Where does the noise gate go?

[Morocotopo]

Well, what the title says:

- Between the guitar and the rest of the pedals?
- Between the dist/OD´s and the rest of the pedals?
- Last in the chain of pedals?
- In the thrash can?

Suggestions and arguments welcome.

[philbinator1]

Well, what the title says:

- Between the guitar and the rest of the pedals?
- Between the dist/OD´s and the rest of the pedals?
- Last in the chain of pedals?
- In the thrash can?

Suggestions and arguments welcome.

some people say last in the chain, so it gates everything/all the noise before it.  but, i usually have my delay then reverb last and
from memory i found it gated the delay - not what i want - so i put it 3rd to last in my chain, before the delay with reverb last. 

oh and that's in my fx loop too..some gates (like the isp decimator) gate differently going through the front than they do in the fx loop.
i have a Decimator, and the manual says for hum reduction go through the loop, for actual gate go through the front (which is what
i'll try next gig).  hope that helps mate! 

[Mark Hammer]

I've written about this too many times to count, but that's not your fault.  Plus, it's my fault for not being able to point you to a thread.  So...

Gates contend with multiple sources of noise.  Among them are:

• hum from single coil pickups and ground loops
• hiss from the pedals themselves, as well as from the amplification of hiss at the input to any gven pedal
• noise produced by "processes" within the pedals, like LFO clicks or HF clock whine

As a result, there is noise which exists at the very start of the pedal chain, noise which accumulates across the pedal chain, and what a statistician might call "the interaction term": noise which was small initially but made worse over the course of the pedal chain.  This would suggest that there are multiple locations for the productive application of noise control.

On top of that, consider that gates use amplitude information to carry out their task.  That is, they differentiate between "signal" and "noise" on the basis of amplitude difference: if it's loud it must be signal, and if its quiet it must be noise.  That assumes that the contrast between the softest and loudest portions of what it sees are substantial enough to be able to reliably and accurately detect the difference.  Of course, a great many pedals have the effect of decreasing contrast rather than maintaining or enhancing it.  What that means is that the likelihood of accurately differentiating between true signal and noise goes down across the pedal chain, even though the amount of noise goes up.

The net result of this is that people are routinely dissatisfied with the application of noise gates because they feel obliged to use them in such a heavy-handed fashion (i.e., threshold set high) that it interferes with the nuances in their playing.

My continuing suggestion has been to locate noise-control devices at both the start and the end of your pedal chain, governed by the signal that has the greatest contrast between transient peaks and quiet parts - the input signal straight from the guitar.  A gate situated at the start of your signal path will mean that any hum and hiss coming from the guitar+cable will be removed before it gets amplified by anything later in the chain.  Since that hiss is not there to be amplified by your compressor or distortions, that means you won't have to set the threshold quite so high at the other end.  Of course, because some of the noise is being generated by the pedals themselves, that means you will also need gating at the end of the pedal chain.  Hopefully, with enough "fuel for noise" removed at the outset, and using an envelope signal derived from the input (but applied to the output), the user can keep the output signal "hygienic" without having to be too heavy handed.

[philbinator1]

OK you win that one Mark   

[Mark Hammer]

Well, it's not a question of "winning" so much as providing some sort of noise control which suits the circumstance.  I can actually think of plenty of instances where the textbook application of a noise gate as the very last thing in line (which is exactly how I used to use mine 30 years ago) is entirely appropriate and perfectly satisfactory.

The trouble is that there are also plenty of circumstances where it isn't. 

It would be nice if someone made a dual-station pedal where you could plug your guitar in to a buffer/rectifier, and that information is applied to two independently adjustable gate, or perhaps a gate and a filter - one at the start and one at the end.  The pedal has a send/return loop, with the type and degree of noise control appropriate to that point in the signal chain applied.  One type of noise control is applied to the "send", and another applied to the "return".  If you have a hum-free input signal then just bypass the send gate and only engage the return filter. So, basically a noise-control "station" that a person could stick on their pedal-board and not have to monkey around with location.

In my list of noise sources, I also forgot other off-beat stuff, like low-frequency rumble picked up by a guitar just sitting there in a stand with the volume pots not turned down.

[Morocotopo]

Interesting Mark, I´m in the planning stages of a noise gate build. I saw one at Tonepad, that seemed easy, no hard to get parts. Could be an opportunity to put your idea to work... So I´d have to isolate the rectifier part of the circuit and duplicate the gate part...and a buffer before all that, possibly JFET... hmmmm...

[Mark Hammer]

Yeah.  The buffer feeds two independent rectifiers.  The intent is to maximize the "dialability" of any gating action.

It also lets you optimize the adjustment range of the corresponding threshold control.  Many gates have threshold controls designed to anticipate anything from 1mv hiss to 200mv hum...or beyond.  The result is that, in the absence of a 10-turn threshold pot, you'll find the control "twitchy".

My suggested strategy suggests designing the threshold control around what the circuit is likely to encounter at that point in the chain.

[Gordo]

Yeah.  The buffer feeds two independent rectifiers.  The intent is to maximize the "dialability" of any gating action.

It also lets you optimize the adjustment range of the corresponding threshold control.  Many gates have threshold controls designed to anticipate anything from 1mv hiss to 200mv hum...or beyond.  The result is that, in the absence of a 10-turn threshold pot, you'll find the control "twitchy".

My suggested strategy suggests designing the threshold control around what the circuit is likely to encounter at that point in the chain.

I use my ISP at the end of the chain but in front of my delay.  You've come up with an interesting point that would work nicely: have a side chain off the gate right after the guitar to trigger a gate at the end of the chain.  That way you really only need to mess with one threshold no matter what signal level various effects are putting out.

[StephenGiles]

Yeah.  The buffer feeds two independent rectifiers.  The intent is to maximize the "dialability" of any gating action.

It also lets you optimize the adjustment range of the corresponding threshold control.  Many gates have threshold controls designed to anticipate anything from 1mv hiss to 200mv hum...or beyond.  The result is that, in the absence of a 10-turn threshold pot, you'll find the control "twitchy".

My suggested strategy suggests designing the threshold control around what the circuit is likely to encounter at that point in the chain.

I use my ISP at the end of the chain but in front of my delay.  You've come up with an interesting point that would work nicely: have a side chain off the gate right after the guitar to trigger a gate at the end of the chain.  That way you really only need to mess with one threshold no matter what signal level various effects are putting out.

ISP?

[Mark Hammer]

The ISP Decimator, a noise control pedal. http://www.isptechnologies.com/

[StephenGiles]

The ISP Decimator, a noise control pedal. http://www.isptechnologies.com/

Ah I see! Thanks.

[Brymus]

I like the idea of using one at the begining and one at/near the end.
I cant stand hiss,even what other people call normal/acceptable really annoys me.
I was once kicked from a band because my gear was to "hissy" at gig levels,(my DSP I used for effects had a noise gate too)and I couldnt afford decent gear at that time(it was REALLY bad noise)
So as a result anything audible that isnt signal irratates me to no end.

[Morocotopo]

On second thought Mark, I hate you.



I was just going to make Tonepad´s Noise Gate, easy, simple... now I have to design a circuit! I´ll never be satisfied with a simple gate now!

Grrr.

That´s the life of the DIYer. Instead of thinking "I have to buy a coffee machine" , you start thinking "I have to make a coffee machine"

Anyone wants to start the design? I know as much about noise gates as I know about brain surgery... more stuff to learn.

[FiveseveN]

have a side chain off the gate right after the guitar to trigger a gate at the end of the chain

While we're on the subject... is that how the NS-2 works? I think I've glanced over the schematic a couple of times but couldn't figure it out. What's the deal with the "loop"?

[Mark Hammer]

The NS-2, and indeed several other noise control pedals, do half of what I've described.  That is, you can plug into them at the start of your pedal chain, where the signal has the greatest ability to discriminate between signal and whatever noise exists at that point using amplitude as the basis for discrimination, but apply the gating action at the point where the signal returns from the loop.

Here is the schematic:  http://www.godiksennet.com/images/sch/NS2.jpg

You will see that Q1 forms the input buffer, whose output is fed to the output driver for the SEND jack (provided by Q2).  If nothing is plugged into the RETURN jack, then the Q1 input buffer is fed directly to the RETURN point.

IC2a sets how much drive there is for the rectifier circuit, and  IC3a/3b/4a/4b form the rectifier circuit and the part used for setting the decay time once the signal has fallen below the required threshold.  IC1 is a seemingly proprietary dual VCA chip, only one half of which is used in this circuit.

So, to repeat, the envelope and triggering of the gate is derived from the input signal, but can be applied to either that same signal, or to the signal occurring at the end of whatever you insert into the send/receive loop.  So, you can stick the NS-2 at the end of a pedal chain like a lot of people do, or you can stick it at the beginning but apply the noise reduction somewhere else.

What I have been advocating is a divide-and-conquer strategy, where you intervene at a point where a particular form of noise can be addressed most effectively in the least invasive manner.  In a perfect world, I suppose, the first half would be a signal-controlled highpass filter that would restore bass when you play but roll off bass at a gradually higher point as you stop playing, so that hum would not find its way into your pedals.  And with the greatest risk of hiss coming from the accumulated sizzles over all your pedals, the second half would apply a sliding lowpass filter to the final output, such that once your signal level drops below some point more and more of the top end would be gently rolled off.  The LM1894 noise filtering chip was designed specifically for that function.  While it is a dual channel chip, unfortunately, as near as I can tell, it was designed only for lowpass filtering and not to be reconfigured for something like dynamic hum filtering.

The 1894 is something we have collectively overlooked here.  It's a nice chip for single-ended noise-reduction.  It still seems to be widely available, even if not in production.  Perhaps we could put it to some use.

[philbinator1]

Well, it's not a question of "winning" so much as providing some sort of noise control which suits the circumstance.

yeah i know just jokes ay   

[Processaurus]

A twist on the two station noise gate, first station isn't a filter, but something like the EH Hum Debugger, which rumor has it is a DSP based pedal that uses the 7v AC power supply to sample the 50/60Hz line and I imagine analyzes the guitar coming in and gets rid of things happening in phase with it.  You could probably make an analog version with by blending in a little 60 cycle or the rectified 120Hz out of phase with what is coming in.  That could takes out the hum, and then the end station is a noise gate or better, a downward expander

[Mark Hammer]

That'd do a nice job too.  Of course, as a digital filter, it starts to evade the skills and "buildability" of many here.  Somewhere out there is a design, or the germ of one, for an analog envelope-controlled 2-pole highpass filter that will efficiently cut the hum out when the signal level drops.  That way, even if one has a compressor and distortion fully cranked in the pedal chain, they would have precious little hum to boiost, which would, in turn, require less heavy-handed gating at the final output.

One of the things I particularly like about the SSM2166-based Q&D compressor IS the inclusion of onboard downward expansion.  I think it takes a bit of tinkering to come up with the right  range of threshold settings to trigger the expansion, and the optimum decay rate (actually, averaging capacitance), but it is a pleasingly unobtrusive form of noise control once you nail it.  People regularly complain about the noisiness of compressors, and this is able to be one of those units that is delightfully noise-free, while at the same time being a damn fine compressor, and a ridiculously easy build.