noise gates?

[blackieNYC]

I'm looking to build an end-of chain noise gate. (I've got some noisy stuff I like a lot.)  Searching noise gates here on the forum, I've figured out a few that won't suit my purposes for one reason or another - the ssm2166, the mxr/tonepad, anything dependent on unobtainable fets, the Boss NS2.  There was mention of a gate that had a fairly minimal signal path using a series LDR.  This is by no means the one I absolutely need, it's just an intriguing idea - a minimum of signal path components when the gate is open.  I'm sure I'll want threshold, release, maybe attack. I have a DOD FX30B that I like, but it's too big- could build a new one, but it has a little difficulty at the end of a chain when the noisy stuff is on. I'd like to see what else is out there. recommendations?

[Kipper4]

I'm sure there's another DoD that is optical iirc

[stallik]

I built a Maplin noisegate which does not have the controls you want, is maybe a little to complicated but sounds great at the beginning, middle and end of the pedalboard. Also works as a slow gears kind of effect if you want it to.
There's a thread on here about it

[Mark Hammer]

The difficulty with end-of-chain gates is that they are forced to make innocent-or-guilty decisions, and have no in-between options.  A gate IS a "decider".  The most satisfactory use of gates comes when the decisions it has to make either have few undesirable consequences (so mistakes are tolerable), or are dead easy and reliable.

At the end of the chain, there are so many accumulated sources of not-the-signal-I-want, that identifying a usable boundary between what you want to keep and what you want/need to eliminate is near impossible.

I've blathered on for years that the ideal noise-control system involves BOTH start-of-chain and end-of-chain devices.  The former is for eliminating hum (since it is rare that guitar outputs provide much hiss, and equally rare that hum is added after the first device in the chain), and the latter is for eliminating hiss and any acquired clock noise.

Gates like the FX30B and the NS2 adopt a compromise position, where the identification of the "need" to gate/fade is made at the start of the chain, on the basis of the raw guitar signal (where the detection of noise, on the basis of level alone, is most reliable), but is applied at the end-of-chain where the accumulated noise is greatest.  In some respects, that is a reasonable strategy.  Unfortunately, if there is any tolerable hum or hiss at the input, and you have  additional gain between start and end of your chain, then the overall "enhancement" of noise in your signal chain can make it hard to identify a gating threshold that cuts out the noise but does not impact on the gentle decay of notes.

It bears noting that, historically, gates have been most reliably and comfortably used when mic-ing up percussion, like a kick drum or snare drum.  In those instances, the drum is either  struck...or not, and nailing a gating threshold that keeps the drum hit, but keeps out the artifacts arising from the PA and other amps making the drum skins vibrate.  Of course, that's very dissimilar to the guitar-into-pedalboard context.  Which is why end-of-chain gates for guitarists is a bit like trying to eat soup with a fork.  Yes, you get the big chunks of stuff, but you end up forfeiting the soupiness you desire.

[Vitrolin]

what you want is a gate with external trigger, the gate i after your gain stages but is triggered by the clean guitar signal.

guitar----splitter---noisebox---more gain----extra gain----noise gate----
                |                                                                      |
                 -----------------------------------------------------------
                                                                           external trigger in

hope you understand my attempt at a block diagram

[blackieNYC]

Thanks Mark. I have put the fx30 at the end of the chain and had some success. I actually have had a deadly after it, which helps, but a long release is ok.  Sometimes I have way too many pedals, and my goal is quiet between songs. I just ordered a noiseless pickup for my only single cool guitar, so hopefully my front-of-chain need will be pretty much over. Sometimes I' driving something with my fuzz face.  The DOD can barely handle it.
The split/triggering is a great thing.  I've been meaning to try it with the fx30B. It's just too damn big.

[Kipper4]

I'm sure there's another DoD that is optical iirc

Found it. its the DOD230 noise gate.
google it up see what you think. Might be worth breadboarding.

edit you note it uses a CLM50 optocoupler
I just had a quikey lookey see the data sheet and it might be worth experimenting with roll your own vactrol
it states 5k light to 1M dark.
you never know it might work.

[blackieNYC]

The DOD 230 has an LDR, and the DOD Fx30B uses a Fet. Hmm.

What do you guys think of this one? Passive signal path with LDR.  I would think maybe a 1meg pot in place of the 100k input pot?  And maybe an output make up stage. Is an input buffer necessary?  I'm sure this can be adapted to 9v. From a 4 channel rack mount unit. The Omnicraft GT-4:
http://smg.photobucket.com/user/uncle_boko/media/GT4.jpg.html

[amptramp]

I would recommend something a little less intrusive than a noise gate.  A level-activated lowpass filter may suit your purposes better.  When the sound level goes down, the cutoff frequency also goes down to eliminate hiss but there is no sudden burst box type elimination of sound.  This is not a new idea - RCA called it the "Magic Monitor" and it was used on the phono input of high-end radios to limit the hiss from a record between cuts or at low recorded levels.  This is a tube high-impedance version:



The circuit block you are looking at is the upper right.  The 6AV6 takes in audio and amplifies it then a diode of the 6AV6 shunt-rectifies it so stronger signals generate a more negative voltage which is filtered to eliminate most of the audio then applied to the 6BA6 to reduce gain in the 6BA6 remote-cutoff pentode stage.  With the 6BA6 near cutoff, the signal proceeds through the parallel combination of a 1 megohm resistor and a 330 pF capacitor that goes from the input phono jack to the output phono plug.  This circuit was intended to be interposed between the record changer and the amplifier, which has a high-impedance input.  When the input signal is low, the 6BA6 is a reactance tube stage that acts like a variable capacitor to ground from the output phono plug.  As the negative bias goes down at low signal levels, the increased gain of the 6BA6 makes the stage act like a higher capacitance, cutting off high-frequency hiss.  Strong signals increase the negative bias, cutting off the 6BA6 and the stage acts like a smaller capacitance.  The small residual capacitance of the reactance stage with strong signals acts like a voltage divider with the 330 pF cap across the 1 megohm resistor permitting flat response somewhat like the input divider on an oscilloscope.  The ouput of this circuit is designed to go to a 1 megohm volume control.

You could make a solid-state equivalent of this for more modern circuitry and it would be easy to use a pair of FET's or an op amp set up as a rectifier with a FET or op amp or transconductance amp reactance stage.  This would give you the effect you want with no on/off clicks or sudden changes like a gating circuit.  But why not keep the tubes - it would look cool!

[Mark Hammer]

Noise filters, like the Burwen, or anything using the LM1894 "Dynamic Noise Reduction" chip, do exactly what amptramp describes.  That is, they don't cut the whole signal, but merely attenuate the treble content like a sort of auto-wah.  Of course, those operate under the assumption that everything else about one's signal path is AOK and the only likely problem is hiss and vinyl/mag-tape surface noise.  On the other hand, that, and a decent hum-reducer at the head of the signal path can give you a nice clean output signal.

[blackieNYC]

Thanks AT. I could see this suiting my needs-band attenuation as opposed to a silent gate. That's pretty cool Mark, but the data sheet says it can reduce up to 10dB. That can't be right. It is a fairly simple audio path - two op amps, like many other gates. What do you think of that Omnicraft passive though? Add a buffer/booster or two?

[Mark Hammer]

I think the diagram oversimplifies.  Keep in mind there is nothing stopping a person from cascading the two halves for greater total hiss-reduction, given that the two halves are "reading" the same envelope.

[amptramp]

I have seen the single-chip noise reduction devices before but they tend to be specialized devices that are not carried by very many people.  Of course you could say that about tubes, but I have baskets of them.  An op amp design would suit many modern designers with the advantage that everyone in the business carries them.

[PRR]

> level-activated lowpass filter

+1.

> not a new idea - RCA called it

Scott had it long before RCA drew that.

1970s Audio Amateur had a rather advanced design. Most hiss reducers settle for the -6B/8ve of a single section, and its soft roll-off. The AA design used a hobbled 2-pole filter which gave a slight peak just before the drop, and nearly 9dB/8ve above that (for a little way). All of this slid based on signal. It was claimed (and my tests agree) that the peak masks the lack of highs, and the steep slope significantly increases the hiss reduction.

This was nit-pick math, but it was simple opamps (and maybe a JFET??), so you could just steal the plan.

My copies of AA are deeply packed, and I may not have that issue anymore.

[Mark Hammer]

There were several articles explaining the principles and offering projects in the various electronics hobbyist mags during the late 70's and into the early 80s.  I'll keep an eye out for them while sifting through the online archives of PE and RE.

[StephenGiles]

I believe the Drawmer DF320 contains all the elements you need, it's a big rack mount beast but works very well. I built a couple of 2 channel units 20 years ago on veroboard!

https://www.dropbox.com/s/ew0gxr5cveollwa/Drawmer_DF320.jpg?dl=0

The problem you have is that it needs 1 DBX2150 and 1 SSM2120 chips, which are probably fairly scarce these days.

[StephenGiles]

http://www.ebay.com/itm/Analog-Devices-SSM-2120-Dual-dynamic-range-processor-dual-VCA-NEW-/271040582551

SSM 2120 chips are out there

http://www.ebay.com/itm/DBX2150-SemiConductor-/161774724686?hash=item25aa86664e:g:J0UAAOSwgQ9VtLln

and so is the DBX 2150!!

[MetalGuy]

The noise gate that proved to work flawlessly for me over the years is the EPFM one. It can be modified in many ways to fit your needs. For example:
1/ Originally it runs from +/- power supply but after some mods it can run on a battery. It's easier to just add a charge pump.
2/ The LDR can be easily replaced with another type. Just get a fast one.
3/ You can replace ICa or ICb section with a FET and get away with only one IC.
4/ The LDR can be replaced with a FET - I did that once using J175.

http://s499.photobucket.com/user/MetalguyPics/media/Noise%20Gate%20EPFM%20.jpg.html

[garcho]

It bears noting that, historically, gates have been most reliably and comfortably used when mic-ing up percussion, like a kick drum or snare drum.

It also bears noting that gated reverb on a snare is a crime punishable by death in many countries

[Mark Hammer]

Not talking about Phil Collins-type gated-drum effects, but about snares and drum skins being rattled by somebody else's amp when the skins are not being deliberately struck.  You may run in different circles, but I don't know anybody who wants to hear drums that are simply vibrating and not being played.