Noisegate "Rauschgift"

Original effects with schematics, layouts and instructions, freely contributed by members or found in publications. Cannot be used for commercial purposes without the consent of the owners of the copyright.
Post Reply
User avatar
toxonic
Information
Posts: 13
Joined: 13 Feb 2024, 20:27
Has thanked: 6 times
Been thanked: 5 times

Post by toxonic »

Hi folks,
it's already nearly two years ago, when I bought my first pedal kit and made my first steps in building my own effects.
Since I have not so much time to practice i'm still a noob but i'd like to upload a circuit I've been working on.
I wanted to build my own noisegate and on my way I stumbled across some circuits to get started, but none of them really satisfied me completely.
Most of them suffer on an unnatural decay behaviour at higher release times.
I now built one by my own but borrowed nearly all the parts from other circuits.
This one is based on a cheap optocoupler IC named PC817 for the gating itself. I don't know, why noone ever uses this thing in a noisegate circuit (at least I couldn't find one on the internet), but for me it works quite well.
I don't use the signal envelope itself for gating the noise, instead I built a schmitt-trigger upon it, fed with with a variable reference voltage to set the threshold of the gate.
After that, i added a simple AR envelope based on two pots, 2 diodes and a 22u cap to set attack and release.
Until now, I only built this one on a breadboard, but it already works fine as far as I can say.
Maybe someone want's to give this thing a chance and rebuild it. I'd appreciate your opinions and ideas for modifications and enhancements.

Edit:
  • Adjusted R3 and R4 to improve threshold sensitivity
  • Added a pull-down resistor after output cap (R12)
  • Adjusted resistor values for R5, R7, R8 and R9 to improve schmitt-trigger hyteresis and threshold sensitivity


Here's the schematic (updated):
Image
Last edited by toxonic on 22 Oct 2024, 22:47, edited 4 times in total.

User avatar
toxonic
Information
Posts: 13
Joined: 13 Feb 2024, 20:27
Has thanked: 6 times
Been thanked: 5 times

Post by toxonic »

I updated the schematic in my first post to enhance the sensitivity of the threshold pot.
I made the gain stage amplify to a higher level by adjusting R3 (1 M before) and R4 (10 K before), and I also corrected the lowest and highest threshold by adjusting R8 (100 K before) and R9 (33 K before).
Works much butter for me.

User avatar
plush
Cap Cooler
Information
Posts: 679
Joined: 08 Dec 2015, 09:29
Location: Moscow, Evil Russia
Has thanked: 37 times
Been thanked: 185 times

Post by plush »

You forgot to bias your opamps and pull-down the output cap.

User avatar
toxonic
Information
Posts: 13
Joined: 13 Feb 2024, 20:27
Has thanked: 6 times
Been thanked: 5 times

Post by toxonic »

plush wrote: 22 Oct 2024, 18:23 You forgot to bias your opamps and pull-down the output cap.
Thank you for your reply! :-)
Yeah, I forgot the pull-down resistor at the output, you are right.
In my first attempt I used to bias the opamps, but i neither see (in the oscilloscope), nor hear any difference up to a input signal voltage of 2.5 volts.
None of my instruments has such a high output, so i thought i'd save some parts.
I think the gain stage at U2A won't need any biasing, since it's saturated anyway and aside from that, i only need the positive wave parts for the envelope stage thereafter (D2, D3, C7, R5). And the LM358 has no output inversion, what makes it suitable for this purpose, as far as i can see.
I already tried to bias the opamps, but couldn't see or hear any significant advantage.
Maybe I oversee something? ;-)

User avatar
toxonic
Information
Posts: 13
Joined: 13 Feb 2024, 20:27
Has thanked: 6 times
Been thanked: 5 times

Post by toxonic »

By the way: I wonder if i could just drop the output buffer, since the input buffer is forwarded directly to the optocoupler and then led to the output.
Seems to me, that there's no need for it... am I right there? Actually I'm thinking about expanding the circuit with a sidechain in and out. But even when the sidechain input was connected to the optocoupler, it would be most likely, that the signal is already buffered, correct?
So i could just use a TL071 as an input buffer and drop the output buffer... what do you think?

User avatar
plush
Cap Cooler
Information
Posts: 679
Joined: 08 Dec 2015, 09:29
Location: Moscow, Evil Russia
Has thanked: 37 times
Been thanked: 185 times

Post by plush »

I don't think this circuit, if made as drawn, should work at all.

User avatar
J4mes
Breadboard Brother
Information
Posts: 100
Joined: 29 Feb 2024, 03:40
my favorite amplifier: Too many to choose...
Location: UK
Has thanked: 34 times
Been thanked: 26 times
Contact:

Post by J4mes »

plush wrote: 23 Oct 2024, 19:39 I don't think this circuit, if made as drawn, should work at all.
No VG and no dual rail supply... I'm still in my infancy with this stuff, but that's what stands out to me.

User avatar
toxonic
Information
Posts: 13
Joined: 13 Feb 2024, 20:27
Has thanked: 6 times
Been thanked: 5 times

Post by toxonic »

J4mes wrote: 24 Oct 2024, 00:33
plush wrote: 23 Oct 2024, 19:39 I don't think this circuit, if made as drawn, should work at all.
No VG and no dual rail supply... I'm still in my infancy with this stuff, but that's what stands out to me.
Yes, no dual rail supply, but the LM358 pin 4 is connected to ground and pin 8 is connected to 9V. Is this what you meant? :?:
However, i re-checked the circuit a couple of times, it works fine like drawn in the schematic.

Nevertheless, i updated the circuit, now it has a JFET input buffer (dropped the output buffer), switchable sidechain support and I kicked out the attack pot (swapped with a static resitor of 100 ohms), since i guess, most people won't use it most of the time. And I added biasing for the LM358, but as I already stated, theres no significant difference to hear or observe via oscilloscope, as far as I can notice.

Here's the schematic:
Image
1. Input Buffer
2. Gain Stage
3. Envelope analysis
4. Schmitt-Trigger with Threshold pot
5. AR-Envelope
6. Gating Stage

EDITS:
- added decoupling cap C5 and a pulldown resistor R8
- corrected the input buffer (drawn wrong)
- decoupled sidechain_in


Here an image of the circuit on a breadboard:
Image
On the oscilloscope, the lower wave is the output of stage 3 while picking some notes on the guitar, the upper one is the output of stage 5 with release pot set to a small - moderate value at a quite low threshold level.
I just can tell, the circuit works quite well, as far as i can say. :wink:
Last edited by toxonic on 26 Oct 2024, 13:31, edited 2 times in total.

User avatar
kusi
Breadboard Brother
Information
Posts: 115
Joined: 15 Aug 2007, 20:45
Has thanked: 2 times
Been thanked: 55 times

Post by kusi »

Interesting design. I'd add a cap before R7 / U1A to decouple the DC .

User avatar
toxonic
Information
Posts: 13
Joined: 13 Feb 2024, 20:27
Has thanked: 6 times
Been thanked: 5 times

Post by toxonic »

kusi wrote: 25 Oct 2024, 04:41 Interesting design. I'd add a cap before R7 / U1A to decouple the DC .
Ouch, yes, you are right. And another pulldown resistor before the sidechain output. I'll update the schematic in my last post in a minute....

User avatar
ppluis0
Diode Debunker
Information
Posts: 931
Joined: 14 Jul 2010, 18:33
Location: Argentina
Has thanked: 128 times
Been thanked: 311 times

Post by ppluis0 »

Hi toxonic,

The input buffer does not work as appear in your schematic: move the condenser C4 to the junction of R6 and fet source pin.

Cheers,
Jose

User avatar
toxonic
Information
Posts: 13
Joined: 13 Feb 2024, 20:27
Has thanked: 6 times
Been thanked: 5 times

Post by toxonic »

ppluis0 wrote: 25 Oct 2024, 11:15 Hi toxonic,

The input buffer does not work as appear in your schematic: move the condenser C4 to the junction of R6 and fet source pin.

Cheers,
Jose
Oops, sorry... i used a wrong symbol for the 2n5457, source and drain are swapped. I'm at work at the moment, will correct it later.

User avatar
J4mes
Breadboard Brother
Information
Posts: 100
Joined: 29 Feb 2024, 03:40
my favorite amplifier: Too many to choose...
Location: UK
Has thanked: 34 times
Been thanked: 26 times
Contact:

Post by J4mes »

plush wrote: 23 Oct 2024, 19:39 I don't think this circuit, if made as drawn, should work at all.
No VG and no dual rail supply... I'm still in my infancy with this stuff, but that's what stands out to me.
Yes, no dual rail supply, but the LM358 pin 4 is connected to ground and pin 8 is connected to 9V. Is this what you meant? :?:
However, i re-checked the circuit a couple of times, it works fine like drawn in the schematic.
It does now that there's a 1/2 voltage ref point, but not like in the first. Really glad you managed to iron out any issues prior to building. Great job

User avatar
plush
Cap Cooler
Information
Posts: 679
Joined: 08 Dec 2015, 09:29
Location: Moscow, Evil Russia
Has thanked: 37 times
Been thanked: 185 times

Post by plush »

1 - this thing needs an output buffer
2 - "sidechain in" is not decoupled, but it should be.
3 - sw1b disconnects signal ground, which is questionable
4 - bjt optron performance near cutoff is atrocious. Thransients will be distorted AF.
5 - this whole design passes the signal only when one's above threshold. Noisegates should do the opposite - they should cut the signal when it's below the threshold, not the opposite.
6 - as already mentioned, input voltage follower is drawn wrong (and in has nothing with swapping drain and source - you can safely swap them w most jfets)

I've simulated this thing and overall performance (signal to noise ratio) is just terrible. No offense.
Last edited by plush on 26 Oct 2024, 08:57, edited 1 time in total.

User avatar
Dirk_Hendrik
Old Solderhand
Information
Posts: 4216
Joined: 03 Jul 2007, 08:44
Location: Old Amsterdam
Has thanked: 235 times
Been thanked: 907 times
Contact:

Post by Dirk_Hendrik »

plush wrote: 25 Oct 2024, 18:52 1 - this thing needs an output buffer
This.
In general it should be a rule of thumb not to make the behaviour of your design dependent on the devices it's connected to. The "fashion" of removing buffers in effects devices comes, IMHO, from some "effects designers" who developed, tested and tweaked only using their own setup and did not even consider the setup of their unknown customers could be different. Therefore the "this device does not work well with this other device" is a sign of weakness.
Sorry. Plain out of planes.

http://www.dirk-hendrik.com

User avatar
toxonic
Information
Posts: 13
Joined: 13 Feb 2024, 20:27
Has thanked: 6 times
Been thanked: 5 times

Post by toxonic »

Dirk_Hendrik wrote: 26 Oct 2024, 08:45
plush wrote: 25 Oct 2024, 18:52 1 - this thing needs an output buffer
This.
In general it should be a rule of thumb not to make the behaviour of your design dependent on the devices it's connected to. The "fashion" of removing buffers in effects devices comes, IMHO, from some "effects designers" who developed, tested and tweaked only using their own setup and did not even consider the setup of their unknown customers could be different. Therefore the "this device does not work well with this other device" is a sign of weakness.
As I stated in my first post, i'm still a newbie. my thought was, that there's no need for an output buffer, because the signal from the input buffer is passed straight to the optocoupler and from there to the output. In the case, the sidechain was activated, the Sidechain_In (which would be usually coming from a kind of fx pedal which is - in most cases - also already buffered) would be passed the same way. It seems, that I'm wrong there? If you could maybe explain that to me, I'd be very thankful. ;-)

Well, first I was kinda proud, that this thing was working so fine on the breadboard (at least for me), but I guess, it's nonsense to post further updates, because it seems, that the circuit is too faulty from the beginning and nearly in every respect.

But I'd be thankful for constructive criticism and - above all - hints on improvements.
Thank you! :-)

User avatar
Dirk_Hendrik
Old Solderhand
Information
Posts: 4216
Joined: 03 Jul 2007, 08:44
Location: Old Amsterdam
Has thanked: 235 times
Been thanked: 907 times
Contact:

Post by Dirk_Hendrik »

Ah. That's a valid comment.

The simple version first.
You use an optocoupler with a transistor output as muting element. As drawn right now it appears that the intention is to use it as a resistive element or a switch in series with the signal. That means that through the output it will also be in series with the input of the next device in the signal chain. What does that device's input stage look like? By adding a decent output buffer between control element and output jack you can assure that your element will always work, regardless of what's connected next.

The less simple verion.
Your muting element will not work. You have an unbiased resistor in your signal chain and expect the C-E barrier to function as a switch. That will not work. Look for a FET type optocoupler and make sure the source and drain are biased at half the supply voltage.
Sorry. Plain out of planes.

http://www.dirk-hendrik.com

User avatar
toxonic
Information
Posts: 13
Joined: 13 Feb 2024, 20:27
Has thanked: 6 times
Been thanked: 5 times

Post by toxonic »

Thank you for your kind explanation! :-)
The first part sound plausible to me, but I still have problems understanding the less simple one:
Dirk_Hendrik wrote: 27 Oct 2024, 10:25 The less simple verion.
Your muting element will not work. You have an unbiased resistor in your signal chain and expect the C-E barrier to function as a switch. That will not work. Look for a FET type optocoupler and make sure the source and drain are biased at half the supply voltage.
As far as I understood the function of the optocoupler, it controls the voltage between collector and emitter depending on the amount of emitted light from the IRED in it. If I got it right, transistors need to be biased, because they can only operate positive voltages. But the envelope generated voltage, the IRED is connected to is positive only, no? Do I oversee something? Thank you for your patience in advance! ;-)

Post Reply