MXR Noise Gate (M-106) Corrected Schematic

[Baran Ismen]

Thanks for all the information you have provided guys

What about the below issue?

Today I've finished this. Seems working, but as I've never used a noise gate before (only on my good old GT-10's integrated one), I'm unsure if it's working properly or not.

It's blocking the hi-gain noise indeed when strings are muted, but the sensitivity pot seems to act a bit unusual. Consider that my volume level is 10 on my preamp;

In the first quarter of the pot range, the volume level is around 7, gating works,
On the 2nd, it is 10 but the attack ripples/distorts a bit and gating doesn't seem to work,
From the 3rd quarter until the end of the range, the volume level is around 3, and gating works.

I've tried some suggestions on Tonepad's build reports, but they seem not to work, it's a bit different schematic anyway so I didn't have much hope for that.

Used 2x 3904's (no idea about their matches or Hfe's, my DMM can't measure that), and 2n5485 in reverse for 5952. Cap's are not tantalum.

[Eb7+9]

first, you'll need to measure the DC voltage across the Zener

second, confirm side-chain activity with DMM while strumming instrument
by measuring voltage across 1uF capacitor (the one that goes across T2)
see that Vc drops to near 0v (GND) during dynamic activity with sensitivity control turned up sufficiently ...
and then re-charges back to Vz upon release

---

Vgs on jFET is set by voltage divider 1Meg+1Meg+100k between Vz and Vc
ie., Vgs = (Vc - Vz ) *1.1/2.1

when cap is fully discharged this becomes -Vz*(1.1/2.1)

let's say roughly -Vz/2

during normal operation jFET goes from full-on (gating during idle) ie., Vgs = 0v
to fully off ... ie., hovering somewhere below Vgs(off)

preferably well below that and not just on the edge

for the jFET to be fully off and therefore not experience volume drop and distortion when gate is open
we thus need to satisfy the following condition:

that |Vgs(off)| be less than Vz/2 and, preferably by some margin

the actual device family is somewhat secondary here
J201, 2n5457, etc should all work, provided above condition is met

---

the easiest and most accurate way we have for estimating this Vgs(off) parameter for n-chan jFET's
is to apply (+) and (-) from a fresh  9v battery (or power source) to Drain and Gate, respectively
then set DMM to volt reading and apply meter probes between Source and Gate nodes on the device
this way the internal resistance of DMM plays the role of (high valued) Rs in biasing jFET

this method assumes that |Vgs(off)| is less than 9v by at least a few volts
and note, using a meter with 1Meg internal resistance will provide an estimate shy by roughly 20mV

---

btw, the sometimes perceivable distortion on the exits can't be fixed
something the original manual mentions ...



good luck

[Baran Ismen]

first, you'll need to measure the DC voltage across the Zener

second, confirm side-chain activity with DMM while strumming instrument
by measuring voltage across 1uF capacitor (the one that goes across T2)
see that Vc drops to near 0v (GND) during dynamic activity with sensitivity control turned up sufficiently ...
and then re-charges back to Vz upon release

---

Vgs on jFET is set by voltage divider 1Meg+1Meg+100k between Vz and Vc
ie., Vgs = (Vc - Vz ) *1.1/2.1

when cap is fully discharged this becomes -Vz*(1.1/2.1)

let's say roughly -Vz/2

during normal operation jFET goes from full-on (gating during idle) ie., Vgs = 0v
to fully off ... ie., hovering somewhere below Vgs(off)

preferably well below that and not just on the edge

for the jFET to be fully off and therefore not experience volume drop and distortion when gate is open
we thus need to satisfy the following condition:

that |Vgs(off)| be less than Vz/2 and, preferably by some margin

the actual device family is somewhat secondary here
J201, 2n5457, etc should all work, provided above condition is met

---

the easiest and most accurate way we have for estimating this Vgs(off) parameter for n-chan jFET's
is to apply (+) and (-) from a fresh  9v battery (or power source) to Drain and Gate, respectively
then set DMM to volt reading and apply meter probes between Source and Gate nodes on the device
this way the internal resistance of DMM plays the role of (high valued) Rs in biasing jFET

this method assumes that |Vgs(off)| is less than 9v by at least a few volts
and note, using a meter with 1Meg internal resistance will provide an estimate shy by roughly 20mV

---

btw, the sometimes perceivable distortion on the exits can't be fixed
something the original manual mentions ...



good luck

Zener reads of 4.6v precisely. I'll try the sidechain thing this evening.

[Baran Ismen]

I made another trial for this and went with the Tonepad version this time, with the mods of 680ohm/1uf to 4.7k/3.3uf (resistor solved the oscillation problem, cap is for gating response), keeping the 1m as it is instead of 1.8k (actually went with 1.5m, as increasing seemed to be better responding, changed the IC with LM358 (rc4558 I had had a feint background distortion),

I wonder if it's possible to increase the attack time more (applied no attack/decay mods), right now gate's activation is around 25-30ms give or take. Release is faster.

[Baran Ismen]

Here's the 10uf "cap-flip" fix,

7) Flipped 10uF to JFET drain.

[Obsoleted see below]

Which is the correct one, Rob? Most of the projects I've found seem to keep this cap as negative side facing to FET, positive to 22k.

[Rob Strand]

Which is the correct one, Rob? Most of the projects I've found seem to keep this cap as negative side facing to FET, positive to 22k.

From a technical perspective, as shown is correct.   The + side of the cap sits at the zener voltage.   The zener voltage feeds the 470k to the base of the transistor Q1.   The base current will cause the base to sit at a lower voltage than the zener voltage.  Then on top of that the emitter will sit at 0.6V below the base.    So, as shown, the negative side of the cap must sit lower than the positive side of the cap which is correct.

Putting that aside I'm pretty sure I would have checked the polarity of the cap on the original board when I re-traced the circuit.

I'm not sure what problem you are chasing at the moment.  It would be wise to put a 22pF to 100pF cap across the 1M resistor at IC1, as mentioned in the notes.   That's especially the case if you aren't using a LM741.   As far as holes in the original design go, you could also add 470ohm in series with the 47nF cap which appears after IC1.   The reason for that is the 470 ohm would help fend-off oscillations due to capacitive loading on IC1.

[Baran Ismen]

Which is the correct one, Rob? Most of the projects I've found seem to keep this cap as negative side facing to FET, positive to 22k.

From a technical perspective, as shown is correct.   The + side of the cap sits at the zener voltage.   The zener voltage feeds the 470k to the base of the transistor Q1.   The base current will cause the base to sit at a lower voltage than the zener voltage.  Then on top of that the emitter will sit at 0.6V below the base.    So, as shown, the negative side of the cap must sit lower than the positive side of the cap which is correct.

Putting that aside I'm pretty sure I would have checked the polarity of the cap on the original board when I re-traced the circuit.

I'm not sure what problem you are chasing at the moment.  It would be wise to put a 22pF to 100pF cap across the 1M resistor at IC1, as mentioned in the notes.   That's especially the case if you aren't using a LM741.   As far as holes in the original design go, you could also add 470ohm in series with the 47nF cap which appears after IC1.   The reason for that is the 470 ohm would help fend-off oscillations due to capacitive loading on IC1.

Tonepad version is actually working now and I've did this mentioned change as well last night, flipped the cap, haven't noticed any difference on gating action but it somehow working now.

I've noticed a weird thing that, with some DC adapters I have, this circuit seems to pulsate on current, its like when activated, it turns on, but you hear a pulsating wave sound around every half of a second, its like a synth sub bass sound, and the indicator LED (which is connected via 3DPT) pulsates accordingly. You can also hear the guitar sound in the background.

It's not like reverse polarity, that's for sure, that case pedal doesn't open at all, but this is something I've never seen before. Polarity is correct.

[Rob Strand]

Tonepad version is actually working now and I've did this mentioned change as well last night, flipped the cap, haven't noticed any difference on gating action but it somehow working now.

FYI, I checked the polarity of the cap by tracing the PCB pics posted earlier in the thread.  I'm pretty sure I checked everything on this one to make the "corrected" schematic.   Anyway the negative side goes toward the emitter.
https://www.diystompboxes.com/smfforum/index.php?topic=122689.msg1278238#msg1278238

It's possible put the cap in the wrong way cause some sort of leakage which then hides some other problem.

While I'm confident the a schematic is correct there does seem to the some general issues with that circuit.   I don't know how MXR got the units out in production.   Possibly a fluke characteristic of the opamps from back in the day.

I've noticed a weird thing that, with some DC adapters I have, this circuit seems to pulsate on current, its like when activated, it turns on, but you hear a pulsating wave sound around every half of a second, its like a synth sub bass sound, and the indicator LED (which is connected via 3DPT) pulsates accordingly. You can also hear the guitar sound in the background.

It's not like reverse polarity, that's for sure, that case pedal doesn't open at all, but this is something I've never seen before. Polarity is correct

That's totally whacked.   The pulsing LED is showing the power is going on and off.  That's not normal.   In fact I can't even see how the pedal circuit could do that.  I'd be thinking something is wrong with the PSU, PSU connectors, 6.5mm sockets.   Maybe even a wiring error on the board - but even then how does it release the fault, seem unlikely.

[Baran Ismen]

Tonepad version is actually working now and I've did this mentioned change as well last night, flipped the cap, haven't noticed any difference on gating action but it somehow working now.

FYI, I checked the polarity of the cap by tracing the PCB pics posted earlier in the thread.  I'm pretty sure I checked everything on this one to make the "corrected" schematic.   Anyway the negative side goes toward the emitter.
https://www.diystompboxes.com/smfforum/index.php?topic=122689.msg1278238#msg1278238

It's possible put the cap in the wrong way cause some sort of leakage which then hides some other problem.

While I'm confident the a schematic is correct there does seem to the some general issues with that circuit.   I don't know how MXR got the units out in production.   Possibly a fluke characteristic of the opamps from back in the day.

I've noticed a weird thing that, with some DC adapters I have, this circuit seems to pulsate on current, its like when activated, it turns on, but you hear a pulsating wave sound around every half of a second, its like a synth sub bass sound, and the indicator LED (which is connected via 3DPT) pulsates accordingly. You can also hear the guitar sound in the background.

It's not like reverse polarity, that's for sure, that case pedal doesn't open at all, but this is something I've never seen before. Polarity is correct

That's totally whacked.   The pulsing LED is showing the power is going on and off.  That's not normal.   In fact I can't even see how the pedal circuit could do that.  I'd be thinking something is wrong with the PSU, PSU connectors, 6.5mm sockets.   Maybe even a wiring error on the board - but even then how does it release the fault, seem unlikely.

That's ok, I mean it's working if the mentioned cap is flipped or not. But this pulsation seems quite weird, leads me thinking of a possible PSU problem because its not happening with ANY psu I use, it happens with some of them (I got 4, 2 are voltage & polarity changeable, 2 are regular smps')

It's a weird circuit indeed, literally no one could manage to make it good to go upon building it, everyone tries something different and fails/succeeds, some use 6.2v zener, and some change the cap/res. combo, some decrease the 1m to 1.8k (in my case didn't work i.e), etc etc. It's not something like a TS or something else. Also, the pot works CCW, I know I can use a C type and connect in reverse, but why would MXR prefer something like this anyway..

I wish NF-1 had a layout, it seems to be a better working noise gate. I can read its schematic indeed and draw anew and there I can prepare a layout, but I don't know how to remove the flip-flop section off of it.

[Rob Strand]

I wish NF-1 had a layout, it seems to be a better working noise gate. I can read its schematic indeed and draw anew and there I can prepare a layout, but I don't know how to remove the flip-flop section off of it.

IMHO, the Boss NF-1 is a better version of the MXR.  You could graft the differences onto the MXR circuit, however, the overall structure of the Boss is better.

Here's how to chop out the flip-flop: