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

[GGBB]

The NE5534 is cheap, widely available in DIP, has very low noise, low output impedance (i.e. strong drive capability), low distortion and is externally compensated for noise gains less than 3. It's one of my favorite audio op amps.

Do you ever have problems or see reduced performance (noise or whatever) with it running off of 9V? Recommended minimum supply for it is 10V. I'd assume the first spec to suffer at lower voltages would be noise, but I really don't know.

[jonny.reckless]

I've not noticed a problem at 9V but I've never put it on the AP to measure it at that voltage. It definitely gets a bit grainy at 5V. It consumes about 4mA at 9V input so I guess the internal bias circuits are OK.

Beware cheap NE5534s off eBay. I've been caught out with some absolute crap. I buy mine from mouser or digikey now.

[jonny.reckless]

What about gate LED indicator?

It's pretty straightforward to rig one up but I just use my ears 

[jonny.reckless]

I finally got the PCB for this debugged and working.

Here is the schematic:


Photo of the PCB:



Video of the unit in operation:

Many thanks to Merlin for sharing his fantastic design.

Let me know if you would like a PCB. I can send you one for $10 plus postage. My email is jonny@recklesstechnology.com

[effexfreek]

Let me be the first to warmly congratulate YOU ! Am-A-Zing job on the integration of the gate !

You mention adding a bit of hysteresis to the gate, preventing chatter (during note fade-out, I gather ?)... does the current schematic reflect this ?


Wow... & love the playing ! (btw, great songs on YT, too !)


P.S. : could a 2SK117GR sub for the J112 with minimal tweaking, you reckon ?

Double P.S. : separate Attack & release (Decay ?) pots. would be GREAT for SlowGear-type volume swell effects ! Hopefully possible ?
                       (please-please-please-please-please-please ?)
                       That'd make it the ULTIMATE pedal for me !

[jonny.reckless]

Let me be the first to warmly congratulate YOU ! Am-A-Zing job on the integration of the gate !

You mention adding a bit of hysteresis to the gate, preventing chatter (during note fade-out, I gather ?)... does the current schematic reflect this ?
Current schematic is what I am playing in the video.


Wow... & love the playing ! (btw, great songs on YT, too !)
Thanks!

P.S. : could a 2SK117GR sub for the J112 with minimal tweaking, you reckon ?

Double P.S. : separate Attack & release (Decay ?) pots. would be GREAT for SlowGear-type volume swell effects ! Hopefully possible ?
                       (please-please-please-please-please-please ?)
                       That'd make it the ULTIMATE pedal for me !

You could replace each of the attack / decay resistors with a 1 megohm log pot.

[Elijah-Baley]

Wonderful updating of this compressor! Though I'm not a great fan of compressor effects, I'm building this ultra settable one using the Sabrotone layout (probably including the two missing LEDs that it doesn't have).
http://www.sabrotone.com/?p=2901
Actually, I stopped for missing parts.

Your circuit gets the schematic more complex, but I'm wondering if I could include your noise gate on a daughter board, I'm still trying to follow the schematics (original and new), I'd need more time.

Thanks!

[effexfreek]

HEY, jonny. !



"You could replace each of the Attack / Decay resistors with a 1 megohm log pot"

-The Decay must be R20 (release ?) in your schem., I'm guessing ?
-Which would be the Attack resistor ? (R16 and/or R15, probably ?) (just after the gate rectifier ?)



Thanks in advance !


P.S. : and please, keep us updated on the hysteresis "tweaks"

...this is getting real exciting !

[jonny.reckless]

HEY, jonny. !



"You could replace each of the Attack / Decay resistors with a 1 megohm log pot"

-The Decay must be R20 (release ?) in your schem., I'm guessing ?
-Which would be the Attack resistor ? (R16 and/or R15, probably ?) (just after the gate rectifier ?)



Thanks in advance !


P.S. : and please, keep us updated on the hysteresis "tweaks"

...this is getting real exciting !

The compressor attack and release are RV4 and R23 respectively.
Gate release is R20.

Gate attack with this design would be tricky. You'd probably need a dual gang pot and replace R15 and R16. I'd suggest trying maybe 1M dual log pot. I haven't tried this.

When I get round to doing the next iteration of the gate design, I'll add hysteresis to the threshold and make it so you can set attack and release of the gate with a single pot. This is probably going to make the PCB bigger so it will no longer fit into a 1590B enclosure. I'd prefer to keep it all through hole to make hand assembly easy.

[effexfreek]

Yes , of course : Gate Attack is what I meant...



I get your point about "...attack... would be tricky." : I figure the Gate capacitors (C13 & C11 ?) need to be discharged, for a quick, new Attack to "register". I'll try different schemes, play around with it.

WRT size : I place components "standing" & use tiny-mini pots, so I can get most all circuits quite "miniaturized", so absolutely NO worries for me...



Thanks-a-million, jonny. for the ongoing R&D : Merlin's ET is such a great platform !

I'll be standing by...

[jonny.reckless]

I think I have a design for a more flexible noise gate with hysteresis and variable attack and decay time. This becomes an 8 knob compressor gate (threshold, ratio, attack, release, gate threshold, gate attack, gate release, volume) which is starting to get out of hand

I have also added an LED which illuminates when the gate is providing gain reduction.

I'll post back here once I have it fully debugged. Give me a couple of weeks.

[rankot]

Why not using Fencepost noise gate design (as I have already proposed here), it uses another half of LM13700 and one more op amp?

[GGBB]

Is there any particular advantage to building the gate into the compressor, instead of having a separate gate pedal in front of the compressor?

[rankot]

Yes, smaller footprint. Especially if they use the same IC.

[GGBB]

Yes, smaller footprint. Especially if they use the same IC.

Right - so what I meant was technical advantage - something other than the obvious.

[jonny.reckless]

Since you already have a VCA inside the compressor, it makes more sense to build a compressor / gate with a single VCA rather than having 2, with twice the noise, distortion, component cost etc. Plus it's probably going to take up less room on your pedal board.

Separate units might not play nicely together depending on settings.

[jonny.reckless]

Why not using Fencepost noise gate design (as I have already proposed here), it uses another half of LM13700 and one more op amp?

Because I can achieve much better noise performance, and greater flexibility, with fewer components.

Merlin's engineer's thumb is a really excellent, economical, simple, high performing design. It's actually quite inspiring how well it works for the circuit simplicity. I really like this philosophy, and try to follow the same principles when doing my own designs and amendments.

[GGBB]

Since you already have a VCA inside the compressor, it makes more sense to build a compressor / gate with a single VCA rather than having 2, with twice the noise, distortion, component cost etc. Plus it's probably going to take up less room on your pedal board.

Separate units might not play nicely together depending on settings.

Thanks, and please excuse my ignorance of the circuit. I've tried comparing the original schematic with the new one that includes the gate, but they look so different that I'm not really making much sense of it yet.

Separate pedal obviousness aside, the "same VCA" bit puzzles me. Is the gate leveraging the VCA used by the compressor - and I mean the same VCA, not the other VCA in the same chip. If it's just using the second unused VCA in the LM13700, then isn't that really two VCAs? And therefore how is two VCAs in one pedal any different noise and distortion wise than two VCAs in separate pedals? I'm really interested in differences that are actually going to matter on the pedalboard playing live, not hifi/studio level differences.

Also - "might not play together nicely" - does the built-in gate prevent that merely because you can't adjust the settings whereas a separate gate could be set up poorly to work with a compressor? In other words, I think you are talking about the obvious, but I'm not sure because of the way you phrased it.

Thanks.

[effexfreek]

Hey, jonny. !



   A coupla' quick questions regarding your updated ET schem. :


(scratch this one  : samhay below noticed you switched to an inverting configuration ! So you could attenuate, as well as compress ?)
A) Merlin's ET has a 10k between OTA out & audio-path opamp's negative input. Your 10k (R6) is in series with RV2 sustain (ratio) pot... Design tweak ? Schem. error ?

B) You omit the 220R between OTA's inputs... better S/N ratio ?

C) You have a dual/parallel network of 10k/1uF & 10k/10n between Input FET & IC1... real curious about that.



Thanks in advance !

[samhay]

Here is the schematic:

Cool. You're using IC1 in an inverting configuration, which means you can, at least theoretically, squish below unity gain. This is quite different to the ET.
If you feed it a hot signal, do you get unity gain out of it? If not, perhaps a make-up gain stage on the end would be helpful - could use an NE5532 to do both stages.