types of response ... to (signal) envelope

Limiters, control-loop (servo) Compressors and ALC's, Expanders are all examples of AGC circuits ...
throw in Gates, Swellers and Duckers, etc ... if you want to be exhaustive

---

Threshold is present in both Limiters and comp/ALC's but the meaning changes in either case
from what I've noticed the problem seems to start when porting over a Limiter interpretation of Threshold to the control-loop scenarios

the servo-loop circuits can either be fed FWD to produce what the 570 datasheet calls an ALC response and fed BCK to produce what the 570 data sheet calls a Compressor response. A canonical set of curves for the ALC case is generally much flatter, reaching perfect flatness as RATIO goes to infinity, while the Compressor set is not so much so.

In both these cases, Threshold has the same meaning

---

THUMBS didn't start messing up this week,

the best case so far, a video that was online as of May of 2020 entitled "Problems" but is no longer there ... (I can post screenshots)
in it we see live scoped wave-forms that are exactly what I found on the bench and on the simulator running the stock circuit

turns out the problem is systemic and has two components to it

I'll be posting about all this more in a few weeks once I'm done this one project ...

[matopotato]

I've tried reading this doc a few times, and it is currently a bit above my head. There is a rectifier cell and a deltaG cell. The rectifier seemed more of an ideal function according to a russian lady on YT and some more digging might reveal more insight. But the delta G cell evades me and I end up in chemistry...

---

THUMBS didn't start messing up this week,

the best case so far, a video that was online as of May of 2020 entitled "Problems" but is no longer there ... (I can post screenshots)
in it we see live scoped wave-forms that are exactly what I found on the bench and on the simulator running the stock circuit

Would that be related to issues with too much distortion? Or is this about some other aspect/issue with the circuit perhaps.

turns out the problem is systemic and has two components to it

Very interesting. I have been feeling like the hobbyist newbie who cried distortion (or wolf or lion in some cultures) so quite interested in this.

I'll be posting about all this more in a few weeks once I'm done this one project ...

I'll do my best to not be the voice from the backseat going "are we there yet?"

[puretube]

One of my suspicions is: maybe using a control-voltage via a 10k-ish resistor into the Iabc-pin, instead of the transistor`s collector-current directly. (Which might be exponentially deformed, at low currents).

[matopotato]

turns out the problem is systemic and has two components to it

I'll be posting about all this more in a few weeks once I'm done this one project ...

I'll do my best to not be the voice from the backseat going "are we there yet?"
[/quote]

Are we there yet?

[percyhornickel]

After reading many posts everywhere I made some changes and I could eliminate the distortion in my circuit. I etched the pcb with the effectslayouts.blogspot.com published file with pdf 2012 layout:

https://www.dropbox.com/s/purgnc68qed3n7t/Engineer%E2%80%99s%20Thumb%20build%20doc.pdf?dl=0

This seems to be the most popular version since many companys based their pcbs on this one.

The changes I did from this layout after a deep search to solve the distortion issue:

Ic: LF353P instead TL072 (Both)

Release B500K
Release lug 1 in series with 220K Resistor
Release 1 and 2 connected

Attack 100K Log
Attack lugs 2 and 3 connected

Ratio 1 inverted (the schematic shows lug 1 the opposite direction showed in the layout).
Ratio lugs 2 and 3 connected
Ratio 2 in series with 4.7K Resistor

Threshold lug 3 in series with 220K Resistor

Now works really good, no distortion at all.

Percy