OT: Does anyone have a good, simple random noise generator?

Random number generator 'improved' - BBC News

Thyratron and a magnet:

http://tubedata.milbert.com/sheets/137/6/6D4.pdf

Tony Rolando wrote:

"Steve Jacobson" <stevelists@jjworld.com> wrote in message

news:c16i94trmro9sh8rl01es6maubk9v9fump@4ax.com...

On Tue, 05 Aug 2008 23:07:28 -0400, Steve Jacobson <stevelists@jjworld.com>

wrote:

 

On Tue, 05 Aug 2008 17:44:33 -0700, Gary Gofstein

<ggofstein@lbl_nospam.gov>

wrote:

 

just reverse bias the base emmitter junction of any normal transistor,

NPN breaks down at about 7V. don't connect the collector. the transistor

will be useless for any other purpose after this, it will appear to work

but detailed measurements will show it is out of spec, so throw it away

if you decide not to use it. 1ma is plenty of current. ac couple the

transistor voltage to an op amp and voila! true analog white noise.

 

Gary,

 

What you are saying is to connect the base-emitter junction of an NPN

transistor

to a source greater than 7V.  Connect the base to +7V and the emitter

to -7V.

Ground the base and take the white noise off the emitter (coupled with a

cap).

Is this correct?  What noise voltage should I expect?

 

Thanks for your help.  I hope I don't appear too dense.

 

I will be running the thing at 9V.  Should I be using a current limiting

resistor?  Also, what would be a good NPN transistor?

 

Steve

 

A BC550 is a good place to start. Grab a handful, and select for highest

noise. I have used 2N3904 as well.

 

However, as Gary wrote make sure it doesn't end up back in the parts

bin. Snipping off the collector lead or tab would be one way, or paint

on some gross color (I've used some of my wife's nail polish for that

but don't tell her ...). If it does end up back in the bin that would be

a perfect recipe for drustrations months down the road.

--

Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.

Use another domain or send PM.

there is no real danger in reverse biasing the base emitter junction of

a standard general purpose transistor for this purpose. at 1ma and 7v

the dissipation will be 7mw, practically nothing. you can experiment

with the amount of current that produces the best noise.

however, as i mentioned, the transistor should not be used for anything

else afterward as it will be damaged in regard to speed and noise

specifications. a zener diode is designed to be reverse biased and will

take orders of magnitude more power - unfortunately, it will not produce

as much or as good quality of noise. the noise voltage will vary from

transistor type to type and within each type from one sample to another,

but i doubt it will be significant unless you need a calibrated accurate

level. the zener noise will also vary from unit to unit.

and yes, you must use a resistor to limit the current. if you run it

from 9v and the junction drops 7v, your resistor should be 2v/1ma or

2kohm. the voltage will be developed between the emitter and base (which

  you will most likely consider ground) so yes, take the output from the

resistor emitter connection.

try a gain of 100 as a first guess, though you may need more like 1000.

it depends upon what final amplitude and frequency range you need. the

whole problem with these simple noise generators is getting enough

output, that's why the transistor is better. the more gain you try to

get from the op amp, the less bandwith you will get. for 25khz, to get a

gain of 1000 you'd need an op amp with 25Mhz unity gain bandwidth.

that's not likely a part you have in your junk box. a better bet is to

use two stages of lower opamp gain if you need it.

it sounds like you are just learning about analog electronics - this is

a great circuit to teach yourself the fundamentals of analog

electronics. simple but full of options, like all analog circuits. do

you think you should use the inverting or noninverting op amp circuits?

where should you put the gain, is it better to have more noise to start

with or more amplification? how about power supply rejection, should you

make a differential connection to the noise source? have fun, try a

zener diode, try an LED, try different junctions and see for yourself

what effects they have!

"Steve Jacobson" <stevelists@jjworld.com> wrote in message

news:c16i94trmro9sh8rl01es6maubk9v9fump@4ax.com...

On Tue, 05 Aug 2008 23:07:28 -0400, Steve Jacobson <stevelists@jjworld.com>

wrote:

On Tue, 05 Aug 2008 17:44:33 -0700, Gary Gofstein

<ggofstein@lbl_nospam.gov>

wrote:

 

just reverse bias the base emmitter junction of any normal transistor,

NPN breaks down at about 7V. don't connect the collector. the transistor

will be useless for any other purpose after this, it will appear to work

but detailed measurements will show it is out of spec, so throw it away

if you decide not to use it. 1ma is plenty of current. ac couple the

transistor voltage to an op amp and voila! true analog white noise.

 

Gary,

 

What you are saying is to connect the base-emitter junction of an NPN

transistor

to a source greater than 7V.  Connect the base to +7V and the emitter

to -7V.

Ground the base and take the white noise off the emitter (coupled with a

cap).

Is this correct?  What noise voltage should I expect?

 

Thanks for your help.  I hope I don't appear too dense.

I will be running the thing at 9V.  Should I be using a current limiting

resistor?  Also, what would be a good NPN transistor?

Steve

A BC550 is a good place to start. Grab a handful, and select for highest

noise. I have used 2N3904 as well.

Tony

On Tue, 05 Aug 2008 23:20:16 -0400, Steve Jacobson

<stevelists@jjworld.com> wrote:

I will be running the thing at 9V.  Should I be using a current limiting

resistor?  Also, what would be a good NPN transistor?

I would not use a transistor at all.  Reverse biasing a base-emitter

junction may simply destroy the junction.  You never really know.  It

may work OK for a hand-built hobby project, but I would not use it in

anything that I was going to sell.

Instead, use a reversed-biased Zener diode.  They are designed to be

reverse-biased.  Here is an app note from Maxim (although you can

probably ignore the use of special low-noise Maxim op amps):

  http://www.maxim-ic.com/appnotes.cfm/an_pk/3469

Robert Scott

Real-Time Specialties

Ypsilanti, Michigan

On Tue, 05 Aug 2008 23:07:28 -0400, Steve Jacobson <stevelists@jjworld.com>

wrote:

On Tue, 05 Aug 2008 17:44:33 -0700, Gary Gofstein <ggofstein@lbl_nospam.gov>

wrote:

 

just reverse bias the base emmitter junction of any normal transistor,

NPN breaks down at about 7V. don't connect the collector. the transistor

will be useless for any other purpose after this, it will appear to work

but detailed measurements will show it is out of spec, so throw it away

if you decide not to use it. 1ma is plenty of current. ac couple the

transistor voltage to an op amp and voila! true analog white noise.

 

Gary,

 

What you are saying is to connect the base-emitter junction of an NPN transistor

to a source greater than 7V.  Connect the base to +7V and the emitter to -7V.

Ground the base and take the white noise off the emitter (coupled with a cap).

Is this correct?  What noise voltage should I expect?

 

Thanks for your help.  I hope I don't appear too dense.

I will be running the thing at 9V.  Should I be using a current limiting

resistor?  Also, what would be a good NPN transistor?

Steve

On Tue, 05 Aug 2008 17:44:33 -0700, Gary Gofstein <ggofstein@lbl_nospam.gov>

wrote:

just reverse bias the base emmitter junction of any normal transistor,

NPN breaks down at about 7V. don't connect the collector. the transistor

will be useless for any other purpose after this, it will appear to work

but detailed measurements will show it is out of spec, so throw it away

if you decide not to use it. 1ma is plenty of current. ac couple the

transistor voltage to an op amp and voila! true analog white noise.

Gary,

What you are saying is to connect the base-emitter junction of an NPN transistor

to a source greater than 7V.  Connect the base to +7V and the emitter to -7V.

Ground the base and take the white noise off the emitter (coupled with a cap).

Is this correct?  What noise voltage should I expect?

Thanks for your help.  I hope I don't appear too dense.

just reverse bias the base emmitter junction of any normal transistor,

NPN breaks down at about 7V. don't connect the collector. the transistor

will be useless for any other purpose after this, it will appear to work

but detailed measurements will show it is out of spec, so throw it away

if you decide not to use it. 1ma is plenty of current. ac couple the

transistor voltage to an op amp and voila! true analog white noise.