RF Broadband Noise Generator: Using wrong Zener diode?

[abitofhelp]

Thank you for all of your contributions and discussions, everyone!  With your information and some time in the lab, I learned a lot about constant current regulators and some common "greenhorn" mistakes, to which I fell prey.  PCB.Wiz's comment about how the ground and tab are used on the LM7805 was spot-on!  I assumed it was named gnd, so it needed to be attached to my ground plane on the PCB.  WRONG!  I switched my studying and research from the LM7805 to the LM317 because I found much more information about how it worked and how to use it as a constant current source.  However, I walked into another "greenhorn" trap...  The order and meaning of the pins on the LM7805 and LM317 are not the same.  Further, the symbolic block used for the LM7805 shows the pins exactly as they are on the chip.  However, the symbolic block for the LM317 does not align with the pins on the chip.  Arggg...  Ultimately, these two issues were the culprits of my implementation of the circuit not working.  I need to update the schematic to match my final version, because I decided to use the LM317 rather than the LM7805 and made a few minor tweaks.  I will post it here in case it can help others. 



Thanks again, everyone, for your help!

[SteveThackery]

Does anyone have an idea why the author of the circuit boosts 12V to 34V and then drops it to 5V?  I cannot see a benefit when the 7805 is a current regulator…

This has been answered a couple or so times, now, but I just wanted to respectfully ask: are you absolutely clear in your understanding?  The 7805 is wired as a constant current source, not as a voltage regulator.  In simple terms, the 7805 will increase its output current until it sees 5V between its output terminal and its reference terminal.   At that point it is "happy" and will maintain that constant current into whatever load happens to be connected to it (within limits).  Please study the diagram.

When I first discovered the idea of using a voltage regulator chip as a constant current source I thought it was very cool and clever. 

[PGPG]

I have never been using Zeners as noise source but I think, I understand this circuit.

Does anyone have an idea why the author of the circuit boosts 12V to 34V and then drops it to 5V?

He doesn't.

I cannot see a benefit when the 7805 is a current regulator…

Current source (ideal) has infinite internal resistance so no loads the noise sorce. If you use resistor to limit current in Zener diode than in small signal circuit model you get that resistor being in parallel to Zener diode so the (probably high impedance) noise sorce model is loaded with this resistor what circuit author wanted to avoid.

The current limiter is a bit uncommon (and power-wasting) with the 7805, but not a problem in itself.

I suppose that the main circuit author goal was to get as high internal impedance current source as possible. I suppose he have tested more common solutions like transistor current source or LM317 current source and may be found that 7805 current sorce have better characteristic. Fo single transistor current source the problem is Miller effect making Ccb being multiplied by something close to beta  and such 'huge' capacitance is seen as parallel to current source making this source being current at DC and relatively small frequencies but not at higher frequencies. Probably 7805 in current source configuration has smaller capacitance parallel to it. Even so at some high frequency this capacitance becomes being 'visible' so to increase the current source impedance at high frequency the ferrite beads were used.
I don't know (and don't want to search it) what MAR6, MAR3 and ERA5 are but look at their loads. In each the ferrite beads were used to increase the load impedance at high frequency. Unfortunately it is not more specified at schematic. As it is old schematic and I remember at the beginning of 2000 years SMD ferrite beads were rather expensive so may be all these are simply small ferrite beads put on wire so having about 50 ohm at 100MHz. In my circuits I am typically using 1k ferrite beads so one far away from the other. May be it is clear from the reast of publication what are these ferrite beads parameters.

Iq varies with temperature and voltage drop, so 7805 is not a great current source.

I think this is the last important parameter here.

[MathWizard]

see many refs on noise diodes and noise gens.


Instead of TRNG, try a PRNG Pseudo random like 

https://www.sound-au.com/project182.htm

Readly made but rare:

  National Semiconductor   MM5837 - an 8-pin PMOS IC that was a complete digital noise source.

https://www.alldatasheet.com/datasheet-pdf/pdf/9284/NSC/MM5837.html

Jon
 
See

I might try something like the project182. But I also have MCU's. So yeah I wonder what's possible with them, if I bare metal some PRNG code. I'm sure there's a bunch of eqn's for all that sort of stuff, and how to output some digital signal into some analog drive circuit, w/o DAC's.

[Grandchuck]

I found some 24 volt Zeners and had a go.  Used a Rigol DP832A and varied the voltage across the diode in series with a 3k ohm resistor.  The best current was around 80 uA (most noise).  Two screen shots below.  The scope was ac coupled and connected across the diode.

Several diodes were tried and all about the same.

The marker at 1 is an FM broadcast station.