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

[Phend]

Very invisible-sounding.

Seems to be on par with Compressors.
However if you have one you will most likely have it turned on all the time.
Because it is there making a difference.

[Mark Hammer]

I should emphasize that "invisible" does not mean it isn't doing anything useful.  I just got tired of seeing and using "transparent".

Nice design, Merlin!   

[PRR]

I should emphasize that "invisible" does not mean it isn't doing anything useful.  I just got tired of seeing and using "transparent".

"I was at an amusement park with my friends. They all said the invisible roller coaster was great, but I didn't see the attraction."

[Eb7+9]

I am not sure how the Ratio control works ...

I wonder if there's something up with varying the RATIO the way Merlin does here ...

One down-side I can see is that high RATIO/gain can lead to signal spiking in the first half-cycle because of the HWave side-chain only responding on the second-half - that is, when the side-chain has recovered and is starting again from rest ...

in contrast, the close-to-theory profile of RATIO variation curves from a single constant THRESHOLD unity-crossing point was shown in EBK's post recently for the THATcorp OneKnobSqueezer : https://www.diystompboxes.com/smfforum/index.php?topic=128654.msg1240074#msg1240074

also, if you run SPICE sims of the NE570 feed-fwd ALF topology you get similar curve sets as the THAT one-Knob topology ... again, from completely different inner-block circuitry

maybe somebody will extract measured RATIO curves on their unit so we can compare ...

[jhsa]

Hi guys, having a problem here, need your help please. Built this on BreadBoard, except for the compressor part at the bottom of the schematic. D7, R17 and LED5 are in place. I just want the noise gate.
The problem is that every time the gate switches there is a thump sound, and also it seems to be working reversed. If I force the gate to be always closed I have sound. But when it opens, it mutes the sound. It should be the other way round. Can't the noise gate work without the compressor circuit?
Used a TC1044S for the charge pump and a BC557 for the charge pump transistor, but I am reading around -7V instead of -8V
Also,  MMBF5457 instead of the j112.
Thanks.

EDIT:

Placing some resistor, say 10k, between the anode of LED5 and 0V allows it to work, but I have the feeling that is not the solution
I might need some similar circuit to the one using TR2?

[jhsa]

Ok, this is what i have so far, and it is kinda working. There are some things I might still change, like replacing the jFET input buffers with another Opamp. The FETs fail quite easily. Just had one failing and it took me nearly half hour to find out what happened. The Opamp might also be less noisy. Also need to tweak some other values.
Added the possibility to place the Noise Gate at the end of the effects chain, but trigger it using the clean guitar sound from the PedalBoard's input buffer. This will be on a switch so I can select to trigger from here or from the Noise Gate input itself.

Not sure the workaround I have found  to make this Noise Gate work without the Compressor circuit is electronically correct, but it seems to be working.

[PRR]

Both your FETs need gate-leak resistors. And the symptom is that they can go full-off and take days to recover.

[jhsa]

Both your FETs need gate-leak resistors. And the symptom is that they can go full-off and take days to recover.

So, just put am 1M resistor from gate to GND?

[jhsa]

Anyway, never mind about the jFETs  I replaced them with an Opamp. This is what I have at the moment on the breadboard. Seems to be working quite well. I am still having a couple little problems though.
a) When the gate is active there is still a very tiny amount of sound leaking. I mean, it is nearly nothing, but I would like to sort it if possible. I could try decreasing R30 and see if that works.. Is there any problem in doing that? i don't want to damage any component.

b) I had to increase the values of the resistors connected to the Noise Gate Attack and Release pots because I could hear some "Thump" sound when the gate switched either way. The attack seems to be Ok and quite fast. but the release seems to have gone a few milliseconds slower. If i decrease the resistor I start to hear this little thump. Is there any other way to mitigate this problem?

Thanks.

[jhsa]

a) When the gate is active there is still a very tiny amount of sound leaking. I mean, it is nearly nothing, but I would like to sort it if possible. I could try decreasing R30 and see if that works.. Is there any problem in doing that? i don't want to damage any component.

b) I had to increase the values of the resistors connected to the Noise Gate Attack and Release pots because I could hear some "Thump" sound when the gate switched either way. The attack seems to be Ok and quite fast. but the release seems to have gone a few milliseconds slower. If i decrease the resistor I start to hear this little thump. Is there any other way to mitigate this problem?

Thanks.

No one?  or is this already a dead thread? It would be a shame if that is the case.

[rankot]

Anyway, never mind about the jFETs  I replaced them with an Opamp. This is what I have at the moment on the breadboard. Seems to be working quite well. I am still having a couple little problems though.
a) When the gate is active there is still a very tiny amount of sound leaking. I mean, it is nearly nothing, but I would like to sort it if possible. I could try decreasing R30 and see if that works.. Is there any problem in doing that? i don't want to damage any component.

You can try to decrease it, but be careful not to exceed maximum current drive of U2.1. Or try to decrease R25.

[rankot]

b) I had to increase the values of the resistors connected to the Noise Gate Attack and Release pots because I could hear some "Thump" sound when the gate switched either way. The attack seems to be Ok and quite fast. but the release seems to have gone a few milliseconds slower. If i decrease the resistor I start to hear this little thump. Is there any other way to mitigate this problem?

Try using TL062 instead of TL072 for U4, and leave the resistors as they are. Thump is probably caused by that IC driving LED. Also try using LTC1054 instead of TC1044, it has more current capability if I remember well.

[Eb7+9]

... anybody got a sampling scope and triggered source they can apply here ?!
would be interesting to see first couple of signal cycles (from dead rest) against span of gain settings

[jhsa]

Also try using LTC1054 instead of TC1044, it has more current capability if I remember well.

I only have the TC1044, so that will have to do.  Also reducing R30 doesn't work, the LM kinda stops working.
I wonder if i should use a rail to rail IC for IC4? It seems that the OTA is not fully turning ON to derease the other Opamp's gain further down.

[rankot]

Did you try to reduce R25 which is actually limiting OTA bias current? It is 4k7, which will limit bias current to 2mA for BOTH OTAs at 9V, which is maximum limit for single one. So you can try to reduce this one to 2k7, or even less. The fact is that voltage at R27 and R28 junction can't be over 5.5V (due to voltage drop of LED and 1N4148 in series), so to achieve maximum current of single OTA, you need 2k7, and for both 1k3. So I suppose 1k5 is the lowest safe value for R25.

[jhsa]

Tried 2k2 for R25, no improvement.
I think I might ditch the LM13700 and use an LDR + LED to control the Opamp's gain.

[jhsa]

As I ended up using an LED/LDR combo instead of the LM13700 chip, I have decided to create a new topic for the noise gate I am building. Here is the link in case you are interested.

https://www.diystompboxes.com/smfforum/index.php?topic=128883.msg1243017#msg1243017

[jonny.reckless]

Anyway, never mind about the jFETs  I replaced them with an Opamp. This is what I have at the moment on the breadboard. Seems to be working quite well. I am still having a couple little problems though.
a) When the gate is active there is still a very tiny amount of sound leaking. I mean, it is nearly nothing, but I would like to sort it if possible. I could try decreasing R30 and see if that works.. Is there any problem in doing that? i don't want to damage any component.
b) I had to increase the values of the resistors connected to the Noise Gate Attack and Release pots because I could hear some "Thump" sound when the gate switched either way. The attack seems to be Ok and quite fast. but the release seems to have gone a few milliseconds slower. If i decrease the resistor I start to hear this little thump. Is there any other way to mitigate this problem?
Thanks.

Try increasing C15 which slugs the switching of the comparator in the threshold detector. It should swing in about 1ms or so. Much faster and the dv/dt transients are likely to be audible unless you have a decent ground plane.
Make sure you have a 1uF or 10uF ceramic cap decoupling the op amp right on the pins.
Keep the ground return of the diode / LED separate from audio ground.
To improve the attenuation when the gate is off, trying decreasing R25 to 1k, you also need a resistor across C8, maybe 47k - 100k for unity to 2x gain.

Hope this helps,
Jonny

[jhsa]

Thanks Jonny, perhaps sometime I will come back to this circuit. Meanwhile I ended up using a Led/LDR combo to mute the audio. I am waiting for the PCBs to arrive
Actually, tomorrow I will play in a place that is terribly noisy, so I have just built a much simplified version of the circuit I used.
It has the same peak detector /comparator circuit of your design, but I replaced the attack and release pots with a couple resistors, and instead of the LDR I have used two BS170 mosfets connected source to source to mute the signal. one of the drains connect to signal, the other connect to GND. It has an input buffer only. It seems to work, but the real test will be tomorrow
Thanks again

[jonny.reckless]

Thanks Jonny, perhaps sometime I will come back to this circuit. Meanwhile I ended up using a Led/LDR combo to mute the audio. I am waiting for the PCBs to arrive
Actually, tomorrow I will play in a place that is terribly noisy, so I have just built a much simplified version of the circuit I used.
It has the same peak detector /comparator circuit of your design, but I replaced the attack and release pots with a couple resistors, and instead of the LDR I have used two BS170 mosfets connected source to source to mute the signal. one of the drains connect to signal, the other connect to GND. It has an input buffer only. It seems to work, but the real test will be tomorrow
Thanks again

You might find that an LDR is not quite fast enough for a gate operation, they typically go low resistance in about 10-20ms, and high resistance in about 0.5s, this is almost ideal for an audio compressor, but might cut off the beginning of notes when used in a fast application like a noise gate. I do like LDRs for compression, but I would generally stick to a JFET, THAT VCA, OTA or something similar for very fast opening of a gate.

Be careful with regards to the body diode if you are using MOSFETs as a variable resistance. I generally prefer JFETs for this purpose. Let me know how you get on!