Two mosfet oscillator

[Groundloop]

@All,

I think the attached oscillator belongs here. The oscillator runs at a very low input voltage.
I have tested the oscillator down to 1,15 Volt and it still runs. The negative mosfet bias
can be from a battery, or an extra coil when the circuit runs. If you use a capacitor instead
of a battery then you will need to pre-charge the capacitor to get the oscillation going.
I use a current limiting resistor on my circuit. The lower the resistance the more power
are you injecting into the circuit. And as a result, harder switching of the mosfet and more
current through the drive coil. This is kind of cool, because you can control how much
current you want to switch by the bias voltage (and bias current).

GL.

[poynt99]

GL.

You may want to try this. Your original circuit will not be able to deliver any appreciable power to a load.

You need stray inductance in the gate, and a good AC return path to make this form oscillate. Hence the 10" of wire, and the capacitor. No guarantee this will oscillate, as I have not simulated your circuit. But you will see that this is getting in line with the circuit I posted before that I will be building for my tests.

.99

[Groundloop]

GL.

You may want to try this. Your original circuit will not be able to deliver any appreciable power to a load.

You need stray inductance in the gate, and a good AC return path to make this form oscillate. Hence the 10" of wire, and the capacitor. No guarantee this will oscillate, as I have not simulated your circuit. But you will see that this is getting in line with the circuit I posted before that I will be building for my tests.

.99

.99,

You circuit "B-Modified" is 100% the same as my  circuit, except that you want to use positive adjustable
bias instead of my negative fixed bias. (Only difference is the added capacitor and labels on the transistors.)

I tested with positive bias and the circuit did not run. So you will need both mosfets and negative bias to get
this circuit up and running.

GL.

[poynt99]

GL,

I don't understand why you say I want to use positive bias.

I have not changed your circuit regarding the bias and MOSFET connections. In your circuit, Q1 receives a positive bias, and Q2 receives a negative bias. My redraw of your circuit shows the same. Agreed?

[poynt99]

The one difference in our drawings is in part B, I show the 12V negative lead commoned with the 9V battery. Your circuit has the 12V negative lead connected as shown in part A of my drawing.

[Groundloop]

The one difference in our drawings is in part B, I show the 12V negative lead commoned with the 9V battery. Your circuit has the 12V negative lead connected as shown in part A of my drawing.

.99

I have - on my input battery common with + on my bias battery.

You have both - common in your circuit. That means that you
get positive bias on your Q1 and negative bias on your Q2.
Now since your Q1 is connected with source to ground then
it will conduct current all the time as long as the voltage is
above the gate trigger voltage. Your circuit will NOT oscillate.

In my circuit Q1 gets a positive bias because the source is
negative compared to the gate. But the circuit will not conduct
any current to ground because Q1 is NOT connected to ground.

Instead Q1 will conduct a small current through the L1 coil and from the battery.
Now because Q1 is conducting then Q1 will provide a small voltage to the gate of Q2.
This will make Q2 conduct current and the oscillation starts.

GL.

[MileHigh]

Groundloop:

So do you see what I said now when you look at Poynt's very simplified drawing of your setup?

The 9-volt battery and the 12-volt battery are in series and their voltages add.  That means the 9-volt battery is also responsible for powering the circuit and power supplied by the 9-volt battery energizes the main coil.

It's exactly the same same case in the RA circuit for her function generator when it is outputting -5 volts.  The -5 volts is added to the voltage provided by the battery bank and helps power the circuit.

Going back to your original circuit and the test with the large supercapacitor, there is an issue to review again.  If the capacitor contributes 1.8 volts and the 9-volt battery contributes 9 volts, then you have 10.8 volts available to power the oscillation and energize the main coil.  That means that the 9-volt battery is supplying much more energy than the capacitor to run the circuit.  That means that the 9-volt battery is supplying most of the energy that couples through to the secondary + FWBR in your circuit.  Therefore that means that the 9-volt battery was supplying most of the energy to recharge the supercapacitor while the circuit ran.

The conclusion is that although the supercapacitor lost voltage very slowly while you ran the test, the test was invalid.  You wanted to see if the capacitor might recharge.  The truth is that the capacitor may actually have recharged, but not because of some alleged RA affect, it could have recharged because the 9-volt battery was recharging it.

A final comment about schematic diagrams.  It can be very confusing to look at some schematics because you can't track what is going on with the voltage potentials.  A way to make schematics much clearer is to use the concept of voltage to help you.   Make a solid line at the across the bottom of the schematic that shows the ground connection.  Make a solid line across the top of the schematic that shows the +12-volt connection.  Then arrange your components between the the +12-volt line and the ground line so that you can "see" how voltage is dropping from top to bottom in the schematic.  Go take a second look at Poynt's schematics and you will see that he does this.   If you have ground, +12 volts, and -12 volts, then make three lines across your schematic diagram.

Finally, a strong suggestion bordering on a warning:  This is not directed at Groundloop, it is directed at all experimenters:  Don't show scope traces without showing EXACTLY where the scope ground and signal leads are connected to the schematic diagram.  Just showing a picture of a waveform or a YouTube clip of a waveform without explaining exactly where the two scope connections are made is pure bullshit.

MileHigh

[Groundloop]

Groundloop:

So do you see what I said now when you look at Poynt's very simplified drawing of your setup?

The 9-volt battery and the 12-volt battery are in series and their voltages add.  That means the 9-volt battery is also responsible for powering the circuit and power supplied by the 9-volt battery energizes the main coil.

It's exactly the same same case in the RA circuit where her function generator when it is outputting -5 volts.  The -5 volts is added to the voltage provided by the battery bank and helps power the circuit.

Going back to your original circuit and the test with the large supercapacitor, there is an issue to review again.  If the capacitor contributes 1.8 volts and the 9-volt battery contributes 9 volts, then you have 10.8 volts available to power the oscillation and energize the main coil.  That means that the 9-volt battery is supplying much more energy than the capacitor to run the circuit.  That means that the 9-volt battery is supplying most of the energy that couples through to the secondary + FWBR in your circuit.  Therefore that means that the 9-volt battery was supplying most of the energy to recharge the supercapacitor while the circuit ran.

The conclusion is that although the supercapacitor lost voltage very slowly while you ran the test, the test was invalid.  You wanted to see if the capacitor might recharge.  The truth is that the capacitor may actually have recharged, but not because of some alleged RA affect, it could have recharged because the 9-volt battery was recharging it.

A final comment about schematic diagrams.  It can be very confusing to look at some schematics because you can't track what is going on with the voltage potentials.  A way to make schematics much clearer is to use the concept of voltage to help you.   Make a solid line at the across the bottom of the schematic that shows the ground connection.  Make a solid line across the top of the schematic that shows the +12-volt connection.  Then arrange your components between the the +12-volt line and the ground line so that you can "see" how voltage is dropping from top to bottom in the schematic.  Go take a second look at Poynt's schematics and you will see that he does this.   If you have ground, +12 volts, and -12 volts, then make three lines across your schematic diagram.

Finally, a strong suggestion bordering on a warning:  This is not directed at Groundloop, it is directed at all experimenters:  Don't show scope traces without showing EXACTLY where the scope ground and signal leads are connected to the schematic diagram.  Just showing a picture of a waveform or a YouTube clip of a waveform without explaining exactly where the two scope connections are made is pure bullshit.

MileHigh

MileHigh,

I have a headache because working and testing on this circuit the whole day now, so I'm in a bad mood right now.
And that is why I'm going to make this post very short.

GL.

[poynt99]

Instead Q1 will conduct a small current through the L1 coil and from the battery.
Now because Q1 is conducting then Q1 will provide a small voltage to the gate of Q2.
This will make Q2 conduct current and the oscillation starts.

Have you disconnected Q2 while the circuit is oscillating to verify that?

If it does stop, and it is indeed operating this way, then that would surely be a unique oscillator.  

[MileHigh]

Don't worry about that Groundloop.  Anybody that has worked on a circuit for a whole day knows how it can make you emotionally and mentally exhausted.  It has happened to me many times.  The best thing is to forget about it for a few hours and get a good sleep and start fresh another day.  - MileHigh

[poynt99]

GL,

If your circuit works that way, then consider the attached drawing.

First, I've redrawn your circuit yet again, but this time with emphasis on Q2 (it drives the load). You may notice that the 9V is simply turning Q1 partially (or fully) ON so as to channel V_D to Q2's Gate.

In the second part of the drawing, I've attempted to simplify the circuit by eliminating Q1, and providing bias to Q2 through Rbias. It would be interesting to see if that would still work.

I would also be interested to see if the first circuit still works with wire "p" removed. I believe it may be unnecessary. There is a balancing act between the negative Q2 bias applied by the 9V, and a positive bias applied by Q1. It's just a question of whether the negative bias is even required, because we know that Q2 ultimately needs a positive bias to operate.

.99

[Groundloop]

.99, MileHigh,

I had my 6 hours beauty sleep now, and it helped on my mood, but I'm still ugly as hell. :-)

I will have time to research this oscillator in approx. 11 hours from now.

.99, I will test your 1 transistor theory this afternoon. I will also see if I can get some good o-scope shots
and measurement for confirming our theories about the way this oscillator works. I have a theory that
the negative bias on my Q1 makes this mosfet behave more like p-type than a n-type. What puzzles me
is that the voltage over the L1 coil looks like only the top part of a sinus signal put together. So it looks
like we operate the switching in the linear area of the mosfet and not as an hard switch on/off. It may be
that RA accidentally has discovered a new and unique oscillator.

MileHigh,

I agree with some of your points in your long post above but disagree on others. Let us just play with
this oscillator and see what we find. My measurements and tests shows me that there is more to this
oscillator than just two batteries adding up in the load. I agree that we inject power from the bias battery,
but in my version of the oscillator, the power injected is only approx. 8 milli Watt at maximum, so this is
not enough to power the circuit. It is barley enough to charge/discharge the gate capacitance of the mosfet
when it conduct current in the linear region, because I see no evidence of hard switching on and off in this circuit.
Just before bedtime yesterday I did a little test to see if I could remove the bias battery and just power the
bias from my L3 coil. The test was partly a success because I got a initial oscillation but it quickly died out
just like a dampened sinus. Maybe we can get this oscillator to run with just one mosfet, I do not know at
this time.

I'll be back...................

GL.

[BEP]

This is kind of cool, because you can control how much
current you want to switch by the bias voltage (and bias current).

Cascode Voltage Oscillator? This circuit (almost the same) was used in the tube days as a VCO.

It reminds me of the little three LED, resistance only oscillator. --  The one used to simulate the alternating red lights at a railroad crossing....

[poynt99]

So it looks like we operate the switching in the linear area of the mosfet and not as an hard switch on/off.

I've been saying this for a long time, but it doesn't seem to sink in. 

[BEP]

I've been saying this for a long time, but it doesn't seem to sink in.  

Hence, the 'astable chaotic avalanche oscillator'. 

1. an oscillator having multiple unstable states near and through the avalanche range of the switching devices.

2. a noise generator with an output so bad it takes more than the average technical ability to understand it or accurately measure the output - if you can get it to work.

3. the circuit shouldn't use MosFETs.

4. It makes more sense if you consider Q2 as the first stage.

Does RA claim she invented this mess? The circuit has been around since the 20's but using the wrong switches doesn't make it new.

I thought cascode circuits were pretty much forgotten and out of vogue. They can be hard to understand.

 

[Groundloop]

Hence, the 'astable chaotic avalanche oscillator'. 

1. an oscillator having multiple unstable states near and through the avalanche range of the switching devices.

2. a noise generator with an output so bad it takes more than the average technical ability to understand it or accurately measure the output - if you can get it to work.

3. the circuit shouldn't use MosFETs.

4. It makes more sense if you consider Q2 as the first stage.

Does RA claim she invented this mess? The circuit has been around since the 20's but using the wrong switches doesn't make it new.

I thought cascode circuits were pretty much forgotten and out of vogue. They can be hard to understand.

 

WW,

RA is not claiming she invented any Cascode Voltage Oscillator.

Do you have a simple drawing of such type of oscillator?

GL.

[Groundloop]

Have you disconnected Q2 while the circuit is oscillating to verify that?

If it does stop, and it is indeed operating this way, then that would surely be a unique oscillator.  

.99,

Now I have tested disconnecting the Drain of Q2 in the circuit (referring to my drawing version 1)
while the circuit was running. When I removed the D on Q2 then the oscillation did stop.
When I connected the D on Q2 the oscillation did start again.

GL.

[poynt99]

OK,

That makes sense in light of my last redraw of your circuit. Of course Q2 will not function with only a negative bias on it.

Try the "p" wire if you can.

[Groundloop]

OK,

That makes sense in light of my last redraw of your circuit. Of course Q2 will not function with only a negative bias on it.

Try the "p" wire if you can.

.99

Will make a bias circuit first, then test the P line.

[EDIT 1] Build done. Now testing P-Line removal.

[EDIT 2] No oscillation was seen when removing P-Line at any bias adjustment.

GL.

[poynt99]

OK GL.

It's looking like a unique configuration.

What about the simplified version I drew up. Could you try that?

Thanks,
.99

[Groundloop]

OK GL.

It's looking like a unique configuration.

What about the simplified version I drew up. Could you try that?

Thanks,
.99

.99

Yes, give me a little time to set it up.

GL.

[Groundloop]

.99

Tested with a 5K pot. No oscillations at any pot meter settings.
Got a linear response of current with regards to positive voltage bias.

GL.

[poynt99]

OK, thanks GL.

Any idea the inductance and resistance of your coil?

.99

[Groundloop]

@All,

What have we learned so far:

- We need both mosfets to get an oscillation going.
- We need negative voltage bias to get oscillation.
- The voltage (and current) in the bias sets the amount of current through the switch.
- The frequency of the oscillations is set by the coil inductance together with the coil capacitance.
- The mosfets is operating in the linear area because of the relative clean sinus output.
- There is no need to inject a lot of power via the bias voltage input to get an oscillation.
- The oscillator can handle a load up to the SOA of the mosfets.
- The oscillator can handle both inductive loads and resitive loads.

So what to test next?

GL.

[Groundloop]

OK, thanks GL.

Any idea the inductance and resistance of your coil?

.99

.99,

No idea.

I do not have any means to measure coils.

I have 100 turns of 0.9mm enameled copper wire onto a bobbin with a Metglas AMCC32 core.

GL.