Today I did some experiments to see what the influence magnets have on a coils flyback.
To keep this short I will focus on the only way I found that increased the flyback voltage.
The best result was placing a 2" cube neo on both sides of a coils soft iron core : SN | SN | SN
I connected a 12vdc car battery to an IRF840 switching the positive line.
I drove the IRF840 directly with a 10v square wave from an HP Function Generator at 5000hz
Bear in mind the IRF840 has an internal free wheeling diode..
I placed a Pearson wide band current probe in line with the coil and a x100 probe across the coil.
From my tests it looks like there is a significant voltage increase with the magnets present in the SN | SN | SN setup
SN | NS | SN and SN | SN | NS both decreased the flyback voltage.
coil 1:
20 awg high ohm coil on soft iron core
168v without magnets
212v with magnets
44v increase
coil 2:
20 awg low ohm coil on soft iron core
395v without magnets
540v with magnets
145v increase
At certain frequencies below 5000hz there were some anomalous current spikes that I need to go back and look at, in fact one killed two of the mosfets.
Anyone want to independently verify these results???
NOTE: This requires a really solid mounting for the 2" neo cubes and big neos are very dangerous to work with! Be careful!!
Thanks..
Darkspeed
Hi ds
Sounds interesting, yes i would be up for a rep, but will need to order some magnets to try, but first i need to get my current stuff sorted, i have loads on at the moment.
Those are large magnets, can it be scaled down a bit maybe 1 inch square magnets.
anyway sounds interesting, keep at it.
Peter
Hello darkspeed
This sounds very interesting! I have a question, what did the amperage do? Did the draw decrease or increase when using the neos? Also did the current change on the output in any significant way?
Peter, I will try it with the next smaller neos I have ( 1 x 0.75 cylinders ) and see if I get anything..
stprue, I kept the current probe visible on one of the scope channels and since the coil is firing with the neos and not against them there was no additional current draw in the coil.
The only strange current issues were the random current spikes below 5000hz.
Im going to do some more work on it today
ds
What are your core dimensions and how many turns was your coil.
Core is a 17mm x 17mm x 75mm soft iron core..
I will have to go measure the windings as these were a couple of coils i pulled off the shelf that I had built a while back.
The only part of the event that is opposing the polarity of the setup is the flyback and I am not sure why a static opposing magnetic field increases the energy in a flyback event??
Replicated the result with 1 x 0.75 n42 neo cylinders
42v increase with two cylinder magnets on each side of core ( compared to the 145v increase with the 2" cubes )
coil : 2.2311mh @ 10khz 2.54ohm @ 100hz / 93.69ohm @ 10khz
I just started from scratch with a different coil / different HP / different scope and was able to reproduce the results
I also just had a friend report that he could get the same results.
Is this a known effect that I somehow just missed before??
Anyone know of this??
Iv been working with earplugs all day because the 5-6khz tone will rive you insane..
Darkspeed:
You may be a bit surprised to learn that the magnitude of the spikes has nothing directly to do with the magnets. Somehow the magnets are causing secondary effects that are influencing the main parameters that affect the maximum spike voltage.
The first parameter is the amount of current flowing through your coil. There is no need to switch the square wave to the MOSFET gate at 5 or 10 KHz. Try 50 Hz instead. This will likely ensure that you have the maximum possible current running through your coil when you switch the MOSFET off and you will not go insane from the high frequencies that you are suffering though now.
The second parameter is the speed the MOSFET switches off. The cleaner and sharper you can make your square wave to the gate the higher the spike voltages you will get.
The third parameter is the inductance of your coil, the higher the inductance the higher the voltage spike.
Your external magnets are somehow affecting the above three parameters to give you increased voltage on your spikes.
The formula for the spike voltage is v = L di/dt. You should try searching on that.
Careful if you start tuning and getting higher and higher spikes I assume that you will likely blow your MOSFET.
MileHigh
Thanks MileHigh!
That reminded me to do something I forgot to do... remeasure the inductance with the magnets in place..
That may be effecting it.. I will check it out..
The best result so far has been feeding the HP through a Wavetek so I could adjust the duty cycle.
Setting it to 5khz on the HP and 10us on the Wavetek, then inverting the output, so 5000 times per second the mosfet turns off for 10us
Ok.. Now I feel like an idiot.. I was not expecting it to change the resistance of the coil..
Resistance without magnets 2.54ohm @ 100hz / 93.69ohm @ 10khz
Resistance with magnets 1.93ohm @ 100hz / 43.14ohm @ 10khz
Inductance without magnets 9.268mh @ 100hz / 2.5528mh @ 10khz
Inductance with magnets 3.844mh @ 100hz / 1.2589mh @ 10khz
The resistance and the inductance drops with the magnets
Cyril Smith (PhD. Physics? - he's associated with MPI) posted something about this a few years ago.
Might be in one his articles here:
http://74.125.113.132/search?q=cache:SAeqsKBrsqUJ:jnaudin.free.fr/html/meg.htm+cyril+smith+magnet+coil&cd=1&hl=en&ct=clnk&gl=us
Is the inducatance and resistance different with the polarity of the magnets reversed?
(I see that the differences with and without the magnet are nonlinear.)
QuoteThe resistance and the inductance drops with the magnets
Generally we speak of resistance at a particular frequency as "impedance", otherwise some might get confused and infer an actual change in DC resistance with the magnets. This is not the case.
Nice to see that your inductance bridge allows excitation frequency selection, makes impedance plots easier.
Nice work ds, you didn't understand the effect and now you do, as do we, thanks
This gives me another test to add to my pulse experiments.
Peter
Quote from: Grumpy on 2010.02.01, 13:59:01
Is the inducatance and resistance different with the polarity of the magnets reversed?
(I see that the differences with and without the magnet are nonlinear.)
Good question.. I tried that and both the resistance and inductance stays the same when i switch polarity... but with opposing magnets on each side both rise..
got me really thinking about the orbo and how this works in a toroid
Haha yes ION "impedance" im bad about that, just like i call neons spark gaps and bulbs resistors lol...
Quote from: Peterae on 2010.02.01, 16:10:47
Nice work ds, you didn't understand the effect and now you do, as do we, thanks
This gives me another test to add to my pulse experiments.
Peter
Thanks Peter, im the master at learning stuff the hard way..
Quote from: darkspeed on 2010.02.01, 17:31:13
Quote from: Grumpy on 2010.02.01, 13:59:01
Is the inducatance and resistance different with the polarity of the magnets reversed?
(I see that the differences with and without the magnet are nonlinear.)
Good question.. I tried that and both the resistance and inductance stays the same when i switch polarity... but with opposing magnets on each side both rise..
got me really thinking about the orbo and how this works in a toroid
write this down and make a diagram
can you change the impedance of a toroidal transformer with a magnet? - sure you can
EDIT:
What "is" the imaginary part of impedance?
write this down and make a diagram ??
are you talking self inductance?
When I think orbo I think ed leedskin magnet
Quote from: darkspeed on 2010.02.01, 18:22:35
write this down and make a diagram ??
are you talking self inductance?
I'm talking about how the magneici effects the impedance and inductance
I wonder if it will effect the permittivity of a dielectric? Hmm - got a doorknob or other plate cap around?
In answer to the question about imaginary part of impedance: http://hyperphysics.phy-astr.gsu.edu/Hbase/electric/impcom.html
I have a case of 40kv 2000pf doorknobs - outline what you have in mind
Quote from: darkspeed on 2010.02.01, 19:02:39
I have a case of 40kv 2000pf doorknobs - outline what you have in mind
actually forget the doorknobs
Do you have vacuum caps?
See if a magnetic field alters the permitivity of free-space within the vacuum cap.
http://74.125.95.132/search?q=cache:nJYwn_GMsLIJ:hyperphysics.phy-astr.gsu.edu/HBASE/electric/elefie.html+permittivity+%22magnetic+field%22&cd=1&hl=en&ct=clnk&gl=us
I actually dont have any vacuum caps..
If I ordered one any advice on pf to get..
http://cp.literature.agilent.com/litweb/pdf/5980-2862EN.pdf (http://cp.literature.agilent.com/litweb/pdf/5980-2862EN.pdf)
since this is jsut a test for an "indication" of a change, any size should do.
You could try pumping out a container and making a simple one. The vac caps that I have a concentric cofiguration rather than plate so the sirection of the gradient between the plates is different.
if a magnetic field compresses space then I would think that the permittivity of a vacuum cap would change and this would be measured as a cange in capacitance
But what if the magnetic field only changes the permittivity of space while the magnetic field is changing?
it was never my intention to start a discussion.
i just wanted to show Grumpy the circuit.
I realize now that i made a big mistake by posting and i appologize for that.
Microcontroller:
You made your first posting, that's great. Honestly I can't really decipher the process that you are describing. You show two diagrams and you description of what's going on is very limited and you don't indicate what diagram you are talking about.
However, let's just talk about dumping a charged capacitor into a coil or vice-versa using a diode. Let's assume that no energy is lost in the diode to make it simpler.
The energy in an inductor is 1/2*L*i-squared.
The energy in a capacitor is 1/2*C*v-squared
Since the energy is going to be the same in both cases, you can say:
1/2*L*i-squared = 1/2*C*v-squared
So if you rearrange the equation for the voltage on the capacitor you get:
v = sqrt(L*i-squared/C)
So so if the inductor discharges into the capacitor, then the smaller the cap, the higher the voltage you can get across the cap.
This should not be confused with the very high voltage spike you can get when you open-circuit an inductor that has current flowing through it. In fact, if you have a large capacitor, then the same coil that can produce a very high voltage when you open-circuit it might only charge the large capacitor to 0.1 volts. There will be no spike to speak of, just a small bump in the large capacitor's voltage.
MileHigh
Microcontroller:
Just for fun, I will tell you what will happen with your first diagram. I saw your comments in the chat so I figure your description in the posting was about the first diagram.
I assume that when you close S1 that current will flow from C1 through D1 into L1.
Let's also assume that the diodes are ideal for the fun of it.
So you start with C1 charged up with voltage. In this case if you close S1 and keep it closed for a long time, then you will end up with C1 charged up again, but the voltage on C1 will be the opposite polarity. So you end up reversing the voltage on C1. That's it, the circuit will do nothing after this if you open and close S1.
Same initial setup again, but this time let's assume that you open S1 the moment the voltage across C1 is zero. In this case let's pretend that the diodes are a bit more like real diodes. So the moment you open the switch L1 will generate super-high voltage that's trying to push its way through C2 and D2. Let's assume that you "blow up" D2 and it makes a firecracker noise. You end up with something like C2 at about half the voltage of C1 and a blown D2 diode. Same ending for S1, the circuit will do nothing after this if you open and close S1.
That's the way the cookie crumbles!
MileHigh
I think perhaps this thread was misnamed from the beginning. There are no anomalies being discussed anywhere in the thread. There are attempts at optimizing inductive switching circuits, and those skilled in the art are certainly advanced at squeezing the last bit of efficiency from such circuits using newer topologies and such things as active snubbing, synchronous rectification etc. etc. to name just a few.
But we see none of that here. There is no anomaly in a boost converter. It is well defined technology.
The thread started with the effects of magnets placed on a switched inductor. There were no quantitative measurements, just a claim that it (the voltage) was higher with one configuration vs. an alternate configuration.
Magnetically biased inductors are also well known and used extensively in video monitors having CRT's. There are numerous patents and literature regarding MBI's. It has been well studied.
Still no anomaly.
a·nom·a·ly (-nm-l)
n. pl. a·nom·a·lies
1. Deviation or departure from the normal or common order, form, or rule.
2. One that is peculiar, irregular, abnormal, or difficult to classify:
Unlike others, I'm jacked into the Matrix...LOL!
This is a lesson in simplicity, not some textbook lesson, or new discovery.
This method requires time to build up to a high voltage, a number of pulses per unit time.
you might switch the low voltage at a high freq and then switch the HV at a lower rate
Remember we are after voltage not power.
if the hv ckt is of high impedance and the pulse fast, very little current flows, you get your pulse effect, and your hv cap charges right back up.
flyback supply,a napkin and shoestring...
did you loop back to tell the low voltage when to turn off?
need a switch that requires very little current like an overvoltage type
free-running avalanche diodes...hmm 1n4007 av at about 1450v lower for the rest of that series down to a few 100v's for 1n4001's. you could use different one for different settings. Very robust if current is low - and small. Avalanching transistors of the TPU times (90's) could do 200v give or take, some to 500v - again - small and robust of current is low.
Triggering requires control and energy, don't need it if you are always triggering the same rate - find the rate - carve it in ckts
Well I know that I am usually out of sorts and I am very textbook. My view is that if you look at a very simple circuit the the best case scenario would be for everybody in the group on the thread to understand it - really understand it. That is "foundation building" and hopefully you can build from that. I am nearly 100% confident that my description of how the now-gone first Microcontroller circuit works was correct. Too bad Microcontroller seemed to take offense. You certainly have to worry about the timing of the opening of the switch.
Somewhere I wrote down on my computer the freebie Spice simulation software that Poynt uses. As you know I am just dabbling here and sticking my nose in. But as a suggestion for the people investigating the AVEC/TPU stuff, it would be great if you passed around some pSpice models and played with them.
I agree with Ion that when you look at small circuits there are no anomalies. And at the same time there are limits to what you can visualize in your head. Even for Microcontroller's remaining second circuit it gets hard to visualize exactly how it works because of the transformer coupling and all that. pSpice would be great for that example. Then when you look at it run in pSpice that tends to makes things 'click' in your brain too and then you can "keep navigating" further.
MileHigh
Mission accomplished.
Seek truth not faults and you will be better for it.