Noise

[j.dilisio]

I'm trying to find the simplest noise source circuit possible.
One jack, noise out.

[fonik]

from the EFM LFO 5A. you may have to try transistors from different batches IIRC. the OpAmp in the schematic is from a quad version. just replace with single OpAmp.

---

EFM_NSE.gif
Description:
Filesize:	4.74 KB
Viewed:	9990 Time(s)

---

[j.dilisio]

Thanks, I'll give it a try.

[Sine]

I have had good results with 2n2222's as "noise" transistors.

[Sound]

I would like recommend the book "The Noise Generator Cookbook" by Thomas Henry.

Regards.

[j.dilisio]

Here are a couple other similar noise circuits I found.
The first is from the CGS Synare and the second from a Thomas Henry tunable noise schematic.

They're pretty much the same as the EFM circuit above. I was just wondering about the differences.
Thomas has the transistor wired to -15 rather than ground.
The EFM op amp is wired differently (inverted?).
Ken omits the capacitor on the output.

---

CGSnoise.png
Description:
Filesize:	28.07 KB
Viewed:	9871 Time(s)

---

---

THnoise.png
Description:
Filesize:	33.21 KB
Viewed:	9871 Time(s)

---

[Tim Servo]

Hey JD,

Yes, those schemos are fairly similar. Most designs I've seen use a non-inverting op amp to amplify the noise coming of the transistor. The non-inverting config has a higher input impedance, so that's the config to use unless you specifically need to invert the signal or you need to sum two or more signals (using an inverting summing circuit).

As far as the different supply voltages, these transistors make noise when they're reverse-biased. Thomas' design increases the potential difference and generates more consistent noise (IMHO). The output cap is simply to remove any DC bias. It looks like the EFM design does this by adding an offset at the op amp to accomplish roughly the same thing.

One plug for my own Magic Smoke stuff - Thomas' Noise Generator Cookbook can be found at:
http://stores.lulu.com/magsmoke
You can also preview the first few pages, including the table of contents.

Tim (let's make some noise) Servo

[j.dilisio]

Thanks Tim. The world of electronics is slowly starting to make more sense.
I was messing with the Thomas Henry circuit today minus the 10k resistor.
It works fine of course.
I went through a bag of NPN switching transistors from radioshack and found that in general the 2n3904s were the least noisy, the 2n2222s were consistently the most noisy, and the 2n4401s fell somewhere in between.

[frijitz]

[fonik]

that is very valuable information. thank you, ian.

[frijitz]

Oh, and guess what the worst one I characterized was? The 2N2712, touted as the all time best ever noise transistor.



Ian

[j.dilisio]

Ahhh, I see what you mean.
The noise from the 2n3904s is much smoother/even sounding.
So in order to amplify the signal I just need to increase the resistance of R13 (on the TH circuit), correct?

[frijitz]

[stolenfat]

what about this one (untested by myself)
found it in the experimentalists anonymous archives

---

Simple White Noise Generator 2.jpg
Description:
Filesize:	52.25 KB
Viewed:	9315 Time(s)

---

[j.dilisio]

The signal is still a little low with the simple circuit so I tried to add another gain stage by wiring the other opamp the same way.

I just get a loud crackling when I try it.
I must be missing something here.

Maybe the gain is too high. I was hoping this would clip off the excess and smooth it out but it seems to be doing the opposite?

---

noise2gain.jpg
Description:
Filesize:	43.48 KB
Viewed:	238 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[blue hell]

[j.dilisio]

Yep, the 1M's were a bit much.
I get a good level with a 200k in one stage and a 10k in the other.
thanKs

[j.dilisio]

.. and here's the board

---

NoisInvertPnP2.png
Description:
Filesize:	8.66 MB
Viewed:	251 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---