Designing a Noise Gate

[Jdansti]

Just throwing this out there in case it might help. I built the Gaines Noise Gate shown on Mark's website and it's pretty good (http://hammer.ampage.org/?cmd=lt&xid=&fid=&ex=&pg=8 toward the bottom of the page-click on the PDF link).

Controls:
Attack
Release
Threshold
Attenuation

Features:
Key (trigger/side chain)
Bypass

Here's a photo of mine. I used an old CB radio chassis for the enclosure, but it could be put into a more traditional stompbox enclosure.



Edit: Disclaimer-This was my first "stompbox" build. I'll eventually replace the faceplate.

[hex_void]

This sounds like fun.  You have to get an envelope of the signal which comes from rectifying it and storing the values with a short time constant then using a slower differentiator to determine how fast the signal is decaying.  Then you set a threshold depending on how fast the signal is decreasing.  This is a take-home exam.  You have until tomorrow to come back with a complete schematic.

I came up with something but it doesn't involve a differentiator. I'll use 2 schmitt triggers (why schmitt triggers? I need to ignore filter cap ripple to avoid false triggering) instead of one and take advantage of the fact that the voltage at the filtered rectifier output drops faster when I mute the string(s) after a note (or whatever) is played. If it drops below a voltage (let's say x volts) that would be the slow decay mode. If it drops below another voltage (call it y, y<x) then it goes into fast decay mode. Obviously, if I let the note ring the envelope detector will follow the note's natural fade-out until it drops below x volts then it starts gradually attenuating it until it reaches y volts when the signal gets muted. I'll start prototyping on a solderless breadboard and come up with a schematic only after everything works as intended. And it's not going to be within a day.

Just throwing this out there in case it might help. I built the Gaines Noise Gate shown on Mark's website and it's pretty good (http://hammer.ampage.org/?cmd=lt&xid=&fid=&ex=&pg=8 toward the bottom of the page-click on the PDF link).

Controls:
Attack
Release
Threshold
Attenuation

Features:
Key (trigger/side chain)
Bypass

Edit: Disclaimer-This was my first "stompbox" build. I'll eventually replace the faceplate.

I have the schematic, found it a few months ago when I started experimenting with noise gates. I always end up designing my own circuit and it nearly always turns out to work better than some other one I find on the internet. I looked at the Gaines Noise Gate schematic and appreciate its simplicity but I'm really picky when it comes to effects I'm going to use which usually complicates things - the two decay modes.

[Jdansti]

Nothing wrong with improvement!  I'm interested to see the end result of your R&D!   

[hex_void]

So far the slow decay mode works fine but I'm still having some issues with the fast decay mode and I think I know why... anyway, I'll keep you up to date.

[hex_void]

Tried a different approach and it works surprisingly well (I did some simulations before and everything seemed to work but that's a different story). There still is room for improvement though...
The gate cuts the signal when I mute the note or suddenly stop playing palm muted notes or power chords but when I let a note or chord or whatever ring it kicks in a bit too late and lets the noise pass through. This is just a matter of adding a schmitt trigger with adjustable threshold so when the note amplitude is too low it forces the JFET on and mutes everything before stuff gets too messy.
I'll post a schematic as soon as I'm done fixing this little issue.

[hex_void]

Finally got that piece of **** figured out. It's working as intended. I'll post a schematic later if I'm not going to be drunk, otherwise I'll post it tomorrow.
This is how it looks on a breadboard.
http://www.aronnelson.com/gallery/main.php/v/Schematics-etc/hex_noise_gate_breadboard.jpg

[Kipper4]

That seems like a loaded circuitboard.
Can you post a sample too as well as a schematic?

[hex_void]

Here's the schematic. Power supply is 20V DC regulated. The sound sample will be available as soon as I get a mic.
http://s1312.photobucket.com/user/hex_void/media/hex_noise_gate_schematic_01_zps9eec66e1.jpg.html

[hex_void]

Had to settle with the the mic built in the PC sound system's controller...  
Here's the sample http://www.aronnelson.com/DIYFiles/up/hex_noise_gate_test_01.mp3
Don't mind the senseless and random playing, it was for testing purposes  
Some issues with the triggering still exist... I suppose some filters in the sidechain may help. Sensitivity for higher frequencies is a bit too low.

[gritz]

Some issues with the triggering still exist... I suppose some filters in the sidechain may help. Sensitivity for higher frequencies is a bit too low.

It's getting there though. I've just about finished prototyping a noise filter / expander and found that the frequency response of the sidechain was very important. I see you're creating a hipass action with R17 + C9 at about 1kHz, but this is only giving you a 6dB / octave hipass action, which will let through a lot of low frequency rumbles to keep your gate open. Try sticking a coupling cap between the input buffer and the input of the gain stage (I can't read the identifier on the schematic - it's between "U something 1 and U something 2"). Obviously the non-inverting opamp input will need a resisror to ground. Try a 4n7 + 100k combination to give a hipass frequency of around 330Hx. Try a coupling capacitor in front of R2 as well - 15n will give you 330 Hz again. Alternatively 10n will equate to about 480Hz and swapping C9 to 1u will give 480Hz too, keeping the rumble out, but allowing a tad more of the mids through.

It does rather depend on the basic spectral output of your guitar, but a bit of trial and error should get you closer.

[hex_void]

Problem solved . 6.8n+100k after the input buffer. I actually have a 100n cap in front of R2... accidentally left it out of the schematic .
Now to make the whole thing run off 12V... shouldn't be a big problem.

[Kipper4]

Sounds promising. mate i cant wait for the final schematic.

[hex_void]

I could use some suggestions for making it run from 12V although I pretty much have everything figured out. The only thing I'm not certain about is what will happen when VGND (1/2 the power supply voltage) will be above the JFET source voltage because I have to keep that properly biased - source at about 7V above ground. One thing that's a must is reversing C6 for 12V (maybe 9V?) operation and change the zener diode to a lower voltage one.

I'm still not sure if I'll have it built into my distortion pedal (the reason for 20V power supply - idiot-proof step-up voltage converter from a standard fx pedal power supply based on a TL494) or if it's going to be a separate fx pedal.

[hex_void]

I've finished experimenting and settled for this version: http://i1312.photobucket.com/albums/t528/hex_void/hex_noise_gate_schematic_02_zpse06de4d4.jpg?t=1366532644
I'll leave the bypass/mute switching to you.
It works with an 18V power supply which must be regulated. It also works with any voltage between 16V and 20V.
Now to stuff all this into a HM-1590B enclosure... using SMD of course.
If there are requests I'll record a sound sample.

Any questions?

[hex_void]

UPDATE: Found a small glitch, power supply related. Using extra filtering to smooth out the ripple created by the sidechain should do the job.
Also, I made a mistake in the schematic, the JFET is not BF245C, it's BF245A.