Self oscillating voltage booster from DC

[TheBaconWizard]

Good afternoon electro-peeps.

Complete noob here and trying to learn.

I have found a jellybean-parts relaxation oscillator circuit, and I would like to alter it to suit my purposes. I'm hoping someone could check my idea over and let me know if I have completely misunderstood.

Here is the original circuit.



In this example, the capacitor voltage rises until it and the DC supply have a combined voltage greater than the break-down value of the transistor. In the working examples of this circuit that I've seen, there was no gain in voltage delivered to the LED.

The LED is switched on, however as the capacitor discharges, the voltage falls below the transistor's breakdown once again and so the current no longer flows through that path. Rinse and repeat.

I am proposing to make the following changes.

1) Replace the transistor with a zener diode.
2) Have a zener diode of a value much higher than the DC supply, and a capacitor with a voltage rating at least equal to the DC supply, such that the voltage rises like in a voltage-doubler and is finally released when the zener permits.
3) to facilitate this, I'd insert a blocking diode between the DC supply and that first resistor.
4) No LED

My aim is to take a 130v smooth DC supply that's already in my device and end up with a pulsed DC voltage of 200v (peak) or greater to run an electroluminscent display. I'm guessing I could have the display in parallel after the zener, because electroluminescent displays act as a capacitor and only work on changing voltages.

Obviously, I'd have to change the resistor and capacitor values to get the pulse to the correct Hz, but that's easy enough.

Assuming I do my maths right, does the principle work? And I get a voltage booster and oscillator all-in-one?

Thanks for any help in advance 

[T3sl4co1l]

At worst, you want a one of these:



Put a diode from collector to capacitor to ground, for a rectified output.  Vary the base bias current to control power output.

This can only get you Vce(max), and I wouldn't recommend trying for more than triple the supply.  To get more, put a third winding on the transformer, and use the turns ratio to your advantage.

You can get still more ratio by tacking on a multiplying rectifier, but this is undesirable for the transistor, giving poor efficiency and capacity.  If all you need is a mA or two, it's fine.

Built with another transistor instead of the feedback winding, and a tapped transformer, and with the control/feedback in place, it looks something like this:



But really, you do have to be rather specific about what current you need...

Also, EL displays run on AC, not DC, so all of this is nonsense anyway, which, I guess you already knew, so, why..?!!?

Tim

[TheBaconWizard]

T3sl4co1l I think you have misread my post.

I already HAVE a 130v DC supply. And the oscillator circuit shown, already DOES work.

I want the relaxation oscillator to increase the voltage again via voltage-pumping (achieved by using bigger voltage-value components) at the same time as it provides me with a nice pulse, and wanted to know if that will work.

Also, if it makes sense to replace the transistor in the oscillator with a nice, fat zener.

EL works just fine on pulsed DC, hence why I am looking at an oscillator. it doesn't have to be true AC, just moving. It's because it acts as a capacitor which I also did say in my post.

My aim is to take a 130v smooth DC supply that's already in my device and end up with a pulsed DC voltage of 200v (peak) or greater to run an electroluminscent display. I'm guessing I could have the display in parallel after the zener, because electroluminescent displays act as a capacitor and only work on changing voltages.

For the sake of clarity I have re-drawn the circuit as I would intend it to be.
[EDIT: Bugger I put 120v supply instead of 130, but I don't think it matters too much]
I don't know the schematic for EL, so I have represented it with a capacitor giving off light. So the question is simply: Will this give me a roughly 200v peak-to-peak pulsing DC at the EL unit?

According to me, the values are correct for 200HZ but that's not my main concern here.



Blueskull, will investigate that chip, but I the only chip I've ever used so far is a 555. i suspect it's going to need a whole host of resistors and caps and stuff to operate it anyway, and I'm aiming for small footprint and low parts count. Still, if it simply plugs into the supply and outputs what I want with just a resistor or 2 to control it, ideal!

[T3sl4co1l]

Huh.. okay...

Well anyway, that a transistor can exhibit latching breakdown, is quite an unusual feature.  Not all transistors will do it, and I don't know of any two-layer (i.e., diode) device that does it.  (I'm not counting tunnel diodes, as that's a low voltage zener breakdown.)  All you'll get from the circuit shown is more DC.

Have you considered a blocking oscillator (reiterating my above post, adjusted for the supply voltage I guess) or resonant oscillator?

Tim

[Richard Head]

Use a PUT (Programmable unijunction transistor). If they are still available! 

[TheBaconWizard]

Huh.. okay...

Well anyway, that a transistor can exhibit latching breakdown, is quite an unusual feature.

I'm sure you're right, but there is only a transistor in the original circuit: the one that already works just fine.
As you can see, in MY circuit there is no transistor at all. Instead, a zener diode with a breakdown voltage of 200v.
I was not aware that the transistor in the original circuit was latching, especially... have I missed something?

All you'll get from the circuit shown is more DC.

Yes. But will it be PULSING DC? And will it be at 200v Peak?

My theory is the same as the original circuit (I can show you the circuit working if you like, pulsing the LED)

1) nice flat DC does nothing to the EL display on its own, but the charging of the 1nF cap provides the upward curve of our output waveform (going from 130v up to eventually 260v if it were allowed-to )

2) However, upon reaching 200v, the zener breaks-down and pulls the EL down to ground, providing the downward curve (spike, really) of our output waveform.

3) It starts again. ie, we have a 200v peak sawtooth wave, from 130v flat DC supply.

I guess I could put the EL display in series with the zener diode, but I don't see the need.

I am specifically avoiding transformers or anything that physically large while still using jellybean parts, with the lowest possible part-count. Simple is king for my purposes. I don't care that the output wave is not true AC or sinusoidal, although I'd prefer not to have a square-wave as the initial jump in voltage shortens the life of EL dramatically. Saw tooth is fine.

[TheBaconWizard]

Use a PUT (Programmable unijunction transistor). If they are still available!

Yeah, my thinking is that zener diodes are easy to find and cheap and take up marginally less room. IF I am right that it'll do the job.

[TheBaconWizard]

Probably because minority carriers in CB junction can not be pulled away fast enough, then open base effect caused CE to turn on again when external voltage recharges fast enough?
For a PN diode, there is no open base effect. Open base effect is such a PITA that in BJT applications where base access is limited, such as in an IGBT, it decreases breakdown voltage by a lot.

So, are you saying that the zener diode in my circuit will be back to its OFF state too quickly to allow the cap to discharge to any useful degree? I'll definitely go with a PUT in that case.

[TheBaconWizard]

So, are you saying that the zener diode in my circuit will be back to its OFF state too quickly to allow the cap to discharge to any useful degree? I'll definitely go with a PUT in that case.

If it works, more likely your EL is acting like a gas discharge tube. The zener is only there as a protection device.
Even if the zener oscillator work, it will be super inefficient because recovery charge of a PN diode in switching application is horrible, especially when you are relying upon it to oscillate.

Um, I didn't put the zener there for protection at all..  It's there to bring the EL back down to a lower voltage and at least partially discharge the capacitor, in place of the transistor that was in the first circuit shown (or indeed a PUT)

I'm afraid I need "recovery charge" explaining though..

[T3sl4co1l]

Yeah, diodes don't oscillate.  There is no lag between rising and falling voltage, and rising and falling current.  Zeners are an excellent way to crudely regulate (stabilize) a DC voltage.

A four-layer (thyristor) type device can be used as a relaxation oscillator.  These have two terminals (diac), or three (SCR, triac), and can be turned on at a fixed or arbitrary voltage (up to some kV for the highest rated devices), and maintain "on" at quite low voltages (1-4V) until turned off.

If anything, you're looking for a neon bulb, which does "turn on" suddenly.  However, the voltages are fixed by the device, and you'd get a sawtooth waveform from maybe 20V to 70V, not 200V peak.

Tim

[TheBaconWizard]

Ok, that explains it then, thanks.

[Richard Head]

Yeah, diodes don't oscillate.

Tunnel diodes do.
Call me pedantic.

[Seekonk]

Go on ebay and search  SELF OSCILLATING HALF BRIDGE.  Think of it as a 500V 555.

[Audioguru]

Just think about all the millions of circuits that use a zener diode as a voltage regulator with a smooth output. If a zener diode oscillated like a reverse-biased transistor emitter-base junction then all those zener diode regulators would not work and your idea also will not work.

[TheBaconWizard]

Just think about all the millions of circuits that use a zener diode as a voltage regulator with a smooth output. If a zener diode oscillated like a reverse-biased transistor emitter-base junction then all those zener diode regulators would not work and your idea also will not work.

This is the beginners forum and I made my noobness known in my original post. What makes you think I know about the millions of circuits that use a zener diode as a regulator?