Dedicated random number generators

[DrDu]

I was looking out for dedicated true random number generators and was surprised that there seem to be only very few around on the market. Anyone has experience here?

 

[chiro]

There is a site www.random.org and the site describes the method to get the numbers by using interference patterns on antennae (electromagnetic signals) to get the data which is obtained by downloading the various data types (text, binary, etc).

I would suggest that you download some of the documents on that site or if you are really keen, do a similar kind of thing and build a device that captures some sort of physical phenomena and use the random-ness apparent at small scales to capture 1's and 0's and save them to a data file.

 

[DevilsAvocado]

The cheapest https://en.wikipedia.org/wiki/Comparison_of_hardware_random_number_generators number generator is $250, but true randomness from atmospheric noise + Raspberry Pi will get you a pretty decent random number generator for $25.

(I.e. a ‘development’ of chiro’s first suggestion ;)

 

[DrDu]

I had expected there to be a chip worth some cents which uses some noise source to generate random numbers. I wonder why they are so scarce and expensive?

 

[DevilsAvocado]

I had expected there to be a chip worth some cents which uses some noise source to generate random numbers. I wonder why they are so scarce and expensive?

Development in progress, if you have Intel 3rd Generation Core Processor and Ivy Bridge you will have access to a new on-chip random number generator codenamed Bull Mountain that uses an on-processor entropy source (let’s hope it’s not Bull! ).

 

[DevilsAvocado]

P.S: Good article on Ivy Bridge Random Number Generator.

The entropy source behavior is determined by unpredictable thermal noise that seems to do a proper job.

 

[Bill Simpson]

Hardware random number generator

http://www.araneus.fi/products-alea-eng.html

I've always wanted to see someone make a cheap easy hardware random number generator project from the decay in a smoke detector, without compromising the detector to the point where people run away scared.

 

[DrDu]

159 Euros for a noise amplifier!

 

[AlephZero]

159 Euros for a noise amplifier!

You can make a cheap and cheerful analog white noise generator much cheaper than that. All you need is a reverse biased semiconductor junction operating like a Zener diode, plus an amplifier. E.g. http://www.eeweb.com/blog/extreme_circuits/simple-white-noise-generator

But if you want "true randomness", the hard part is making the circuit immune from any external influences - e.g. periodic fluctuations on the power supply rails etc. That's what costs real money.

Apart from cryptographic security they are non necessarily very useful either. For software simulations, not being able to repeat a run exactly (for debugging, code performance optimization, verifying porting to different computer hardware, etc) can be a big DISadvantage!

 

[Bill Simpson]

All you need is a reverse biased semiconductor junction operating like a Zener diode, plus an amplifier

Plus the analog to digital conversion circuitry to convert to a digital stream.

Apart from cryptographic security they are non necessarily very useful either

You could just save the digital stream to a file and replay it if you need that.

And on that 159 Euro price, if the device was like a cell phone which costs far more to manufacture, but has 6 billion customers willing to pay $40 or $80/month forever to be allowed to use it, then I imagine they would be paying you to to take their random number generator hardware in exchange for your putting your name on the two year service contract.

 

[DrDu]

But if you want "true randomness", the hard part is making the circuit immune from any external influences - e.g. periodic fluctuations on the power supply rails etc. That's what costs real money.

I don't doubt that it is not entirely trivial. But the electronics in even the smallest microcontrollers isn't either, yet they cost at best a handfull of euros.
The question is probably why there is so small a market for this kind of technology.

 

[DevilsAvocado]

The question is probably why there is so small a market for this kind of technology.

That’s obvious, isn’t it? X billion people who love to fiddle around with their xPad + FB vs. the very few who know anything about true randomness (and QM) = tiny market = low sale = higher charge.

What exactly do you want to do? And why is free random.org not good enough?

http://www.random.org/integers/

 

[Zarqon]

If you're actually interested in true random numbers I suggest you check out the quantum random number generators (e.g. http://www.idquantique.com/random-number-generators/products.html ), which are based on letting single photons pass through beamsplitters. I'm not sure about the price, but they are dedicated, with both PCI and USB versions, and are obviously one of the few real true RNG's.

 

[Bill Simpson]

http://www.idquantique.com/random-number-generators/ordering/online-shop.html

1305 Euros, 990 Euros, 1135 Euros and 2230 Euros, depending on which model! That is expensive even for my tastes.

I'm still hoping for a hobby project modified smoke detector (without compromising the package to keep from freaking people out about safety sillyness) true random number generator that would cost about 20 U.S. All that is needed is to detect each spike in current from the alpha particle decay.