THE ENGINEER'S THUMB... At last, a better compressor!

[rankot]

Which is the version of schematic? Which version of pedal - Merlin's, or Jonny's?

[Rob Strand]

FWIW, most of the schems, including later ones,  don't show how to wire the Threshold clearly.    The configuration in Reply #617 is perhaps the most detailed but it has the socket. and two 1M resistors and two caps.     In the original circuit (any version ISS3, ISS4) , the 1M resistor is *replaced* by a 1M pot in series with 220k and the threshold circuit "IC B" taps off the pot wiper.

https://www.diystompboxes.com/smfforum/index.php?topic=97123.msg1162622#msg1162622


Merlin's Engineer's Thumb Compressor
Version summary and links


Original Thread
"THE ENGINEER'S THUMB... At last, a better compressor!"
merlinb, April 21, 2012,
https://www.diystompboxes.com/smfforum/index.php?topic=97123.0


Issue   Year   Link / Changes

(Ver.1)   2012   http://www.valvewizard.co.uk/engineersthumbschem.jpg
      http://www.valvewizard.co.uk/engineersthumb.html

      PDF version
      http://www.valvewizard.co.uk/engineersthumb.pdf
      - Part Designators added
      - 10uF caps shown explicity as electros
      - 1uF caps shown explicity as non-polar
      - Pot taper for LEVEL pot shown as 10k log

Ver.2   2016   http://www.valvewizard.co.uk/engineersthumbschem2.jpg
      http://www.valvewizard.co.uk/engineersthumb2.html

      Changes:
      - Opamps changed from 2xdual to 1xquad
      - 1uF and 10uF caps shown explicitly as Electros
      - Attack pot added
      - 33R added to +9V PSU rail
      - +4.5V divider cap increased from 100n to 1uF
      - 1k feedback on U1a (now IC A) removed
      - Ordering of 1k and 10u output network swapped

Ver.2   2016/7   PDF version
      http://www.valvewizard.co.uk/engineersthumb2.pdf
      - first PCB
      - 2x47uF caps (C6 & C9) on +9V
      - 1uF caps shown explicity as non-polar
      - Pot taper for LEVEL pot shown as 10k log

Iss.3   2018   PDF version
      http://www.valvewizard.co.uk/engineersthumb3.pdf
      http://www.valvewizard.co.uk/engineersthumb3.html
      - Both OTA's used to reduce noise  (following Jonny Reckless's suggestion)

Iss.4   2019   PDF version      
      http://www.valvewizard.co.uk/engineersthumb4.pdf
      http://www.valvewizard.co.uk/engineersthumb4.html
      - BJT output buffer added (part of OTA)
      - Effect LED powered from +9V input supply
      - Optional compression LED
      - Optional Threshold pot now has 220k limit on min setting.   
      - 4.5V bias dividers changed to 100k's and cap increased to 10uF electro
      

Other Circuits:

A0   2018   The Engineer's Thumb Redux, CT030, Jonny Reckless
      Reply #480, page 25 of main thread
      https://www.diystompboxes.com/smfforum/index.php?topic=97123.msg1124984#msg1124984

      - Both OTA's used to reduce noise
      - NE5534 opamp used to reduce noise
      - Input buffer added; as part of noise reduction scheme   
            - Noise gate added

Iss.3 mod 2019   merlinb, reply #617, Separate side-chain   
      https://www.diystompboxes.com/smfforum/index.php?topic=97123.msg1162622#msg1162622


Compression and Gain Curves for Ver.2 2016/7
https://www.diystompboxes.com/smfforum/index.php?topic=130630

[Gandalf_Sr]

I'm a retiring EE design engineer that's tinkering for fun and I just started to learn electric guitar after messing around 40+ years ago.

Anyway, I came up with a 1590B enclosure PCB version of the Engineer's Thumb (vn 4) and thought I'd share the project here.  To undertake this, you have to be up for surface mount work on SOIC ICs and 0603 discrete components.  I work under a stereo microscope and have all the SMT tools so, to me, this is easy.

The version 1a that I show in the pictures has an error that there's a missing resistor that's added as a wire-leaded part .  This is corrected and added in the 1b schematic attached.  The PCB is double-sided and uses an approach that involves fitting the 1/4" jack sockets and then cutting the protruding pins flush on the Pot side, then mounting the Pots using 2-sided tape.

I play through it for the first time yesterday and I think I want to add a release Pot to the design (that could be omitted).  I am still messing with the Pot values and think the Attack isn't working very well.  Any input or suggestions for change can be tested on the vn 1a PCB and then rolled into the vn 1b PCB.

I'm not aiming for a commercial product and I'm willing to provide PCBs at cost if anyone wants them.

[ThermionicScott]

What a way to make an entrance!  Welcome, Gandalf_Sr.   

[merlinb]

Greetings, fellow wizard...

I think the Attack isn't working very well. 

In what way?

[Gandalf_Sr]

Greetings, fellow wizard...

I think the Attack isn't working very well. 

In what way?

Well, the volume level seems to drop quite quickly which is what I think Attack does.  I just don't seem to get a smooth progression from no effect to the pedal backing down the load sounds.  The Attack Pot is a 10k Log so maybe I have it wired up backwards, your notes did say something about it working backwards and I struggled with which end was the clockwise end on your diagrams (marked cw perhaps?), in my schematic, the arrow points to the fully cw position.

Eh up! Lancashire eh?  I live in the US now but I was born and raised in Morecambe/Lancaster.

[merlinb]

Well, the volume level seems to drop quite quickly which is what I think Attack does. 

In a normal compressor, even the longest attack is 'quite quick'. Longer attack times do not change the volume drop as such, rather it changes the 'snappyness', because transients are better preserved. Sounds like it may be working normally...

[Gandalf_Sr]

Well, the volume level seems to drop quite quickly which is what I think Attack does. 

In a normal compressor, even the longest attack is 'quite quick'. Longer attack times do not change the volume drop as such, rather it changes the 'snappyness', because transients are better preserved. Sounds like it may be working normally...

Thanks Merlin 

Can you clarify how the Attack pot should be connected in terms of pins 1,2,3 as shown in the attached diagram?

Thanks in advance.

[merlinb]

Can you clarify how the Attack pot should be connected in terms of pins 1,2,3 as shown in the attached diagram?

For the smoothest range of attack control you want to be useing pins 1 & 2. Short pin 3 directly to the pin 2.
In other words, referring to the circuit diagram you posted, swap 2 and 3. This only changes how fiddly the pot is to set, it won't change the actual range of tones available.

[Gandalf_Sr]

Thanks Merlin.

I can see that by shorting pins 2 & 3 I'm changing the voltage on C2 (my reference designators) because the output of U2.2 (pin 7) feeds through the diode/resistor network to charge/discharge C2.  What's confusing me is that I assumed that the current flowing through R5 is the attack current and what you suggest means that the Pot is in series with R5 before it gets to C2 but that would also mean that the release current would flow through R1 and also the Pot so I'd be affecting both the attack and release times when I varied the Pot setting.

Is this right or have I missed something?

[merlinb]

the release current would flow through R1 and also the Pot so I'd be affecting both the attack and release times when I varied the Pot setting.

Yes it does affect the release, but the release time is mainly set by R1 which is so large in comparison that the variation is too small to be noticeable.

[Gandalf_Sr]

Was that a change you specifically wanted? In your original version you had the release path as a 470k resistor that was connected directly to C4 (your reference designators) - see attached.  It would be simple to wire up the release path in my setup to match this fixed-value release resistance.

[merlinb]

Was that a change you specifically wanted?

It was a by-product of wanting to get two usage options out of the same pot position, and it ended up in that specific configuration because it was the easiest to lay out. You can do it differently if you prefer.

[Gandalf_Sr]

Was that a change you specifically wanted?

It was a by-product of wanting to get two usage options out of the same pot position, and it ended up in that specific configuration because it was the easiest to lay out. You can do it differently if you prefer.

OK, thanks Merlin

[PRR]

....the release current would flow through R1 and also the Pot so I'd be affecting both the attack and release times when I varied the Pot setting.

Like from 470k to 480k. And it isn't like 470k is some magic value.

[Gandalf_Sr]

....the release current would flow through R1 and also the Pot so I'd be affecting both the attack and release times when I varied the Pot setting.

Like from 470k to 480k. And it isn't like 470k is some magic value.

How can it NOT be a "magic" value if Merlin came up with it? 

[Gandalf_Sr]

I ran some more tests and I think it's working.  If you look at the attached scope screen capture, the yellow trace (1) is the input signal running at 1 kHz; it starts at 15 mV pk-pk and then steps up to 500 mV pk=pk.

The cyan trace (2) is the output and it shows that, with a 15mV input, the output runs at about 400 mV and then, when the amplitude step comes along (ignore the junk inbetween, it's my signal generator changing and settling as I punch in a new value on the keypad) the output is clamped down to a max of about 600 mV.  For these tests, the Level pot is set to about 90%, the attack pot is set fully clockwise (max time of about 30 mS, and the ratio pot is fully to the Wet (compressor signal) side.

Part of what I'm missing is how it's supposed to work - I think it's supposed to keep the output at a constant(ish) amplitude regardless of the input level.

Does this seem about right?

[PRR]

> starts at 15 mV pk-pk and then steps up to 500 mV pk=pk.

That's a 30dB jump. VERY large in audio terms. I can't fault it for a few cycles of clipping under that abuse.

[Eb7+9]

Does this seem about right?

yeah // matches my sims, even the one-sided clipping ... btw, this has nothing to do with "signal jump" lol - a good limiter circuit should catch that stuff easily without barfing // that's what they're made for ... not to mention that 500mVpp is typical output for a humbucker, so nothing out of the ordinary here ...

this limiter is nothing but a discrete variant of the 570/571 circuits and suffers the same fate under largish signals, exactly why the 570/571 was generally avoided in broadcast and eventually died a quiet death ... it's pretty obvious why an OTA never belonged in there in the first place - not to mention the excessive noise a trans-impedance amp and active feedback combo tends to produce when not careful ...

one guy I know had them work at modest 2:1 ratios and made a living in the 70's installing them inside guitars - without controls, similar in overall function to the Schaffer Replica ... but I'm told they still clipped on buckers // likely because he followed the datasheet examples to the letter ...

the headroom fix is actually rather simple if you think about it, the noise is as well ... in this case the circuit is simply cobbled together incorrectly, where the author overlooks an obvious constraint in the topology ... a careful study of the 570/571 AGC datasheet examples can help show why, where and how to fix this ... but it's unlikely you can ever get a general AGC circuit following this approach // historical evidence suggests it's really meant as a simple one-IC guitar squeezer/gate wonder and no more ...

[rankot]

this limiter is nothing but a discrete variant of the 570/571 circuits and suffers the same fate under largish signals, exactly why the 570/571 was generally avoided in broadcast and eventually died a quiet death ... it's pretty obvious why an OTA never belonged in there in the first place - not to mention the excessive noise a trans-impedance amp and active feedback combo tends to produce when not careful ...

one guy I know had them work at modest 2:1 ratios and made a living in the 70's installing them inside guitars - without controls, similar in overall function to the Schaffer Replica ... but I'm told they still clipped on buckers // likely because he followed the datasheet examples to the letter ...

the headroom fix is actually rather simple if you think about it, the noise is as well ... in this case the circuit is simply cobbled together incorrectly, where the author overlooks an obvious constraint in the topology ... a careful study of the 570/571 AGC datasheet examples can help show why, where and how to fix this ... but it's unlikely you can ever get a general AGC circuit following this approach // historical evidence suggests it's really meant as a simple one-IC guitar squeezer/gate wonder and no more ...

You compare NE570 to Engineer's Thumb or something else?