Inductive Circuits - The "Classical" Approach

[poynt99]

Oscillating Circuits

Mosfets are used in many applications specifically because they oscillate so well. I have done the research after learning about Rosemary's circuit and it is all over the web from industry references, massive amount of patents on self oscillating mosfet circuits, audio circuits, etc...

Even on IRF's website in all the technical docs, they give you many scenarios and values of components for the mosfet to oscillate at different frequencies. It is all right there. I know because I took the time to look it up.

I agree the oscillation should not be required to achieve a gain according to how the circuit is supposed to work.

As Hoppy pointed out, MOSFETs are not designed to "self-oscillate". In fact, I'm not even sure what you mean by that term. I can assure you that MOSFETs in and of themselves do not oscillate on their own.

ANY active device can be used in a circuit specifically designed to incorporate that specific device's properties and amplifying characteristics in order to achieve stable oscillation.

Active devices utilized to implement oscillators include: BJT's, JFET's, MOSFET's, Vacuum tubes, SCR's, PUT's, Avalanche diodes, operational amplifiers, etc. etc.

Oscillations in audio are only used for creating test tones. Other than that, oscillation is absolutely not desired in any audio circuit unless specifically required as in for certain effects devices etc. Audio power amplifiers employing MOSFET output stages are notorious for breaking into oscillation in the 50 to 100kHz range, especially if not properly designed of bias not properly set. This is a huge no no, and is highly undesirable. Power MOSFETs have relatively large input capacitance and as such introduce a lot of phase shift in the audio signal, particularly when the MOSFET's are included within the amplifier's feedback loop, which 99% of them are. With gain >1 and 180º phase inversion, you will have oscillation unless the amplifier is "compensated" internally to prevent this.

This is how some oscillators are purposely designed, and this technique goes back to the days of Hewlette and Packard when they came up with the first (patented btw) and now well-know Wein-bridge oscillator topology.

So don't equate purpose-designed oscillators made with active devices (including MOSFETs) with parasitic/undesirable oscillations that sometimes occur due to poor design, failed components, or mis-adjustments.

MOSFETs don't have the market locked up on oscillators if that is what you are thinking. Give me a Schmitt Trigger gate, a capacitor and a resistor, and I'll make you a nice simple square wave oscillator at any frequency you desire.

Here's a doc I put together some time ago for the folks at OU (inspired by Ash's need for one to use in the MRA). It employs this very circuit as well as the Wein Bridge design I mentioned. Square and Sines available, cheap and easy.

.99

[MileHigh]

Zealots:

I have something that might get some hearts beating, the Exar XR8038 and XR2206 voltage-controlled oscillators. Some of the hair-pulling over using oscillators as function generators could be made easy with these chips. Be sure to check out the links, and then search on "<chip> application notes" and have fun.

Both of these chips can produce sine waves with about 1% distortion, which is pretty decent. Some may not know that a pure sine wave, and a voltage-controlled one at that (as easy as pie) is a very important "research tool" for looking at circuits. A pure sine wave has a frequency spectrum that consists of two spikes, okay - one spike, but it is really two for those in the know. Don't be confuse this with a "voltage (time) spike", this is a "frequency spike."

You even get true triangle waves. ooohh... lol

And now for the almighty links:

https://www.jameco.com/Jameco/Products/ProdDS/34972.pdf
XR8038 pdf, XR8038 description, XR8038 datasheets, XR8038 view ::: ALLDATASHEET :::

To get real power out of it, you would need a DC-coupled linear power amplifier circuit. You could probably use any stereo amplifier, but be bandwidth-limited. You might have to hack into a stereo amplifier to get pure DC-coupling. For sure there are these power MOSFET or BJT modules that go into car audio amplifiers. A brave soul could buy one and play with that. It's a palm-sized amplifier module, all you have to do is give it a beefy power supply and add a few support components.

You could end up with one high-class voltage-controlled signal generator with adjustable gain and bandwidth and high power output. Use 10-turn pots for the voltage control.

MileHigh

PS: From glancing at the pdf files, one, probably both, run on up to a +/-15-volt supply, so the output waveforms are very high in amplitude. That means the gain setting on your DC-coupled power amplifier can be set to unity or less for an output waveform with the same or less amplitude. This is _very_ good news. The power amplifier output will be very stable and low-noise because the voltage gain is unity or less. You are effectively lowering the output impedance of your waveform, giving it the ability to source or sink a large amount of current into a load and maintain the same output voltage. Typically the output from an amplifier module is a differential output (difference in voltage between two separate outputs), so one of your differential outputs becomes the ground for your circuit. Just be aware that this ground is NOT the same as your battery and/or power supply ground that powers the amplifier module and you should never connect the two together. However, a separate and independent battery can still be added to the circuit if need be. The independent battery's ground can connect to the ground derived from the differential amplifier output. This independent battery will "hitch a ride" on the differential amplifier output circuit ground and not be "aware" of this.

[Joit]

Milehigh, when you found one Ying-Yang Part, it doesnt mean, you can take that for everything

And if you think, Coils replace Capacitors, well,
then we then now we can build all E-Motors with Capacitors, and save all this Coils because the EM Field is the same. Ahaha-ha.

Point is, to not to say its impossible to build it with Capacitors,
i dont think, it works with Capacitors, you need at last one Inductive Element,
and not only a Cap, what cause a Short at a certain Point and has changing Resistance down to zero like a Cap has.

The Answer is allready given at Post #40 from Altair.
Thanks for that, Altair.

[poynt99]

Do a bit of study rather than ranting on with opinion

Milehigh, when you found one Ying-Yang Part, it doesnt mean, you can take that for everything

And if you think, Coils replace Capacitors, well,
then we then now we can build all E-Motors with Capacitors, and save all this Coils because the EM Field is the same. Ahaha-ha.

Point is, to not to say its impossible to build it with Capacitors,
i dont think, it works with Capacitors, you need at last one Inductive Element,
and not only a Cap, what cause a Short at a certain Point and has changing Resistance down to zero like a Cap has.

The Answer is allready given at Post #40 from Altair.
Thanks for that, Altair.

Joit, are you listening at all?

Altair is actually wrong in his "guess" that the capacitive circuit would be inefficient. Opinions are welcome but they don't hold much water I'm afraid. To say "I think this one would be awfully inefficient if built." is just that, an opinion. There is no basis in fact.

When comparing the efficiency of a circuit employing an inductor vs. one with a capacitor, the capacitive circuit could be more efficient. What makes inductors inefficient is their inherent DC resistance. It's something that just can not be eliminated. In comparison, capacitors of large value generally have significantly less series resistance (called ESR for caps), and therefore less energy loss. We can make more efficient coils using high permeability cores and large wire, but there are distinct limitations here as well.

When and if either of you gents actually take the time to understand how the circuit works, then you will have a leg to stand on at least.

ASK if you don't understand. We're all here to learn (including myself) and help each other.

.99

[Altair]

.99, I took a second look at your interesting circuit.
I just realized that I1 is a current source, so it will (theoretically) supply any required voltage to keep the charging current flowing in C6.
When the shunt mosfet turns ON, C6 discharge current flows up through the load. A big spike of current will be generated, of course. If the load is a capacitor, its voltage will get pumped-up by the action of the 2 diodes and C6. Somewhat like a boost regulator.
However, I don't see OU here... please enlighten me.
P.S. Ideally, I1 should be gated, as to not continuously try to feed current through a shorted mosfet.
P.S. 2 If the load is inductive, there will be a recirculating current established through the load. Maybe OU could be attained here ?

Thanks for the brain teaser.

[poynt99]

Tough Crowd

.99, I took a second look at your interesting circuit.
I just realized that I1 is a current source, so it will (theoretically) supply any required voltage to keep the charging current flowing in C6.
When the shunt mosfet turns ON, C6 discharge current flows up through the load. A big spike of current will be generated, of course. If the load is a capacitor, its voltage will get pumped-up by the action of the 2 diodes and C6. Somewhat like a boost regulator.
However, I don't see OU here... please enlighten me.
P.S. Ideally, I1 should be gated, as to not continuously try to feed current through a shorted mosfet.
P.S. 2 If the load is inductive, there will be a recirculating current established through the load. Maybe OU could be attained here ?

Thanks for the brain teaser.

I wish folks would read what is actually posted. Sure would save a lot of trouble.

Anyway, again, if y'all read the posts, I said clearly that this is not about OU per se. In fact the whole point of this exercise was to establish the parallels between the inductive and capacitive circuits, and that if it can be clearly shown that although the familiar spikes can be produced in the capacitive circuit (similar to but "transposed" in comparison to the inductive circuit), that one would come to the conclusion that neither circuit exhibits OU!

Boy, you guys are a tough crowd sometimes. LOL.

.99

[Altair]

Hi .99, please relax, your first post (#1) SPECIFICALLY states CAPACITIVE OVERUNITY CIRCUITS.

Here it is:

Before discussing anything directly about this subject, I would like to ask any individuals that have a "new age" slant on these topics, a question pertaining to "Capacitive Overunity Circuits".

To start, surely all would agree that inductors and capacitors are true opposites, or as one person says "mirror images of each other". Also, check out his circuit half way down the page....looks quite familiar. Odd he hasn't discovered any overunity in his testing

Here's the question:

If inductive circuits (particularly the inductive kickback type) are perceived and apparently metered to exhibit COP>1, then please illustrate an example of a complementary capacitive circuit that does the same.

.99

[poynt99]

Hi .99, please relax, your first post (#1) SPECIFICALLY states CAPACITIVE OVERUNITY CIRCUITS.

Here it is:

Right,

I was asking the new age folks from their point of view the question.
It's not something I would be asking a classicist.

It was a challenge out to all, and in particular the new age folks, to show the "effect" in a capacitive circuit...is there OU with caps?

Anyway, a small oversight on my part. Apologies. I thought it was pretty obvious the viewpoint I take based on the title of the thread.

Cheers,
.99

[poynt99]

There's still one component missing in the capacitive circuit posted here to make it 100% "equivalent" to the inductive circuit. Presently, as the circuit stands there is no "flyback diode" to curtail the big spike as we have (now abandoned in the RA circuit) in the inductive version.

Anyone know what it is and where it would go?

.99

OK, now on to some rebuttals...

[poynt99]

1) All claims the spikes will damage the mosfet and that the ringing should be stopped (FACT - this mosfet IRFPG50 is designed EXACTLY for this kind of application)

2) All claims that the spike would be too small to be significant (FACT - on a decent circuit the voltage is 4 times the input voltage, it charges batteries or caps - it is VERY significant)

3) All claims that when the mosfet is off, the battery cannot conduct and therefore won't receive a charge (FACT - the diode in the mosfet allows just this exact current conduction as it is designed to do this!)

1) You are correct in that a "ruggedized" (i.e. avalanche rated) MOSFET can take the hammering. However, it's not quite correct to state that they were "designed" to be used this way. In fact using a MOSFET as an avalanche device is not very efficient. There are better devices (such as the diode version) that should be used for avalanche applications, such as pulse generators and voltage suppressors.

In general, active devices such as MOSFETs, BJT's etc, aren't "designed" for any specific application (yes there are exceptions). If for example you know that the switch in your application will experience periodic pulses exceeding it's rated VDS, then yes it would be a very wise choice to use an avalanche-rated MOSFET for this application.

2) The energy in the kickback pulse through the flyback diode (causing current to go back into the coil or external battery) is equal to the energy stored in the inductor prior to the switch opening (turning OFF), minus the energy loss in the inductor's resistor (10 Ohms) and the energy loss in the flyback diode itself. The fact that the voltage is higher than V+ is obviously good if we want to charge a battery, but the voltage itself is not an indication of the energy available in that pulse. When you load that pulse down (as you do when charging a battery), the voltage will drop quite a lot.

3) I have already identified the possible current paths through and around the MOSFET when it is OFF. These are via capacitances in and around the MOSFET. In the case where the flyback diode is removed, and there are large kickback spikes hitting the MOSFET, then the voltage would have to be in excess of 1000V (for the IRFPG50) to cause reverse breakdown of the body diode (actually a NPN transistor) and thus allow conduction current for the battery. From what I've seen in all the tests done by Aaron and TK (and my sims) the kickback voltage is well below 1000V.

.99

[FuzzyTomCat]

1) You are correct in that a "ruggedized" (i.e. avalanche rated) MOSFET can take the hammering. However, it's not quite correct to state that they were "designed" to be used this way. In fact using a MOSFET as an avalanche device is not very efficient. There are better devices (such as the diode version) that should be used for avalanche applications, such as pulse generators and voltage suppressors.

In general, active devices such as MOSFETs, BJT's etc, aren't "designed" for any specific application (yes there are exceptions). If for example you know that the switch in your application will experience periodic pulses exceeding it's rated VDS, then yes it would be a very wise choice to use an avalanche-rated MOSFET for this application.

2) The energy in the kickback pulse through the flyback diode (causing current to go back into the coil or external battery) is equal to the energy stored in the inductor prior to the switch opening (turning OFF), minus the energy loss in the inductor's resistor (10 Ohms) and the energy loss in the flyback diode itself. The fact that the voltage is higher than V+ is obviously good if we want to charge a battery, but the voltage itself is not an indication of the energy available in that pulse. When you load that pulse down (as you do when charging a battery), the voltage will drop quite a lot.

3) I have already identified the possible current paths through and around the MOSFET when it is OFF. These are via capacitances in and around the MOSFET. In the case where the flyback diode is removed, and there are large kickback spikes hitting the MOSFET, then the voltage would have to be in excess of 1000V (for the IRFPG50) to cause reverse breakdown of the body diode (actually a NPN transistor) and thus allow conduction current for the battery. From what I've seen in all the tests done by Aaron and TK (and my sims) the kickback voltage is well below 1000V.

.99

Hi poynt99,

I'm not aware that you may have seen this from "International Rectifier" Application Note AN-1005 - Power MOSFET Avalanche Design Guidelines this has some good information everyone might like.

http://application-notes.digchip.com/014/14-15377.pdf

Also another good one from "Advance Power Technology" Understanding the Differences Between Standard Mosfet's and Avalanche Energy Rated Mosfet's

http://www.nalanda.nitc.ac.in/indust...PT/APT9402.pdf

Regards,
Glen

[poynt99]

Thanks Glen.

Good info!

.99

[boguslaw]

I don't understand many things.
Could someone explain me how magnetic field around capacitor is different then magnetic field around inductor and how it is changed over time during inductive and capacitive discharge ? Are they any scientific documents about it ? For example : take inductor, charge it and disconnect completely from the circuit instantly, do the same for capacitor. Measure magnetic fields around : shape ,direction and changing over time.
I'd really like to know how they looks like...

[morpher44]

physics

I don't understand many things.
Could someone explain me how magnetic field around capacitor is different then magnetic field around inductor and how it is changed over time during inductive and capacitive discharge ? Are they any scientific documents about it ? For example : take inductor, charge it and disconnect completely from the circuit instantly, do the same for capacitor. Measure magnetic fields around : shape ,direction and changing over time.
I'd really like to know how they looks like...

Capacitors are not designed to respond at all to magnetic fields.
C = Q/V were Q is charge and V is volts. Magnetism is not in this function.
The charges, however, will respond to magnetic fields, but if they
are inside the capacitor at the time, it probably won't make much of a difference. If you can PULL charges into your capacitor using magnetic
fields, that would be cool.

Inductors are a more complex machine. They have capacitance and inductance really, and hence have a self resonance.

re: scientific documents
A good physics book .. or electronics book might get you going.

If you disconnect a capacitor, it might hold a charge .. but it will
decay because there is an implicit resistance... and hence
it is an RC circuit.
For a suddenly disconnected inductor, you have the LRC circuit
were C and R are implicit values as well.
The inductor will ring-down to equilibrium.
If L is HUGE, this ring-down can take a while.

[morpher44]

seeing magnetic fields

I'd really like to know how they looks like...

Check this out:

Dailymotion - Magnetic fields exposed - a Tech & Science video