looking for the noisiest non smd zener diode that is less than 7.5 volts

[Brian Marshall]

im working on a new circuit that needs a noise generator.

essentially here is the circuit.
Most of the zeners ive tried dont make enough noise... the noisiest one ive found is an 8 volt zener, but im scared to use it because im trying to build this at 9 vlots.



Ive also had some luck with pn2369 transistors as well.  Any help is appreciated.

[gez]

the noisiest one ive found is an 8 volt zener, but im scared to use it because im trying to build this at 9 vlots.

If you're using a current limiting resistor it shouldn't be a problem, it's the power rating of the device you have to watch.  Try reducing the current, might give you a little more noise?

[MartyMart]

would a schottkey 5817 work ? it's supposed to be a bit like a "Ge" diode..
probably not .... 2N2222  !! very noisey in the garage fuzz !

[Steben]

would a schottkey 5817 work ? it's supposed to be a bit like a "Ge" diode..
probably not .... 2N2222  !! very noisey in the garage fuzz !

Noise can come from anything in that fuzz!

[gez]

the noisiest one ive found is an 8 volt zener, but im scared to use it because im trying to build this at 9 vlots.

If you're using a current limiting resistor it shouldn't be a problem, it's the power rating of the device you have to watch.  Try reducing the current, might give you a little more noise?

Sorry Brian, it's first thing in the morning.  I see what you mean now - if the battery flattens there won't be enough supply voltage to cover the threshold.

Try one with a lower threshold and starve the current (as much as you can get away with).  Might make a difference.

[Brian Marshall]

the noisiest one ive found is an 8 volt zener, but im scared to use it because im trying to build this at 9 vlots.

If you're using a current limiting resistor it shouldn't be a problem, it's the power rating of the device you have to watch.  Try reducing the current, might give you a little more noise?

Sorry Brian, it's first thing in the morning.  I see what you mean now - if the battery flattens there won't be enough supply voltage to cover the threshold.

Try one with a lower threshold and starve the current (as much as you can get away with).  Might make a difference.

Well, ive noticed the larger resistor i use the less noise i get... i suppose that could be because of the impedance of the noise rather than the actual volage (or mV) swing of the noise at the zener.  The lower higer the impedance the less gain i get out of the opamp since its inverting.  I tried it with a non inverting opamp, but had little luck with it... the noies had too much of an obvious frequency to it for some reason.

I'm starting to think that I shouldnt rely on parts to do things they werent meant for, so my fix may be to use a smaller zener... like in the 6 to 7 volt range, and follow the existing circuit with a 100K trim pot with one side to the opamp out put and the other to the bias voltage... and the whiper going to a small resistor to another inverting amplifier...

If i do this, it will be more parts, and more work, but at least i can deal with the inconsistancies as they arrise.

[George Giblet]

Some years ago I spent a lot of time playing around with avalanche noise sources.   The best noise sources were reverse biased BE junctions of transistors (they tended to be in the 8V breakdown region at the operating current) and 8.2V to 16V zeners (the actual breakdown voltage is lower than the labelled voltage because of the low currents).  You want to operate them at around 10uA; you can go a little higher to get a buit more noise level.  The quality of the noise varies a lot especially at high and low currents, and with the device.

Below 8V the breakdown mechanism isn't entirely an avalanche break down and the noise quality and level is poor.

The avalance breakdown circuits barely make it on a 9V rail, with 12V things are sweet.  I think you would be better off with the transistors, perhaps hand select one - try BC547, BC548, 2N3904.  You know expensive noise diodes tend to be sorted at the factory.

For the opamp you don't want to add a resistor between the "zener" and the -ve input.  You want the diode's noise current to pump into the amplification stage.

Lastly make sure you add some decent decouping on the supply.

[Paul Perry (Frostwave)]

http://www.ciphersbyritter.com/NOISE/NOISRC.HTM I havn't tried this, (the first schem on that page) but at least it is said to be reproducable, and low voltage. I'm interested in the possibility of (using modern low offset op amps) of just amplifying hte intrinsic noise of a 1M resistor. Anyone have experiences or opinions on this?

[DDD]

1. Try non-invertimg Op-amp circuit. Your circuit is a kind of inverting amp with gain=1M/(10kOhm in parallel with dynamic resistance of Zener). Also there is a compromise: to get higher gain of the amp you should lower "resistance" of Zener by increasing its current. Otherwise Zeners have higher noise figure at low currents. So the product of "gain by jZener noise" is approximately constant and hence limited.
2. If you feed supply voltage to Zener via big resistor and use non-invertimg Op-amp circuit to amplify the noise you can get better results. For example: 1MOhm from output to inverting input, 1 kOhm from inverting input to ground through capacitor will give you gain about 1001 instead of 100-200 from your inverting amp.
3. As far as I know low-voltage Zeners have higher noise at 10-30 uA.

[Brian Marshall]

2. If you feed supply voltage to Zener via big resistor and use non-invertimg Op-amp circuit to amplify the noise you can get better results. For example: 1MOhm from output to inverting input, 1 kOhm from inverting input to ground through capacitor will give you gain about 1001 instead of 100-200 from your inverting amp.

i pretty much tried this exactly, but the noise produced, as stated above, didnt seem random... it seemed have a frequency to it... i dont know... ill sit down with it again tonight.

[Vsat]

I've had good results with the Motorola MPSA05 transistor with reverse-biased BE and 15 volt supply. This transistor has a larger reverse-BE breakdown voltage than 2N3904 etc however, so might not work so well on 9V supply. You might also want to look at the noise generator in the Boss DR-55 Dr. Rhythm schematic.... operates on 6V supply, uses two transistors, both operating as amplifiers.
Mike

[Brian Marshall]

took a look at teh Dr55... i wish they had the transistosr marked on the schematic... doesnt show a part number, or if they were hand selected or what not.

it seems to me it is amplifying the natural noise of the first transistor... i wonder how true that noise actually is.

[Vsat]

Brian,
Sounds good for cymbals!
The transistor used in the DR-55 noise gen is the 2SC945 (this is mentioned in service manual), same as used everywhere else in the unit.
Zener diodes vary widely in performance as noise generators, one of the sdiy-list guys did a bunch of spectrum analyzer tests and found that largest noise output usually does not go hand-in-hand with "closest-to-textbook" white noise spectrum. Let your ears be the judge. If low voltage operation is a necessity, then amplifying resistor noise may be the simplest approach.... or use a microcontroller. Pseudorandom sequence generator built from CMOS chip takes a lot of parts.
Mike

[Tim Escobedo]

The DR55 style circuit is a good start. I made a nice noise source by using a 2N2222 with the emitter grounded, a 10k collector resistor and a 500k C-B resistor. Amplified by AC coupling the collector to the inverting input of a opamp with a 10M feedback resistor. This setup ran at 6V giving a nice full hiss.

The reverse biased transistor might be a "better" source of noise, but they can get tricky to operate at 9V and below. Old, otherwise useless Ge transistors seem useful for noise applications. I have a box of bad Ge transistors only useful as diodes or noise sources.

[Paul Perry (Frostwave)]

Pseudorandom sequence generator built from CMOS chip takes a lot of parts.

Built from a PIC, it takes about ONE part  Or so they tell me..... (yeah, OK, dependign what you are doing wiht the output you might want a RC filter on it.)

[Vsat]

Might need a 5V regulator for the PIC, unless there is a regulator already present on the board. Some of these uC need only about 10 microamp if you clock them at 32 KHz... might be able to run it directly from the 4.5V fake ground in a 9V unit. BTW that DR-55 circuit needs shielding - it is very sensitive to amplifying stray pulses, etc. With the uC you directly generate a p-p large output swing.
Mike