Tweet
As this deals primarily with the inductive spike I placed it in this forum though it might possibly fit
better in a more general forum. Like many others I have given some thought to the inductive spike for a bit of time now (hopefully the expression "there you go thinking again" won't fit too well here). In any event, I have a model which is conventional and appears to fit quite well for many observations though with at least one glaring omission and another problem. If valid, at least to a point, the model might be useful in design issues.
So in explaining the inductive spike we look first back to Faraday, (and leaving aside the Faraday disc issue) we conventionally note that electricity is produced in a conductor when exposed to a changing magnetic field. So we now charge an electromagnet and abruptly cut off the current supply. The abruptness of the cutoff will depend on the abruptness of the switch (i.e relay, BJT, Mosfet) and possibly, I don't know (I am not being sarcastic), on the resistance and turns of the coil. So what happens now? Well starting from the Bloch wall of the electromagnet the coil sees a very rapid change in electromagnetic flux, as the wire is a conductor, per Faraday, we see then a one off high voltage spike. the collapse of the N pole leading to one spike and the S to an opposite spike. What happens next? The two voltage spikes head down the coil in opposite directions to the end of the wire "bounce?" off the end of the wire and head back, repeating this until damped out. This is the "ringing" of the coil and unless I misunderstood, somewhere I believe I saw Bedini show a close-up of the inductive spike on an oscilloscope showing it "ringing" as numerous decreasing spikes. So you have an AC current which "rings" internal to the coil and which dies out at a rate dependent on the Q factor of the electromagnet. This AC current is also what shows up on a meter.
Now if one puts a diode on one end of the coil what happens? The diode will allow either the spike from the collapsed N field or S field to pass but not both. So you can then ala the SSG run one of these spikes into a cap thence to circuit ground and charge the cap. What if you put two diodes facing opposite directions (ala AV plug or bridge rectifier) at one end of the coil? The N spike will pass down one diode and S spike down the other and from the end of the two diodes (from the single end of the coil) you now have positive and negative poles that you can again hook a cap up to and charge the cap. This model also explains a couple things which almost had me questioning my sanity (well not really but it was weird) 1) when I measured the time to charge a specific cap using the one wire set-up it charged much faster when I left the meter hooked up then if I charged it then later hooked up the meter (what was this some sort of Heisenberg nonsense) 2) then I saw that the cap charged faster even if I just hooked an alligator clip jumper lead to either of the diodes. If I hooked up more jumper leads it charged faster (as long as the lead wasn't itself coiled) and not small changes either, it might charge 3x faster with enough jumpers hooked up. So what was going on? Possibly two things but the more important I suspect might be if you think of the diode as a one way road sign, when you just have the diode there is only a small space left for the one way traffic to flow on. When you lengthen the end of the diode there is more road for traffic to flow on as it were and the cap can capture more of this. The second thing going on is that when you add more wire, a meter or a load like a cap you are changing the circuit. It might be noted then when one captures the inductive spike, consider for instance a single pulse of the coil, all of the energy being captured from the spike happens after the electromagnet has been shut off. The captured energy does not directly affect the input energy. However, if you now charge the coil a second time, the circuit is different, the cap is hooked up and there is a certain voltage in the cap. This affects the properties of the electromagnet. I don't know what it is really changing (inductance? resistance?) all I find is that input energy changes (increases/decreases) without any rhyme or reason that I can find depending on what the load is. That's something I would like to understand better.
Okay before going to objections, design issues that I see from this model. 1) You want a coil with the highest inductance for the least amount of input energy (i.e. lots of turns, high permeability core) 1a) I need to think more about litz wire 2) You want to charge the coil just to the point of being fully charged then discharge it, leaving the circuit on with a charged coil is wasted energy (i.e. optimizing duty cycle). 3) With a duty cycle of 50 percent the amount of input energy in charging a coil for a half second then discharging for a half second should be the same as charging and discharging 5,000 or 50,000 times a second while the number of spikes to be gathered would be different. The pulse hertz is dictated by maximum switch speed (relay, transistor, spark gap) and the "relaxation time" don't know if that is a real term, of the electromagnet. An iron core apparently has a slower maximum discharge rate than an air core. Now the magnetic strength from an air core electromagnet is much less than an iron core, however, interestingly, the inductive spike is not proportionately less, though it is apparently the change in magnetic flux which causes the spike. I attribute this to magnetic field strength decreasing as a function of the square of distance. So a disproportionate amount of the spike is caused by the change in field strength right next to the conductive wire not at the center of the core. I don't know how coil inductance and wire resistance (aside from the core) affect electromagnet relaxation time, i.e. what tells you how fast you can pulse an air core electromagnet. I also may be quite wrong on some or all of this but it is my best guess. Lastly, as a real quick aside, I wish I understood what Tesla was getting at with his style pancake coil where he talks about inter wire interactions leading to a very large capacitance. While I don't get it, I would also just note that when you look at the original design of the joule thief the wires are wound side by side in a manner analogous to a Tesla Pancake coil so I want to look more at that and maybe try an air coil or Tesla pancake coil JT and see what happens.
Objections:
1) If you pulse a coil with a triangle or sine wave you still get, though smaller, an inductive spike. I suppose that could be explained that the abruptness of the change in magnetic flux is not as important as I thought or more likely that even with a square wave pulse you still are getting a less than instantaneous change in flux because of the properties of the electromagnet.
2) Not really an objection but with one wire transmission if you hook the other end up to earth ground you get a noticeable improvement (i.e. as Tesla recommends on the Tesla coil). Have no idea what is going on there.
3) This one seems a deal breaker. So if you use two diodes in opposite orientations at one end of a coil you are letting positive electricity through one diode and negative through the other and can charge a cap. Now if you take say the diode which let positive through and attach two diodes in opposite orientations (AV plug) to it, you again get a positive and negative off this second AV plug which will charge a cap. How the heck does that happen? In fact if you take a diode attach it to the end of a coil, then in series attach a second diode pointing the other way -><- then attach an AV plug to this, you still get a positive and negative pole that can charge a cap. What electricity flows through two diodes in series in opposite orientations? So as I said I have no idea, but I suppose I may start calling the inductive spike a radiant spike from now on.
better in a more general forum. Like many others I have given some thought to the inductive spike for a bit of time now (hopefully the expression "there you go thinking again" won't fit too well here). In any event, I have a model which is conventional and appears to fit quite well for many observations though with at least one glaring omission and another problem. If valid, at least to a point, the model might be useful in design issues.
So in explaining the inductive spike we look first back to Faraday, (and leaving aside the Faraday disc issue) we conventionally note that electricity is produced in a conductor when exposed to a changing magnetic field. So we now charge an electromagnet and abruptly cut off the current supply. The abruptness of the cutoff will depend on the abruptness of the switch (i.e relay, BJT, Mosfet) and possibly, I don't know (I am not being sarcastic), on the resistance and turns of the coil. So what happens now? Well starting from the Bloch wall of the electromagnet the coil sees a very rapid change in electromagnetic flux, as the wire is a conductor, per Faraday, we see then a one off high voltage spike. the collapse of the N pole leading to one spike and the S to an opposite spike. What happens next? The two voltage spikes head down the coil in opposite directions to the end of the wire "bounce?" off the end of the wire and head back, repeating this until damped out. This is the "ringing" of the coil and unless I misunderstood, somewhere I believe I saw Bedini show a close-up of the inductive spike on an oscilloscope showing it "ringing" as numerous decreasing spikes. So you have an AC current which "rings" internal to the coil and which dies out at a rate dependent on the Q factor of the electromagnet. This AC current is also what shows up on a meter.
Now if one puts a diode on one end of the coil what happens? The diode will allow either the spike from the collapsed N field or S field to pass but not both. So you can then ala the SSG run one of these spikes into a cap thence to circuit ground and charge the cap. What if you put two diodes facing opposite directions (ala AV plug or bridge rectifier) at one end of the coil? The N spike will pass down one diode and S spike down the other and from the end of the two diodes (from the single end of the coil) you now have positive and negative poles that you can again hook a cap up to and charge the cap. This model also explains a couple things which almost had me questioning my sanity (well not really but it was weird) 1) when I measured the time to charge a specific cap using the one wire set-up it charged much faster when I left the meter hooked up then if I charged it then later hooked up the meter (what was this some sort of Heisenberg nonsense) 2) then I saw that the cap charged faster even if I just hooked an alligator clip jumper lead to either of the diodes. If I hooked up more jumper leads it charged faster (as long as the lead wasn't itself coiled) and not small changes either, it might charge 3x faster with enough jumpers hooked up. So what was going on? Possibly two things but the more important I suspect might be if you think of the diode as a one way road sign, when you just have the diode there is only a small space left for the one way traffic to flow on. When you lengthen the end of the diode there is more road for traffic to flow on as it were and the cap can capture more of this. The second thing going on is that when you add more wire, a meter or a load like a cap you are changing the circuit. It might be noted then when one captures the inductive spike, consider for instance a single pulse of the coil, all of the energy being captured from the spike happens after the electromagnet has been shut off. The captured energy does not directly affect the input energy. However, if you now charge the coil a second time, the circuit is different, the cap is hooked up and there is a certain voltage in the cap. This affects the properties of the electromagnet. I don't know what it is really changing (inductance? resistance?) all I find is that input energy changes (increases/decreases) without any rhyme or reason that I can find depending on what the load is. That's something I would like to understand better.
Okay before going to objections, design issues that I see from this model. 1) You want a coil with the highest inductance for the least amount of input energy (i.e. lots of turns, high permeability core) 1a) I need to think more about litz wire 2) You want to charge the coil just to the point of being fully charged then discharge it, leaving the circuit on with a charged coil is wasted energy (i.e. optimizing duty cycle). 3) With a duty cycle of 50 percent the amount of input energy in charging a coil for a half second then discharging for a half second should be the same as charging and discharging 5,000 or 50,000 times a second while the number of spikes to be gathered would be different. The pulse hertz is dictated by maximum switch speed (relay, transistor, spark gap) and the "relaxation time" don't know if that is a real term, of the electromagnet. An iron core apparently has a slower maximum discharge rate than an air core. Now the magnetic strength from an air core electromagnet is much less than an iron core, however, interestingly, the inductive spike is not proportionately less, though it is apparently the change in magnetic flux which causes the spike. I attribute this to magnetic field strength decreasing as a function of the square of distance. So a disproportionate amount of the spike is caused by the change in field strength right next to the conductive wire not at the center of the core. I don't know how coil inductance and wire resistance (aside from the core) affect electromagnet relaxation time, i.e. what tells you how fast you can pulse an air core electromagnet. I also may be quite wrong on some or all of this but it is my best guess. Lastly, as a real quick aside, I wish I understood what Tesla was getting at with his style pancake coil where he talks about inter wire interactions leading to a very large capacitance. While I don't get it, I would also just note that when you look at the original design of the joule thief the wires are wound side by side in a manner analogous to a Tesla Pancake coil so I want to look more at that and maybe try an air coil or Tesla pancake coil JT and see what happens.
Objections:
1) If you pulse a coil with a triangle or sine wave you still get, though smaller, an inductive spike. I suppose that could be explained that the abruptness of the change in magnetic flux is not as important as I thought or more likely that even with a square wave pulse you still are getting a less than instantaneous change in flux because of the properties of the electromagnet.
2) Not really an objection but with one wire transmission if you hook the other end up to earth ground you get a noticeable improvement (i.e. as Tesla recommends on the Tesla coil). Have no idea what is going on there.
3) This one seems a deal breaker. So if you use two diodes in opposite orientations at one end of a coil you are letting positive electricity through one diode and negative through the other and can charge a cap. Now if you take say the diode which let positive through and attach two diodes in opposite orientations (AV plug) to it, you again get a positive and negative off this second AV plug which will charge a cap. How the heck does that happen? In fact if you take a diode attach it to the end of a coil, then in series attach a second diode pointing the other way -><- then attach an AV plug to this, you still get a positive and negative pole that can charge a cap. What electricity flows through two diodes in series in opposite orientations? So as I said I have no idea, but I suppose I may start calling the inductive spike a radiant spike from now on.