Announcement

Collapse
No announcement yet.

Investigations into the Radient Spike

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Investigations into the Radient Spike

    Posted this first video on a previous thread detailing efforts on a solid state charger but realized I am heading off into basic research about the spike so am devoting a thread to the topic. If you are curious from a theoretical standpoint why shorter pulses are good this video is for you.
    https://www.youtube.com/watch?v=5wzK5DzX8Jo

    The second video covers more theory and spreadsheet work and sets the stage for gathering data.
    https://www.youtube.com/watch?v=BATRjDO-rbA

    Ideally, I will look to post some experimental data in near future.

  • #2
    Another video on inductors as I am mapping out the sorts of experiments I will look to do. One correction, I was saying the absolute ratio of input charge to output emf would be four times as good when you cut the pulse duration in half. I forgot to factor in that the emf would persist for twice as long with the twice as long pulse so "only" twice as good not four times. Some jeez, years back now I guess, I picked up a large prewound bifilar coil from r-charge or whatever it was called. What I consistently saw was I could never get a great inductive pulse to end up in a cap from that coil, about the same as most other coils, at the same time whenever I pulsed the damn thing it would always end up knocking out my wireless mouse and keyboard and whatnot so I knew there was a tremendous emf pulse occurring that I just wasn't getting at. With this video I would say I think I understand more clearly what was going on and I am looking forward now to gathering data. https://youtu.be/VXdYqVrQJF8

    Comment


    • #3
      An update: The spreadsheet work in the above video, based from the equations provided by the Texas Instruments sponsored Electronics Tutorials, is correct from all I can see. Nonetheless, I have gone to town on gathering data, and it is wrong (the spreadsheet work, not the data). Yet, once again things go off the rails into a parallel universe "your agonizer please Mr. Kyle" type scenario. https://www.youtube.com/watch?v=1ydGv5XId4g. It is indeed again confirmed that things do improve with shorter pulses, yet it does not follow the absolute time value for coils that are not close to saturation as postulated in the above video. Slower coils, with greater inductance and less resistance do better for a given absolute length time pulse and I am not talking about oh is that a few percent better, I mean an order of magnitude type difference better. I looked at batteries, dc power supply and capacitors as input power, always the same trend, looked at paralleling the diode and transistor, no appreciable change. Could still be missing something what with the arduino, but no I think >inductance < resistance improves things. This also "feels" right from previous time I have spent tapping out spikes on different coils by hand. So I shouldn't be overly dramatic the concept of shorter pulses better is once again confirmed. The hypothesis that how much better things get with shorter pulses is independent of the inductor if you are not close to saturation is spectacularly wrong, though that is what one comes up with from the textbook equations. The Li-ion batteries also didn't behave at all the way I thought they would, basically they stunk for low amp fast pulses, don't know if this is just them or would be similar with other battery types. Basically when you got down to real quick low amp draw the batteries were just saying meh, get back to me when there is a real job. I imagine they could switch quicker if they were seeing a higher amp draw, but that is dependent on the coil you are using. Really to be on the safe side, might be best just to throw the whole kit and kaboodle in a cap before pulsing the coil. At least with Li-ion you either need a higher amp draw or a longer pulse time before the thing decides to get busy. Will try and post some video detailing what I looked at later, and for the "heck" of it (can't curse on this site, amen) may look at things at 36 volts instead of 12, though am pretty sure I have a decent idea what I will see. Have a "darn" good idea what two coils I will order next and after they arrive will have more data to share.

      Comment


      • #4
        A synopsis of the first tranche of data. I might have made it more clear at the start of the video I am doing a single pulse of a coil, "Give me a ping Vasily, one ping only please" https://www.youtube.com/watch?v=lw5Tv8pTrfY The inductive spike from this pulse is rectified with a diode and placed in a cap, that is the data being gathered. Otw, should all be clear.

        https://www.youtube.com/watch?v=2BZgKEMAD-0

        Comment


        • #5
          A lot of things are falling into place as it were and I just want to get them down in writing for myself and others, there is no appended video. Put something important together recently, in fact I would say this is the most important thing concerning energy, some might say only thing concerning energy, (gosh I hope it's correct) that I've to date contributed to this forum. I don't know if there is a clique where this is all old news and they are either unwilling or unable to share it or if they never formalized the concept. But you're wrong about me, I do share, I'm nice that way (Nick Fury: Winter Soldier). Seeing as no one has come out and just said the following, though I suspect some people know this, certainly Tesla and John did, and seeing as my middle name is Duffy, I hereby decree the following to be "Duff's Law". Not contrariwise, I would find reference to "The First of Law of Duff" acceptable as well. So this applies to a single pulse of an inductor where the the inductive/radiant spike off the inductor is rectified by diode and placed in a capacitor or other storage medium.

          "For a given time length pulse of sufficiently brief duration as to not approach saturating the coil, and for a given input voltage, the ratio of the input charge expended to the output charge collected is directly proportional to the ratio of the inductance of the coil to its resistance"

          As L/R defines the time constant of the inductor another way of saying this is the efficiency of your input charge to output charge is determined by the time constant of the coil. Aside from seeing this showing up experimentally, I'll confess only on two coils that I've kept data on so far, how do I know this is true? It follows directly from Faraday's Law of Induction. EMF is the product of magnetic flux per unit time, change in flux density determined by coil Henries, rate of change by coil resistance. This is nothing more than a corollary of Faraday's law of Induction. What do we see from this plugging in some real world numbers. Consider two coils from Remington Industries (aside: Remington Industries great wire, don't have both ends of the wire available on their shipped coils grrrr, Temco FTW) then:

          Coil 1) weight 1 lb, gauge 22, wire length 507 ft, resistance per 1000 feet 16.14 ohms.
          Coil 2) weight 1 lb, gauge 24, wire length 803 ft, resistance per 1000 feet 25.67 ohms.

          Now if each length of wire is put onto a coil where one wind equals one foot, coil 1 will have 507 winds and coil 2, 803. Inductance is proportional to the square of the winds. There are 803/507 1.58 more winds in coil 2 then coil 1, hence if the inductance of coil 1 is declared 1, coil 2 will have 1.58*1.58 = 2.509 times more inductance. Now let's consider the resistance in each coil, for coil 1 (507/1000)*16.14 = 8.18 ohms, for coil 2 (803/1000)*25.67 = 20.61 ohms. Let's revisit each coil

          Coil 1) Inductance "1" resistance 8.18 ohms 1/8.18 = 0.122
          Coil 2) Inductance "2.509" resistance 20.61 ohms 2.509/20.61 = 0.122

          These two coils have the same ratio of inductance/resistance, they have the same inductor time constant. For a given pulse length and a given voltage the ratio of input charge to collected charge will be identical in both coils, the only difference is there will be a larger pulse with the lower resistance coil, but input/output ratio identical. Are there any questions?/there should be no questions at this time. Sorry, haha, that was a favorite line of an Army ranger I went to Med school with, I would imagine referring to his time in training.

          Now let's do the same exercise and double the weight of coil 1.

          Coil 1) weight 2 lb, gauge 22, wire length 1014 ft, resistance per 1000 feet 16.14 ohms.
          Coil 2) weight 1 lb, gauge 24, wire length 803 ft, resistance per 1000 feet 25.67 ohms.

          Doing the same calculations as previously:
          Coil 1) Inductance "1.595" resistance 16.36 ohms L/R = 0.0975
          Coil 2) Inductance "1" resistance 20.61 ohms L/R = 0.485

          Please note I've set induction as "1" as two different arbitrary figures we are concerned with the ratio of inductance between the two coils as determined by turns squared, however in this case the ratio of inductance to resistance is twice as high in coil 1 (0.0975/0.485) = 2.0. Coil 1 is now twice as slow, the time constant is twice as high, and per Duff's Law the ratio of output charge to input charge is twice as high. Don't quote me but I think I read either Tesla or someone talking about Tesla saying it is the mass of the coils that is important. That is what is seen here. With each doubling of mass the ratio of input charge to output charge halves.

          I previously posited that for each halving of the time of the input spike the ratio of input charge to output charge should halve and this finding is only constrained by the limits of ones components. With a single spike I have confirmed this down to 3 uSec and one sees an absurd excess of output charge. The problem is with a short spike and a brief single pulse one ends up with seeing 0.01 V in a cap with an input of 12V. Well no problem just have a constant pulse train. Nuh, uh, uh, ahhhhhh. In the two coils I have looked at there is a sweet spot around 1/16th to 1/32nd the inductor time constant. Faster than that, things first start to go off the rails then just go kerfluey, meaning while one pulse looks great, you try pulsing fast there, 1) you can't transform voltage and 2) amp draw goes through the roof compared to what it "should" be theoretically. I've been puzzling about this like the Grinch on Christmas morning, well, okay let's not get carried away, maybe not like that, but I've been thinking about it and what I strongly suspect now is the component that prevents you from having fast pulse trains is the inductor itself, namely the skin effect and proximity effect within the inductor. Are there ways to mitigate this yes, Litz wire, pancake coil, perhaps, perhaps, eliminate with counterwound pancake. I'm not going there yet. I have had two cols where the coil starts to go kerfluey about at the same point in terms of the pulse length in relation to the coil Tau. Now then, if the place where the skin effects/proximity effects doesn't start migrating with a more massive coil, as it didn't seem to with the two coils I've looked at so far then good. That is to say if the proximity effect checks in twice as soon, doesn't matter if your coil is twice as good, I don't think the proximity effect will check in twice as soon from the little I've seen experimentally and from considering things like Big Eureka/Wydencliff tower. While there may be more elegant ways to approach it, if the "kerfluey point" in relation to the Tau, doesn't migrate you can really just brute force the thing. "Big Eureka!" Jeez man zip it up, I think he had a half ton of wire or something though, so would be hard for me to have my Big, Big, Big, Big Eureka, so maybe a its "really pretty good sized and works damn well Eureka" machine.
          Last edited by ZPDM; 05-15-2020, 09:30 PM.

          Comment


          • #6
            Well Duff's law is falling apart, that didn't take long. I have a coil twice as heavy that is working maybe 20% better than one half its weight, unless some component issue shows up which I am doubting there is no way to pretend it is twice as good. Duffman! Making Mistakes, Duffman sad. Well its not bad news it just means we learned something, what did we learn? We learned I'm full of it and have no clue what I am talking about. More seriously I would still say it should be a general trend that a heavier coil works better. EMF is magnetic flux/time. Unless I am wrong induction, for a given amp input, should determine B field strength, i.e flux density. You have a coil with more induction you have a larger magnetic flux that occurs over the inductor's transient. Resistance should factor into how fast the flux occurs. Again being serious, I would say what this means is proximity effect and skin effect are in play with a single DC pulse of a coil. Where I may have made my mistake was I was thinking inductive reactance occurs with sinusoidal AC current, so if you look at a single DC pulse you should avoid all that. However it is current flow and change in magnetic field strength that is causing skin effects/proximity effects so it would make sense they are showing up even with a single DC pulse and apparently these effects are far from insignificant. I don't know, that's all I got to work with. So moving on and getting over, need to read on skin effect, proximity effect and look at litz wire and different coil shapes/designs, winding and rewinding coils, oh what fun!

            Comment


            • #7
              To think out-loud a bit and jot down where I want to go with this next, as I understand it, might be wrong:
              Proximity effect is primarily the inter-layer effect in a multilayer coil, though I believe also the effect between adjacent turns. It has the effect of pushing conduction to the outside of the coil.
              Skin effect is that which pushes the conduction in a single strand of wire towards the periphery of the wire.

              These terms are used with respect to a coil being powered with alternating AC current. Both are frequency dependent and more predominant at higher frequencies. Though generally mentioned with AC, when you think about it what is causing each is the current flow and the magnetic flux. Even with a single DC pulse that doesn't spend most of its time powering a saturated coil you have just like with AC, magnetic flux. So for a pulsed coil, analogous to what is said about AC frequency, one might say at briefer and briefer pulses, proximity effect and skin effect begin to predominate in comparison to DC resistance. I previously pointed out that for briefer and briefer pulses the ratio of output charge to input charge improves theoretically ad infinitum until one runs into real world component issues. The component issue I think one is most likely to run into is the coil itself, namely proximity/skin effects are altering coil resistance.

              I'll pause for a moment here to say quite politely I find this stuff so darn confusing. When I am pulsing a coil there is what I now call a kerfluey point. It is when I am running with a 10% duty cycle and the predicted amp draw no longer matches the observed amp draw. The observed amp draw first begins to creep up versus the previously perfectly accurate predicted then soon is ten times or more higher. That has got to be where the skin effects and proximity effects are beginning to become predominant. This kerfluey point varies with different coils both in regards to absolute time length of the pulse and in regards to the pulse length as a function of the inductor time constant. So far so good, the problem is the skin/proximity effects are increasing the coils functional resistance as compared to its DC resistance, shouldn't amp draw just start choking off completely and the coil sort of die off at very fast pulse trains? I don't understand this but will throw out some pretzel logic. This is weak, but who knows maybe its right, it does seem clear after the kerfluey point things are off the rails (though oddly if you keep going faster and faster, things do start to right themselves a bit, I just haven't had a slow enough coil yet to see if that continues to hold up as I can't go faster than 1uSec with the arduino). So when a coil is fully saturated the wire serves as a short, current flow limited only by the DC resistance of the wire. The time it takes to transition to saturation is determined by coil induction and resistance, if resistance increases the coil reaches saturation more quickly. Well, with shorter pulses, if S/P effects are kicking in the functional resistance in the wire is increasing, maybe the amp draw as compared to predicted gets worse because the increased functional resistance means it is seeing something more like a saturated coil. Again, I don't know, maybe a weak explanation but what I've got at the moment. I mean I get it, the coil isn't saturated, but the functional wire resistance is also dynamic during the inductor transient.

              Want to look at three things, really the last one is the one I am really interested in. So the first two are lets go the extremes.
              1) A single layer coil, this should get rid of 90% of the proximity effects, though not sure if it has much effect on skin effect. Did wind about a quarter lb of 26 gauge wire on two feet of one inch diameter PVC pipe and results were interesting. Getting back to Duff's Law the ratio of inductance to resistance of this coil was more than times worse than a 5 lb coil of 20 gauge wire I have. So didn't look good at first, however, again faster pulses better. This one layer coil could switch about ten times faster than the other before things went kerfluey. So if you picked the sweet spot for pulse length for each coil, the 0.25lb one was nearly as good as the five pound one. The confounding variable is the 26 gauge versus 20 gauge wire, I don't know I'm pretty sure the improvement in switching speed was nearly all due to the one layer not the difference in wire gauge, would have been nice if they had matched though.
              2) A very thin coil, I'll use the same length of wire from 1) and wind it, not as a one layer pancake cause I don't want to go nuts trying to make a great pancake coil from 26 gauge wire, but a real thin coil with lots and lots of layers. I have a sneaking suspicion that for this same given length of wire it will work better than the one layer coil from 1) but I would also guess it will be slower. Don't know, haven't done it yet.
              3) This is what I am most interested in. I will skip the math unless there are inquires but both from Texas Instruments Electronics Tutorials equation for self induced voltage and from Wheeler's formula for induction it is seen that if one has 2 turns of wire on a one inch diameter coil or 1 turn of wire on a two inch diameter coil the inductance (and self-induced voltage) is identical. Well that really is quite interesting isn't it. That means if you had 500 feet of coil in a 20 layer coil you could just wrap it wound the house in one turn and have the same inductance. As their is only one turn by definition you have gotten rid of proximity effect, no adjacent wire is proximate now is it? There would still be skin effect but oh well, we can worry about that another time. If you could go from a 12 layer coil to a 2 layer coil the difference in proximity effect would be enormous. I would guess there has to be a practical limit to this as there must be many 100 mile plus loops in the power distribution system. The thing doesn't blow up every time they turn it on and off, so I would guess the amps through the loop perhaps have to be sufficient to form a coherent thing in space. That said if you can go from inches to feet while letting the coil diameter make up for lessened coil turns, you can use that to mitigate/eliminate proximity effect. So I'll take the same length of wire from part 1) and wrap it on a coil with a 1/4 inch diameter and then on a coil with a 2 inch diameter. If the inductance measures the same with fewer turns ... Yea Baby! ... Yea!! ... Groovy!!!
              Last edited by ZPDM; 05-17-2020, 11:21 PM.

              Comment


              • #8
                I've completed this set of investigations and it has been a success. There is a lot to cover and this may be lengthy, for those who enjoy experimenting I would say there are at least a couple nugets here if you are willing to wade through the dross concerning the work involved. Though a success, by no means am I saying I understand things entirely, I just understand more clearly certain consistent patterns. So I may not know what I am talking about, but the more I read and experiment the more I am convinced neither does anyone else, hahaha, including (maybe especially) the career guy with the PhD. So I feel free to B.S. ad infinitum and enjoy it.

                So much to cover, so first I was talking about looking into different coil shapes, sizes wire gauges ... doesn't matter, how's dat? Hardly a whit, (as a caveat this applies only to being powered with batteries or DC power supply, more later). You will do a bit better with more wire, what you run into is as winds increase (inductance increases) or wire is thicker (resistance decreases) the coil slows down. What I have been calling the kerfluey point, anyone whose diddled around with this has seen this as you go faster there comes a point where amp draw increases, decreases a bit if it is not shorted but no longer is decreasing smoothly. I think, not adamant about this, but would guess this kerfluey point is the self resonance frequency of the coil. That is the frequency to tune for and as it slows down with more or thicker wire, you pretty much are close to the same in/out efficiency in any case. If you tune for this frequency you can recapture 80 maybe 90% at 24 volts, as a very generalizable theme things improve at higher voltage you would likely be at CoP >1 somewhere before hitting 600- 800 volts the limit that you could switch with transistors, but at 12- 24 volts nuh uh and doesn't matter your coil. This is not to dispute the success people obviously have with CoP>1 battery charging at this voltage only to reiterate what Bearden and Bedini have said that much of the magic is going on in the batteries and your meters won't be able to pick this up if you are calculating charge in/charge out with capacitors.

                Beyond the Kerfluey point/SRF, the inductor is acting less and less like an inductor and more like a capacitor. I spent some time going over how the effective resistance of the inductor is varying during its transient time. At this point I am fairly well convinced the effective capacitance and inductance are likewise varying during this time as well. This is unconscionable! "How dare you!" haha. So I have my new $200 LCR meter from a company that supplies sensors to BMW, the U.S. military etc., it samples at 120 Hz and 1 kHz. Now if you measure the inductance of an inductor at the two frequencies it is stable, the capacitance of the inductorr varies by an order of magnitude depending on the frequency sample. Conversely if you measure the capacitance of a capacitor it is stable at the two frequencies but the caps inductance varies by an order of magnitude. Good meter from reputable company.

                Now we get to some good stuff though more to follow. A first important generalization is that capacitors are not batteries or DC power supplies, you got that? If you look at what is going on at the SRF with a single pulse from a DC power supply, yes got it makes no sense just consider the pulse duration as half the sine wave if you were to go into AC frequency. So again at this sweet spot one might see 5 fold excess charge though at 1/4 the input voltage. Doesn't matter what you do with your coil, you'll see something around there, not orders of magnitude type improvement no matter what you do with your coil. Now if you pulse faster it appears better until you actually run more than one pulse and see you are drawing much more power than would be predicted once past the SRF.

                This doesn't happen with capacitors into a coil, I'd seen it before and I confirmed it earlier today. I saw 25 fold excess charge with a capacitor straight to another (granted at much lower resultant voltage than the input), real hard to argue with it and it is better than from a DC power supply or battery, also the transient time for the cap is like 0.1 uSec (way, way faster than the coils SRF) but there is still this improvement also from a larger slower cap. What is going on? Well I will tell you something, not sure what it means but I am only sometimes a moron. When I sample a 0.01 uF cap at 120 Hz the reading is 2.61 Henries at 1 kHz the reading is 172.37 Henries!!! That's 4 orders of magnitude more inductance than in the inductor! Whateva, I don't think I am using the meter wrong and this is why you can go faster with caps as you get past the self-resonance frequency of the coil the inductance is now coming from the cap. If the reading is correct could you imagine if you could capture that 172 big Henries as it fluxed? So caps and coils play well together in ways that DC power supplies and batteries do not. You can go faster with caps but to Alice in Wonderland absurd as you speed up the cap is the inductor and the inductor is the cap, as you slow down (the amp flow in a single pulse) the inductor kicks in with what it usually does. So again they play well together and one also sees a nicer more smooth improvement with increasing voltage.

                I started this investigation partly because I was curious and also because I wanted to have a more clear picture of how things behaved before I started on the Mark 3.14 Solid State battery charger. I had an idea for that charger which I also confirmed today so I, ... I'll just give it to you. At this point I know that capacitor discharges can work equally well to far superior to pulses from a battery or DC power supply. Now, like Gandalf and Bilbo in the Shire, stick this in your pipe and smoke it. The radiant spike off a capacitor being charged is of equal magnitude to the radiant spike off a capacitor being discharged. You should be dreaming of the undying lands, are you dreaming of the undying lands?

                The implementation of this with transistors is one of those things which appears so trivial but quickly looks like "no not the agonizer Mr. Spock". You just want to charge a cap then discharge it, how simple. But you are reversing what is the positive and what is the negative when you go from charging to discharging. So really this is another job for the Bedini-Cole bipolar commutator circuit. However, I have a cheat, work around, or perhaps very elegant solution depending on how you look at it. Depending on whether that
                work around solution works figures into how quickly I'll release the Mark 3.14. Again, and you're welcome, "The radiant spike off a capacitor being charged is of equal magnitude to the radiant spike off a capacitor being discharged."
                Last edited by ZPDM; 05-21-2020, 12:10 AM.

                Comment


                • #9
                  Progress continues. I mentioned two insights. 1) Caps don't behave the same as pulses from a battery or power supply and are at least as good to superior. This isn't a stretch by any means, I have seen people put a bridging cap over their battery supply on an SSG or similar to improve the output from the radiant. It makes sense to me, as a first approximation, that the most abrupt discharge you can get is from a cap where the charge is just sitting on plates as opposed to circuitry involved with a DC power supply or ions moving through electrolyte with a battery. So will say again caps aren't batteries and you can see things with cap pulses that you do not see with battery pulses. I attempted to discern if for a given expenditure of coulombs of charge one gets worse the same or better a radiant charge from a cap versus a battery. For reasons I won't rehash this is not an easy question, but again, empirically as far as I can tell, yes you can. What is a little surprising is that when charging a cap from a battery it also, empirically, appears you can get a radiant as good as from the battery.

                  I suspect it is important not to overgeneralize this. Let's consider if you had a nine volt DC motor. Now lets say you expend 1 coulomb of charge powering that motor from a nine volt source for ten seconds. If you now instead took the nine volt battery and discharged it into a super cap such that the cap was empty at the start and at nine volts after a coulomb was expended, and you put the nine volt motor between the battery and cap as it charged what would happen? Well at first the motor would run great but as the voltage between the cap and battery equalized it would run slower and slower. The average voltage during the time the cap charged would be 4.5 volts. If you then turned around and discharged the cap you would get the exact same behavior. Add the two together and you would get would happen if you just ran the motor from the battery in the first place. This is also why there is no "magic" in and of itself with a split positive, common ground set-up, it is the same issue.

                  Howeva, what if your desired effect isn't related to the average power but is related to an abrupt discharge. An analogy here might be what if to save energy you decided to pulse an Led 50% on 50% off 400 times a second. You could save energy as the perceived illumination, I suspect even from a lux meter, would be identical. Now what if for 1/2 the time you pulsed the Led by charging a cap and for half the time pulsed the led by discharging the cap. I haven't done this but I suspect I again you might get a perceived and lux meter illumination similar to the first set-up but at half the cost.

                  Empirically, from what I can tell, and again it is a difficult question with varying current, you can pulse into and out of a cap and capture a similar radiant to simply a pulse out from a battery for the same coulombs. Of course as I get down in the weeds I find it is best not to discharge the cap entirely. Let's be clear, I'm a physician, I may have never really got what the engineering texts were saying in the first place and with this crazy stuff whether I got them or not I think they're out the window. You just have to play around with the timing and look at trends. Just to be confusing, this is different, but is it really that much different from a node, I don't node? Is this conceptually similar to the Tesla switch, I would say maybe, really quite possibly. In any event, I've been playing around with it for some time now and as Sam Gamgee said in the Lord of the Rings, when hitting orcs with a frying pan, "I think I may be getting the hang of this".

                  Now then you may likely wish to know how to charge and discharge a cap with transistors. My previously mentioned thought for a work-around didn't work-around at all. My other initial thought that you needed an H-bridge such as the Bedini-Cole bi-commutator was wrong as well, too complicated. After diddling for awhile and reading I came to understand that what I needed was something akin to a "Class B Amplifier" circuit aka a "Push-Pull" circuit. What I ran into was that all these circuits were assuming an AC trigger source. Without an AC trigger the PNP transistor never turns on. There is not an easy way around this issue triggering from an arduino and no one seems to have circuits on the internet for a push-pull circuit triggered by a square wave or DC pulse. What I came up with, I find it a bit hilarious, but the circuit remains two transistors and it has worked beautifully and consistently. The power through the PNP flows from base to collector to ground, not emitter-> collector ground. At low power this has worked great and I think you can spec out where you might run into problems. To do this "correctly", I don't know, maybe half an H-bridge, coupled to an NPN/PNP if any one still does this sort of stuff. You could run it all with solid state relays, the problem with SSRs is the on time will be determined by the Led and will be something like 1 millisecond. Anyone who starts fooling around with caps and coils may be like me and feel a need for speed. 1/1000th of a second is slow oh so very very slow. You could get around this with big honking slow coils and perhaps I may look into this at some time, but my jury rigged system works fine and is fast.

                  So then, Mark 3 should be coming soon looks like it should be better with two coils and two transistors than the earlier versions I showed with more coils, i.e. lots of branch and node currents.
                  Last edited by ZPDM; 05-28-2020, 12:05 AM.

                  Comment


                  • #10
                    Mark 3 demo version still a little ways off and writing just to decompress from a few days work. 1) The push/pull idea of charging and discharging and gathering the inductive spike from both instances is valid. I have found greatest efficiency on either end of the push or pull of the cap when it is not fully discharged. In such an environment adding the second inductor ends up affecting the first inductor. This is why I said I am not certain if it is much different than a node. I suspect it is at least a bit different and I want to pursue it further. I am really quite intuitive but how do I or anyone else know if my intuitions are complete bul****? Hmmmm. My intuition tells me Bedini did what I will soon do for you with nodes. Could be wrong but just considering what I did with nodes previously, I think I may have been in the "zone" and just never bothered to optimize and look. Still I want to try and follow this push pull idea further. Why am I decompressing then? Well I can take the spike from charging a cap through a coil and recapture easy, 75-80% of the input energy. I cut through all the BS and just take a 1000uf cap, capture the spike into a 100 uf cap and send it back to the 1000 uf cap. Should have done this long ago as what I am seeing is different than what the meters show and this strikes as about as acid an acid test as you can have. I can also on the other end when discharging the cap again recapture 75-80%. I can do this when either of them is connected or disconnected. The problem is the two caps don't play nice together when I discharge them both back to source!!! Aaaaarrrgh. Some people like to say there is a conscious element to all this. So I will tell you, there was a simply beautiful symphony of cursing involved, punctuated with many whats?!, hmms. hrrrmpphs! maybe even a whos dat? don't recall at this point only it was ugly, haha. I diagrammed it out twice and on the second time I may just be missing a single additional diode but had to call it quits for what should be clear reasons. If I can't get it from looking at the circuit diagram I may just take the inductive spikes collected in the two caps and send them to two more caps through opto-isolated solid state relays wait a bit then discharge those caps. Hopefully, and I don't think it will, it won't come to that. If it did, I do suspect you could get a similar though not quite as good sort of product from using a node or two on either the push or pull cycle of the cap and only using one half the cycle.

                    When I was sitting with John Bedini at his last conference I was telling him, you know I think if you rectified the high voltage spike off a Wimshurst machine and sent it back to batteries spinning a motor driving the Wimhurst machine it might be over-unity and he looked at me kinda bored and said "Yea, that'd work". Haha, still haven't worked on that one, Now I have done only very few interesting things in this field but that sort of attitude of any foo will see overunity at 5000+volts stuck with me. When what I am working on is finished you have OU from 16 to 600 volts. But right now, as Loki said, "Oooooh, it burns doesn't it, to have come so close." I can see the 80% recovered in each leg of the push/pull circuit the darn thing just goes kerfluey when I discharge both back to source.

                    As a last thought, for those who work on the Tesla Switch, and John demonstrated OU at what 12 volts maybe less, you aren't going to get there without the inductors. People, I suspect, far, far more competent than me at circuit building put time into this thing, but if you dig around there is a version of it showing the inductors in place. What I am working on with this push/pull circuit may just turn out to be an arduino controlled two transistor version of what was disclosed previously.
                    Last edited by ZPDM; 06-02-2020, 11:52 PM.

                    Comment


                    • #11
                      Am working towards the next Solid State battery charger but this is still sort of "bench research" so I'll post it here. Most previous videos get maybe a half dozen views, which I am happy with especially if I sometimes go back and watch them, haha. This video documents a CoP of 1.05-1.1 going from capacitor to capacitor, from a circuit with one opto-isolated relay switch, one inductor and one diode. I might not want 60 views but I might be disappointed with six. https://www.youtube.com/watch?v=VWGi...ature=youtu.be

                      As a last point I haven't actually sent the gathered spike back to source I am just suspecting that would not work from previous investigations, will maybe look at that but have an idea what to expect. Also was a bit sort of inside baseball towards the end. I am discussing the Bedini-Cole bicommutator circuit, which among other uses is used to power the Bedini Window motor. This circuit allows one to change polarity of a power source through a coil. While I have no data yet, there are some things which might be considered re the possible efficiency of voltage transformation and the results documented herein.
                      Last edited by ZPDM; 06-17-2020, 03:44 PM.

                      Comment


                      • #12
                        Update: Built a two relay H-Bridge. Efficiency using a 0,1 uf cap with shooting the spike back to source ~70- 80%. Could look at as discouraging in that I've achieved 80% + without an H-bridge set-up. Or look at, as I am, as encouraging, in that you can't switch relays faster than 50 ms or so and I am at 70-80%. Next step is to build a Mosfet H-bridge, good tutorial here if any interested https://www.youtube.com/watch?v=IjgvKJrAOS4 . With better control of timing what will I see with this set-up?? So need to order parts. A concern, which I suspect is silly, in fact now I can think of no way to make this a valid concern. But when the cap that is reversing polarity and connects, essentially in series, to the reversed batteries, is it losing half its capacitance. But again, as I think of it who cares, (it most likely isn't losing capacitance) but if it loses capacitance on each leg of the cycle, no harm no foul. Again I am 99.9% certain that a bipolar commutator provides 100% efficient 2x transformation of voltage and that is no small thing. For what I am going after with this project, efficient transformation of voltage is in a sense the end-all be-all. To wit, and I've gathered stuff like this six ways to Sunday, tapping out by hand, transistors, now here with relays, this is something that is consistent.

                        12 pulses, into cap at Input Volts 10 output voltage gathered in second cap 20.8
                        6 pulses, into cap at Input Volts 20 output voltage gathered in second cap cap 29.4
                        3 pulses, into cap at Input Volts 40 output voltage gathered in second cap 37.9

                        The cap being pulsed and the cap gathering the induction are the same two caps in all three examples. Hence the coulombs expended in all three cases is the exact same yet at 40 volts nearly twice as much voltage was gathered in the second cap. Hence again, you can see why if you can transform voltage without losing charge you are doing well indeed. Never gonna do this with caps in series, they lose half or more of their capacitance once placed in series, output voltage didn't double going from 10-20 volts so you are moving backwards by attempting to transform voltage with caps in series. An H-bridge (bipolar commutator) set-up does give one a 100% efficient 2x transformation of voltage. Is that enough to end up being able to feedback more coulombs to source than were initially expended? Don't know yet, not with a relay version, but with a BJT or Mosfet version where you can really poke around with timing, who knows, maybe. As I said at the end of the last post though, there are other ways to transform voltage. This may be why you so often here about first get a microwave transformer, haha. Okay so here are two links the first is to a Texas Instruments reference design for a boost converter that goes 5 volts in to 80 volts out with nearly 50% efficiency. https://ti-content-syndication.elect...-tutorials.ws/ That got me thinking. Now while I only looked at it over 10-40 volts with this recent relay set-up I have looked at it further tapping out by hand and with transistors, so I would say that just as going from 10-40 volts was nearly twice as good, going from 10 -80 volts would be nearly 4x as good. This TI transformer goes from 5-80 volts with loss of half the charge, that's moving in the right direction. Then I've found this on digikey https://www.digikey.com/product-deta...PI-ND/10287689 it says it is 90% efficient going from 9-54 volts. If you are 90% efficient going from 9-45 volts, given the data posted above ... (and keep in mind that would become 10x transformation that is 90% efficient once throwing in the H-Bridge ...) Now I am looking very hard for other ways I might transform voltage myself. Is there any way to chain that 2x transformation with an H-Bridge to a 4x transformation that is also 100% efficient, I haven't been able to see it. Why don't I build a boost converter myself. Yea fine, but there seems to be very little documentation let alone an "instructable" on how to build an 80% efficient buck boost converter that transforms voltage 5 fold. Hell the Bedini backward diode set-up is a buck boost converter, if I understood it, I could build one that was efficient with very significant voltage transformation. I got no problem with buying an off the shelf solution, I just, oh and I hope this doesn't happen, don't enjoy ending up learning, "Ohhh, huh, so that's what they meant by 90% efficient, not into a varying load like charging a cap, how interesting". Nonetheless, I don't care if it is an AC transformer there are ways to transform voltage 5x with 90% power efficiency.

                        In any event, the first part is building a BJT or MosFet H-Bridge because I feel a need for speed. If that is not enough (which you know it just might be enough, but if not), there needs to be a preprocessing of the voltage that needs to be done with efficiency.

                        Zero views of the last video, just how I like it, will keep posting unless or until things become too terribly interesting.

                        Comment


                        • #13
                          Hi ZPDM,
                          Interesting thread. I've been (trying) to following along a bit and perhaps learn something.

                          I had a thought about your "kerfluey" point. While it may be the coil's self resonant frequency, I was wondering if perhaps it is related to the coil's discharge time. For example, at slow speeds the coil charges with the pulse and discharges after the pulse. But at higher frequencies, the next pulse may start while the coil is still discharging. So the current may go from 0 to 1 on the first pulse, then .25 to 1.25 on the second, etc.

                          Anyway, that's the thought that flew past while reading your posts.

                          Comment

                          Working...
                          X