Definitive MXR Noise Gate Thresh/Sensitivity pot fix?

[MrStab]

hi,

every so often i build something out of boredom and happen to like it enough to wanna use it live. i'd attempted an MXR Noise Gate before, but i kinda half-rectumed it and didn't bother since. so i made one the other day, following a layout based on the modded Tonepad version + fixes, and it works great. it doesn't screw up my attack as much as i thought it would.

this isn't such an issue, as it's usable as-is, but my Sensitivity/Threshold pot is quiet for the first 1/10th of the turn then suddenly it just lets all the noise through at full volume. no gradual change, it's night or day - which may or may not be similar to an issue reported by others. now, i'm awaiting some of the recommended 500k Log pots as i don't have any, but i have a feeling that's not the issue as it occurs with the 1M Linear pot i tapered down to 500K pseudo-Log, a bare 1M Log pot, a 500K Linear pot and a 50k pot (which confirms that it's at about 20K where the gate suddenly lets everything through). i've tried all the suggestions, but no luck. worst-case scenario, i could just get a 20K pot and leave it at that, but that'd limit functionality somewhat.

points of note:

-i've changed the 680R resistor to 4.7k, and the following cap to 2.2uF
-already performed the 1M -> 1.8k resistor change
-i have 60pF (2x30pF caps in parallel) between opamp pins 1 & 2
-i'm using matched 2n5088s (~370 or something hFE - some say this is too high, others say matching matters but not hFE - either way i dont want to change these, as apart from the pot, i like how everything else works)
-2n5952 JFET
-i'm using this layout: http://www.sabrotone.com/?attachment_id=1233 (i'm going on the assumption people will know about the Tonepad version already, apologies if not).

any ideas?

off-topic, but to avoid making another thread, i've been screwing around with a Schmitt Trigger and an LM386 for a cheap VU meter and it's pretty cool, i recommend it even though i'm still working on it. lol

cheers!

[PRR]

Schematic?

[MrStab]

well, this is the schematic of the original noise gate, sans mods: http://img40.imageshack.us/img40/2674/mxrnoisegatev2.jpg, but upon trying a reverse-log pot wired backwards (which is the rotation i want anyway), something has just occurred to me: maybe that sudden "jump" i'm hearing is just the threshold suddenly hitting the ceiling of my noise (dunno the correct term)? ie. maybe there just isn't much noise above that point and the majority of it comes in suddenly as opposed to gradually. i mean, there is some change when you adjust the pot within either the loud and quiet halves of the rotation. does this sound feasible?

i'm used to dynamic-altering stuff on a VST level (not that i'm exceptionally experienced with that either), where there's often some visual indication of the signal level & threshold. not sure where i'd take the signal from for some kind of indicator, but i guess that's not really important.

[peterg]

I built an MXR NG a few weeks ago and ran into the same issues. It seems that each NG has it's own quirks. I used matched 2N3904's and a 2N5486 FET and stock pot. I replced the 680 resistor/1uF cap with a 22k/10uf - not sure if this is relevant given the different components. The unit works well except for some minor gating(?) Issue if I let aa note ring too long. For hiss noise I set the pot at full CW and for chord and note gating a setting at about 2 - 8:00 I works.

The cap at the FET source is reversed on the Tonepad scheme. The + side shhould be at the source of the FET. Installing this way eliminated attack clicking 

I suspect that using the 500k pot should help but you might have to play around with the 680 resistor/1uf cap some more.

[MrStab]

thanks for the input, Peter. i'm starting to feel pretty sure that mine actually does work properly - i mean, if you picture a wave, the peak's gonna be at a fixed, sudden point, and not so much gradually (unless there's noise from various sources with different amplitudes). um... right? lol. i guess one way to test it would be to raise/lower the noise in my rig and see if the pot kicks in at different times - it'd be a subtle difference if i'm using the same equipment, but worth a shot. i'll check that out later. i've only actually tried with my full-on distorted setup, maybe i'll notice more subtlety with some of that turned off.

i used a layout in which much of the original Tonepad issues were already fixed, apart from the 680R. i tried 680R/1uF at first, before reading about most people settling for 4.7k/2.2uF. one guy on Tonepad even skipped the resistor altogether, but he seemed to be in the minority. re. that cap you mentioned: there's actually a post on this forum which mentions uncertainty about orientation, and the following post got it wrong, but luckily the layout i used had already corrected that as well.

while it makes more sense on paper to have the pot go CCW to lower the threshold, i feel as though in guitar-land it feels more intuitive to lower it the opposite way. it was the Engineer's Thumb compressor that corrupted me in this way.

[MrStab]

just curious:

The whole point of the loop is to extract the envelope while it's in it's purest (?) form and use that envelope to control the gating of the modified return signal.

Q1 splits the signal to the envelope detector and to the attenuating circuit. Looking at the circuit, by inserting a serial effect loop between Q1Emiter and the 22k resistor going to the two caps, you'd be able to extract the envelope from the guitar before it becomes 'dirty' and use that to gate the return.

All you'd need is a capacitor to couple the effect send. It would probably be a good idea to add an input buffer, but that just complicates things.

(http://www.diystompboxes.com/smfforum/index.php?topic=29051.msg197046#msg197046)

can't be THAT simple...? anyone know if it's meant to be before or after the signal goes to the Threshold (500k) pot? here's the modded schem: http://www.tonepad.com/getFile.asp?id=77. not sure where's meant with the input buffer either.

Edit: after some thought, i'm not sure i'm actually bothered to make/deal with more patch cables for what doesn't seem like a huge benefit in retrofitting this circuit to do something it wasn't designed to. unless someone talks me back into it.

[DIMstompboxes]

You may check on these, from the other forum hope might help...
- Input cap is in reality 10n input resistor 10k.
- The 10µ electrolytic between the 22k and the FET is drawn in the wrong direction - plus should be on the Fet-side.
- Output cap is 10µ instead of 1µ.
- the capacitor before Q2 in parallel to the 100k is a 10n instead of a 1µ.
- The source of the FET is connected to Vb and the drain to the 10µ electrolytic

[MrStab]

You may check on these, from the other forum hope might help...
- Input cap is in reality 10n input resistor 10k. i have 47n, 1k but no perceived tone loss
- The 10µ electrolytic between the 22k and the FET is drawn in the wrong direction - plus should be on the Fet-side. check
- Output cap is 10µ instead of 1µ. still 1uF, but i'm guessing that wouldn't impede function. i have a lot of FX with smaller output caps
- the capacitor before Q2 in parallel to the 100k is a 10n instead of a 1µ. check
- The source of the FET is connected to Vb and the drain to the 10µ electrolytic  check

hmm, well there appears to be some discrepancy with the in/output caps, but the circuit itself seems fine. i think i've just been an idiot and not really visualised the wave tbh lol. thanks though! i'll have a ponder over those caps.

[peterg]

https://www.dropbox.com/s/yrxsusxrwjeefk0/1.jpg
https://www.dropbox.com/s/yrxsusxrwjeefk0/2.jpg
https://www.dropbox.com/s/wya6a7312aphqe1/reversed%20back%20of%20pcb.jpg

Take a look at these gut shots of the real thing for confirmation. (I found these on another forum - Sorry Aron!). I created a reverse image of the back for tracing.

Not to say that MXR didn't change things over the years they were making this pedal.

[MrStab]

thanks, Peter. the 2nd link is just 404-ing. i'll double-check with my layout when i can (i'm sure this could help a lot of people, with the inconsistency of builds) - i've barely seen references to the actual thing, all the build reports seemed to be a bit more hearsay.

i'm dead-certain mine works now - i boxed it up and screwed around with the level of noise going into it - the threshold kicks in at different points. turning the pot eliminates hiss gradually, but for actual hum & feedback etc, it's more sudden. this thing is awesome.

so, future builders: get your NPNs matched, get rid of that 680R/1uF for a 4.7K/2.2uF, change the 1M from the Gate of the FET to a 1.8k (i used 1.5), check Peter's pics (when the 2nd one is back up!) and DIMStompboxes' list, and if it still doesn't work, i'll eat my metaphorical hat. and use a reverse-log pot for threshold if turning anti-clockwise to reduce noise seems counter-intuitive.

[peterg]

https://www.dropbox.com/s/hjhovm23d71arb6/2.jpg

This should work.

[DIMstompboxes]

Hey there's another diode connected to both 741 IC 
That blue one maybe the zener
The input cap is 103 with 10k but  there's no difference using 473 and 1k I think.

[MrStab]

i did this without glasses on, and the pics were taken at different angles, but i screwed around with it quickly. this might help (although id imagine some of the components were deemed unnecessary by the dude(s) at Tonepad):

[DIMstompboxes]

MrStab your eyes is up to par and fine than mine, I can't do soldering at night even with well lighting 
This one used two separate single opamps
I think that red diode is a zener reverse biasd with Vcc and Vee?

[MrStab]

I think that red diode is a zener reverse biasd with Vcc and Vee?

i'm no expert but it looks like something along those lines - could be regulating voltage to the opamps back before 78**-style regulators were common? maybe just a voltage stability or protection feature. i wonder why you'd need that. any idea if other MXR pedals from the same period used something similar?

a 32" monitor compensates for my poor eyesight! lol

[duck_arse]

I'm looking at a circuit w/ a 1n5331 zener. it looks to me to be providing a stable Vref line for both the op-amp biasing input transistor and the fet. are you using a 2n5952 for the fet? do you have something like a 2n5457 you could try as a substitute?

[MrStab]

the only other FET i have in my stash is some MPF102, but as i say it works fine and i'm happy with it!
interesting about the zener - i guess the question is either why did MXR put it in, or why did Tonepad take it out...

[Kipper4]

Goddam this is the second time i've had a go at this and it's still not working.
This time i did callgate1 without the mods.
I replaced the 680ohm with a 4k7 and the 1uf cap with a 2u2 in the op amp feedback loop no joy.
I'll try reversing the 10u cap to the fet gate tommorow night and i'll try the 10 input resistor too.
but i'm not holding out much hope.
Mr stab does yours work?

[animal]

Hi guys i just want to share my experience with this pedal
I see a lot of post about the problema with this so i decided tell you whats works for me.

I made all mods: 1M to 1.8k for volume drop / 680 to 4.7k for the oscillation.

In first place the pedal didn't work so then i tried to adjust the attenuation with a trim pot and its works with value over 15k but then i found the sensitivity pot works almost full i mean 5 o'clock so i change from Log pot to a RevLog pot and now works perfect!! 

(2:40 min)

Sorry for my english i'm use google translate

Cheers !!