tonepad's mxr noisegate mods

[Al Heeley]

I'm looking at the layout for Tonepad's MXR noisegate & mods - there's a mod where you replace the jumpers (in version II) with pots to control attack, release and attenuation - can anyone explain to me what attenuation is?

[Sanguinicus]

Probably like a volume control given that attenuation means to reduce signal amplitude.

[Mark Hammer]

The FET is in parallel with the 1M fixed resistor.  When the FET is turned off, its resistance goes high...much higher than the 1M resistor, so that there is little bleedoff of signal.  When the envelope signal turns the FET on, the FET's resistance goes lower than the 1M resistor, and with the FET in parallel with the 1M resistor, there is lots of bleed-off of the signal to Vref (Vb).  Placing a varibale resistance in series with the FET path to Vb, means that no matter how low the FETs resistance goes, it can never go lower than the sum of FET+pot.

So, the attenuation pot sets the minimum volume the gate would produce.  If you want absolute silence when not playing, then you set the pot for zero resistance (gate-off resistance = 0 ohms in parallel with 1M).  If you don't mind a bit of sound, such that when the gate "cuts out", you can still hear the tail of the signal, set the attenuation pot for max resistance.  Some folks may find that 250k works better, but I suspect with that value, you may end up wanting some exotic taper.

Some gates will provide a rotary switch to select between predefined amounts of maximum attenuation.  If a person felt like it, they could simply use a 3-position on-off-on toggle for three levels of cut (hard, medium, soft gate).  To do that, one might replace the attenuation pot with a 220k fixed resistor (soft gate).  The toggle's common would be tied to the junction of the FET and 220k.  One side position of the switch would completely bypass the resistor and go to Vb (hard gate).  The other might go to a 56k resistor, tied to Vb.  In that second side position, the switch would put 565k in parallel with 220k, yielding an effective resistance of 44k6, for a "medium" cut, when the gate turned off.

[Al Heeley]

Thanks for the detailed explanation, if I understand correctly this could address the issue some Peavey Vypyr amp owners complain about, that the built-in noise gate kicks in a little too early to silence the slow decay of a note - by preventing the noise gate from closing 100% you would not get that sudden cut off but might get a bit more low level hum or buzz when not playing?

[Mark Hammer]

The cut-off is a function of two things: the lag or decay of the gate, and the amount of attenuation when the gate is off.

Think of it like have a soundman with amazing reflexes.  The instant you aren't playing, that soundman turns down yur channel on the mixer, and the instant you even think of playing, you're back on.  When you stop, the soundman can simply flick a switch to cut you off entirely, or they can just ride the fader and turn you way down, OR they can turn your fader down a bit so they don't miss anything.

Adding more decay/lag time is like the difference between pulling a fader down and flicking a switch.  Reducing the amount of attenuation (a double negative) is the difference between pulling the channel fader all the way off, and only partly off.

Of the 3 mods (attack, decay, attenuation), decay is the most useful, followed by attenuation, and attack much further back.  Keep in mind that most of the naturally-occurring harmonic information in a guitar is in those first 50msec or so, making a fast attack useful.  Unfortunately, guitar signal is not a steady-state phenomenon, so the decline in amplitude is substantial.  The perpetual struggle with gates and guitars is identifying that magical point where it's not a guitar signal any more, but just background crap.  A longer decay, and modest attenuation allows you to hedge your bets a little.

As I've written here before all too many times, I'm a fan of using noise-control at two points.  Most folks complain about gates because they use them in one place and expect a single gate to accmplish far too much under adversity.  The signal leaving your guitar will have the highest dynamic range at any point in your signal path, and highest contrast between signal and noise.  That makes it much easier to identify a useful gating threshold that won't interfere with your sound.  At the same time, hiss accumulates over that signal path from all the "stuff" along the way, so some artful gating at the start of your chain won't solve everything.  Best to use some gentle gating at the start, and some additional gentle gating at the "exit", before going to the amp.  If you're etching one board, etch two.  ideally, the second gate should be fed by the envelope extractd from the first, but there's no harm done by having them be two entirely separate units.

If you find yourself having a dickens of a time dialing in just the right threshold, the place to turn is likely the 680R resistor.  Ratchet it up to 1k, or even 1k2.  With 680R, the gain of that stage is just under 1500x.  With 1k, it's around 1000x.  reducing the gain of that stage allows the threshold pot to dial in gate-points with a little more selectivity.

[Al Heeley]

...identifying that magical point where it's not a guitar signal any more, but just background crap.

this made me laugh out loud; I have friends who say all my guitar sounds are purely background crap! Ho hoo.

[Mark Hammer]

yeah, my wife thinks that a "noise gate" is the power bar, or maybe the breaker switch in the basement.  .