Why can't I breadboard this oscillator?

[Lightium]

I hate communicating.

I've simulated a nice circuit for my purposes and find I can't breadboard it. I've checked continuity on all connections before powering, but I've failed to get it running.

Any thoughts?

 

[JimB]

Any thoughts?

Yes,

There is no decoupling capacitor from the +ve supply to the 0v (Ground) line. Try a 10nF.

At a quick and dirty calculation, the oscillator frequency appears to be of the order of 2MHz.
Some breadboards can be a bit funny at high frequencies, although at 2MHz I think that you may "get away with it".
What frequency does the simulation say?

The simulation is saying 20kV on the secondary of the transformer, that is a bit much for any plug-in breadboard with 0.1" contact strips.
For 0.1" plug-in breadboards, 50vV would be a sensible maximum voltage.

I suggest that you build your circuit in an "ugly" or "Manhatten" style on a piece of circuit board with a continuous copper ground plane.

JimB

 

[Lightium]

It's 2MHz pulses at 38KHz. Shoot, I didn't try a decoupler.

I'll try it in the morning, thank you.

 

[AnalogKid]

I hate communicating.

In that context - which we know *nothing* about - what is it you are trying to achieve? (both practically and electronically)

ak

 

[Tony Stewart]

Where did you come up with this design?

What made you think you could pulse such a high reverse voltage in both Vbe's? ( generally rated for Vbe=-5V max)

LTspice I believe does not simulate component failures from voltages or currents or power. Nor is it psychic to see your layout for stray wire inductance, and capacitance.

Anything less than 20 MHz is trivial on a breadboard . I measured ~35 ns risetime or ~ 0.35/0.035us = 100 MHz with no sparkplug ( wrong lib ref.) This means parasitic wire length and mutual coupling is more significant to be included in simulation and why 3D Manhatten style can work to reduce ESL ~ 0.8nH/mm

So either you have a wiring error or Vbe is broken (diode test) or parasitics with feedback are suppressing ( NFB) rather than reinforcing (PFB) oscillations or mutual and C coupling

 

[schmitt trigger]

I seriously doubt that a blocking-type oscillator would actually achieve a 2 Mhz output.
Blocking oscillators depend on the transformer core’s characteristics; namely saturation flux density.
And LTSpice, to my knowledge, doesn’t have a means to model saturation.

 

[Lightium]

I am trying to make use of these 15KV coils on ebay.

Sure the circuit is stressed, but it should still work for a time.

As my last-ditch effort I put in some 1N750 zeners from the bases to ground and reduced the inductance of the primary to 80u with resistance of 410m ohms.

Haven't tried to breadboard yet, but I should make a circuit on copper board.

 

[schmitt trigger]

The website you linked shows the suggested drive circuit, and it is indeed a blocking oscillator. The core’s saturation characteristics play a role in its operation.

Why don’t you start with the provided circuit, rather than attempting to re-invent the wheel with a modified circuit that doesn’t include the most important characteristics?

 

[Lightium]

I don't like the provided circuit.

Many improvements of the wheel have accured since the stone age.

 

[Nigel Goodwin]

I don't like the provided circuit.

But presumably it works?.

Many improvements of the wheel have accured since the stone age.

A square wheel doesn't, make the original circuit, then (assuming all is well) try and make your own driver.

 

[schmitt trigger]

Nigel beat me to a similar response. You don’t improve the wheel by making it square.

If you don’t like the original circuit, and the truth is that it is very crude and not very efficient, at least attempt to design a circuit understanding the essential theory behind it.

I have already mentioned a couple of times that core saturation characteristics play an essential role in a blocking oscillator. You asked for help. But the response I am only getting from you is pushing back on the fact.

 

[ZipZapOuch]

But presumably it works?.



A square wheel doesn't, make the original circuit, then (assuming all is well) try and make your own driver.

A square wheel works if you're driving on the right surface.

Or, if you have a well designed square...

This Square-Wheeled Bicycle Works Like a Pedal-Powered Tank

Is it a practical design? Nope. Does it work? Absolutely.

gizmodo.com