UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Other Discussions > Homebrew Equipment

Notices

Homebrew Equipment A place to show, design and discuss the weird and wonderful electronic creations from the hands of individual members.

Closed Thread
 
Thread Tools
Old 11th Jun 2006, 4:30 am   #1
adibrook
Retired Dormant Member
 
adibrook's Avatar
 
Join Date: Dec 2004
Location: Weymouth, Dorset, UK.
Posts: 377
Default Noise Gate

I am trying to convert my entire stage setup (everythgin from bass to power amp) to valve sutff. So far everythign is goign well, exept from one thing.

My digital effects pedal has an adjustable noise gate on it. This makes sure that feedback doesnt happen at high distortion levesl, and also gets rid of the anoyign sounds that happen when you have very high distortion and you tap the guitar or move your hand abit or soemthgin.

I know a very crude method: a lightbulb across the speaker. Then at low power (such as noise) the filiment isnt passing much current, and is cold and low resistance. So most of the current flows through the lighbulb and not through the speaker. However, at higher volumes, the filiment heats up, becomes high resistance, and the speaker works.

This is pretty crude tho, because unfortunatly lighbulbs arnt adjustable. I want to create soemthgin adjustable with a pot.

I'v only ever herd of digital methods using processor chips. But there must be an analogue method of noise gating.

I suppose one rout woudl be to feed a rectified voltage from the inut signal into a comparetor with the other input connected to an adjustable reference voltage. The coparetor can then turn a solid state relay on and off , or control the power to the 1st stage amp.

But is there anythgin that can be done with valves? I think some pentodes can be ''switched'' on and off by varyign their screen grid voltage. But a comparetor is still needed.

Any ideas?
adibrook is offline  
Old 14th Jun 2006, 11:46 am   #2
Merlin
Retired Dormant Member
 
Merlin's Avatar
 
Join Date: Apr 2006
Location: Shrewsbury, Shropshire, UK.
Posts: 67
Default Re: Noise Gate

Perhaps a variable resistor in series with the light bulb in the speaker circuit?
Merlin is offline  
Old 14th Jun 2006, 2:16 pm   #3
adibrook
Retired Dormant Member
 
adibrook's Avatar
 
Join Date: Dec 2004
Location: Weymouth, Dorset, UK.
Posts: 377
Default Re: Noise Gate

Hmm. Would that work? A variable resistor would change the overall resistance of the circuit, but it wont change the properties of the bulb filliment.

Has anyone actually used a bulb noise gate before? Do they work well?

I really dont liek the idea of the properties of a bulb filiment affecting how this works. It would be alot better to have an electronic device.

I came up with this block diagram:

http://img.photobucket.com/albums/v4...brook/gate.gif

The AF is rectified and smoothed into a flat (ish) dc voltage, which is fed into the comparetor and compared to a stable reference voltage, which is variable.

When the signal is less than the ref voltage, the device must be fully off. As soon as the signal rises, the device must be fully on. This is what valves are bad at. I think AGC systems on radios are very proportional. They gradually turn down the gain as the signal gets stronger.Is it possible to build shmitt triggers from valves?

The comparetor then turns on the active component (here shown as a transistor) which supplies current to the amp which controls the af passing through the circuit. The active component must pass no current when in the off state

Buildign this from valves will be a challenge. Valves arnt very good at being digital. Unless some are?
adibrook is offline  
Old 14th Jun 2006, 3:01 pm   #4
Merlin
Retired Dormant Member
 
Merlin's Avatar
 
Join Date: Apr 2006
Location: Shrewsbury, Shropshire, UK.
Posts: 67
Default Re: Noise Gate

I see the theory, but you're right, building it from valves wont be easy. Are you saying your design actually turns the amp off when gating??

I know of people who have used the lightbulb trick, including the THD 'Univalve' amp. Putting something like a 100R variable resistor in series with it would allow you to control the gating level I think, and would probably soften the response some- I'm only theorising here mind. At low setting the blulb will get hot pretty quick, eg, only V low levels are gated. With a resistor in series, the bulb would stay cold longer allowing more signal to be dissipated in the resistor. Ok, not perfect, but an interesting concept I think.
Merlin is offline  
Old 14th Jun 2006, 4:08 pm   #5
adibrook
Retired Dormant Member
 
adibrook's Avatar
 
Join Date: Dec 2004
Location: Weymouth, Dorset, UK.
Posts: 377
Default Re: Noise Gate

Yeah.

It's inspired by my DOD FX7 pedal. I love the gating on it because it's really sharp.

The preamp is actually turned off when the sound level is below a preset amount.

I think i may know how to provide a shmitt trigger type action.

http://img.photobucket.com/albums/v4...rook/noise.gif

I only just thought of this...i'v never actually seen anythgin liek this before.

Anyway. Device X can be a diac, a transient supressor, or even a neon. A voltage will only appear across the resistor when the anode voltage of the valve is high enough to make device X conduct.

BTW, i know that an anode loaded valve is an inverting amp, so the output will need to be inverted again (or taken from the cathode) to be used. But the general theory should work...right?

Device X wont work too well as a neon, because once neons are fired the voltage needed to maintain discharge is smaller than the firing voltage. However, diacs and transient supressors are alot sharper i think. Is there any types of gas discharge devices that dont have much difference between firing and arc vltage?
adibrook is offline  
Old 14th Jun 2006, 5:14 pm   #6
Merlin
Retired Dormant Member
 
Merlin's Avatar
 
Join Date: Apr 2006
Location: Shrewsbury, Shropshire, UK.
Posts: 67
Default Re: Noise Gate

Hmm, I'm worried that your design would introduce crossover distortion if there was a minimum voltage drop across X. There must be something that would work though, you're right. Maybe "X" could be an LDR controlled by an LED, or would that be against your "no solid state" plans? I suppose it'd be tricky to get the fine line between compressor and expander correct in this design.

Wouldn't it be easiler just to use lower noise equipment so you dont need an expander!
Merlin is offline  
Old 14th Jun 2006, 6:04 pm   #7
adibrook
Retired Dormant Member
 
adibrook's Avatar
 
Join Date: Dec 2004
Location: Weymouth, Dorset, UK.
Posts: 377
Default Re: Noise Gate

Crossover distortion? How would that be produced?

There wont actually be any usefull signal goign through X. The grid of the valve will be connected to the comparetor, and othe output will feed into anothwer valve which switches the preamp on and off. I'll try to come up with a more complete schematic.

Umm..lowering noise? That wont work too well i'm afraid. I use tons of distortion and thers 3 valve circuits in the chain. Each with their own heater hum. And at higher distortion noise gets a real problem. Even with a digital processor controlled fx pedal noise is a problem...and everythign in it is very neat and it's made by a proper company who know what theyre doing. So i'm guessing the noise in an all avlve setup would be worst.
adibrook is offline  
Old 14th Jun 2006, 7:02 pm   #8
Merlin
Retired Dormant Member
 
Merlin's Avatar
 
Join Date: Apr 2006
Location: Shrewsbury, Shropshire, UK.
Posts: 67
Default Re: Noise Gate

Well if X were say, a diode, then it will only conduct if the signal voltage exceeds 0.7V, so you'd be shopping out the crossover bits of the AC signal, sam applies if it were a different component that had a min voltage drop.
I may have misunderstood what the circuit does though.
How come you need so many distortion boxes? Any more than one high gain one, and it'll be a distorted mush surely??
Merlin is offline  
Old 14th Jun 2006, 9:01 pm   #9
adibrook
Retired Dormant Member
 
adibrook's Avatar
 
Join Date: Dec 2004
Location: Weymouth, Dorset, UK.
Posts: 377
Default Re: Noise Gate

Ok.

Hers a rough idea.

This is ONLY an idea, and will need improvement. The circuit may not work in practice, but it will help to illustrate my design phylosophy for this device.

http://img.photobucket.com/albums/v4...rook/ngate.gif

V1 is a normal class a preamp. It amplifies the input signal enough to work with. It also buffers it so the non-liniar action of the diode doesnt load down the guitar pickup unevenly.

V2 then rectifies it, and the capacitor smoothes it. What should come out of it is a roughly smoothe DC voltage, it's magnitude being proportionaly to the volume of the input.

This is connected to a grid of a pentode, V3 , with its screen grid connected to a variable voltage source. I'm not sure how to make one yet. Maybe a neon regulator and a pot?

This controls the pentodes screen voltage, which controls it's gain.

When the input is high enough, the pentodes cathode voltage will rise high enough for the X device to conduct. And this will turn on triode V4, and provide power for the actual preamp, V5.

The red arrow shows the path the actual usable signal takes. As you see, only V5 handles the actual audio signal. Only the output of V5 will be further amplified. All the other valves are just for processing.
adibrook is offline  
Old 15th Jun 2006, 11:25 am   #10
Merlin
Retired Dormant Member
 
Merlin's Avatar
 
Join Date: Apr 2006
Location: Shrewsbury, Shropshire, UK.
Posts: 67
Default Re: Noise Gate

Nice idea, you've basically designed a sort of voltage controlled SRPP amp (V4 & V5).
The problem is, when X is not conducting, the grid of V4 will be grounded, so V4 will actually be conducting quite a lot- close to saturation, you really need to make V4's grid really negative to cut it off.
With a few tweaks in design though, you should be able to remove the anode resistor from V5, which will make for a very low output impedance like a true SRPP.
Pentodes make pretty crappy cathode followers, you might want to take the control voltage off V3's anode somehow.
Something that might work in a similar way to how you've described, is to set up V4 and V5 as a cascode, and have the grid of V4 go very negative during low "noise" levels. This system would also make for a VERY high gain stage, which I know you like! (Essentially I mean add an anode resistor to V4 and remove the anode resistor from V5, and take the output signal off V4's anode)
Merlin is offline  
Old 15th Jun 2006, 4:18 pm   #11
YC-156
Retired Dormant Member
 
YC-156's Avatar
 
Join Date: May 2005
Location: Aarhus, Denmark
Posts: 281
Default Re: Noise Gate

Making the actual signal gate is simple enough if you don't mind making a negative bias voltage to cut off a class A buffer stage. Using valves for the 'control logic' this would also require another floating heater supply tough.

Best regards

Frank N.
YC-156 is offline  
Old 15th Jun 2006, 4:39 pm   #12
YC-156
Retired Dormant Member
 
YC-156's Avatar
 
Join Date: May 2005
Location: Aarhus, Denmark
Posts: 281
Default Re: Noise Gate

Actually I think this can be done without an extra, negative supply rail. However the trouble appears to be that all designs I can come up with are DC coupled, and DC coupling and valves doesn't mix very well.

Read: Valve drift might ruin the day for a circuit like this, once it has been put into operation.

Best regards

Frank N.
YC-156 is offline  
Old 15th Jun 2006, 7:24 pm   #13
adibrook
Retired Dormant Member
 
adibrook's Avatar
 
Join Date: Dec 2004
Location: Weymouth, Dorset, UK.
Posts: 377
Default Re: Noise Gate

Floating heater supply? Why is that...please explain.

I'm also workign on another design....based onall the comments.

Drift shoudlnt be a problem unless it happens very quickly. The cut off level will be readjusted every time anyway by the variable voltage source. So as long as the charectaristics of the valve dont change in a space of 2 hrs, it hsould be allright.
adibrook is offline  
Old 15th Jun 2006, 7:47 pm   #14
adibrook
Retired Dormant Member
 
adibrook's Avatar
 
Join Date: Dec 2004
Location: Weymouth, Dorset, UK.
Posts: 377
Default Re: Noise Gate

Ok. Hers what i was thinking of.

http://img.photobucket.com/albums/v4...ook/ngate2.gif

Maybe if the HT is quiet high, and V5 needs quit low voltage and low current, this woudl be possible?

As soon as V4's anode voltage is high enogh to fire X, makign it conductive, the voltage drop across the valve starrs feeding the HT to the preamp.

If V4 has liek 400V on it's anode than there shoudl be a point when thers about 150 on the anode...right>? And thats enough for an ecc83 preamp. It only needs a few mA if that.

But as i said, this is only a thought.

What woudl really be cool is a relay in the anode circuit of V5. Thats what i'm trying to acieve. but a mechanical relay wont work. It introduces movign parts and delay time...

If if as building this out of ''dirty silicone'', than a solid state relay chip could replace V5 very nicely. I think solid state relays (based on triacs on a chip?) work really well.

Are there any valves which are easy to turn off completly? A negative supply is no problem. A half wave recified small shaver transformer is small, light, and if it's + rail is connected to shassis it will give a -130v ish supply at a few mA.
adibrook is offline  
Old 15th Jun 2006, 8:22 pm   #15
YC-156
Retired Dormant Member
 
YC-156's Avatar
 
Join Date: May 2005
Location: Aarhus, Denmark
Posts: 281
Default Re: Noise Gate

Some comments: You are right, of course. Valve drift shouldn't be a problem, since you already stated that you wanted a variable 'cut' level, which is adjusted regularly. So DC coupling should be OK.

Secondly, your current design won't work, since V3 and V4 will never stop conducting. There is no way a negative bias can develop on either valve. Even a cathode resistor would be fairly useless here, since it will only cut off a valve if it is so large as to inhibit normal operation.

Adi, what program are you using for making those nifty valve circuits, please?
YC-156 is offline  
Old 15th Jun 2006, 8:46 pm   #16
adibrook
Retired Dormant Member
 
adibrook's Avatar
 
Join Date: Dec 2004
Location: Weymouth, Dorset, UK.
Posts: 377
Default Re: Noise Gate

Sorry to dissapoint you...i'm using a very very advanced program that can be found in the start menue of any pc built after win3.1 . MS paint!

Ok. I think i need to get one fundermetal fact right here....

When a valve with a 220k resistor (for example) on the anode has a positive voltage applied to it's grid...the valves resistance drops, so more current flows, and so it's anode voltage drops. is this right?

So, when thers a low voltage on v3's grid, it's anode voltage will be high. (well...relitive of course) so v4 will be conducting and have a low anode voltage.

However, when the voltage on v3's grid is higher, v3 ill conduct more, which will make v4 conduct less. This will rise V4's anode voltage.

V3 and V4 will allways conduct...in fact V4 will conduct when the rest of the system is OFF i think....

BUT component X is what gives it a sudden on/off type action. As soon as v4's grid is low enough, and v4's resistance is high, the voltage on v4's anode will rise enough to make X start conducting.

I think connecteing the preamp directly to X is a bad idea. Lets say...whenever a positive voltage appears on the output of X, the preamp needs to be switched on. Whenever thers 0v on the output of X, the preamp needs to be switched off.

I think i'm beginnign to confuse myself .

Woudl the circuit up to X work correctly? Disregarding the preamp for now...will there only be a positive voltage on X's output of the input signal is high enough?
adibrook is offline  
Old 15th Jun 2006, 9:11 pm   #17
YC-156
Retired Dormant Member
 
YC-156's Avatar
 
Join Date: May 2005
Location: Aarhus, Denmark
Posts: 281
Default Re: Noise Gate

Quote:
Originally Posted by adibrook
Sorry to dissapoint you...i'm using a very very advanced program that can be found in the start menue of any pc built after win3.1 . MS paint!
MS Paint. Snazzy!

Quote:
Originally Posted by adibrook
Ok. I think i need to get one fundermetal fact right here....

When a valve with a 220k resistor (for example) on the anode has a positive voltage applied to it's grid...the valves resistance drops, so more current flows, and so it's anode voltage drops. is this right?
In principle, yes, but.

Most normal valves have the useful part of their characteristics, Ie. where they are designed to amplify without exceeding their design parameters, with the grid negative. When the grid is run positive, it will draw grid current, of course.

But the bad part about it is that if you try to add a significant anode resistor, you will find that the anode impedance is very low, relatively speaking. And it doesn't change much if you go from having the grid a bit positive to a lot positive. The tube is already conducting for all it is worth. You might even burn up the grid from excessive grid current as well.

Have a look at the anode curves for any valve of your choosing. For a triode the positive grid territory is to the upper left of the Eg = 0V curve. That way lies madness and destruction.

Quote:
Originally Posted by adibrook
So, when thers a low voltage on v3's grid, it's anode voltage will be high.
Nope, it will be low-ish, as the valve will be conducting for nearly all it is worth. With the exception of class C (and AB2) amplifier stages, you pretty much never venture into positive grid territory, since the valve will easily die from excessive grid or anode dissipation unless you are really careful.

(Snippety.)

Quote:
Originally Posted by adibrook
I think i'm beginnign to confuse myself .
That is OK, you are also confusing me.

I don't think the circuit around V3 and V4 will work as you expect, though it probably won't hurt to try. Just be careful with grid and anode dissipation.

Frank N.
YC-156 is offline  
Old 16th Jun 2006, 12:10 am   #18
Sam
Hexode
 
Sam's Avatar
 
Join Date: Apr 2003
Location: Higham-on-the-Hill, Leicestershire, UK.
Posts: 338
Default Re: Noise Gate

Quote:
Originally Posted by adibrook
MS paint!
Don't knock Paint! I drew all the diagrams for my project reports at Uni in Paint.

This may be a Heretical suggestion, but would Diode Logic (with Crystal Diodes ) breach your ideals, and be useful for some of the Digital/Logic stuff, rather than Valves?

Sam
__________________
Can he lead a Normal life, Doctor?
No, he will be ... an Engineer.
Sam is offline  
Old 16th Jun 2006, 11:17 am   #19
adibrook
Retired Dormant Member
 
adibrook's Avatar
 
Join Date: Dec 2004
Location: Weymouth, Dorset, UK.
Posts: 377
Default Re: Noise Gate

OOOOh.

riiight.

I think i understand now.

Valves are not what i thought they are.

So a valve with 0v on the grid is actually nearly ''open''?

To get it to ''close'' you make the grid negative by lifiting the cathode. And the input signal varies the NEGATIVITY of the grid.

So if you have -10v on the grid from the cathode resistor, the valve would pass no current. But then if you put in a 3v signal, this will rise the grid voltage to -7, which will make the valve partially conduct.

Damn. I keep treating valves like transistors.

But i think i understand now.

very negative grid = no anode current
slightly negative grid = a bit of anode current
positive grid = max current.
adibrook is offline  
Old 16th Jun 2006, 1:04 pm   #20
YC-156
Retired Dormant Member
 
YC-156's Avatar
 
Join Date: May 2005
Location: Aarhus, Denmark
Posts: 281
Default Re: Noise Gate

Yep, you got it now. Valves are voltage controlled devices, not current controlled like bipolar transistors.

It is not that a valve doesn't amplify with the grid positive. But the cathode has a maximum current capacity, which is quickly reached when the grid is allowed to run positive. Ie. the anode current saturates, so there is nothing gained from making the grid even more positive.

Since there is no inherent benefit from running the grid positive, small signal valves are usually designed such that their proper operating area is, as already stated, with the grid mostly below zero.

Say the average negative grid voltage is -5V, a typical value. Then if the input signal increased to more than 5V peak, the grid would of course run positive for part of the AC cycle. Current, Ie. power, would be drawn by the positive grid, and this would result in unwanted distortion due to the uneven loading of the driver circuit.

Conversely, if the grid is never positive, no power would be drawn from the driver circuit and the valve input impedance would be very high. (This is actually only perfectly true under idealized conditions, but you get the idea.) The large resistances encountered in valve input circuits for audio are only possible due to the high input impedance of negatively biased signal grids.

So now you might also see how you can easily make a signal gate.

Say you have a normal grid bleed resistor of 100Kohm going from grid to ground in a class A buffer stage. Imagine the cathode resistor ensures that the average negative bias voltage is -3V and that we expect the signal input to have a maximum peak-to-peak amplitude of 2V. This way the grid will never be able to go positive, just like we want.

From the datasheet we might see that when the first grid is biased to -8V, the anode current is essentially zero. So if we lift the 'cold' end of the 100K grid bleeder resistor away from the ground and connect it to a fixed bias source of -20 or -25V, the valve will essentially be completely cut off, signal or no input signal.

So cold end of grid bleeder to ground: Gain stage works as usual. Grid bleeder to negative fixed bias: Audio cut off.

Try designing a bit around that idea.

Frank N.
YC-156 is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 9:45 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.