noise gate by david haylock

[davidhaylock]

i tried a couple of designs that have been around the net for ages and quite honestly they were rubbish so i had a go at making my own its simple to build uses cheap parts found on eBay and sounds really good no distortion and no extra noise added as there is no active circuitry between input and output I've also added a visual indicator on the box so you can see when the gate is opening the photo resistor i have used is sold on eBay 100 for 99 pence and the led's are cheap white ones

[BetterOffShred]

That's pretty dope man

[duck_arse]

on your circuit dia, the two leds have their Kathodes pointing to V+. they won't produce any light.

[blackieNYC]

Do you have a measurement on the photo resistor's resistance with the light on and off?

[noisette]

Yes, LEDs wrong way round and will use less current if put in series.
The Perkin Elmer Datasheet contains a simple noise gate which is not unlike yours...
http://www.datasheetcatalog.com/datasheets_pdf/V/T/5/0/VT500.shtml
Check page 66

EDIT: although a passive attenuator can work well, you´re making a voltage divider of R1 against the input impedance of the following circuit (as it´s drawn), which will be only working inconsistently. Actually a low pass filter R1/C10... Also why is U2 there?= You could have enough gain with U1, I would put it in positive configuration.
Sorry, don´t mean to say it´s not working for you, but there are still a couple of things to tweak

[davidhaylock]

thanks for the feedback ive obviously put the leds in the diagram the wrong way around i was using a program on digikey to draw it and was new to it. the ristance on the ldr when light on is about 200 ohms and when off above 40 megohms the reason i used two stages of the tl072 was i wasnt getting enough isolation of the amplified singnal getting back on to the input and distorting it but i will carryon and hopefully make future improvement allthough the sound on this is pretty amazing in my opinion regards david. any revised improvements that anyone can suggest i will try thanks in advance

[davidhaylock]

i have modified the circuit diagram so that the led's are the right way around

[noisette]

the reason i used two stages of the tl072 was i wasnt getting enough isolation of the amplified singnal getting back on to the input and distorting it

In now way I´m an expert, but that is very unlikely to be the cause of the distortion. Pin2 of U1 is at vrtual ground so to the input signal of the sidechain looks like R2 in series with C1 (which is really small) to ground, which can be driven by most sources. While imho it would be better to make R2 the threshold pot into noninverting opamp and use the 2nd opamp for a better rectifier...

Maybe more to do with undefined (read unbufferd) source impedance into "one-legged" voltage divider into undefined (unbuffered) load impedance.

I think even the most experienced circuit designers, have to tweak their conceptual circuits on the breadboard for reallife compatibility.

[davidhaylock]

i am waiting for more parts to arrive and i will try adifferent aproach as suggested by noisette and see if i get improved results thanks for your sugestions and i will continue to improve on my desighn regards david

[Mark Hammer]

Sometimes, what can sound like distortion is really envelope ripple during the decay portion of a signal.  I recall when Jon Gaines published his noise gate in Modern Recording magazine, one or more letters noted clicking and "distorted" sounds.  The recommendation was to slow down the attack and decay times.  You use a 33nf cap for C7 to smooth the envelope.  Fast response, to be sure, but not a lot of smoothing going on.  The trouble is that , as notes decay, there are natural perturbations in the envelope that can very quickly turn the gate on and off again.  It ends up sounding like distortion.  Maybe less so with an optical than with a FET-based design, but still a risk. You may want to add some lag to that envelope by sticking a small-value resistor between D2 and C7, as well as by increasing the value of C7.

You can read some of the letters and responses here: https://www.americanradiohistory.com/Archive-Modern-Recording/80s/Modern-Recording-1982-02.pdf

[davidhaylock]

ive done some mods and replaced the ldr led with a transistor optocoupler used in most switch mode power supplys it responds a lot quicker and goes open circuit a lot quicker as well so heres the modified diagram

[Mark Hammer]

When it comes to noise gates, faster is not always better.  Certainly it makes the gating action more responsive, but that responsiveness needs to be complemented by a very high quality rectifier circuit, whose ripple is inaudible.  Otherwise you get the perceived "distortion" I mentioned earlier.

The first step in minimizing such noise is use of a full wave rectifier circuit, rather than the half-wave used here.  This will effectively double the frequency of any envelope ripple.  The other good practice is to use a photocell whose transition time from full on/bright to full off/dark is sluggish enough that any remaining ripple simply comes and goes too fast for the LDR to respond to it; the LDR will "smudge" the ripple.  Finally a slightly higher-value averaging cap (which I see you've done) helps to add a wee bit of lag to assist in reducing ripple.

I take it that what was shown as C10 in the original drawing played a role in reducing noise even further.  R1/C10 form a single-pole lowpass filter.  With C10=39pf, if the LDR had a maximum off resistance of 5meg, that would provide a 6db/oct rolloff starting around 816hz when the gate is off.  If the on resistance of the LDR were 100k, that rolloff would rise to over 40khz, so well out of the way of the audio signal.  In which case, depending on the properties of the LDR you were using, you could probably get away with an even higher value for C10, like 220pf.  Using that value would give a rolloff at 144hz when dark=5meg and 7.2khz if light = 100k.

[noisette]

Yes but why not using a regular voltage divider???
Or build a Buchla Low Pass Gate with both options?

[Mark Hammer]

I was kind of wondering that myself.  At least the first part of that.

[amz-fx]

ive done some mods and replaced the ldr led with a transistor optocoupler used in most switch mode power supplys it responds a lot quicker and goes open circuit a lot quicker as well so heres the modified diagram

I'm not sure that the EL817 is a good choice for this project. It looks like it has a bipolar transistor as its photo element, which is not usually a substitute for a photo-resistor. Try it out on a breadboard and let us know how it works in this project.

Best regards, Jack

[amptramp]

See Figure 3 here for the RVA Magic Monitor, a noise reduction circuit that worked a little differently (better) than a normal noise gate:

http://pacifictv.ca/schematics/rcavra123data.pdf

The circuit has a 6AV6 amplifier/rectifier tube that feeds a 6BA6 reactance tube.  If the signal level is high, the amplifier output is high and the rectifier cuts off the 6BA6 and the signal goes straight through.  If the signal level is low, the 6BA6 acts as a reactance tube and multiplies the value of C78, the 27 pF capacitor which cuts the level of high frequency signals.  There is a sliding scale where the lower the level is, the lower the rolloff frequency.  This could easily be redesigned as an op amp circuit to fit modern sensibilities.

This circuit does not have a snap on / snap off functionality, it just lowers the rolloff frequency as the signal gets lower to cut high-frequency noise.

[davidhaylock]

both circuits work well the reason i tried the el817 is that it goes high resistance much faster than the ldr which mutes the signal better when you want silence the ldr went above 40 meg ohms but took a few seconds the el817 goes above 40 meg ohms what seems like instant the on resistance in both circuits is about 200 ohms. some asked about the 39p capacitor in the original circuit this was put in originally to stop any rf getting on the signal i hope someone out there has a go at building one and then tell there opinion. as the opinions im getting are mainly based on electronic theory where as everything i make is a bit of theory but mainly building it and then fine tuneing it with the results i find in practical use. regards david