PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2024-11-27, 22:37:10
News: A feature is available which provides a place all members can chat, either publicly or privately.
There is also a "Shout" feature on each page. Only available to members.

Pages: 1 2 3 4 5 6 [7] 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 ... 29
Author Topic: The Rosemary Ainslie Circuit  (Read 477193 times)
Group: Guest
Poynt:

Those are some very interesting comments.  I admit my interest is slipping and I would have to see a schematic and refresh myself on things and look at your comments step by step so that I could truly follow what you are saying.  I don't have the energy to do that.  Such is life.

But I will address the following to Rosemary:

Quote Rosemary:

Quote
And you're rather keen on using your standard DMM. If you put the setting to AC it will show that negative battery voltage.

This doesn't make sense, and I am not going to try to decipher what you really mean here, if indeed there is something sensible between the lines that I am not getting.

Not wanting to be nasty but it's quotes like the one above that disqualify you from being able to analyze and make judgment calls about your circuit and Poynt's comments.  And the truth is that although you try to get up the learning curve, you have disqualified yourself hundreds of times in the past with similar types of comments.

So I suggest that you pass Poynt's comments along to the hidden "experts" that you often make reference to for their consideration.

This is becoming like Steorn, it's just taking too long and nothing is happening and there is no merit to either proposition.  Free energy schemes don't die, they just fade away with an exponential decay that never reaches zero and seemingly goes on forever.

MileHigh
   
Group: Guest
http://newlightondarkenergy.blogspot.com/2011/02/59-close-but-no-cigar.html

In Rosemary's post #59, after parsing and ignoring all the bullstuff and idiotic statements and teasing, I do find one thing I have to agree with, Poynt.  I don't think the body diode is broken.  As I said in an earlier post, I was able to get continuous oscillation during the gate low time by adding some inductance in series with the gate and lowering the gate resistance and tinkering with the load inductance.  I have to agree with her, though, that the drain voltage (which she keeps wrongly calling the battery voltage) doesn't go below zero by more than a volt or so.  This indicates an intact body diode to me.  

In general, MOSFET body diodes can hack the same levels of pulsed current that the enhanced D-S channel can hack and even with her now higher battery voltage, I think she's still inside the "safe area" of operation in both directions.

In order to model this oscillation, you must use a genuine PSpice Level III MOSFET model complete with the parasitic capacitances.  I used an IRF740, since that's the closest one in my library.  I got it to ring like a bell, full time with no decay.

But I have tired of playing Rosemary's teasing game of hide and seek.  She seems to want desperately to carry on a dialog, reading our every post and responding almost immediately, but she also seems completely unwilling to show her latest circuit, test lashup, layout and component values.  She is enjoying our guessing games immensely and, since she provides no facts, she can always say we are wrong and that we cannot model or replicate her magical circuit.

If Rosemary had any real interest in legitimate analysis or testing or replication, she would stop playing hide and seek and be more open about the particulars of the day.  But that has never been her way.  Rather she enjoys keeping a shroud of mystery and "dark energy" draped closely around her private fiddlings.  It has gotten boring for me at this point.

Read my byline, Rosemary, one more time:  REASONABLE CLARITY IS REQUIRED, AT MINIMUM.  You are a master of obfuscation, partly out of ignorance of the technology and its terminology and partly out of pure girlish spite.

Forgive my sexist remark, but you are acting like a silly little girl with a silly little secret and loving every minute of it in a spiteful and deceitful emotionally-based teasing way.   Scientific investigation in an open "cards on the table face up" forum is the furthest thing from your mind, it's obvious.  Your secrecy reminds me of a Bob Dylan lyric:  "She tries to hide what she don't know to begin with".

Humbugger
« Last Edit: 2011-02-19, 01:23:13 by humbugger »
   

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Humbugger,

If you could, please post your schematic and sim scope traces showing how you achieved the same oscillations as what Rose showed.

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
I was able to get continuous oscillation during the gate low time by adding some inductance in series with the gate and lowering the gate resistance and tinkering with the load inductance.  I have to agree with her, though, that the drain voltage (which she keeps wrongly calling the battery voltage) doesn't go below zero by more than a volt or so.  This indicates an intact body diode to me.
How much inductance are you using in the Gate, and what Rgate value? What load inductance and load R value?

Quote
In general, MOSFET body diodes can hack the same levels of pulsed current that the enhanced D-S channel can hack and even with her now higher battery voltage, I think she's still inside the "safe area" of operation in both directions.
That's ok in general, but you can not rule out the possibility with 100% confidence, I trust? I look forward to seeing your simulation, because without making some very drastic changes and/or using unrealistic component values, I was not able to get that oscillation.

Quote
In order to model this oscillation, you must use a genuine PSpice Level III MOSFET model complete with the parasitic capacitances.  I used an IRF740, since that's the closest one in my library.  I got it to ring like a bell, full time with no decay.
I do in fact use Level3 MOSFET models in PSpice, AND I am using an IRFPG50 model. It's a nogo for me, but I haven't tried anything too drastic that would not be possible in the real circuit. I did try some inductance in the Gate, but still nothing.

As I said, looking forward to your sim schematic. I'll see how it looks in PSpice.

.99



---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
Group: Guest
Part One:  Simulation showing full oscillation of drain voltage with MOSFET body diode fully intact

Note that this only reveals the full possibility of getting sustained MHz oscillation (real, not artifact) by adding series gate inductance to the model.  

Part two, which will have to wait for tomorrow's (or next week's) episode, will show that you are likely quite correct about the large oscillations showing on the current trace and the enlargement of this oscillation to include apparent negative excursion are indeed false measurement artifacts caused by probe and probe ground (remembering the "blue wire" movie) placement errors in conjunction with wiring inductances.

Humbugger

Admittedly, I had to make L2 about 1/5 of L1 (L1=10uH) and drop R2 to around 6.5 ohms to get this and it is quite finicky until you hit it (three variables) but I suspect there are numerous combinations that will oscillate like this and also that changing the MOSFET model will greatly effect the results and requirements for oscillation.  2uH is admittedly large for a stray gate inductance but, if you consider that she is probably using a coax cable of unknown wavelength that is improperly terminated from her function generator to the MOSFET gate, it's anyone's guess what the equivalent circuit and impedance presented to the gate is.
   

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Incidentally Humbugger,

If the body diode is working properly, Vbat should not go more than a few Volts or so below the battery voltage, not ground (0V). The fact that the Battery voltage oscillation is going down to below zero, indicates that either she is measuring the Drain voltage, or the body diode is severely damaged.

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Part One:  Simulation showing full oscillation of drain voltage with MOSFET body diode fully intact

Note that this only reveals the full possibility of getting sustained MHz oscillation (real, not artifact) by adding series gate inductance to the model.  

Part two, which will have to wait for tomorrow's (or next week's) episode, will show that you are likely quite correct about the large oscillations showing on the current trace and the enlargement of this oscillation to include apparent negative excursion are indeed false measurement artifacts caused by probe and probe ground (remembering the "blue wire" movie) placement errors in conjunction with wiring inductances.

Humbugger

Throw me a bone here Humbuggy,

Give me some component values. ;)

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Also,

In your diagram, you are measuring VD, not Vbat.

The measurement is of Vbat. ;)

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
Group: Guest
See the edit for the changed values.  Load R is 10 ohms anf Vbatt is 26VDC, shunt is 0.25 ohms.

"...either she is measuring the Drain voltage, or the body diode is severely damaged."  BINGO!

I am assuming that Rosemary is measuring Vdrain and calling that Vbat, incorrectly.

There is no way in hell that the battery has that much AC on it.  She even states clearly that when she puts a DMM on the battery, set for AC, she reads zero volts.  Didn't she say that?  Think about it:  there is always at least ten ohms pure resistance in series with the battery (Rload) and the reactance of Lload adds to that, so in order to get 100V pp excursions at the battery terminals, either the battery has an enormously high impedance or the actual drain voltage excursions would have to be thousands of volts if the battery impedance is as it should be...less than an ohm or so.

So, I'm figuring her negative excursions are scope/probe/ground inductance artifacts, as you suggested originally as the second part of your analysis.

Hum

P.S.  I'm not saying for sure that you're wrong about the body diode.  Just saying you can get continuous oscillation without a broken body diode.  The rest is so hopelessly confused by Rosemary's innane statements and terminology that only God knows what she is actually probing when she says it's the battery voltage.
   

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
See the edit for the changed values.  Load R is 10 ohms anf Vbatt is 26VDC, shunt is 0.25 ohms.

I am assuming that Rosemary is measuring Vdrain and calling that Vbat, incorrectly.

There is no way in hell that the battery has that much AC on it.  She even states clearly that when she puts a DMM on the battery, set for AC, she reads zero volts.  Didn't she say that?

So, I'm figuring her negative excursions are scope/probe/ground inductance artifacts, as you suggested originally as the second part of your analysis.

Hum

P.S.  I'm not saying for sure that you're wrong about the body diode.  Just saying you can get continuous oscillation without a broken body diode.  The rest is so hopelessly confused by Rosemary's innane statements and terminology that only God knows what she is actually probing when she says it's the battery voltage.

Humbuggy, with all due respect, I can assure you that the gate inductance won't be that high, and the gate resistance not that low. You need to factor in the 50 Ohm output impedance from the FG as well. Having considered these factors, the Gate circuit will have little to no effect on creating the oscillations we are seeing in Rose's wave forms.

Also, your schematic is incomplete, which is why you are not seeing the oscillations on the BATTERY vs. the Drain. She IS measuring the "battery" voltage, not the Drain. It is plain to see on the bench shot I provided (see short red wire).

See the attached drawing for a more complete (and necessary) rendition of what she actually has there, and what should be included in the simulation.

The fact that the Battery voltage already goes well below 48V is indication that the body diode is damaged. It appears to be trying to work by evidence of the distortion present there.

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
Group: Guest
Well, my brain is burning out with all the guesswork here.  You show 2.5uH for each battery wire.  At 20nH per inch thats 125 inches or over ten feet for each battery wire.  Admittedly, I see no battery in the picture, but do you really think it's ten feet away?  Could be, I suppose.

I also don't see the function generator cable coming to the MOSFET.  Actually, not much is very clear on that optics slab.  Could be she's driving the gate with another 10 feet of plain wire, for all I know.

I didn't mean to challenge your body diode theory to the point of a heated battle here Poynt.  I may be wrong.  I'm definitely getting burnt out trying to second-guess in the dark here, though.

I thought for sure Rosemary claimed that the scope waveform looked exactly the same when she was directly on the battery terminals.  Of course that begs all credulity, just like 95% of everything else she says.

I'll have to think about all this body diode stuff a little more before I really take a hard stance one way or the other.  God, I'll bet she's just loving all of this!   :D

Humbugger
   

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Well, my brain is burning out with all the guesswork here.  You show 2.5uH for each battery wire.  At 20nH per inch thats 125 inches or over ten feet for each battery wire.  Admittedly, I see no battery in the picture, but do you really think it's ten feet away?  Could be, I suppose.
I use 60nH/in. (20nH is for 24 GA.) and I estimate roughly 48 inches (x2 of course). So I am being conservative.

Quote
I also don't see the function generator cable coming to the MOSFET.  Actually, not much is very clear on that optics slab.  Could be she's driving the gate with another 10 feet of plain wire, for all I know.
The cable is there (black jacket, white and black wires), and it appears to be a twisted 18GA. or so. It doesn't appear to be shielded.

Quote
I thought for sure Rosemary claimed that the scope waveform looked exactly the same when she was directly on the battery terminals.  Of course that begs all credulity, just like 95% of everything else she says.
Agreed. If she had the other scope probes in their place, then the probe was probably never placed right on the battery terminals. She did say "IF and when...". As far as the rest about the DMM, I haven't a clue what she is talking about. I hazard a guess that she is referring to my claim that using a DMM set on DC volts can provide a solid measurement for average DC voltage....despite all the transients.

Quote
I'll have to think about all this body diode stuff a little more before I really take a hard stance one way or the other.  God, I'll bet she's just loving all of this!   :D

Humbugger
One other thing to consider, is if the body diode was doing its job, there could be no continuous oscillation, because the first excursion below zero conducts a current pulse through the body diode into the battery, and that eats up at least 50% of the energy.

If you perform the simulation as per my diagram with an ideal switch (RON=15m, ROFF=1M), you will quickly observe and realize that the only way those oscillations can exist as shown, is if the body diode is either non-existent, or it is severely damaged. Also, almost as important, is the large inductance between the battery and the circuit. Trust me, there is a lot of wire there. ;)

These two conditions need to be met or there is no wave form like we see.

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
Group: Guest
Humbuggy, with all due respect, I can assure you that the gate inductance won't be that high, and the gate resistance not that low. You need to factor in the 50 Ohm output impedance from the FG as well. Having considered these factors, the Gate circuit will have little to no effect on creating the oscillations we are seeing in Rose's wave forms.

Wait just a minute now.  Undamped oscillation requires gain and positive feedback.  You've emphatically stated that there is no way the MOSFET is switching at the oscillation frequency in a post to Milehigh.  There are obvious Q-destroying resistances in the circuit (Rload).  So, if it's not positive feedback into the gate that is causing the oscillations, where is the gain coming from?  How does the circuit oscillate unless there is something causing an added 180 degrees of phase shift to form positive feedback from drain to gate?  

Without use of the gain of the MOSFET, a passive circuit with low Q couldn't ring out like that.  Something is causing the battery to keep pumping energy into the LC tank compriised of the various series inductances and the Coss of the MOSFET.  It is my theory that the MOSFET, while the gate is driven low by the FG, and the drain-source channel is less than fully hard on, is operating in a linear mode as a feedback oscillator.  Otherwise, what could explain the non-diminishing ringing in such a low-Q circuit?  Energy is being added somewhere and it ain't from zipons! I say the gate has a lot to do with it.

Quote
Also, your schematic is incomplete, which is why you are not seeing the oscillations on the BATTERY vs. the Drain. She IS measuring the "battery" voltage, not the Drain. It is plain to see on the bench shot I provided (see short red wire).

See the attached drawing for a more complete (and necessary) rendition of what she actually has there, and what should be included in the simulation.

It looks to me like the point you have labelled Vbatt is suspended between two inductances and is thus neither Vbatt nor Vdrain.  I understand why you drew it that way, assuming her battery wires are ten feet long and that she's got her scope hooked up near the load unit and not the battery.  In the case of continuous oscillations, the volatge there is sort of floating about, isn't it? So why would you be so surprised and sure the body diode is blown if it happened to go below the battery positive terminal?  I don't get that part either.

Quote

The fact that the Battery voltage already goes well below 48V is indication that the body diode is damaged. It appears to be trying to work by evidence of the distortion present there.

.99

So, there is no hurry.  I can live with this confusion.  I obviously need some gentle instruction and hand-holding here.  I am getting lost.  Throw me a lifeline, Poynt!  At your leisure (I'm not actually drowning...just lazy). :P

Humbugger

P.S.  This thing oscillates like a bitch from hell...150W out at 45MHz.  It hasn't blown any body diodes.  So what's the major difference?  The voltage at the drain swings between ground and twice Vdd.  The load out there at the other end of the quarter-wave line is a dynamic plasma resistor inside a laser tube.  The entire trick in getting this to work involved massaging the impedances and phase angle at the gate.  The quarter wave line provides 90 degrees and produces 1KV to light up the load in the first few microseconds then, as the plasma forms, its impedance drops and the quarter-wave line serves as a constant-power regulator while always giving 90 degrees phase shift.  The L and C in the feedback loop plus the Ciss add the final 90 degrees of phase shift.
« Last Edit: 2011-02-19, 04:38:20 by humbugger »
   

Group: Tinkerer
Hero Member
*****

Posts: 3055
Wait just a minute now.  Undamped oscillation requires gain and positive feedback.  You've emphatically stated that there is no way the MOSFET is switching at the oscillation frequency in a post to Milehigh.  There are obvious Q-destroying resistances in the circuit (Rload).  So, if it's not positive feedback into the gate that is causing the oscillations, where is the gain coming from?  How does the circuit oscillate unless there is something causing an added 180 degrees of phase shift to form positive feedback from drain to gate? 

Without use of the gain of the MOSFET, a passive circuit with low Q couldn't ring out like that.  Something is causing the battery to keep pumping energy into the LC tank compriised of the various series inductances and the Coss of the MOSFET.  It is my theory that the MOSFET, while the gate is driven low by the FG, and the drain-source channel is less than fully hard on, is operating in a linear mode as a feedback oscillator.  Otherwise, what could explain the non-diminishing ringing in such a low-Q circuit?  Energy is being added somewhere and it ain't from zipons! I say the gate has a lot to do with it.
 
It looks to me like the point you have labelled Vbatt is suspended between two inductances and is thus neither Vbatt nor Vdrain.  I understand why you drew it that way, assuming her battery wires are ten feet long and that she's got her scope hooked up near the load unit and not the battery.  In the case of continuous oscillations, the volatge there is sort of floating about, isn't it? So why would you be so surprised and sure the body diode is blown if it happened to go below the battery positive terminal?  I don't get that part either.

So, there is no hurry.  I can live with this confusion.  I obviously need some gentle instruction and hand-holding here.  I am getting lost.  Throw me a lifeline, Poynt!  At your leisure (I'm not actually drowning...just lazy). :P

Humbugger

P.S.  This thing oscillates like a bitch from hell...150W out at 45MHz.  It hasn't blown any body diodes.  So what's the major difference?  The voltage at the drain swings between ground and twice Vdd.  The load out there at the other end of the quarter-wave line is a dynamic plasma resistor inside a laser tube.  The entire trick in getting this to work involved massaging the impedances and phase angle at the gate.  The quarter wave line provides 90 degrees and produces 1KV to light up the load in the first few microseconds then, as the plasma forms, its impedance drops and the quarter-wave line serves as a constant-power regulator while always giving 90 degrees phase shift.  The L and C in the feedback loop plus the Ciss add the final 90 degrees of phase shift.


Is it possible for a body diode to fail without impairing
MOSFET integrity?

What is the nature of the vast majority of MOSFET failures?

THIS LINK is a good practical discussion.

Very often, spurious oscillations are very hard to nail down
but are easily prevented.

Unless, the spurious oscillations serve a useful purpose...


---------------------------
For there is nothing hidden that will not be disclosed, and nothing concealed that will not be known or brought out into the open.
   
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down
It seems to me that in the real world there will be some air core coupling or transformer action between the inductance of the driven resistor and any inductance in the gate wiring. I've played with FET's with a little bit of coupling in this manner and it can oscillate very well and at considerable distance between the inductors.

I don't see any air core transformer type coupling in the sims presented but maybe I missed something. Also consider stray capacitive coupling to further muck up the works.


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
Group: Guest

Is it possible for a body diode to fail without impairing
MOSFET integrity?

What is the nature of the vast majority of MOSFET failures?

THIS LINK is a good practical discussion.

Very often, spurious oscillations are very hard to nail down
but are easily prevented.

Unless, the spurious oscillations serve a useful purpose...

But then they wouldn't be spurious anymore, would they?  If oscillations are desired for a purpose, then the circuit should be designed to oscillate at the desired frequency with the least amount of heating in the MOSFET.  No?

Even Rosemary has given up crediting her "semi-random aperiodic oscillations" for her fabulous infinite COP demonstration unit and now has found the far preferred staedy fixed frequency oscillations!   C.C

I think that link is pretty darn complete.  It seems he wrote it well before the current generation of "Avalanche-rated" fifth and sixth generation parts were around.  I say that because he talks about destroying the parts by even a small hit against the breakdown voltage for a brief few nanoseconds.  Today's MOSFETs can handle a fair amount of abuse that way.

Tanks fer da link,

Humburger

P.S.  My vote is that it is fairly unlikely to have a body diode fail open and still have good gate control of the channel and impossible to have it fail shorted and not lose all function.  After all, the body diode isn't called "intrinsic" just for fun.  It's not as if it were a seperate diode soldered in there on purpose, after all.  I've busted a lot of MOSFETs in my day.  The only ones that keep limping along have been gate-oxide punctures due to brief excesses in gate drive voltage.  

Every hard failure has resulted in a D-S short, at least that I've seen.  Except the ones that literally exploded, spraying shards of epoxy in my face and hair.  Those ones often had no trace of a chip left on the lead frame, so we'll never know about them!  I'd say they shorted first and then vaporized.
« Last Edit: 2011-02-19, 13:43:16 by humbugger »
   
Group: Guest
It seems to me that in the real world there will be some air core coupling or transformer action between the inductance of the driven resistor and any inductance in the gate wiring. I've played with FET's with a little bit of coupling in this manner and it can oscillate very well and at considerable distance between the inductors.

I don't see any air core transformer type coupling in the sims presented but maybe I missed something. Also consider stray capacitive coupling to further muck up the works.

Trying to pin down the exact source of MOSFET oscillations without knowing the exact layout and all the parasitics and possible couplings is virtually impossible, in my opinion.  It's a lot easier to just make them go away than to precisely understand why they happen, unless you are setting out to make an oscillator.  Even then it can be quite tricky, that's for sure.  Especially if you want high efficiency and high frequency.

That's why it's hard to converge on agreement regarding Rosemary's particular setup.  Too many unknowns.  Too many possibilities.

Humboogert
   
Group: Guest
http://newlightondarkenergy.blogspot.com/2011/02/60-ta-muchly.html

Blog #60 (link above)

Einstein said "Everything should be made as simple as possible, but not simpler"

I know some don't hold much truck with Einstein here, but I find this saying to be golden.

Oh, how simple it would be, Rosemary, if all sorts of little irritating realities could just be blithely ignored.  If a simple theory could instantly explain all electronic and physical interactions.  If endless energy could be tickled out of hiding with twelve bucks worth of Radio Shack parts.

Rosemary, if you have any real appreciation for all the effort and attention you're getting from us here...and if you are sincere about learning even a tiny bit of anything from all of it, please do us all a big favor and show us your setup.  Up close and personal, with the component values, hookup details, measurement points and physical layout.  Leave nothing to the imagination.  Nail it down.  Pick you best and favorite duty cycle and oscillation pattern so far.  And make some straight-up simple measurements of equivalent DC input power and equivalent DC heating power.

You have a rapt audience, despite our poking of fun and criticisms of your bench prowess and lingo-lacks.  Use us as guinea pigs for a dry run of your "hopefully unequivocal" demonstration/presentation/proof or whatever you want to call it.  Lay it out here...no holds barred, cards on the table.

Sincerely,

Humbugger
   
Group: Guest
Okay...couldn't get to sleep until I played a little more and updated my silly sim to agree with Poynt's as best I could (not having the exact Spice Model for the MOSFET so still using IRF740).

Here are two interesting situations that tend to show that the gate drive arrangement may have more to do with the oscillation tendency than some gave credit for.  Also shows plenty of excursion above and below battery voltage with so-called "Vbatt" scoped at same mid-point of positive terminal battery lead inductance and load inductance as prescribed by Poynt as his best estimate of what Rose is doing.  Body diode not broken.

Top figure shows the system driven by a source square wave gen with a 50 ohm output impedance and using a piece of 50 ohm cable to get to the gate.  The cable can be any length and nothing changes.

Bottom figure shows exact same setup but now the drive cable to the gate is a twisted pair with 150 ohm characteristic impedance.  Note the greatly changed duration of the ringing.  Those who want to simulate this will also find now that the length of the cable has a large effect on the ringing duration and can cause some quite odd modulations of the oscillations at certain lengths.

I couldn't readily get sustained oscillation with these values and no additional couplings or parasitic L's and C's and my MOSFET parasitic C's are not the same as Rose's circuit or Poynt's sim, but I think it can be said tht it's quite clear that as small a detail as the impedance of the drive cable and its length can have a very dramatic effect on the stability and tendency to oscillate for a longer or shorter duration (i.e. the damping factor is radically changed).  

So the impedance placed on the gate by the drive cable reflection is purely resistive when the cable impedance matches the generator impedance.  But when those do not match, the gate end of the cable acts like an L or a C and varies as the length of the cable is adjusted.  This seems to have a big effect on the damping, at least, which implies strongly that the gate impedance definitely changes the stability criteria by way of changing the phase angle of feedback into the gate.

Humbugger
« Last Edit: 2011-02-19, 13:10:45 by humbugger »
   
Group: Guest
It seems to me that in the real world there will be some air core coupling or transformer action between the inductance of the driven resistor and any inductance in the gate wiring. I've played with FET's with a little bit of coupling in this manner and it can oscillate very well and at considerable distance between the inductors.

I don't see any air core transformer type coupling in the sims presented but maybe I missed something. Also consider stray capacitive coupling to further muck up the works.

I think that's all that is needed to get those big continuous oscillations happening.  The simulator shows that its on the verge and minor tweaks of the impedance at the gat can move the feedback phase angle.  add some coupling from that big old coiled up heater element to the drive cable twisted pair and off we go!  Or a myriad of other possible feedback augmentation/phase shifting parasitics and couplings.  Right on ION.


DaBugger
   
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down
POYNT

Could you explain what we are looking at in this picture?....Looks like a steel cased immersion heater.


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
Group: Guest
POYNT

Could you explain what we are looking at in this picture?....Looks like a steel cased immersion heater.

Poynt seems to have the inside track on that photo.  I think that's her new inductive load resistor.  Sure looks kind of stock immersion to me.

It also looks like she's using a slew of real long "sandbox" style wirewound highly inductive resistors for her shunt.  Bad girl!  At least they look like they're in parallel...

The very best (I mean if you can't get a real Kelvin-sensing shunt purpose-designed) is a slew of 1W 2512-size chip resistors in close parallel, like stacked and soldered together.

Hum
   
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down
Poynt seems to have the inside track on that photo.  I think that's her new inductive load resistor.  Sure looks kind of stock immersion to me.

It also looks like she's using a slew of real long "sandbox" style wirewound highly inductive resistors for her shunt.  Bad girl!  At least they look like they're in parallel...

The very best (I mean if you can't get a real Kelvin-sensing shunt purpose-designed) is a slew of 1W 2512-size chip resistors in close parallel, like stacked and soldered together.

Hum

I use carbon rod strapped with copper to get any milliohm shunt I desire. I also use a center tapped arrangement so the resistance temperature drift cancels. Works very well. But I've been known to make my own resistors. Once had a job in a resistor manufacturing plant.

So how do you model all the capacitive coupling and distributed capacitance from the outer steel sheath to the swaged heating element, not to mention the heater ground back into the circuit.

And whatever happened to the special very carefully wound "Ainslie" resistor of precise dimensions?


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Poynt seems to have the inside track on that photo.  I think that's her new inductive load resistor.  Sure looks kind of stock immersion to me.

It also looks like she's using a slew of real long "sandbox" style wirewound highly inductive resistors for her shunt.  Bad girl!  At least they look like they're in parallel...

The very best (I mean if you can't get a real Kelvin-sensing shunt purpose-designed) is a slew of 1W 2512-size chip resistors in close parallel, like stacked and soldered together.

Hum

Yes, I am assuming this is the setup she is currently testing on. This was a photo she uploaded some time ago to OU or somewhere, I don't recall.

Yes, 5x 1.25 Ohm resistors in parallel would yield 0.25 Ohms for her CSR. They do look wire-wound, but there are non-inductive versions available using the Aryton-Perry winding method, so I will giver her the benefit of the doubt on that one.

The load seems to be some type of water or liquid heater. I wonder if anyone could dig up the typical specs on such an item? Being in South Africa, I suppose their line voltage is 220V, so that would help perhaps. I took a guess at 4 Ohms and about 20uH to 40uH inductance, which includes the wiring to and from the resistor. All in all, the actual values are not too critical for this discussion as I am quite confident we are in the ballpark in terms of the component values. This real issue is what is causing these apparent constant amplitude oscillations.

3 probes; one on "Vbat", one on the CSR shunt, and one on the function generator output before the gate resistor. Pretty basic.

Note that there is a considerable length of wire leading to the batteries on the floor (not seen but assumed). If she has 4 batteries wired in series, then there can easily be 48 inches of wire in each leg. Even if we use 30nH per inch, there is a significant amount of inductance there, and this is one of the contributing factors explaining what we are seeing in her wave forms.

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
@Dumped:

That link is poor in terms of his description of the MOSFET internal body diode and his description of a free-wheeling (flyback) diode. He claims that the MOSFET body diode is a free-wheeling diode "built-in" for free. INCORRECT. They are not the same and do not perform the same function. A MOSFET driving an inductive load without a flyback diode across the inductor, will be hit by huge voltage spikes on its Drain pin. The body diode ONLY conducts under two conditions; 1) if VD goes negative wrt the Source, and 2) if VD exceeds the body diode (really a NPN transistor) avalanche voltage. In that latter sense, it performs a similar job to what a fb diode would do, but it clamps the voltage at the avalanche voltage (very high) vs. at a few volts higher than V+ as a fb diode would.

.99



---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
Pages: 1 2 3 4 5 6 [7] 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 ... 29
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2024-11-27, 22:37:10