Push-Pull Output Transformers - Part III, The Final Countdown:

Discussion in 'Amateur Radio Amplifiers' started by KD2NCU, Sep 28, 2017.

ad: L-HROutlet
ad: l-rl
ad: OK1UUad-1
ad: FBNews-1
ad: Left-2
ad: Subscribe
ad: L-MFJ
ad: Left-3
  1. KD2NCU

    KD2NCU Ham Member QRZ Page

    No, it's your turn to finish explaining and answering questions first. Ask WW1WW for help if needed.
    Answer all of the following questions.
    Question 1: Go back and answer the simple questions. (Without getting into a convoluted discussion, of course.)
    [​IMG]
    Figure 1: What does the current coming out of the source look like? Check only one answer below. (Keep it simple now)
    •____ 1 amp pulses matching the resistor current exactly?
    •____ A constant flat DC current of 0.25 amps?
    Figure 2: What will the current coming out of the supply be? Check one. Keep it simple.
    •____ A full wave rectified sinewave of current with peak value of 50 mA matching the collector current exactly?
    •____ A flat DC current equal to the average value of the collector current or 0.637 x 50 mA = 31.85 mA?
    Question 2: Just tell me (no convoluted discussions) how a DC supply works. I'll start you off. A DC supply will work very hard to keep the voltage constant. Now you tell me what a DC supply does regarding current. Keep it simple!

    Question 3: Describe what the current looks like just as it enters the center tap of the bifilar feed coil. ie; full wave rectified pulses of current essentially replicas of the individual transistor current pulses, or a dead flat DC current equal to the average of the transistor current pulses.

    Question 4: If you say the current entering the center tap is a dead flat DC equal to the average of the transistor current pulses then explain when Q1 first turns on but the collector/drain current has only risen to a few milliamps but the current coming out of the DC supply is a flat DC of about 10 amps, all those extra amps and electrons are going (fill in the blank) ____________________ .

    Question 5: Does WW1WW's assertion that all those extra electrons representing the unexplained 10 amps of question 4 are converted into stored energy in the magnetic field sound plausible to you? Keep it simple now, like maybe YES or NO. If so, can you or WW1WW provide a valid physics reference that supports this new "discovery"?

    Question 6: Does WW1WW's assertion that all those extra electrons representing the unexplained 10 amps of question 4 are leaking through the off transistor seem plausible to you? Simple now, should be a YES/NO answer. Either you agree with him or you don't.

    Question 7: What is "Pure DC" (no convoluted discussions) You and WW1WW throw it around like it has some special meaning. I have been taking it to mean the voltage is constant but the DC supply will provide whatever current the connected load demands or asks for because that's all a DC supply can do. I think you and WW1WW think that "Pure DC" has some extra meaning. So please explain in simple terms what you are assuming when you say at point B you have Pure DC.
     
  2. K7JEM

    K7JEM Ham Member QRZ Page

    Your diagrams are interesting, but not representative of an RF power amplifier. Go back to the simple RF power amplifier I posted and show that "pulsating DC" is being drawn from the PS, at a rate equal to the instantaneous value being drawn at the collector. After we get this simple matter resolved, we can move to other topics. I would submit that the current being provided by the PS is a "pure DC" with a very small trace of RF.

    If we can't agree on this, there is no reason to go any further.
     
  3. K7JEM

    K7JEM Ham Member QRZ Page



    1 amp pulses matching the switch closure.

    Well, this must be a trick question, because neither answer is correct. We don't know what the waveform is, so hard to predict what the current waveform will look like. The second answer has values provided, but no values are given on the schematic, so who knows? We don't know the voltage, the input voltage or current, the Re value, or anything else. So, neither answer is correct.
    The DC supply will supply the current needed for the attached circuit, up to its maximum output.

    Assuming standard amplifier circuit with bypass capacitor, it will be dead flat DC input, with a few milliamps of AC ripple. The DC will be equal to the current draw of the amplifier, assuming steady constant power out during measurement.

    Charging capacitors and inductors throughout the cycle.

    Yes. Energy is stored in inductors and capacitors. This is not a new concept.

    I don't think he said that. He mentioned that there was some capacitance in the other transistor.


    Pure direct current would be a constant flow of current, not pulsating at RF. An easy way to see this is to insert a non inductive resistor of low resistance, right into the DC feed point. Something that would drop 1 volt or so at full amp power out. Now activate the amp, and use your scope to look at the voltage drop across the resistor. You will see a 1 volt (or whatever) drop. You may see a small amount of RF, but probably a mV or so. Won't be a lot in any normally designed amplifier of any configuration.

    I think this is something you need to do, to prove to yourself that the current supplying these amps is not "pulsating" to any degree beyond the feed choke.
     
  4. WA1GFZ

    WA1GFZ Ham Member QRZ Page

     
  5. WA1GFZ

    WA1GFZ Ham Member QRZ Page

    you are confusing a transistor with a tuned tank circuit with broadband amplifier. Apples and oranges
     
  6. K7JEM

    K7JEM Ham Member QRZ Page

    No confusion going on here. There is an inductor between the DC supply and the transistor. It has inductive reactance. That reactance is what isolates the DC from the RF.
     
  7. WA1GFZ

    WA1GFZ Ham Member QRZ Page

    KD2DCU,
    The flux would be balanced but there would still be common mode inductance with no offset in the old BH loop, in a perfect world. In the real world at least in all my testing the flux is not canceled and even a 1 turn third winding picks up the offset by the flux transformer action not some leakage mumbo jumbo. I've spent hundreds of hours making measurements and concluded T2 is not necessary if the transformer is designed correctly. We just bought a 2500 watt AR amp at work and the first thing I checked was the output transformer circuit. NO T2s. BTW we have a number of AR amps and none of them have a T2.
    T2 was added in commercial hammy hambone rigs because the half ass flux transformer had big offsets and ran hot. I duplicated the effect. I then converted the exact same cores to a balanced TLT and the cores ran stone cold. My case I din't have room for T2 and T3 on my MRI amplifier boards so found a better way. That allowed me consistent performance 160 through 6 meters. (The AR amp runs 10khz to 220 MHz)
    I had this same battle with myself a few years ago so more fun to watch gfz
     
  8. KD2NCU

    KD2NCU Ham Member QRZ Page

    WA1GFZ, I'm not disagreeing with you that in the real world there will be "some" inductance due to imperfections. I know all about the real world and Ham Radio is not the real world. Ham Radio is a fantasy sandbox where we get play and blow stuff up and empirically try stuff over and over until it seems to work and where there are really no consequences to not really knowing physics and circuit analysis and such or what heck you are talking about because usually the worst penalty for talking out your rear end is a blown transistor or maybe a shock. The real world is operating a nuclear submarine loaded with nuclear missiles where if you don't know what you are talking about or try to BS, or don't know physics and how circuits really work you and the missiles and 100+ men end up on the seabed. The real world is designing and building communication satellites that have to work perfectly for something like seven years just to break even and if they don't you can't fix 'em so you better know your physics and circuits and theory because you're dead meat if it fails prematurely. The real world is the electronics industry where results do matter and you get about 1.5 chances to get something right and then you are out on your rear end so you better know your physics AND your practical stuff or you are done. You don't get to try crap over and over empirically. You get it right quick and it works or it's the ovens. I've operated in the real world very successfully for 44 years since engineering school. Yeah, I know the real world better than most, actually.

    Now I'm not arguing for or against using the T2/T3 approach and never have been. The start of this entire chain of threads was me asking why the T2/T3 approach is even used and what is T2 doing that T3 can't do by itself. A lot of answers I got were absolute B.S. that defy simple physics and basic circuit analysis. Initially, in gentlemanly fashion, I provided physics and circuit analysis to show which areas I thought were fallacy and why. Ever since then, people have shrieked, "you're wrong", "it's a choke", "but it's pure DC", "this guy doesn't understand" and took it personal asking about my education, hurling insults, etc. But that's OK, because all the shrieking, asking about my education, and hurling insults is the surest sign of weakness, self doubt, and admitting they don't really have an argument. Besides, with my experience, I've been worked on by experts and real engineers, physicists, etc. and these guys with the mouths are amateurs who seem unable to use physics and circuit analysis. I've never lost a technical argument yet because I do my homework first while most people responding on this site think they already know everything and start running their mouths.

    The posts I have provided have used solid physics and well established circuit analysis to show which claims I believe are false and why. Prior to you, WA1GFZ, not a single person actually tried to refute ANY of the analysis based on physics and circuit analysis. They just shrieked "you're wrong" etc. etc. and hurled insults. Gee, I wonder what that means. Now you've said a couple of things that lead me to believe you have at least looked at some of my prior analysis. At one point you did say you agree with my analysis but I don't know which analysis you were talking about.

    So what about you, WA1GFZ, do you believe in physics and established circuit analysis techniques?

    So here are some of the responses I originally got. Let's see what you think about them.
    1. There is absolutely no RF of any kind in the coils of T2. What say ye WA1GFZ?
    2. It is absolutely impossible for the core of T2 to saturate "because it's not grounded". Well, typically it is AC grounded and the core isn't aware and doesn't give a @#$% whether it's grounded. Imbalance will staircase up the BH curve until either the core saturates or until something in the external circuit finally limits the current. What do you say about this?
    3. The bifilar feed coil keeps RF out of the DC supply. This defies basic physics and basic circuit analysis. In the circuit below I've said that the current entering the center tap has to equal the drain current of the on transistor regardless of whether T2 is providing any inductance at all simply from conservation of charge (Kirchhoff's current law applied to the circuit.) The current enters the center tap, splits into I1 and I2, and regardless of whether they are equal or not, they just wander around a bit and then recombine to become the collector or drain current except for some leakage current through the off transistor. Yes, in the real world there will be some leakage current through the off transistor. But this is a red herring. Regardless of the other real world realities of imbalance, imperfect windings, non-linear BH curve, and so on, other than the small leakage current through the off transistor, there is only one place current gets into the circuit (the center tap) and one way for the current to get out of the circuit and that's the collector or drain of the conducting transistor. Unless, of course, electrons are elastic and can bunch up in the wires or something. Example below. At the very start of Q1's conduction cycle, say the current has only risen to a few milliamps, the current going into the center tap must also be just a few milliamps. K7JEM holds that the current at the center tap will be essentially flat with minor ac ripple and its value will be the average of the collector current pulses. So if the collector currents are positive sine pulses peaking at 16 amps then the current as it enters the center tap will be essentially flat DC current of about 10 amps. I asked where all those extra electrons are going. IE: the collector current is 1 mA or some other very low value, but the current going into the center tap is 10 amps flat DC. What accounts for this difference? Where are all those amps and trons going when the collector current is so low compared to the incoming current. K7JEM and WW1WW think the electrons are being stored in the magnetic field of the T2 coils/cores and maybe even T3's coil/core. So electrons are somehow jumping out of the wires of T2 and T3, somehow transforming to some sort of magnetic ..... goo or maybe magnetic goo-ons of some sort and hanging around drinking covfefe until the collector current rises to 10 amps at which time there is now a shortfall so covfefe break is over, put down your cups, all you goo-ons turn back to electrons and jump back into the coils to make up the shortfall between the constant flat 10 amps and the collector current which is about to peak at 16 amps. But don't worry all you trillions of goo-ons, because you'll be back again shortly when the collector current falls below 10 amps again and there is an excess of amps and trons.
    At the last minute, K7JEM threw in that the mismatch of trons and amps also goes into charging capacitors during the normal cycle. Problem is, there are no capacitors present INSIDE the circuit for the trons to get stored in.
    What do you think about the electrons turning into magnetic goo-ons theory WA1GFZ?
    It's a pretty fantastic theory, but I think he's screwing with me, I don't believe for a minute that anyone would really believe such a fantasy. What do you think?
    Below right is K7JEM's theory of magnetic goo-ons. Whaddaya think about this WA1GFZ? Could maybe THIS explain the imbalance you always find? Imbalance in the electron to magnetic goo-on conversion plasma?
    upload_2017-10-8_9-33-31.png
     
  9. K7JEM

    K7JEM Ham Member QRZ Page

    I think your image is rather humorous, so I offer a video to rebut:



    Yes, inductors can store energy, it is a well known phenomenon. I'll let you do some research on that.

    It is interesting that you place a current probe inside the bypass capacitor. But the DC supply does not feed at that point, it feeds to the bypass capacitor, so readings need to be taken at that point, if we are talking about DC that is coming from the supply. So, move the current probe to a place where the DC actually feeds, and explain what the DC current will look like at that location. I have already made my explanation, has to do with inductors and capacitors that store energy, if only for a few nanoseconds. But that's what is going on.

    It would be nice to work with schematics and diagrams that were actual circuits with real parts and values listed. That is why I provided the link to the 2M class C amplifier. The main thing is that we know what is going on, rather than try to figure out some main issues, like inductor values, capacitor values, voltage and current, etc.

    The diagram and photo that I posted are easy to reference, and a good starting point. If you don't like them, you can post any schematic that you like, as long as it has part values and is actually a working amplifier. I chose the simple class C design with a single transistor, since it eliminates a lot of parts, possible current paths, etc. If we can't agree on how this works, it will be much more difficult on a more complex amplifier design.
     
  10. WA1GFZ

    WA1GFZ Ham Member QRZ Page

    KD2NCU,
    Nice to find someone in the know. I did my simulation in LT spice. Send me an email address and I will send you a copy of my Sept/Oct 2015 QEX article. I would love critical comments. I have received some interesting emails from Apps engineers. Anyway I'll start with a little history. Back in the day BB amplifiers had a pair of ferrite sleeves in the drains with DC fed into the common point between the two sleeves. This is a flux transformer and the DC feed is unbalanced. Basically you have two single ended stages fed out of phase driving a common secondary threaded through both sleeves inside the 1 turn tube primary. As power increased the sleeves started getting hot due to the DC offset driving the cores close to saturation. Then someone had the bright idea of adding T2 and taking the DC off the transformer. This can be successful but it does add reactance to the drain circuit which degrades BB performance. I look at T2 as a 180 degree hybrid. When a winding is in conduction the other winding flows current because of the coupling between windings. If the t2 is wound as a TLT the core just provides enough reactance to provide isolation from the Drains and DC supply. As long as the T2 core is big enough with enough turns everything is happy. The only magnet field set up in the T2 core is due to the imbalance in the two winding phase currents. Since no transformer is perfect the primary has more power than the secondary. Erbteck wound their t2 transformers with two 25 ohm coax cables in parallel. I found this configuration superior to twisted wire. Twisted wire tends to be around 100 ohms impdeance. A guy from Erbteck told me they found this was a better impedance match than twisted wire. I found the longer the wire on T2 the worse it performed at high frequencies. (SWAG there may have been some delay in the circuit) So when you want to have a short cable you need a big core to avoid saturation due to the unbalanced currents. My Driver board at one point, a 3 stage modified Erbteck amplifier with a pair of mRF150 finals had a t2 larger than the output transformer. This was never going to fit on the 1200 watt MRI board because the existing transformer was a 7/8 type 61 toroid. In the end I used a 1 inch by 1 1/8 inch type 43 sleeve sitting on its side to do both T2 and T3 functions wound with heavier cable than the stock transformer. These transformers run stone cold 160 through 6 meters. The flux density is pushing 200g on 160 but core losses are low at 2MHz. I considered using type 61 material but the inductance may have required another turn on the core for each phase. The winding lengths are 10 inches per phase of #20 single shielded teflon wire. Two hole Baluns worked well with even shorter cable but The cables had sharp bends that may have shorted over time. I've been doing this long enough to be humble and open to new ideas. Frank WA1GFZ
     

Share This Page