my own original oscillator

[krayvonk]

Hey ArthurDent,  what is the mystery component.  Im not experienced enough to know yet.
And loved the vid thanks,  felt all warm and fuzzy inside.

[joeqsmith]

No Joe! Its a go ahead,  I say its equivilent.     I like your new spin on the design.

PPl that just dont want to try things...   

Yeh but the buzzer has a diode in it, it just went in by accident.  The other side is diodeless (double dioded) but I dont think it matters.

I actually want it without the diode as well.

If there is a buzzer, add it to your drawing.  Clear pictures, etc.   I have no problem trying to replicate your circuit but the first thing is I need a VERY clear understanding of what you have built.  As it stands now, I do not have that.   What you have shown in your first schematic shouldn't oscillate.   Then again, I have made a few oscillators that I would have thought wouldn't have as well.

[Alex Eisenhut]

Hey ArthurDent,  what is the mystery component.  Im not experienced enough to know yet.
And loved the vid thanks,  felt all warm and fuzzy inside.

Gunn diode, which behaves just like a neon bulb, just much faster and operates at much lower voltage.

Tunnel diode. Gunn diode is another story.

[krayvonk]

How does it show it doesnt oscillate.   Theres just another way to look at the circuit than the way I was when I designed it.

Heres the issue at hand.


charges capacitor
+ _w__________| |____out__ -


capacitor discharges on release of the positive pole,  battery bypasses capacitor, i dont want this.
                      _______     
                      |           z
+ __w_______|__| |__|__out___ -


extra negative connection  supplied, now current that takes the discharge path will be diverted from the output.
                             -
                             z
                      ____|__     
                     |           z
+ __________|__| |__|__out___ -


so, the issue u were having with me is this is equivilent:
       -                   
       z                     
       |               _______     
       |              |           z
+ __|________|__| |__|__out___ -


Im saying no worries bro you might be right.



And why it works more info:

this is ONE HALF.
  INVERTER WIRE
    -
    |  -
    |  z                     
    |  |               _______     
    |  |              |           z
+ |_|________|__| |__|__out___ -


Then the second half is the first halfs inverter, and vice versa.

So they are inverting each other.


    -
    z
    |------------------ feedback inversion
    |-----------------+ (gutter pole, what i called it,  in the position you put it)
    |------
   ___    |
   ___    |
    |------
    |
    |  -(gutter pole - where u put it.)
    |  z                     
    |  |               _______     
    |  |              |           z
+ |_|________|__| |__|__gotofeedbackinverter___ -



So you can see quite humourously,  that each side is inverting the other.

[alsetalokin4017]

Gaaahhh!  Please don't do that. Just sketch and photograph, like you did before.

Yes, there is a slight element of "hazing" or "jumping in" going on here but what did you expect, really? There are thousands of examples of correctly sketched simple schematics on this site to look at, there is even "DaveCAD" .  So when you come onto a new block you really should at least try to speak the idiom, feel me? Use standard symbols so we can understand what you are communicating, and be ready to put up with a little mockery, as you are learning.

What does your "buzzer" element do if it is just connected directly to the battery?

[joeqsmith]

How does it show it doesnt oscillate.   Theres just another way to look at the circuit than the way I was when I designed it.

Heres the issue at hand.

Based on your original drawing, if you followed along, there is no feedback, only two LEDs in parallel.  Both LEDs will behave the same.  So there is something wrong.

I'm sorry but I do not understand your last drawings or explanation.  Take some clear pictures of your setup and we can start there.

A few people were trying to make some simple oscillators a while back.  G0HZU posted one that was pretty interesting.   I took a crack at it as well.   

https://www.eevblog.com/forum/projects/el-cheapo-oscillator/

 

[krayvonk]

Warning its got my cringey voice in it, but here it actually doesnt osc at first here and you can hear the buzzer by itself.
https://www.youtube.com/watch?v=YHiS8HLVkFs.


When the top row goes off, the bottom row discharges,  when the bottom row goes off, the top row discharges... if you make it discharge quicker than it charges, then you get this action happening.

I know its funny when you take it from a steady state, how does 1 side rule over the other,  but its equally true that it oscillates in this fashion as just stands still.   

[joeqsmith]

Sorry, I can't tell anything further from that video.   Is there a reason you are reluctant to show a decent picture of it or to draw a complete schematic? 

[krayvonk]

 

ah dont worry about it.  I shouldnt be letting my tech get stolen anyway.


Your better off with transistors anyway, who cares.

Didnt even try it...

[ArthurDent]

"Hey ArthurDent,  what is the mystery component.  Im not experienced enough to know yet."

The mystery component is indeed a tunnel diode which doesn't behave like a normal diode or other 2-terminal device. Having a negative resistance region allows it to oscillate without feedback. I had a high end scope years a number of years ago that used a tunnel diode in the trigger circuit. Here's a link to a video that will explain its operation.

[joeqsmith]

ah dont worry about it.  I shouldnt be letting my tech get stolen anyway.


Your better off with transistors anyway, who cares.

Didnt even try it...

Correct, I did not try it.   Mainly because I know the outcome by just looking at your circuit.  It's possible you have something else going on that you don't understand.  This buzzer you mention on the side may be part of it. 

The circuit would be easy enough for you to model in SPICE to help you understand what is going on.  Maybe once you have a better idea of why it oscillates, you can report back and we can try to replicate your findings. 

I'm not seeing the whole "stolen tech".    There is nothing novel about turning on a couple of LEDs, which is what you currently are showing.  If you could come up with something new and wanted to try and patent it, you would need to explain it.   

[krayvonk]

It looks funny how there is a positive right near the negative and it sneakily looks like voltage loss.

When the capacitor is empty it charges through,  only once that pole is blocked does it dump out the back pole, and thats so i dont turn off the other side UNLESS its using the through cap going pole.(*)

the resistor balance 4,2,1,   makes sure the discharge "loop" is never taken over 2&1.  and the 1, through capacitor line will always overscede the 2,  the voltage loss pole.  the loss pole is activated to make sure i dont run around the cap and cause it to still the inversion at a constant power!(*)

the 4 resistor, the heavy one, cant be too heavy because its the discharge speed, it was constantly stressing me out because it might not have worked.

You temporarily had me befuddled when you simplified my topology, and it even made me understand it better in the end thanks!     ur circuit is equivilent i was making a mistake.


(*) same point in different words.

[joeqsmith]

It looks funny how there is a positive right near the negative and it sneakily looks like voltage loss.

When the capacitor is empty it charges through,  only once that pole is blocked does it dump out the back pole, and thats so i dont turn off the other side UNLESS its using the through cap going pole.

the resistor balance 4,2,1,   makes sure the discharge path is never taken over 2&1.  and the 1, through capacitor line will always overscede the 2,  the voltage loss pole.

You temporarily had me befuddled when you simplified my topology, and it even made me understand it better!     ur circuit is equivilent i was making a mistake.

You should use common terms.   Poles and zeros may mean something different to me than you.   The way you use poles and back poles, it means nothing to me. 

There is no discharge path with what you show, shy of removing the power.  The caps will both charge at the same rate and reach a steady state value.   At that time, they will both have the same value.  That much should be obvious.   As a mater of fact, you could calculate all of this easy enough.   

[krayvonk]

Your thinking is correct on that one,  and both caps will fill and will not toggle, it does actually do that sometimes.  and it just turns off, even with the battery connected, it is one of its possible states.  Also probably even blended 50/50 travel happens too,  and pessimisticly you can not bother with it.

There is a way to relay-start it on so one cap starts charged and one starts empty, but funnily it tends to oscillate without it, and u can connect and itll find the imbalance itself.

When I say discharge path, i literally just mean putting the capacitor in a little loop.   and it needs this guard "positive" to make sure it never takes the route of 4 resistance in the loop, over 2 resistance to the guard "positive".   (therefore its variable what hits the opposing "positive" to load thieve from the other side/phase.)  - from this sides negative.  which puts the other cap into discharge - (down its 4 resistor "loop",  and its lucky it wasnt too much resistance to slow its discharge down so slow that it didnt work.)

So effectively the battery is changing which side its going,  one side always has a bigger percentage than the other, and it oscillates.

[joeqsmith]

It would be easy enough to short one cap manually on startup.  The other will reach steady state and when you remove the short across the other, it too will reach steady state.   

If you don't own a decent camera, no problem.  You obviously have a PC of sorts.  Load up Linear Tech's free SPICE.   Enter your schematic.  You can include switches and what not to try and get it to start. See if you can show any case where it will oscillate.  If you do, then just post that circuit and we can go from there. 

If you followed that link to look at some of the oscillators,  an easier one is just using your resistor and one LED.   A bit faster than 1Hz.

[krayvonk]

Controlled shorting yes.

I dont know if it works in your simulator,  but if I coded mine myself It would probably work. 

That Falstad one looks like it could do it,  needs to be real time.

https://www.falstad.com/circuit/ <- i dont know how to do it...

[joeqsmith]

".... and pessimisticly you can not bother with it."

That's a common comment I will see with the uneducated perpetual motion crowd.  I was willing to give you the benefit of the doubt and try to help you understand what you are seeing but comments like this put you in a different bracket.

[krayvonk]

it said wire loop detected?   geeze this doesnt handle junctions very well...

[Rerouter]

I can make it work in a simulator, but not with LED's. as such I would assume It would not work if you replaced the LED's,

It does require the LED component to act like a neon bulb, draw no current until it reaches a certain voltage and then suddenly turn on and drop there voltage, the capacitance then increases the voltage seen by the other element and trips it,

This sounds like the LED vendor may have done something non linear inside the LED to prevent it faintly glowing until enough voltage was present on the input.

Based on your current design, a 5V supply should give about a 55Hz flash rate, however it does require an external stimuli such as the switch in the picture to trip it into action.

A self flashing LED would actually be a good example of a device that could cause this behavior, the switch that controls the LED would only power on once the voltage rises high enough, draw a sudden gulp of current, and increase the voltage across the other element

There may be some very marginal case where the voltage gain of a normal LED may be able to cause this, but not yet been able to simulate this.

To the OP: http://tinyurl.com/yxpezhc3 a falstad circui link for the simulation I have run

Long link below for later visitor:

Code: [Select]

http://www.falstad.com/circuit/circuitjs.html?cct=$+1+0.000005+10.20027730826997+50+5+43%0Ag+64+256+64+272+0%0Ag+-144+144+-160+144+0%0AR+256+144+288+144+0+0+40+5+0+0+0.5%0AR+64+48+64+16+0+0+40+5+0+0+0.5%0A162+64+96+128+96+2+default-led+1+0+0+0.01%0A162+-16+208+64+208+2+default-led+1+0+0+0.01%0Ar+64+96+64+48+0+1000%0Ar+64+208+64+256+0+1000%0Ac+128+96+128+208+0+0.000001+0.2285886398764465%0Ac+-16+208+-16+96+0+0.000001+-0.22858863987272304%0Aw+64+96+-16+96+0%0Aw+64+208+128+208+0%0Ar+128+96+192+96+0+4000%0Ar+-80+208+-16+208+0+4000%0Aw+128+208+192+208+0%0Aw+192+96+192+144+0%0Aw+192+144+192+208+0%0Ar+192+144+256+144+0+2000%0Aw+-16+96+-80+96+0%0Aw+-80+96+-80+144+0%0Aw+-80+144+-80+208+0%0Ar+-80+144+-144+144+0+2000%0Ar+368+208+368+256+0+2000%0Aw+368+160+368+208+0%0Aw+368+96+368+160+0%0Aw+448+96+448+144+0%0Ar+512+96+512+48+0+2000%0Aw+512+144+512+208+0%0Aw+512+96+512+144+0%0Ar+368+208+432+208+0+4000%0Ar+448+96+512+96+0+4000%0Ac+432+160+368+160+0+0.000001+-1.0145511033526469%0Ac+448+144+512+144+0+0.000001+0.3143925517936923%0Ar+512+208+512+256+0+1000%0Ar+368+96+368+48+0+1000%0AR+368+48+368+16+0+0+40+5+0+0+0.5%0AR+512+48+512+16+0+0+40+5+0+0+0.5%0Ag+368+256+368+272+0%0Ag+512+256+512+272+0%0Aw+432+160+432+208+0%0As+432+208+432+272+0+1+true%0Ag+432+272+432+288+0%0A187+432+208+512+208+0+1000+1000000+1.2+0.000009999999999999999%0A187+368+96+448+96+0+1000+1000000+1.2+0.000009999999999999999%0Ao+42+64+0+12290+1.2000880957814164+0.0001+0+2+42+3%0Ao+43+64+0+12298+1.2000880957812254+0.0001+1+2+43+3%0A


Key points:
If the circuit is not triggered, the LED's would be sitting at roughly 210mV each,
the circuit is relying on one LED to start conducting at a certain voltage, draw a chunk of current to discharge the other sides capacitor, then for that LED to fall under a certain voltage and stop conducting
When it stops conducting, the now partially discharged capacitor gives enough of a delta across the other LED to turn it on and form an oscillation

So there is something to it, just not sure what in an LED would give such a sharp transition,

[krayvonk]

Is there any place you can go on the circuit to get the oscillation and get rid of diode?

[Rerouter]

Please draw the correct diagram, I have followed your original diagram as best I can, and played with it until I could come up with something that seemed to make the circuit repeatable.

I made the simulation because I am honestly trying my best to understand how this is working.

Even if you just edit the values or connections on the simulation I have made, please use "File - Export as Link" and post the long link here for the corrected version

[krayvonk]

Well thankyou.   Im too confused already,  the neonbulb blasta was insane.

[bsfeechannel]

I managed to quickly reproduce the krayvonk "oscillator" in my lab. What is causing the oscillation is the buzzer itself.



It works just like a neon oscillator. The buzzer has a region of "negative resistance". It starts buzzing with a voltage that is higher than the voltage with which it stops. So the buzzer draws a very low current while the capacitor charges up. Then the buzzer buzzes and suddenly draws more current. The capacitor discharges through the buzzer until it stops buzzing and the cycle goes again and again.

Mine is a 3v buzzer. But it buzzes with less than that as seen in the schematic.

The frequency of the "oscillator" is dependent on the supply. The LEDs of our clever engineer must be somewhere in series with the current drawn by the buzzer.

Have a good night and stay beautiful.

[Alex Eisenhut]

Great. You stole his tech.

[Ian.M]

Another cause of oscillation in a simple LED circuit with no other active devices, is if one of the LEDs is faulty with a cracked bond to the die forming a thermal oscillator.  As long as the bond wire makes contact when the LED is at room temperature, current can pass and the die and bond wire will warm up, and heat the surrounding epoxy.  If differential thermal expansion opens up the crack, it will break the circuit until the LED has cooled down again.  The result is a faulty LED that when fed through a normal current limiting resistor will flash or blink with a frequency anywhere from a couple of hundred millihertz to several hertz.

You most often see this effect in one of an array of paralleled cheap LEDs that have been mechanically abused e.g. by bending and/or twisting their leads too close to the encapsulation, e.g the typical construction of dollar store LED torches.  However its far more common for the damage to simply result in a dead, open circuit LED rather than a self-flashing one.  It also tends to be fairly sensitive to further abuse - straightening or re-bending the leads tends to change the flashing fault to totally dead, but its vanishingly rare for the reverse to happen.