PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2018-12-10, 02:28:42
News: Forum TIP:
The SHOUT BOX deletes messages after 3 hours. It is NOT meant to have lengthy conversations in. Use the Chat feature instead.

Pages: 1 2 3 [4] 5
Author Topic: Self running coil?  (Read 68788 times)
Group: Moderator
Sr. Member
*****

Posts: 386

First shot is of dual coil toroid at eighth H returning -39uA and next is single coil toroid at eighth H returning -27uA



« Last Edit: 2010-04-01, 04:46:34 by gotoluc »
   
Sr. Member
****

Posts: 253
Luc:

About your main toroidal coil where you have it in two separate parts.  I am assuming that you have the two parts connected in series and they are helping each other and not canceling each other out.  In other words if the current runs clockwise through the first part for a given current direction, when you make the connection then the current will also continue to run clockwise.

When you put the two parts together like that you have a standard toroidal coil where the inductance is proportional to the square of the total number of turns.  With this understanding your comments about "same resistance (implying the same total wire length) and double the inductance" don't make sense.

Your only real concern is the number of turns of wire you wrap around your toroid.  The tightness of the winding means nothing, as long as you are reasonably neat you will be fine.

To wire a two-part coil in series where the second part acts to cancel out (counterclockwise instead of clockwise turns) is not a logical thing to do.  You don't gain anything like that, you loose inductance.

I looked up common mode chokes and I don't think you are doing anything like that in your experiments.  A common mode choke is wired like a transformer with two separate parts on the same core but with independent connections.   The choke will impede the passing of high frequencies.  I don't think you are doing anything like that.
MileHigh

@MileHigh

I put your text above in bold.  Nobody stated Luc uses a common mode choke in the sense as such choke should be used.

What I wrote (and Luc quoted) clearly states the winding technique is just like a common mode choke has but the connection of the coils is different.

Here is what I wrote to a member, Rob and Luc copied for you:

Quote

Hi Rob,

Re on common mode choke: you are correct the winding technique Luc used for his toroidal coils is really the one as the so called common mode chokes are made BUT the big difference is the way how they are connected: Luc connected the two coils in series aiding i.e. the MUTUAL inductances of the two coils add to the sum of the individual inductances in series,  so that the resultant inductance is nearly the 4 times of a single coil, Lresultant=(L1+L2+2M where L1=L2=L in the equation (the two coils are assumed to have the same inductance which is nearly true and M is nearly L because in ring cores the coefficience of coupling is nearly 1 ).  (A useful link on this is here: http://www.daycounter.com/LabBook/Mutual-Inductance.phtml )

I hope this is fully understandable now.

Respectfully,
Gyula
« Last Edit: 2010-04-01, 14:39:45 by gyula »
   

Group: Elite
Hero Member
******

Posts: 1859
Luc,

I like your bungee cord analogy.  :)

You should also see when a magnet is added to a cylindrical coil with the polar axis matching, the Q is increased. In other words, the tuning of the resonant frequency is sharpened.

Small button magnets are added to filter coils and coupling transformers to increase the Q or selectivity.
Magnets added with the magnet polar axis normal to the coil polar axis decreases Q and increases bandwidth. I don't know of this being used.

The same thing is true for toroidal coils but it is more complicated.

A toroidal coil is just a cylindrical coil providing mutual inductance (series aiding) to itself.


---------------------------
"As far as the laws of mathematics refer to reality, they are not certain; as far as they are certain, they do not refer to reality." - Einstein

"What we observe is not nature itself, but nature exposed to our method of questioning." - Werner Heisenberg
   
Sr. Member
****

Posts: 253
Luc,

I like your bungee cord analogy.  :)

You should also see when a magnet is added to a cylindrical coil with the polar axis matching, the Q is increased. In other words, the tuning of the resonant frequency is sharpened.

Small button magnets are added to filter coils and coupling transformers to increase the Q or selectivity.
Magnets added with the magnet polar axis normal to the coil polar axis decreases Q and increases bandwidth. I don't know of this being used.

The same thing is true for toroidal coils but it is more complicated.

A toroidal coil is just a cylindrical coil providing mutual inductance (series aiding) to itself.

Hi WaveWatcher,

Do you mean air core coils or coils with ferromagnetic cores whose Q increases?  Have you checked such personally?

Thanks, Gyula
   
Group: Guest
Gyula:

Got it, thank you.

Luc:

With respect your negative current readings I will just tell you what I would do:  I would question it because logic is telling you that the current should always be positive.  Therefore you assume that something is amiss somewhere.  So my next move would be to scope the current waveform across the shunt resistor to see what was going on.

Question for everybody:  What is the relevance of the Q?

Thanks,

MileHigh
   
Group: Guest
   
Group: Guest
Darkspeed:

Please expand on that.

MileHigh
   
Group: Guest

The greater the Q the more narrow the tuning sweet spot and thus the greater the amplitude that can be acheived

Low Q = wide range of amplitude but lower peak amplitude

High Q = narrow range of amplitude but greater peak amplitude

To increase the Q space the windings further apart

   
Group: Guest
Darkspeed:

I am taking somewhat of a devil's advocate role here.

That's for the link and the synopsis.  I suppose the real question is how does the Q apply to what Luc is doing?

Full devil mode here:

Quote
To increase the Q space the windings further apart

Please explain this or if you were simply told it then that's fine too.

Thanks,

MileHigh
   
Group: Guest
No problem MileHigh devils advocate is a good position!

The Q definition was just a general answer to the question of Q

I have found from personal experience that if a coil has a given Q,  that slightly spacing the windings will lower the Q , but if you continue to increase the spacing of the windings there is a point that the Q will swing the other way and reach a point that is higher than the inital Q value. I am guessing that the increased spacing reduces the self capacitance, self inductance, etc..
   
Group: Guest
Hey Darkspeed thanks for the comments.

With respect to your answer to my question, if you go back and read the Wikipedia link you will see that a stand-alone coil has no property of "Q."  It takes the combination of resistance, capacitance, and inductance to derive the Q.  Same thing applies to resonance, a stand-alone coil does not possess a resonant frequency.

With respect to changing the spacing of the windings (I'll assume that we are talking about a toroidal coil here), that will have almost no affect at all on anything.  We are not even talking about Q here, it's not on the table.  The next logical thing to discuss is the inductance.  Changing the spacing on the windings will have almost zero affect on the inductance.  In terms of the types of experiments that you are doing here, the only thing that counts is the number of turns of wire wrapped around the core for determining the inductance.

The capacitance associated with the coils in these types of experiments is almost completely irrelevant.  The capacitance of a coil only comes into play at very very high frequencies.

Anyway I hope that you got some new ideas to ponder.

MileHigh
   
Group: Guest
Hey Darkspeed thanks for the comments.

With respect to your answer to my question, if you go back and read the Wikipedia link you will see that a stand-alone coil has no property of "Q."  It takes the combination of resistance, capacitance, and inductance to derive the Q.  Same thing applies to resonance, a stand-alone coil does not possess a resonant frequency.


MileHigh

A stand alone coil does have a natural rf as a function of its natural L C R

Q in a stand alone coil is relative to the width of its natural resonant frequency threshold.

Or that is what I was taught...





   
Group: Guest
Darkspeed:

You are right and it's my mistake.  I should have said that an inductor does not possess a resonant frequency.  The implication being that you are talking about an ideal inductor with no associated capacitance.

Going back to a real-world coil, it has inductance, resistance, and inherent stray capacitance, hence you can derive a Q for it.  However, and it is a big however, this is all data that never comes up in the real world.  Nobody is concerned about these parameters because they have no real affect on whatever circuit that you a likely to be using your real-world coil in.  Chances are the frequencies in your circuit will be low relative to the stray capacitance, so the stray capacitance will do nothing and have no effect.  Chances are the real-world capacitor you use in your circuit will be between thousands and billions of times larger than the stray capacitance of the coil, so the stray capacitance of the coil will do nothing and have no effect.  Chances are the self-resonant frequency of the coil will be between thousands and millions of times higher than the operating frequencies of your circuit.  Thus the self-resonant frequency of the coil will never affect the circuit.

More food for thought...

MileHigh

   

Group: Elite
Hero Member
******

Posts: 1859
All good informative conversation....

I doubt Q has much to do with Luc's circuit.

Yes, magnets have an effect on Q and resonance of an air or ferrous core coil (yes, ... tank circuit). One example I mentioned was made by RCA for Grid-Dip meters and IF cans for communications receivers. I know I have some somewhere. I'll post a pic when located. I bought my last batch at the local Ham Fest.
Other examples are a proximity detector, certain types of magnetometer coils and 'mag-pickups'.  'Mag-Pickup' is not a common device so I'll explain.

A mag-pickup is a simple iron rod wrapped in a high impedance coil. A magnet is placed at one end with a magnetic pole in contact with the end of the rod(core). When metal passes near the other end the magnetic field is distorted and a current is induced into the coil. You can find these devices counting ring-gear teeth on the flywheel of an engine.
The reverse is also used. Other uses include monitoring a change in resonance due to a magnetic field changing in proximity to a coil.

My understanding is the magnet 'concentrates more flux lines' (to use a classic description) in the area of the coil.
A better analogy is Luc's bungee cord or it stiffens the local field.

The latter is a bit of a dumb description but accurate. 

In any case, a magnetic field cannot add or remove energy. It can only cause acceleration(change of direction on moving charge) or act as a transfer for energy, without heat loss.


---------------------------
"As far as the laws of mathematics refer to reality, they are not certain; as far as they are certain, they do not refer to reality." - Einstein

"What we observe is not nature itself, but nature exposed to our method of questioning." - Werner Heisenberg
   
Sr. Member
****

Posts: 253
Hi WaveWatcher,

Thank you for the answers on coil + magnet. If you do not mind I would like to return to two sentences from your earlier answer for Luc,  I quote them here and added #1 and #2 in their beginning:


1) You should also see when a magnet is added to a cylindrical coil with the polar axis matching, the Q is increased.

2) Magnets added with the magnet polar axis normal to the coil polar axis decreases Q and increases bandwidth. I don't know of this being used.


English is my second language so I ask how a magnet is positioned when added to a coil "with polar axis matching" ? And how is it positioned when "the magnet polar axis normal to the coil polar axis" ?

I have uploaded a picture found at random and did some editing on it: let's say the inner rod in the coil is a cylinder magnet and positioned exactly in the geometrical center line of the coil axis.  About the poles of the magnet: have it your way, ok?

So as you put it, which position: #1 or #2 matches the picture I uploded?  IF it matches one of them, then what is the other position like? 

Other notice I would like to add:

I had access to an inductance meter in the past at my workplace and I placed magnets into air core coils and watched their inductance.  I did not find any significant change at all, ceramic magnets had about  .1 - .3% effect on the L and Neo magnets slightly more, maybe  .5% or so.  And I am not sure on the Neo's stronger field caused the bigger change,  rather I suspect the presence of metal caused the change, just like inserting copper or Alu cores into air core coils: they increase air cored coil's inductance when inserted. (Philips and others too manufactured car radios in the 70's with Alu cores tuning in the FM (88-108MHz) band.)

I have not tested ferromagnetic cored coils how a magnet affect them but I do think the magnets influences the core's permeability very much so this surely involves changes in both Q and inductance. So what you wrote on using small button magnets for IF cans or GDO coils: I can accept them when they include ferromagnetic cores but would mean a big surprise if they have only air core...

IF you could share some further thoughts on the above, I would be interested to learn.

What you wrote on the Mag-Pickup coils is ok with me.

Thanks,  Gyula
   

Group: Elite
Hero Member
******

Posts: 1859
"...It appears therefore that the stress in the axis of a line of magnet force is a tension, like that of a rope..."

J.C. Maxwell[1861]

Or like a bungee cord  :D

gyula,

I think your English is better than mine...

#1 matches your photo.
#2 would be the magnet's polar axis perpendicular to the polar axis of the coil.

The inductance of a coil with a magnet near to, or as, the core would not change much unless that magnet had a high content of iron.
I have never known a magnet to change the inductance of a coil simply because of the increase in magnetic force due that magnet, unless that magnet brought the coil near saturation.

The IF cans had ferrite slugs for tuning.

Inserting brass rods into an air-core RF coil will cause the inductance to decrease. I don't think I ever saw a radio with aluminum rods for tuning.


---------------------------
"As far as the laws of mathematics refer to reality, they are not certain; as far as they are certain, they do not refer to reality." - Einstein

"What we observe is not nature itself, but nature exposed to our method of questioning." - Werner Heisenberg
   
Group: Moderator
Sr. Member
*****

Posts: 386
Hi all,

thank you all for your interest and informative debate ;)

I attached 2 scope shots for your viewing and in hope they will help the debate on the effect of adding a magnet to a high Inductance ferrite core toroid coil.

The first scope shot has no magnet so it is at full Inductance. Please note that it is also creating the gate switching current for the MOSFET by using a coil of tuned Inductance across the gate and source. So please consider this fact in your debate.

The next scope shot has a magnet so the inductance is less hence resonating at a higher frequency. Please note that it is also creating the gate switching current for the MOSFET by using a coil of tuned Inductance across the gate and source. So please also consider this fact in your debate.

The power source is 2 old D cell batteries in series giving exactly 3vdc and connected to a 3900uF capacitor bridged by a 1% 10 Ohm 50W resistor connected to another 3900uF capacitor which is connected to the circuit. A Schlumberger 7150 plus high resolution volt meter is connected between the resistor to calculate the current consumed by the circuit.

My question is: why is the non magnet Toroid using more current then the magnet Toroid?

Thanks for sharing :)

Luc

ADDED: Please note that both tests were tuned to the minimal point (just before circuit stops operating due to not enough gate voltage)

Here are the details of this test: IRF640 at 3.79KHz  main toroid 1074mH pulse coil 583mH 3vdc at 10uA

http://i944.photobucket.com/albums/ad290/gotoluc/IRF640at379KHzmaintroid1074mHpul-2.png
Self running coil?


Here are the details of this test: IRF640 at 5KHz  main toroid 604mH pulse coil 370mH 3vdc at 7.3uA

http://i944.photobucket.com/albums/ad290/gotoluc/IRF640at5KHzmaintroid604mHpulsecoil.png
Self running coil?
« Last Edit: 2010-04-03, 05:30:25 by gotoluc »
   
Sr. Member
****

Posts: 253
Hi WaveWatcher,

Thank you for clarifying my questions.  In the next couple of weeks I will have access to a Q meter and can check how the Q of an air core coil may change when a strong Neo magnet is inserted or brought very near to it.  Will return to this then.

You are right,  the inductance decreases when a copper slug is inserted into it (I erred) and the explanation is that copper is diamagnetic and has a relative permeability of just below 1:  ur= .999994   
For aluminum which is paramagnetic, the ur=1.000022 

I mainly found Alu cores in the oscillators of Philips FM tuners of table, portable and car radio sets (manufactured mainly for the European market in the 70's and 80's).

Thanks,  Gyula
   

Group: Elite
Hero Member
******

Posts: 1859
Luc,

As pointed out by Gyula, permeability is key in inductance. Since the inductance varies, I believe the cause is the presence of the magnet in your coil.

I need to look at your earlier posts again before I try to answer your question on current decreasing. Typically current decreases and voltage increases as you approach resonance.

>>Edit...

This may sound a bit idiotic to some. Luc, have you measured capacitance of your coil? Connected as it is in your circuit but disconnect the leads to the FET and supply. first.
The reason I ask... bifilar(series opposing) coils with good turn-to-turn spacing can exhibit a high capacitance measurement. Since your coil is bucking with series opposing connection, I'm wondering if you are seeing the varying capacitance effect.
I've had coils of lamp cord measure as high as 1,600pF. When built so a magnet will effect the inductance...that magnet also effects the capacitance of these odd-ball coils.
« Last Edit: 2010-04-03, 14:33:13 by WaveWatcher »


---------------------------
"As far as the laws of mathematics refer to reality, they are not certain; as far as they are certain, they do not refer to reality." - Einstein

"What we observe is not nature itself, but nature exposed to our method of questioning." - Werner Heisenberg
   
Sr. Member
****

Posts: 253


My question is: why is the non magnet Toroid using more current then the magnet Toroid?



Hi Luc,

I have edited your scope shots and uploaded them again.  You can see I indicated a 2V threshold line above the zero line for CH2 that shows gate-source voltage. I simply assumed the IRF640 has a 2V threshold, this is the lower manufacturer limit for this type, it would not matter from my explanation point of view if it actually had any higher, you will understand.

Under the 2V threshold line for the gate soure voltage the MOSFET is OFF, above it is ON. (Please next time use DC coupling for both channels when displaying gate and drain voltages, it would be more informative, especially if you overlay the two zero lines to cover each other. When you use AC coupling the waveform shifts up (or down) wrt the scope zero line, you can see this in the drain source voltage form which starts from under the zero line, while your scope grnd clip is (correctly) at the source, the most negative point in the circuit.)

So the key answer for your question is that the lower current consumption comes from the fact that the MOSFET has a much shorter ON time in the 5kHz case (about 56 usec) than in the 3.79kHz case (about 78 usec).

Unfortunately this explanation did not consider any other effect of the magnet than inductance reducement on the toroidal coil.

I uploaded a schematic on your circuit so that others could clearly see what is the oscillator like.  It is a tuned gate Class C type oscillator,  the capacitances are being that of the MOSFET interelectrode caps (this is why I did not use any cap symbol).
In the gate source there is a parallel resonant LC circuit, the C is the (Ciss) input cap of the MOSFET, this LC determines where the oscillator mainly oscillates. Feedback comes from the drain side via the gate drain capacitance (Crss).

rgds,  Gyula
   
Group: Guest
Just some generic comments about the nature of the investigation Luc is doing.

The generic idea here is that the circuit is a "filter."  The filter has an input side and an output side.

The input could be some form of on/off pulse train at a certain frequency, or a sine wave at a certain frequency.

The response of the filter to the input stimulation gives you the output.  The output can be characterized by how it responds to an on/off pulse train or how it responds in amplitude and phase for a given input stimulus frequency.

Certainly when you add or remove the magnet from the side of the toroid you will change the nature of the filter and it's resultant output characteristics.

For what it is worth, this is an area of study that has been thoroughly investigated and the mathematical modeling has been perfected.

For a pulse train on the input, you are interested in looking at the time domain response of the filter.

For a sine wave on the input, you are interested in looking at the frequency domain response of the filter.  As you sweep through a given frequency range on the input, you observe the amplitude and phase response of the output.  Peaks in amplitude response at a given frequency are known as "poles" and minimums in response at a given frequency are known as "zeros."  Hence the filter can often be characterized by its "poles and zeros."  What is often referred to as the "resonance" is in indeed a "pole" of the filter.  There may be one pole, or there may be multiple poles for a given filter.

I hate to be a downer but the notion that you can get something "extra" at resonance is false.  Not that what I just said is going to stop people from trying.

With respect to the MOSFET, I think there is a fair amount of misunderstanding with respect to how it works.  The gate input does have some capacitance, but otherwise it is insulated.  The current input on the gate is zero.  This is in direct contrast with the base current of a transistor which is non-zero.

You typically operate a MOSFET in one of two possible modes:

In switching mode you put enough voltage on the gate input to switch the MOSFET completely ON.  To switch the MOSFET completely OFF you bring the gate input back to zero volts.  That's the only thing that you want, ON or OFF.  When the MOSFET is ON, there is a measurable low-valued resistance between the source and the drain.  In many cases it does not matter which MOSFET you use in your circuit if all that you want to do is use it as an ON/OFF switch.  You will notice however, that a square wave from the signal generator sometimes gets turned into what looks like a sine wave at the MOSFET gate input.  This could be because the frequency is too high or it's more likely that the signal is not being properly terminated at the gate input.  This will screw up any switching analysis for the MOSFET and must be avoided.

In linear mode the MOSFET gate voltage is somewhere between the ON and the OFF voltages.  Here the MOSFET is partially conducting and in many ways resembles a typical transistor in non-switching operation except for the fact that the gate current is zero.  The MOSFET can get hot and even burn up here because it is partially conducting and acting like a resistor.  Linear mode can be used to act as an amplifier, amplifying a small AC voltage on the gate input and turning that into a larger AC voltage on the drain output.

It gets very very complicated when you operate the MOSFET in linear mode and you are looking at the response from an inductor that is charging and discharging like Luc is doing here.  Using the gate coil and setting up a resonance with the gate capacitance is causing the MOSFET to operate in linear mode.  However, a sine wave at the input will not necessarily produce a sine wave at the output.  You have to carefully study the input characteristics of the MOSFET gate input to make sure that the sine wave input swings in the proper voltage range so that the MOSFET operates in linear mode.

My suggestion would be to exclusively operate the MOSFET in ON/OFF mode and sweep the input frequency and observe the output response from the filter.  As the frequency changes you should see changes in the "amplitude" and "phase" of the response from the filter.  I am using "amplitude" and "phase" in parenthesis here because you are really looking at the pulse response of the filter.  A few days ago Luc posted some traces where the "phase" difference between the input and the output was 90 degrees.  This was for a different type of resonance as compared to a normal LC resonance.  In Luc's more recent postings you now see an apparent 180 degree in phase shift between the input and the output and the MOSFET is operating in "linear" mode.  All of this is related to how the filter responds when you sweep the input pulse frequency from low to high.

I put "linear" in quotations because most likely when you are putting a sine wave on the gate input parts of the sine wave are putting the MOSFET in switching mode and other parts of the sine wave are putting the MOSFET in linear mode.  Again, you have to take a very careful look at the specifications for the MOSFET to know how it operates for a given voltage input at the gate.  This is also dependent on the drain voltage and the drain current.  The charts that you see in the MOSFET datasheet are typically for a pure resistive load at the drain pin, and not for a reactive load (coil) at the drain pin.

With all that taken into account, when you see a higher output voltage at resonance for one setup compared to another setup, that is implying that the setup with the higher voltage has a higher Q factor.  This typically means the amount of resistance in the oscillating circuit is lower, so the voltage can get higher.  Higher peak-to-peak voltage in setup A as compared to setup B does not really tell you much.  It does not mean that you are getting any gain associated with over unity from the circuit.  You have no frame of reference to establish any kind of over unity gain from one setup to the other.  Measuring voltage means nothing without measuring the corresponding current.

There are probably hundreds of engineering textbooks about filters and analyzing filters and I am just giving you a brief overview.

MileHigh
« Last Edit: 2010-04-04, 19:36:26 by MileHigh »
   
Sr. Member
****

Posts: 253
Hi MH,

Generally I can agree with most of what you wrote above but there are two statements I disagree with. I qouted below in bold the first problem:

....
With respect to the MOSFET, I think there is a fair amount of misunderstanding with respect to how it works.  The gate input does have some capacitance, but otherwise it is insulated.  The current input on the gate is zero.  This is in direct contrast with the base current of a transistor which is non-zero.   

Unfortunately input current is not zero on the gate, the gate source input capacitance (Ciss) is continously charged up and discharged whenever you drive it either with a square or sinusoid wave.  Luc nicely showed the charging and discharging pulses going back and force via the input capacitance, see his scope shots in the link below, he used a series 100 Ohm resistor in one of the wires that brought the drive pulses from his signal generator, the voltage drop across the series resistor nicely show how the current spikes look like. The gate source capacitance has to charge up first when the input square wave rising edge appears,  then the cap discharges when the square drive pulse falls.  Here is the link:

http://www.overunityresearch.com/index.php?topic=205.msg2237#msg2237   

So unfortunately, driving a MOSFET does cost input power you have to furnish from somewhere. It is true a MOSFET gate source is an open circuit but this is true for a DC point of view, for AC it is a capacitor (that even has a series resistance too).

My second problem is when you wrote what I qoute below also in bold:

Quote
...
It gets very very complicated when you operate the MOSFET in linear mode and you are looking at the response from an inductor that is charging and discharging like Luc is doing here.  Using the gate coil and setting up a resonance with the gate capacitance is causing the MOSFET to operate in linear mode.  However, a sine wave at the input will not necessarily produce a sine wave at the output.  You have to carefully study the input characteristics of the MOSFET gate input to make sure that the sine wave input swings in the proper voltage range so that the MOSFET operates in linear mode.


If you have a look at Luc scope shots that I edited to indicate the MOSFET on-times (see my last post just above yours) you can see that a full cycle last for about 264us (3.78kHz case) but the FET is on only for about 78us during this one full cycle and in the 5kHz case when one full cycle is 200us long, the FET is on only for about 56us.  Linear mode (I mean Class-A) is normally defined when the device is on for the total full cycle  i.e. continuosly.  (For a MOSFET like the IRF 640, you have to supply DC bias to the gate to run it in linear mode.)
AND the gate source resonant reactive power charges and discharges the gate cap as normally a parallel LC circuit behaves when excited, this means the FET is able to conduct whenever its charging gate source cap reaches the FET threshold voltage and exceeds it.  This happens only in 1/3 to 1/4 of a full cycle, this could be called a Class-C mode. This is why so efficient Luc finds it.

Respectfully
Gyula
   
Group: Guest
Hi Gyula,

I agree with your first comment.  I was implying that the gate current was zero for a DC input.  Sorry for the omission of that detail.  Certainly you need current to charge and discharge the input capacitance.  Whether or not the input capacitance charging/discharging is going to affect what you are trying to do depends on your excitation frequency and what kind of ON/OFF slew rate you need.  You are probably aware that there are MOSFET driver chips that take care of the annoying capacitor charging and discharging for you if that is critical and your signal source can't handle it.

The other thing to keep in mind about the charging and discharging of the gate input capacitance is that it is propagating AC energy into the circuit.  A spike of current to charge the input capacitance will generate a spike of current on the source and/or drain pins.  That energy will continue going into the circuit somewhere.

Going back to the capacitive effects of the gate input itself, it can be part of the fun of the investigation into a given setup.  You want to be aware of the capacitive effects of the gate and decide if they are going to affect what you are trying to accomplish or not.  For example, if you want to use the MOSFET in a switching application at 60 Hz and you can tolerate the slew rate limitation from the gate input capacitance, then you can ignore it altogether.  On the other hand, if you want to do a switching application at 20 KHz, then it may come into play and you may need the MOSFET driver chip or do it yourself with transistors, etc.

For your second comment I just have a few things to say without getting too deeply into all of the design issues.  For the gate input let's say that you can define an "OFF" region, a "linear" region, and an "ON" region.  This is a function of the source-drain current and source voltage.

So for example if the source-drain voltage is 10 volts and the current is 100 milliamperes, then "OFF" might be 0-2 volts, "linear" might be 2-6 volts, and "ON" might be 6-10 volts.  Therefore if you excite the gate input with a sine wave, it all depends on the voltage sweep of the sine wave to determine how the MOSFET is going to react to this stimulus.

Just as a caveat, I am being pretty general and probably oversimplifying things here.  Certainly if I was testing a similar circuit I would be referencing the datasheet for the MOSFET and investigating all of these aspects.

I don't know how the gate input LC resonator is acting in Luc's setup, and just putting a scope probe on it will probably change it.  It would still be nice to see anyways to get a sense of the voltage sweep on the gate input.  I am making the assumption that the MOSFET is operating partially in switching mode and partially in linear mode by observing the waveform.

Anyway, with respect to Luc's observation of negative supply current to this setup, that could be an interesting investigation.  I am assuming that the multimeter is averaging the AC component superimposed on the DC current and returning a value that is slightly incorrect.  We are talking microamps here.  A suggestion for Luc would be to replace the power supply with a large capacitor and see what happens.  I am assuming a self-resonant setup here because you absolutely must remove the signal generator from the equation.  I saw some references by Luc about the a capacitor as the power source charging while the setup was running, and I will assume that he was also using a signal generator in those cases.  You cannot forget that the separate signal generator is injecting AC power into the circuit.  Going back to the case where the capacitor powers everything and the current consumption is only microamps, then you may be able to get away with using a smallish capacitor to power the whole thing so that the observed voltage drop takes place in a reasonable amount of time.

It make sense that the current consumption is very small because an RL (R very small) resonant circuit is constantly storing and returning energy to the source for a net current consumption of zero.  The very small R is dissipating energy and hence that energy ultimately has to come from the power supply (or capacitor acting as the power supply).

And just to be complete an LC resonant circuit is constantly transferring energy back and forth between the L and the C, you don't need a power source at all.  Luc may be interested in taking a charged capacitor and connecting it to one of his coils to observe the oscillation on his virtual scope.  He can capture the waveform in "one shot" mode and then measure the frequency.  That can then be compared to his measured L and C to see if the expected resonant frequency is in accord with the measured resonant frequency.

MileHigh
« Last Edit: 2010-04-04, 21:44:06 by MileHigh »
   
Group: Moderator
Sr. Member
*****

Posts: 386
Hi Luc,

I have edited your scope shots and uploaded them again.  You can see I indicated a 2V threshold line above the zero line for CH2 that shows gate-source voltage. I simply assumed the IRF640 has a 2V threshold, this is the lower manufacturer limit for this type, it would not matter from my explanation point of view if it actually had any higher, you will understand.

Under the 2V threshold line for the gate soure voltage the MOSFET is OFF, above it is ON. (Please next time use DC coupling for both channels when displaying gate and drain voltages, it would be more informative, especially if you overlay the two zero lines to cover each other. When you use AC coupling the waveform shifts up (or down) wrt the scope zero line, you can see this in the drain source voltage form which starts from under the zero line, while your scope grnd clip is (correctly) at the source, the most negative point in the circuit.)

So the key answer for your question is that the lower current consumption comes from the fact that the MOSFET has a much shorter ON time in the 5kHz case (about 56 usec) than in the 3.79kHz case (about 78 usec).

Unfortunately this explanation did not consider any other effect of the magnet than inductance reducement on the toroidal coil.

I uploaded a schematic on your circuit so that others could clearly see what is the oscillator like.  It is a tuned gate Class C type oscillator,  the capacitances are being that of the MOSFET interelectrode caps (this is why I did not use any cap symbol).
In the gate source there is a parallel resonant LC circuit, the C is the (Ciss) input cap of the MOSFET, this LC determines where the oscillator mainly oscillates. Feedback comes from the drain side via the gate drain capacitance (Crss).

rgds,  Gyula


Hi Gyula,

I was away for the weekend and now back.

Thanks for taking the time to make a schematic of my test setup and posting it with your explanation as to possibly why the 3.9KHz circuit uses more current then the 5KHz circuit with magnet. That is a good point!... however if we multiply the  78us X 3.9 = 304  and if we take the 56us X 5 = 280  so we have a possible difference of 25Us

I don't know if this small a difference would account for the total current differences?

Thanks for your explanation and time

Luc
   
Group: Moderator
Sr. Member
*****

Posts: 386
Hi Gyula,

I added some circles and lines to the 3.9KHz scope shot from my above post.

If we look at the 2 red circles I added (tops of the sine wave) we will notice the negative is perfect and the positive has a flat top to it. If we look at the main coil sine wave I added a purple circle were a bump appears at every pulse. Could that be the pulse coil being charged?... if it is, then I added a vertical blue line at the point it starts to fall and it coincides with the beginning of the flat area of the + sine wave. I then added another vertical blue line at the end of the flat area and then a blue line horizontally between the two.

I believe this is the on time of the MOSFET (between blue vertical lines) and the purple circle (main coil bump) to be the charging period of the pulse coil. What do you think?

Thanks for sharing.

Luc

http://i944.photobucket.com/albums/ad290/gotoluc/IRF640at379KHzmaintroid1074mHpul-1.png
Self running coil?
   
Pages: 1 2 3 [4] 5
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2018-12-10, 02:28:42