"Smoking Gun" - finally!

[gyulasun]

Hi Paul,

There was a thread here in 2005 active for 2-3 months dealing with trying to utilize the Johnson noise in diodes and among the contributors to that thread there were some members with diode manufacturing backgrounds so it would seem worthwhile to look for them.

The title of the thread is Diode Array, many small parallel diodes to aggregate rectified Johnson Noise   and here is the link: http://www.overunity.com/index.php?topic=26.0

The topic is two pages long only and has not received any contribution since 2005, wonder what happened to Charles Brown's idea since then.

rgds,  Gyula

[Mk1]

If you like diode use 100 led in series use the clear casing one red light work good , white light don't know.you will have 120 volt if you put it in the sun light.A single of those gives 1.2 to 1.5 volts,this is good but 6 micro volts is still not much.Cheap solar cell but difficult making sure getting the most light in but maybe using flat ones and glue to mirror . There are uv led and also infrared so test could be made for different light spectrum .If i put my led in series it raises the amps but getting decent current will make it much more costly and regular solar panel cheaper.

[PaulLowrance]

I occasionally talk to Charles. He's more of a thinker. Thinking is fun, but in the end someone has to do something. 

PL

[Goat]

Hi All

Thanks for opening the thread PaulLowrance interesting subject.

I've been following this thread and noticed that Tom Schum did some work on the Jhonson noise effect while surfing the web, seeing he hadn't been mentioned I thought maybe some people would like to see his work and results from last year.

Array: http://mysite.verizon.net/vzesfls5/sitebuildercontent/sitebuilderfiles/array_mounted_e.jpg
Schematics:  http://mysite.verizon.net/vzesfls5/sitebuildercontent/sitebuilderfiles/sms7621_array_sch.pdf
Test results: http://mysite.verizon.net/vzesfls5/sitebuildercontent/sitebuilderfiles/7630_data.pdf

Regards,
Paul

[PaulLowrance]

Hi,

I have Tom Schum's diode array in my lab. I still have not tested it, yet. One thing I don't like about his diode array is that it's not 100% in-series. As far as my understanding of the semiconductor mathematics, paralleling diodes increases the junction capacitance. So in Tom's case, he has 32 diodes per row in-parallel, which equals 0.5pF * 32 = 16pF. The total Johnson noise is equal to the kTC noise, which equals sqrt(k*T/C). So the kTC noise is sqrt(16) = 4 times less than a single 1N34A diode, which means the DC voltage (due to rectification) is 4^2 = 16 times less-- diode square law.

My recommendations are -->

* If one is going to test for voltage, then place all of the diodes in-series.

* If one is going to test for current, then place *groups* in-parallel. Each group should consist of at least 5 diodes in-series. The reason for not placing diodes in direct parallel connection is because it increases the junction capacitance.


If anyone's interested in the mathematics, you can find it at my website -->

http://greenselfreliantenergy.com/physics/dirtydetails/

PL