Maker Pro
Maker Pro

White Noise Source

C

Chiron

I need a source of white noise (audio). It doesn't have to be perfect.

I seem to recall that you could use a zener diode as a white noise
source. Is this true? Are there any other possibilities? I anticipate
using a 9-V battery for my power, and a few transistors to amplify the
signal to drive earphones.

You may ask, "why the heck would he want to do that?" Good question:

http://en.wikipedia.org/wiki/Electronic_voice_phenomenon

Since I believe that the chances of there being anything to this are just
about zero, I don't want to invest a whole lot of effort in doing it.
 
B

Bob Masta

I need a source of white noise (audio). It doesn't have to be perfect.

I seem to recall that you could use a zener diode as a white noise
source. Is this true? Are there any other possibilities? I anticipate
using a 9-V battery for my power, and a few transistors to amplify the
signal to drive earphones.

You may ask, "why the heck would he want to do that?" Good question:

http://en.wikipedia.org/wiki/Electronic_voice_phenomenon

Since I believe that the chances of there being anything to this are just
about zero, I don't want to invest a whole lot of effort in doing it.

Download my Daqarta software... it's free for signal
generation use. Midway down the Generator dialog, click the
left Waveform Controls button. Click Wave on the dialog that
appears and select Noise: White. Toggle the Generator
button on in the main toolbar. The first time the Generator
goes on the volume control dialog will pop up, plus a
cautionary message about loud levels. You'll need to cancel
the message before the sound starts. (In WinXP you may also
need to unMute, at the bottom of that dialog.) Later, you
can open the volume dialog via the F9 key at any time.

Now I'm going to give away a secret: In the next version of
Daqarta (v7.00, in a month or so) there will be a whole Help
section on EVP-related stuff, and a demo on how to hear it
for yourself. Well, not exactly what the woo-woo folks are
talking about, but a phenomenon I call "Phantom Signals"
that I think is at the root of the EVP nonsense... as well
as "dental radio".

The trick for phantom signals is to use headphones and put
the white noise in *one ear only*. The loudness should be
about 30-40 dB above your detection threshold. (You'll need
to be in a fairly quiet room, with no music, TV, or voices.)


I'll provide a simple way to set that level. Let me know if
you want me to explain a manual approach here. It's not
super-critical, but it needs to be loud enough to hear
easily, but not too loud.

Listen for a few minutes, and you will start to hear a
"radio station". You might hear music or voices, but you
won't be able to tell exactly what the song is, or
understand what they are talking about... nevertheless
you'll be convinced there is a "signal" there. You might
categorize the station as "Country", or "Talk Radio", or
whatever, based on what you hear. You'll probably have an
overpowering feeling that if only you could just improve the
S/N a *teeny* bit you would understand everything. But you
can't.

And if you switch the noise source to *both* ears, the
phantom signal vanishes and you just hear noise. (Sometimes
you can get the effect with both ears, but it may take a
*lot* of listening.)

I actually discovered this a long time ago, while designing
a diode-based noise generator. Since that sort of circuit
involves a lot of gain, when I heard the "radio station" I
naturally assumed that's *exactly* what it was hearing. I
tried all kinds of filtering before I finally got rid of
it... coincidentally having switched from a single-ear
lineman's test set to stereo phones. Problem solved!

Then I built a completely digital version, which doesn't
involve gain and is not susceptible to AM pickup... or so I
assumed. Yet it was back again, but only with the mono
headset, not stereo or speakers. Took a lot of testing to
convince myself that it wasn't a circuit problem... at least
not in my hardware circuit, only the "wetware" between my
ears!

Then when I showed the phantom signals effect to others they
were absolutely convinced it was a real radio station, and
that there was some defect in my hardware. Had to go
through all the steps with them to prove it was only in
their heads. The really convincing proof was that it
vanished instantly when heard through both ears... hard to
explain if it was AM pick-up!

For a long time I didn't want to announce this to the
general public because I hoped to recruit subjects for an
unbiased test. But that got put on the back burner for so
long that I've given up on it.

So enjoy!

Best regards,


Bob Masta

DAQARTA v6.02
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Scope, Spectrum, Spectrogram, Sound Level Meter
Frequency Counter, FREE Signal Generator
Pitch Track, Pitch-to-MIDI
Science with your sound card!
 
O

Oppie

Chiron said:
I need a source of white noise (audio). It doesn't have to be perfect.

I seem to recall that you could use a zener diode as a white noise
source. Is this true? Are there any other possibilities? I anticipate
using a 9-V battery for my power, and a few transistors to amplify the
signal to drive earphones.

You may ask, "why the heck would he want to do that?" Good question:

http://en.wikipedia.org/wiki/Electronic_voice_phenomenon

Since I believe that the chances of there being anything to this are just
about zero, I don't want to invest a whole lot of effort in doing it.

Zeners do work. I had used a 6V zener feeding it through a 1 meg resistor
from 15V. The spectrum was lacking in much of the audio range though. I had
better success reverse biasing the base-emitter junction of any garden
variety transistor. The NPNs that I tried break down in the range of 8
volts, so again, a 1 Meg feed from 15V is a good starting point.

the '15V' can be either from two 9V batteries or a boost converter.

OR you can look up "white noise schematic"

Pseudo Random Noise can be digitally generated with a recirculating shift
register that has the feedback tapped at a few points and XORed as shown in
this diagram
http://www.discovercircuits.com/DJ-Circuits/noisegen.htm
It shows 12V power but modern logic elements will work at lower voltages. I
used a similar circuit for calibrating a sonar system (signal of interest
burried in 90 db of noise).
 
C

Chiron

On Thu, 09 Feb 2012 09:32:55 -0500, Oppie wrote:

<snip>

Thanks, Oppie. I might try the idea with the transistor. Seems like it
might be what I'm looking for...
 
C

Chiron

On Thu, 09 Feb 2012 13:28:04 -0800, George Herold wrote:

I really like 20V 1/2 watt zeners for this... bias from +/-15 volts
through a 1 meg resistor. Lotsa noise and flat out to ~500kHz or so.
The voltage is a bit asymmetric, but that can be fixed..... by adding
two of opposite polarity.

Thanks, George. Is there anything special about the zener being 20V? In
other words, would a 20V zener be better than using, say, a 5.1V?
 
B

BeeJ

On Thu, 09 Feb 2012 13:28:04 -0800, George Herold wrote:



Thanks, George. Is there anything special about the zener being 20V? In
other words, would a 20V zener be better than using, say, a 5.1V?

Yes, the higher voltage zeners have a little negative resistance in
part of their curve. That is why you can build an oscillator using
just the diode a capacitor and resistor. Try it. Fun.
Look closely at the curves on the higher voltage zeners.
 
C

Chiron

Yeah a big difference. In low voltage Zeners the current is from
tunneling electrons (one at a time) and only have shot noise. (Which is
pretty small). High voltage 'Zeners' do this avalanche break down...big
random current pulses.
Lot's more noise 'amplitude'.
I looked at a bunch of different voltages. (Above somethiing like 6V is
where you start to get some avalanching.) 10 and 12 Volt Zeners showed
more than shot noise, but there was a lot of part to part variation in
the amount of noise. For some reason* 20V zeners all give about the
same signal levels... I posted a 'scope shot some time back.
If you use glass encapsulated diodes then wrap some black electrical
tape around them... else the room lights leak in and ruin everthing.


<snip>

This is interesting. Who knew there was so much to something as
(seemingly) simple as a zener diode? Not me, anyway.

Thanks for the info.
 
C

Chiron

On Thu, 09 Feb 2012 09:32:55 -0500, Oppie wrote:

OR you can look up "white noise schematic"

Never even thought of that; I'll do it.
Pseudo Random Noise can be digitally generated with a recirculating
shift register that has the feedback tapped at a few points and XORed as
shown in this diagram
http://www.discovercircuits.com/DJ-Circuits/noisegen.htm It shows 12V
power but modern logic elements will work at lower voltages. I used a
similar circuit for calibrating a sonar system (signal of interest
burried in 90 db of noise).

I think I recall something similar from Don Lancaster's CMOS Cookbook...
but IIRC, the "noise" repeated. Not sure whether that would be a
problem, though...

Anyway, thanks for the ideas.
 
M

Michael Black

On Thu, 09 Feb 2012 09:32:55 -0500, Oppie wrote:



Never even thought of that; I'll do it.

I think I recall something similar from Don Lancaster's CMOS Cookbook...
but IIRC, the "noise" repeated. Not sure whether that would be a
problem, though...
It depends on what you want, and how long the sequence is.

Sound generator ICs like the AY-5-8910 (I think I got that number right)
or those from Texas Instruments used that scheme. If the sound generator
in the Commodore 64 included a white noise source (I can't remember), then
it used that scheme.

For a while in the seventies, National Semiconductor had an IC that
generated noise using that scheme, it was sold as a "pink noise
generator", where you'd add an external filter to make the noise "pink".
Another company had a similar device.

Of course, PAIA just used a 2N2712 for the noie source in their projects.
It was never clear if they found that device was better at it, or if they
just found a cheap device.

Michael
 
C

Chiron

On Thu, 09 Feb 2012 22:49:29 -0500, Michael Black wrote:

It depends on what you want, and how long the sequence is.

Sound generator ICs like the AY-5-8910 (I think I got that number right)
or those from Texas Instruments used that scheme. If the sound
generator in the Commodore 64 included a white noise source (I can't
remember), then it used that scheme.

For a while in the seventies, National Semiconductor had an IC that
generated noise using that scheme, it was sold as a "pink noise
generator", where you'd add an external filter to make the noise "pink".
Another company had a similar device.

Of course, PAIA just used a 2N2712 for the noie source in their
projects. It was never clear if they found that device was better at it,
or if they just found a cheap device.

Michael

IIRC, the C-64 did have some sort of noise function... we're going back a
ways now. That was about the last time I played around with electronics,
too. Things have changed a wee bit since then - what, 1988 or so?

I found a circuit that looks promising. I'll have to refer you to the
Website, since I don't yet know how to draw schematics using ASCII:

http://www.seekic.com/uploadfile/ic-circuit/2009624223240862.gif

This circuit uses 2 resistors, 2 capacitors, a diode and a 2N3904. I
doubt I'm going to find anything simpler, unless it's an IC like you
described.

I think my main problem was I didn't hit upon the right Google search
words, so I was chasing down all kinds of stuff that had nothing to do
with what I wanted.

Thanks, all, for your help.
 
C

Chiron

You give them a solution - and for free that they can use right now -
and they continue to talk about using Zeners to make noise. I have a
horse. I showed him where the water is and he drinks it. Not sure about
these people though.

G²


How about, I ask for a cheeseburger; someone offers me a free glass of
water; but I keep talking about the cheeseburger, because I'm hungry, not
thirsty. Besides, your horse is drinking out of the glass, and that's
just plain gross.

To be honest, I never saw the post you quoted. It didn't get past the
killfile - too many lines. Unfortunately, one of the problems with a
killfile is that you may wind up blocking perfectly legitimate posts, as
you try to get rid of all the crap.

OK, I'll have a look on Google and see what he said...
 
C

Chiron

On Thu, 09 Feb 2012 23:58:58 -0800, stratus46 wrote:

You asked for an audio noise source and mentioned Zeners but you have to
build something to actually use Zeners. You obviously have a computer
which very likely has audio capabilities. You can build something in a
few days or you can have your noise source in minutes. You were offered
a noise source. What's your time worth?

G²

Yes, of course you're right. However, there are several considerations
here. First, as I said, I had no idea Bob Masta had responded to my
post- a problem I hope to fix very soon. Second, I don't use Windows,
which is the OS he writes programs for. Third, I do this stuff for fun,
so I have not the slightest objection to reinventing the wheel or doing
things the "hard" way.

The circuit I found contains six parts. If I had the zener I could slap
it together in about ten minutes.

But mostly I just didn't see Bob's post. Even if I didn't like his idea
(and I do like it), I'd have been courteous enough to reply, had I seen
the post.
 
B

Bob Masta

Yeah, well, as I was saying, it's an interesting question, but not one
I feel up to answering.

I've listened to recorded samples where people thought they heard a
voice, but I never got anything other than the noise. So I figured
I'd just try it myself and see what happens. Since I am doubtful, I
didn't want to invest a whole lot of effort building a circuit.

Thanks for your offer. I don't think I can use it, because I'm not
using Windows. But I do appreciate the offer, and I again apologize
for seeming to ignore you.

No apology needed. If you go with a hardware solution you
will need to be able to convince yourself that it is not
picking up an AM station, so be sure to compare one-ear
listrening to two-ear listening if you get an effect.

As for Zeners, I haven't tried the 20V units George
mentioned... but that sounds like a good aproach if it
avoids the need for a lot of gain. I used a reversed-bias
EB junction from an NPN back in the "olden days"... I think
this was pretty common. You ground the base and feed the
collector through 100k from +15, and tap off the noise via a
big cap (0.47 or whatever) and a bunch of gain.

If you go with a pseudo-random generator, you have to be
careful: "Back in the day" the single-chip units were
notorious for having a short repeat pattern. They were
designed for transient uses like snare drums, and if you ran
them continuously you'd hear "shoop, shoop, shoop" from the
repeats. Make sure whatever you use has at least a minute
or more, and for your purpose best if it's hours or beyond.


In the past, when computers had easily-controlled hardware,
it was fairly easy to do this with code that drove the
printer port. Alas, no more.

Best regards,


Bob Masta

DAQARTA v6.02
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Scope, Spectrum, Spectrogram, Sound Level Meter
Frequency Counter, FREE Signal Generator
Pitch Track, Pitch-to-MIDI
Science with your sound card!
 
C

Chiron

On Fri, 10 Feb 2012 00:03:15 -0800 (PST), chiron613
No apology needed. If you go with a hardware solution you will need to
be able to convince yourself that it is not picking up an AM station, so
be sure to compare one-ear listrening to two-ear listening if you get an
effect.

You know, you reminded me that my system has a random function built into
it - not sure exactly how it works, but it spews out (pseudo)random bytes
at a pretty fast rate (A few seconds worth created a file > 100M). Now
the trick is to convert that to sound... I suppose I can figure that
out. I might pass it through an FFT program just to see whether it's
even close to being white noise... I just don't know how this "random"
function works. As John von Neumann said, "Anyone who attempts to
generate random numbers by deterministic means is, of course, living in a
state of sin."
As for Zeners, I haven't tried the 20V units George mentioned... but
that sounds like a good aproach if it avoids the need for a lot of gain.
I used a reversed-bias EB junction from an NPN back in the "olden
days"... I think this was pretty common. You ground the base and feed
the collector through 100k from +15, and tap off the noise via a big cap
(0.47 or whatever) and a bunch of gain.

I don't see much problem with gain - if I can't do it with a few
transistors (maybe in a Darlington arrangement?), then I can always use a
couple of op amps.
If you go with a pseudo-random generator, you have to be careful: "Back
in the day" the single-chip units were notorious for having a short
repeat pattern. They were designed for transient uses like snare drums,
and if you ran them continuously you'd hear "shoop, shoop, shoop" from
the repeats. Make sure whatever you use has at least a minute or more,
and for your purpose best if it's hours or beyond.

I'm probably not going to try this. I don't have the parts handy.
In the past, when computers had easily-controlled hardware, it was
fairly easy to do this with code that drove the printer port. Alas, no
more.

Sure, back when you just about had to write your own drivers for
everything. I had a lot of fun back then, but I wouldn't really want to
go back to it now.



--
Fortune's current rates:

Answers .10
Long answers .25
Answers requiring thought .50
Correct answers $1.00

Dumb looks are still free.
 
E

ehsjr

You asked for an audio noise source and mentioned Zeners but you have
to build something to actually use Zeners. You obviously have a
computer which very likely has audio capabilities. You can build
something in a few days or you can have your noise source in minutes.
You were offered a noise source. What's your time worth?


Sometimes you want (or need) to use a noise source (or whatever)
when there is no computer close enough. A laptop solves that
problem - but the zener source solves it a whole lot cheaper.

Ed
 
M

Michael Black

You asked for an audio noise source and mentioned Zeners but you have
to build something to actually use Zeners. You obviously have a
computer which very likely has audio capabilities. You can build
something in a few days or you can have your noise source in minutes.
You were offered a noise source. What's your time worth?
You can throw together something that will generate white noise in
minutes. Yes, longer if you need to put it in a box, but it's a very
simple circuit, a dioded to generate the noise, and then amplification.
It won't take "days".

On the other hand, having a little circuit can be pretty useful. If one
is using it for testing, it kind of beats trying to balance that laptop on
your knee as you kneel behind the stereo system. You can take the noise
source to the problem, rather than bringing the problem to the noise
source.

It really depends on what is needed.

But these are discussion groups, to debate what someone has said, to offer
other solutions. The discussion can go on long after various things have
been offered, it doesn't negate the validity of those solutions offered.

But lots of times, people ask questions that are too vague to get an
absolute answer, so while a computer generated white noise source is a
valid answer, it may not be what's actually needed.

Michael
 
M

Michael Black

I don't see much problem with gain - if I can't do it with a few
transistors (maybe in a Darlington arrangement?), then I can always use a
couple of op amps.
Then there was the circuit in "Electronics", Designer's Casebook or
Engineer's notebook, that just used a noisy op-amp and amplified it, as a
noise source. No diode needed.

Michael
 
J

Jasen Betts

On Thu, 09 Feb 2012 21:58:06 -0800, stratus46 wrote:

To be honest, I never saw the post you quoted. It didn't get past the
killfile - too many lines. Unfortunately, one of the problems with a
killfile is that you may wind up blocking perfectly legitimate posts, as
you try to get rid of all the crap.

Your filter blocked a noise source,
kind of ironic.
 
J

Jasen Betts

On Fri, 10 Feb 2012 13:37:29 +0000, Bob Masta wrote:

You know, you reminded me that my system has a random function built into
it - not sure exactly how it works, but it spews out (pseudo)random bytes
at a pretty fast rate (A few seconds worth created a file > 100M). Now
the trick is to convert that to sound... I suppose I can figure that
out. I might pass it through an FFT program just to see whether it's
even close to being white noise... I just don't know how this "random"
function works. As John von Neumann said, "Anyone who attempts to
generate random numbers by deterministic means is, of course, living in a
state of sin."

if linux, this is easy:

cat /dev/urandom > /dev/audio

(dev/audio is (IIRC) mono 8 bit Ulaw and runs at a fixed sample rate
of about 8000 samples per second so it's not propper white noise.

for something better try sox or audacity.
 

Similar threads

Top