Just curious. Is there a way to wire a 7805 which guarantees that it will oscillate?

Thinking of using a couple in a random number generating circuit, by clocking counters with two oscillating 7805s through an AND/NAND gate.

Thought since I didn't need a stable clock that this might be simple to implement.

 

Darn, these things will only oscillate when you don't want them to, never when you do.
Guarantee? that's a tall order

If you connect the sense pin through a capacitor, rather than directly or set a trigger to discharge as the sense pin resistor voltage rises can you not create a relaxation oscillator?

 

If you connect the sense pin through a capacitor, rather than directly or set a trigger to discharge as the sense pin resistor voltage rises can you not create a relaxation oscillator?

I don't know, can I?

Which is the sense pin? I thought the pins were input, output and ground.

 

Sense = ground it is the control pin that needs connecting to ground via some circuitry if you wish to modify the action of the regulator.

I would suggest a little experimentation, simply connecting to ground through a cap may be enough.

 

I think it will latch up with just a cap between pin 2 and ground.

 

Just curious. Is there a way to wire a 7805 which guarantees that it will oscillate?

Thinking of using a couple in a random number generating circuit, by clocking counters with two oscillating 7805s through an AND/NAND gate.

Thought since I didn't need a stable clock that this might be simple to implement.

Apparently they can get upset about ultra-low ESR multilayer ceramic chip capacitors to big and/or to close to the device.

There are appnotes devoted to this topic on its own.

 

Use a Schmitt trigger inverter with resistor and capacitor to make a relaxation oscillator.

http://electronics-course.com/schmitt-trigger-oscillator

Bob

 

Regulators are designed to be stable and not oscillate. Why do you want to use a regulator as an oscillator? Just use a 555 in an astable configuration.

 

Hello,

In the datasheet of the LM317 there is a schematic for a switching regulator:

This is the use of a standard regulator as an oscillator.

Bertus

 

Regulators are designed to be stable and not oscillate. Why do you want to use a regulator as an oscillator? Just use a 555 in an astable configuration.

As stated, this question is purely out of curiosity. I imagined that it might be possibly to generate a circuit that oscillated with fewer components than an astable 555 circuit. As mentioned, I am not concerned with an oscillator at a specific frequency; just one that oscillates. And whereby I might have a 7805 in the circuit in a more typical configuration, it would be convenient to use a couple more as oscillators.

 

Going back to your original post, if your goal is to build a random number generator (simple random as in I dunno what's gonna come up... and not as in quantum weirdness unpredictable kind of thing) I think that you would be best suited with combining two square wave oscillators working at very odd frequencies, and then have their outputs go through a logic gate (and, nor, xor, etc...) and then retrieve each bit with an MCU working at another frequency... that way you can generate your number as long as you want (8, 16, 24 bits) etc, and it would almost certainly be random enough for your use...

 

Going back to your original post, if your goal is to build a random number generator (simple random as in I dunno what's gonna come up... and not as in quantum weirdness unpredictable kind of thing) I think that you would be best suited with combining two square wave oscillators working at very odd frequencies, and then have their outputs go through a logic gate (and, nor, xor, etc...) and then retrieve each bit with an MCU working at another frequency... that way you can generate your number as long as you want (8, 16, 24 bits) etc, and it would almost certainly be random enough for your use...

The second paragraph of my OP should imply as much. I'm thinking more like component minimization.

 

Just curious. Is there a way to wire a 7805 which guarantees that it will oscillate?

Thinking of using a couple in a random number generating circuit, by clocking counters with two oscillating 7805s through an AND/NAND gate.

Thought since I didn't need a stable clock that this might be simple to implement.

Noise generators are good in random number generators, there's the logic approach with a shift register and a few XOR gates, another is the white noise generator by almost zenering a reverse biased B/E junction - most published designs I've seen use 2x PP3 battery for 18V.

 

The second paragraph of my OP should imply as much. I'm thinking more like component minimization.

How are you planning to retrieve the generated number? using an MCU? How many bits per number would you like to generate? or do you just want random individual pulses (as in single bit numbers) to be later used in a different circuit?

 

How are you planning to retrieve the generated number? using an MCU? How many bits per number would you like to generate? or do you just want random individual pulses (as in single bit numbers) to be later used in a different circuit?

Thanks for you post. I was thinking along the lines of the latter. Random individual pulses.

 

Thanks for you post. I was thinking along the lines of the latter. Random individual pulses.

Well... if your goal is minimization... I can't think of anything more minimal than this:



I got that image from this link.

If you combine two of them through an nand gate, with different RC values you might approximate what you need. Since this thing works with a fairly wide power window (10 to 15V) you might do away with the 7805 altogether and run its output through a 5V zener...

Edit: on the other hand... as Ian said, a noise generator is no more complicated than what I just proposed...

 

Well... if your goal is minimization... I can't think of anything more minimal than this:

View attachment 76034

QUOTE]

If you've got a *REALLY* fast CPU - you can use a tunnel diode.

 

How about this circuit? It could be optimized and can work at fairly low voltages.
It needs a certain large inductor however or very thin wire, and it produces high voltage, enough to light up a neon bulb. It needs to be started too.

If you use a MCP1640, you can clamp off from the inductor with a resistor, works for clocking.



the 7805 isnt meant to be used as oscillator, while it is possible, you need quite a few external parts.

Another way for "random" noise is to use a CXO Quartz oscillator for the clock input and drive a coil with the controller oscillator. The ratio will be odd, as well the coil will drift over time and at different temperatures, so number sequences will never repeat.

 

I like that basic oscillator.

When I first posts this, it was more like a thought experiment. As you all posted great ideas, my thinking became more focused. Lately, I am honking ( d*m autocorrect) thinking about a specific application - simulating a burning candle. (No, don't send your favorite circuits!).

Excepting mods/consolidation with a commercial LED candle, there are 1,000s of circuits available. (I know, I used to build haunted houses). A simple circuit or two like cmartinez's would probably do the trick. All that would be needed is a simple RTL or DTL AND gate and a driver for the output (LED).

The original thought was an out of the box reaction to 7805s would oscillate and I saw a board with three of them neatly lined up.