This video presents the development of automatic antenna tuners for extremely low frequencies (640m, 2200m, and 8900Hz) using Litz wire and printed toroids to achieve high Q-factor inductors. The presenter, Ross Ballantyne, explains that traditional Z-match tuners require high-voltage capacitors (1kV, 160nF) and suffer from power loss when using ferrite toroids. By using Litz wire (30 bundled filaments) wound on printed toroids with multiple series windings, the inductance can be increased 25-fold while maintaining high Q (approximately 20 at 8900Hz). The design requires accurate LTSpice modeling to calculate the required capacitances for tuning and loading, with the model validated by comparing simulated voltages to actual measurements. The tuners can match impedances from 50Ω to 10,000-15,000Ω, making them suitable for challenging antenna situations like simple wires coupled to power lines.
Inmersión profunda
Prerrequisito
- No hay datos disponibles.
Próximos pasos
- No hay datos disponibles.
Inmersión profunda
FDIM 2026 - The development & construction of automatic antenna tuners for 640m, 2200m & 8900Hz opsAñadido:
All right. Good evening. Good afternoon.
Bonjour. I'm not sure what time it is, but that's another matter.
Um Little bit about me. Ross Ballantyne.
Number of call signs. UN missions around the world.
Uh early adopter of Whisper. And I started Whisper in 2008, March. So, I was a very early user between and uh um Bob Sutton, a New Zealander who was in Eritrea Ethiopia at the time. And he was He and I did Whisper between Helmand, Afghanistan and Ethiopia.
Just a little bit of history. I like Whisper because it's a set and forget type um mode that you can actually do a lot of work in, but you don't have to be there monitoring it all the time.
And given my work in the UN where I had to go off and do things, um it was a great mode to take with me.
I did try voice mode in Kabul before Whisper, but I ended I used to end up with rather large pileups because at the time Afghanistan was a little bit of a difficult place to get any radio in and out of.
What I'm going to be doing is talking about the development design development and implementation of low band tuners.
And low band tuners are quite difficult.
Um but I did do a lot You can do LC matches. You can do other all sorts of matches, but the one thing I do like is the free match. Now, the free match is a modified Z match.
Uh a twin gang tuning a a either a solenoid or a toroid inductor and a another twin gang loading it to get the output. You can tune a very large range of frequencies on that.
But of course, certain poles having two twin gangs and an inductor that you're going to be able to resonate at whatever frequency.
Which you can't buy off the shelf.
They're very difficult. So, I spent a lot of time uh looking at the options between using a Z match and all you're familiar with a Z match. It's pretty old tuner round, but uh a Z match works quite well. Um but I found that I had great trouble modeling a Z match on LTSpice.
Why did I need to model it? Because I need to know capacitances that I would use to load and tune, the voltages involved, and um the values involved.
And uh when you look at some of the models, you'll get 800 V on the tuning uh part of the circuit. So, you need specialized high-value capacitances. In other words, a tuning uh gang capacitor that's uh 1 kV and 160 uh nF peak, which is you don't get easily, not even in uh large um vacuum variable capacitors.
Anyway, um I've done quite good success with Z matches and manual free matches. And my manual free matches, I used Oops, wrong one. Um my manual free matches I used uh were with um switched capacitors. And what I did was I got two 12 12-position rotary switches. Did I series array of capacitors in a calculated step? And so, I'd have first be uh you know, a base capacitor of say 560 picofarad. The next uh capacitor would do if you were series step of say 80 picofarad down, and so on.
I had a spreadsheet that calculated that nicely.
But, of course, when you're dealing with uh frequencies below 160 m, you need capacitors that are fairly high volt high value capacitors, and you need high value inductors.
So, there's the basic Z-match circuit.
I'm sure you've seen them all. Uh but, my problem was I could not get a a uh design that would work on a uh an LTSpice model. What I couldn't do is get the model to match the actual free match.
And another factor was I started using um Z-matches many years ago, and I would use T68 toroids uh T68-2 or -6.
But, when I did an LC match of the right values and my best free match, and I measured the RF voltage on the wire, I found that the LC match gave me twice the RF voltage on the wire than the Z-match.
Which means that it was only putting out a quarter of the power.
So, I then said, "Well, I'm not going to use uh toroid iron powdered or ferrite toroids or any of those because I'm wasting power."
So, I started using um T130-0s or printed uh toroids and uh effectively I wound toroids on a PC And indeed, this is a pair of my uh Z-matches. The one left one is a 10 to 200 m Z-match, and the right one is the 630 m head match.
And if you uh note there, these capacitors these inductors uh that's a 76 mm inductor, that's about a 50 mm inductor. And these step da- twin gang switches were effectively twin gang capacitors with a large step value and a smaller step value. So, I'd have 12 steps uh with small 12 steps.
Um it just took a bit of I've got a spreadsheet that calculates out the step value uh series to make the next capacitor step down of however many picofarads you want.
They worked relatively well, but I found that when I measured the output voltage on the wire um of those, I wasn't getting anywhere near what I should have been getting to get out. I got spots with them but on whisper, but they weren't particularly efficient.
Here is the basic free match circuit, and you will see it's a five inductors in series, and it is tapped input on the um on the uh input there on the first, then two twin gang capacitor tuning peak the the large inductor and the smaller tap.
And the same here, the two uh tapping out.
Now, I don't know precisely the theory of how they work.
I've asked Claude describe how a Z match works or does I've asked Claude how a free match works, and of course I've got no answers because there's not been the research into it.
So, what I wanted to do was produce a model for this uh free match that would allow me to calculate the peak voltage capacitances that I would need and the voltage and the capacitances that I would need to tune the thing in and load it.
Here's a bundle of free matches that I've made. I've made it literally probably a couple of hundred different free matches and I've found them very very successful.
And indeed the one on the left here is one of my HF free matches. It's a a twin 130 one toroid zero toroid air wound so it's very low loss with two twin gang uh twin gang 330 picofarad. Now this is another free match that I've made and you will see that I'm using a multi-turn multi-wire twisted wire in the um in the you probably can't see it there but I've got five wires twisted around the stator of the um of the um of variometer printed variometer and you may have seen the printed variometers I bought previously in the last few FTIIMs.
But that one is that one is a very effective match but if you consider auto tuners you can have to have auto tuning the capacitance and auto tuning the the actual inductor which is difficult because you've got to do a lot of controls that selects whatever capacitor etc. But so what I in my various free matches and things I decided to use Litz wire and Litz wire I'm sure you know what Litz Anybody who doesn't know what Litz wire here is, and I'll explain it.
Okay, Litz wire is when you take a multiple bundle of woven smaller filaments to make and you use the uh the uh skin effect of the in small wires and those those bundles of 30 wires at.2 all have very close coupling.
Um so, it's it's basically a a wire which is split into 30 bundles. Now, and it's insulated. They're all uh um uh enameled.
Um but, winding a Litz wire, there is there is very little information about Litz wire toroids and Litz wire windings again for LTSpice.
It's another problem.
Anyway, so, I decided to use the free match uh configuration with a toroid a fixed toroid, but with um each of the windings split the bundles of 30 into five bundles of six and series them up. So, I've got six series a five series windings all in the one winding, but five in the same actual geographical physical space, which means the inductance of the fifth point, if they're in series, is at least 25 times, I e n squared, 25 times the basic inductance of that toroid. It's actually quite a lot more than that because along with the uh the very close coupling and the capacitive and other effects, you get a multiplication of the inductance, so it's actually uh probably about 20% at least more in inductance than a pure 25 times the base the one coil.
As I said, the next step was to make models on LTSpice that would work.
So, I did some modeling and calculated, okay, I use an inductor and put the inductance of each of the sections as series.
But, that's not actually how it works because you get um the first is going to be 1 * the 4.0. The next time is going to be 4 * the next is 9 * the next is 16 * and the next is 25 *.
But, as I said, there is no information on the internet or Claude or any AI that gives a proper model for how a Litz spaced series toroid can be treated in a in a model.
But, what I wanted to do was use fully electronic capacitors uh to actually tune the circuits and load the circuits.
And unfortunately, I'm in Bishkek, Kyrgyzstan. Anybody who doesn't know where Kyrgyzstan is?
Central Asia, north of India, between Tajikistan.
Electronic shops there at the end of the food chain at food chain. So, there's not a lot of things there that are available there. So, I have to order from AliExpress or I have to think laterally and create my components or I have to search for old um Soviet components. And I actually have got some very nice um variometers and some very nice vacuum variable capacitors, but they're big.
And if you wanted if you think you understand that a a uh an inductor in a 630 mic a tuner mic actually not be big, it might be big.
So, um it's a problem. And and again, capacitors, you might have a nice vacuum variable of 1,000 picofarad like this, uh couple of thousand volts, but I need 160 nanofarad.
So, you can imagine, uh taking a tuner with 160 nanofarad vacuum variable through security when you're traveling. Doesn't work.
So, um my problems were that I had to get some form of model designed so that the model would accurately reflect the components that I had or needed.
It was difficult. So, um the other thing is, I can't do printed circuit boards easily. I have to use the uh multi-dotted um development boards, which means that I've got an awful lot of uh parasitic capacitance. Mind you, parasitic capacitance at 160 nanofarad is not that relevant, but I'm a purist, so I like to get rid of all of the extra capacitance.
And um there are obviously uh PIN diode designs for uh HF, but PIN diodes will not work below about 1 MHz.
They just rectify. And so, again, I couldn't use a PIN diode. So, I had to think, well, what am I going to use instead of a PIN diode to do a switched capacitor?
And what I was going to do was generate a switched like what's on a rotary capacitor with series and just switch with some form, either relay or electronic switch.
Um and the other thing is China would not export any tape money or export due to Russian sanctions for any printed circuit boards that that was going to be sent to Bishkek. So, I was stuffed there. So, I had to actually make the boards. And making the boards in an apartment with a an ex-wife who you want to use the kitchen table and that you're not making boards on that table.
Um So, there are there are a lot of little issues, but uh we had to solve it and we did. Okay.
Here's an example of a switch capacitor.
You set the step size and you start on with the uh the initial capacitance and then you step down by seriesing the value down in steps.
It's a very it effective it works quite well.
Um, and that's with a rotary switch. You just have the capacitances in series out on the outside of the rotary switch and you step down to each of the steps.
But of course, you look at that and my basic capacitance there is 56 560 picofarad and 56 steps.
Well, I'm actually probably going to need 150 nanofarad with 100 with 1,500 picofarad steps. So, it's it's issues.
Okay.
Those worked and I did but then I was thinking there must be a simple way.
And then I I I decided that the coarse bank and dual bank were not really going to be the end of it all.
So, I then looked at electronic switching. And I did looked at hey, hang on now. What about light switches that use an opto a triac?
And let's use if the opto triac can switch 250 volts at 50 hertz, why can't it switch at RF frequencies?
So, I built up and tested and I I got the MOC Oh, it's not MOC, it's MOC 3023 and set up using the MOC 3023 to switch down the tap points on the on the switch capacitors. So, you just switch number one tap, now switch number two tap, etc. Which is fine, but the number of steps is relatively small. Because I can use a 16 uh sorry, an 8-bit shift register and therefore I only get eight steps.
I can do eight steps at coarse and eight steps at fine, but that gives 64 steps total.
But, it's awkward because you've got to have twin gang of of for tune and twin gang for the load.
And as you see, I did a test using an MOC 3023, fed the RF through it at and it was from a QDX QMX, so the third harmonic was fairly low at 63 minus 63, but fed through the triac it actually raised it to 40 minus 40 odd dB, which is okay, but it's not good.
I'd like to be perfect and the reason is the the opto-triacs have a a voltage across that it switches on and then drops and then switches off and then switches on again. So, there is a switching transient which degrades the signal a little bit, gives distortion because you've got some little bit of inrush of current and therefore almost square wave start.
And so, you get this distortion.
I did find a solution for that, but I'll I'll explain that a bit later.
So, there was an example of RF into the MOS without the MOC, and RF through the track. And you can see that there's uh quite a bit of more um third harmonics there, which I didn't want. But, at 6:30 m, minus 42 dB, I could I could deal with that. I didn't want to have to, but I could deal with it. So, I looked at doing uh real real ICs as well.
Um but, if you look at the uh the shift register, um a normal 74HC or AAC won't deal with reasonably high currents. It'll only deal with 20 milliamps odd.
But, there's the TPIC, the high-power Motorola stuff, which has an open drain um uh FET on the actual switch, which meant that you could switch up to an amp or so on the control line, which for that were great. And I thought, okay, I'll design circuits to use um designed a a uh a uh a Python routine, which would switch numbers into the register and select capacitors. And I had a pair of um uh rotary encoders, and you could tune up tune down manually.
And you could select that capacitance by specifying the capacitance value, or the tune up tune down. So, it it could be done fully uh electronically.
Okay. So, I looked at these uh parallel lines, and as I said, I thought, "Now, hang on.
Why am I bothering to use a series array? Why don't I use step values and just parallel them up?
Because I can have the shift register parallel all eight bits.
It might be 360 milliamps of triac opto-triac current, but I would get all I could go from one step up to 112 steps continuously. So, I did that. And I I got some very effective capacitors. And this was the look-up table that I had using a step value, and then it would just send the the data value out to the shift register. And so, you could actually say to the computer, "Okay, I want a 333 or 300 Sorry, yeah, 330 picofarad capacitor. So, I send that value to the shift register."
And it's that's a very important feature. And I'll when I get to the automated side, how why I would do it that way.
And that's just the complete look-up table. And you can see the shift registers could do 111 steps include of any size step capacitor that you wanted.
I could put a thousand picofarad in the step.
And so, you'd have 150 111 * 1,000 picofarad. In other words, 100 111 nanofarad capacitor. Quite a decent capacitor.
Still not quite enough for 8,900 hertz, but that's another matter.
So, uh it was a much more efficient approach that allowed you to step up, step down, or select directly.
Here is my first effort in constructing a twin gang fully switched. And the beauty of these these MOC 3023s they required 60 microamps to fully drive the the thing on.
And what I did was I found I modeled it in spice and I found that a a red LED in series with the two MOCs in series with the voltage drop across the open drain would give just nice 61 or 62 milliamps with the two and which means I had the two sides of the capacitor the twin gang already with two switches. And so what I would do is I I can't show I've got these by the way in my suitcase there and you can look at them all tomorrow night. But I would put a series of capacitors there series of capacitors across the bottom and there's a dual gang that would go 100 111 steps at whatever capacitors and not only that you can put whatever voltage SMD capacitors in there and I'm using 18 1808 1000 volts. So I've got a 1000 volt capacitor there. I can put a 5000 volt SMD cap if I can get them not in Bishkek but if I can get them in there. So but the beauty of these things is you can put any step value and you can specify to the python routine select this value.
Of course as long as you put the value in the look up table because you it's got to look up the table and then select capacitance or as I said you just use the rotary encoder and manually tune up or manually tune down. They worked very well.
I didn't like the idea of having any distortion. So I built miniature relay version of this. Um these are tiny little relays that I've got off AliExpress.
My biggest problem is I can get them off AliExpress. Uh but it's usually 1 or 2 months after I think about what I want, you know.
And you've got to think strategically ahead what components you might need and what you're going to get. And then you get them and go, "Oh, what did I get this for?" you know.
As you get the deliveries, but it's it's a problem because it's always this delay of I want to build it now, but I haven't got the bits. So, I start collecting lots of relays. I've got about 10 or 12 different types of relays, little miniature ones. I've got reed relays, 3-V reed relays, 5-V relay relays, 12-V I've got I've got all sorts of things. But the the bottom line is I've actually probably spent two or three or five times what I really need to spend, but I've got a great check uh component chunk box.
Anyway, that's the uh TPI 60 um 6B595 1-A.
Oh, and the one thing I forgot to mention, the um the opto isolator triacs are good for 600 V.
So, you they've got a peak voltage of 600 V. So, and they will transmit 1 A through. So, RF of 600 V and 1 A means that I can do reasonable amount of power through these capacitors. Actually, less than the relays because the relays probably have a peak voltage of a 2 or 300 V.
Um anyway, that's another matter.
And here is my Raspberry Pi 2040 uh with two um two rotary encoders. They work quite well.
Okay.
There's another one two twin gangs on the same board.
So, this one is really and the I calculated the resistance equivalent of the diode was 24 ohms and that gives another at 61 62 milliamps through uh sorry, in this case it doesn't it actually gives Yeah, 63 milliamps. These are opto triacs, but I put them in the vertical mode to try and get the capacitance away like split the RF side from the the DC control side because you will get capacity Mind you, as I said at 160 nanofarad, it's not that critical, but I I was designing these things for small tuners and large tuners.
So, I wanted it to be suitable.
Um So, there is the circuit of the opto triacs and I thought I read through the um the specs on it and I found that the if you put a 100 microamp of DC in either direction, they stay on.
So, you don't get the switch on switch off. So, you lose the I I believe that we will lose the uh um the degradation of the signal, the square wave, the noise, the whatever.
Yet to be fully tested, but um the uh the zero The only question is if you're putting a current through, are you changing the baseline of the turn on or turn off or are you actually allowing it to turn on the entire time?
And as you see, I've got the uh the three resistors. This is on the twin gang version. The three resistors which are uh about 10 ohms uh 10 K, 10 K, and 10 K.
Now, think about it. At 8,900 Hz, 10,000 ohms of reactance is going to be a bloody big uh choke, isn't it?
At 8,900 Hz, even at 630 m, a choke of 10 10 K reactance is going to be quite large. So, by doing it this way, I actually get rid of large components that I need to build.
And of course, with a choke like that, it's going to be on very small wire, so I'm going to get a high DC resistance as well.
Okay.
I think we might fix the uh the third harmonic um distortion of the signal, but that's to be determined. It's It's As I said at the start, -62 to -42 45, it's barely acceptable. And let's face it, below 20 K uh Hz, it doesn't matter.
Because below 20 K is not regulated. The FCC rules don't apply to under 20 K, do they?
Okay.
Next thing are the inductors.
You need bloody big inductors.
And what you don't need is a rubbish tin with a half a kilometer of wire wrapped around it just to get three or 400 uh microhenries.
So, what do I do? I use Litz wire, and I series the windings. I split them into windings of uh bundles of six wires on a 30 uh wire Litz, six wires in five bundles.
Which means that you have to match 30 wires front to back.
There is a strategy that it's actually not difficult and you see here the various sizes of inductors that I've got and you'd see here on this bottom one I've actually labeled the windings because you can get yourself very confused with these things when you've got 30 wires, six wires in bundle and five bundles. You go, "Have I got the right one?"
Okay.
This big toroid is 16 cm in diameter.
Sorry, 18 cm in diameter.
And by series joining up the five windings, I get 7.534 millihenries of inductance.
And it's relatively small and the Q of it is actually quite high. A reactance at 8900 hertz to 7 millihenries is about 420 ohms.
The DC component resistance is about 12 ohms. So, you do a Q calculation, it's actually 20.
And a Q of 20 on a very large inductor like that is not bad. It's not bad.
Okay.
Here's what it looks like when you're matching the bundles.
And for an old bloke with whose eyes are now arms are too short for his eyes, I have to put on two pair of glasses.
Make sure that I've soldered each of the windings and then match it. And the way I do it, it's it's really quite simple.
One end, you separate them into five bundles of six wires. You just any five, any six, it doesn't matter.
Then you tin those together. So, you've got all the wires at one end tinned and joined.
Then you tin the 30 wires the other end and you start with a multimeter on on sound. You choose a wire and then you find the matching wire that has zero resistance forward and back.
Now, that means in the first set of wires you're looking at, you've got one wire out of 30. So, you've got 29 left, five of which you'll continuous.
Right?
Because of five of those wires is is a bundle, six of them.
Right? So, it becomes easier to find as you find more wires, then you've got a lot fewer to check and a lot fewer to find for the next bundles. So, but takes a bit of time, and as you see, that's if you're doing a little toroid, it's it's quite annoying.
Okay.
Here we go. The um I decided that it is very prudent to measure the inductance of each of the Well, first of all, the overall inductance of a single turn when you before you split it, you measure one turn's inductance.
And then when you split and match the five bundles correctly, uh checking resistance between each of the windings to make sure you haven't cross-connected because you build the thing and then you find you've got one wire out of opposite bundles connected and you got to you know, got to do the whole thing again. Um do you understand what I mean there? It can be a real pain if you cross-connect two of the wires.
Um then you check each of the inductance at the various tap points, like zero, first in first first winding, second winding, third. And as you'll see, they're not exactly uh 25 times the the minimum or and each of them are because what you've got is an awful lot of coupling between the windings. And as you increase the number of uh windings that are coupled with the fifth, you get more inductance.
I know that's probably a bad explanation, but you do.
Okay.
So, here you also do a little spreadsheet and you calculate out initial in um the inductance at the first tap, the inductance of the second tap. Then you calculate how much the difference inductance between the first and the second. This is vital for the new model that works in in uh LTSpice because then you can actually specify how much inductance is there.
Okay.
So, as I said, you've got the five windings.
You split it out into the five. You calculate the additional inductance between two points like the third point and then it's the first plus the second and the difference between the second and third and so on.
Is Do you understand what I mean there?
Anybody not understand what I mean?
Yep.
Okay, again.
If you measure the inductance at the third tap, you need to calculate the inductance between the third tap and the second tap.
So, you take the taps below, you subtract them from the third tap, and then that is the difference between the second and the third or the third and the fourth or the fourth and the because you need that figure for the model.
Right? And then you add them all up so that you make sure that you've got the real values. So, you know, you need to make sure that this figure and that the the sum of those figures are the same.
Okay, and this is a 21 uh Henry micro Henry tapped a toroid and I used this one on 160 m and it works quite well.
Okay.
I then looked at my original uh tuner that works really well with a variometer and the the two gangs, but I realized that I'd never put the additional capacitance on the tuning side of the gang for the lower tuner.
I'd had value variable capacitor and I hadn't put the same value and I thought, "Well, if that's the case, why do I need two gangs?" So, I tested with um uh instead of having this gang and this gang and this gang and this gang, I tested with the model with single gangs.
All right. This is an actual model we have, but I never found that the voltages that I was measuring uh with the capacitances, with the inductances, with the frequencies matched the actual voltages on the the model.
So, there was not quite it until I did this calculation of inductance. When I did the calculation of inductance and put the differences in all these rather than the actual value, then the model worked.
So, in order to do the um the automatic side, what I wanted to do was to do a calculation of the the load [snorts] that is seen on the output of the tuner.
And I can do that with a three voltage meter at three-volt meter method that's been described in calculating impedance.
Anybody not familiar with the three-volt meter induct- impedance measurement?
Okay, there is a paper or two where you you measure the voltage across the a resistor a um a standard load, the voltage across the actual load, and the voltage across the uh the source voltage, and you can actually calculate what the load impedance is.
What I wanted to do was have the three-volt meter measure the load.
I want in a model that I could look at the model and say, "Okay, with a model of 5,000 ohms on the load, I need this capacitance and this capacitance for tuning and load." Set up a lookup table.
And then if if the tuner measured that the load was going to be 6,000 ohms, it would choose from the liquid lookup table the values of the capacitances for the load and tune on 6,000 ohms and start from there.
And who is not familiar with the Wheatstone SWR LED uh indicator uh indicator? Anybody not familiar with that?
Wheatstone bridge load uh LED in the middle to detect if there's any imbalance, and you measure the the load that you're So, what I would do is have if I was using a 50-ohm load, fine, I'd have 50 ohms, and then I'd measure if and make sure that the input to the tuner measured 50 ohms or matched 50 ohms, not measured but matched.
Um if I was doing 8,900 hertz, I'd be using 4 ohms uh loads. Why? Because um I can use audio amps with a 4 ohm uh load uh to power the transmitter. So, I'd I'd have a a a Wheatstone bridge of 4 ohms, not 50.
That's quite logical, really.
Okay.
And at the end of the output, I would have a toroid current antenna current sensor.
So, what I'd do is I'd switch on the thing, it would measure the uh load.
It would set the capacitors.
It would then look at the Wheatstone bridge and say, "Yes, it is matched." Or up and down, maybe it's not, maybe it is. And in the Wheatstone bridge, I'd put an optoisolator would feed just back as a a resistance measurement, and I want minimum resistance or maximum resistance.
I.e., zero lead up on that.
Um and check the current that is being fed out the antenna.
And if you've got maximum current to the antenna, what have you got? Best signal.
Right. That was the strategy in what the tuning, and that way I'd I'd look at set it, measure it, confirm it.
Okay.
Now, I did a good model, and here's an interesting little This is 62 turns of um of Litz wire on a 56 mm No, 54 mm toroid.
Calculated from my my my uh Spice model, which said I needed this picofarad on the tune and this picofarad on the Well, nanofarad, actually. And this nanofarad on the on the load, and I put it on the thing, and guess what?
It got a pretty good tune just from pure calculation. So, what does that mean? My model works.
And if you look at it, you can't quite see it there, but it was 1.12 or something to 1 SWR.
Now, I put in a a series tuning cap here so that I could adjust the load value, and I got it down to 1.06 to 1, which is probably equipment error, not necessarily match error. But, it's a pretty good match. I I like to go for less than 1.0.
Usually, in 1.01, 1.02.
I don't like having 1.2. I don't transmit on 1.5.
Let me put it that way.
Okay, so, there is a a 220-m tuner that is only this big with a 54-mm toroid um that is a good match.
I haven't got any auto tuners yet because I would just It took so long to get these models correct, and so long to get the values, etc. I really wanted to have two tuners and have a station work here with an 8900-Hz whisper signal between them here. Maybe next year.
But, um what I do have is four four four uh micro tuners, 160, 630, 2200, and 8900 Hz. Now, 8900 Hz, the tuner is actually about this big, but when you consider the wavelength is 33 km, it's a micro tuner.
Um all of my time was really taken getting an effective spice model and design that would actually work.
And this last side, that's the demons- that's the validation. And what I did do is I had a spice model with.353 5 V in the trans input fault peak to peak.
I then had I measured the voltages on the model on the peak of the tuning circuit and on the output circuit and they matched the real thing very closely.
But when I on the model, when I put 5 W into it, 22.5 V peak, I get 800 V on the the tuning circuit. So, you need capacitors that are high value and you need to know that value because there's no point in putting a polyvaricon in there which has got about 120 150 breakdown voltage or a an NP0 50 V capacitor in there. Just won't work.
Um I have a 630 m antenna tuner like this using two polyvaricons because I've actually calculated that I don't you get high voltages on the model. It's all due to the peak inductance of the and model etc. The same size as a QDX QMX case.
So, that's a tuner that'd be for 630 m.
I've got a 160 m tuner which is that big and I've got 2200 m the tuner that that's big and a an 89 100 Hz tuner that's that big. I have all these I've got samples of the printed toroids, etc. And you'll see them tomorrow night if you come to the home brew thing. They'll all be a table full of bits.
Okay, any questions?
Yes.
Uh I'm curious, what what sort of antennas are you matching?
Okay, let me give you an example. Were you here last year? No, I wasn't. Okay.
Then you would have known precisely what antenna I was matching because last year I I gave my presentation on stealth operations from hotel rooms and other unlikely QTHs on 630 and 2200 m. Which seems almost impossible I got 112 spots at up to 2500 km with a 12-m wire out the window.
With just a just a simple wire? Well, it wasn't simple because it was coupling into it extra high tension power lines, but that's another matter.
When I say coupling, it was like the stub on a on a Yagi. I had a stub underneath the antenna line the my antenna the high tension power lines.
Now, I cannot explain any other way that I got spots 2500 km to the north of England with two to three watts on a 12-m antenna.
Can you explain?
No.
Are going from 50 ohms to some extreme Hi, yeah. Yeah.
Free matches are very good for doing up to 10 or 15,000 ohms. no problem.
Yeah.
Would some train tracks also work? With what? Train tracks.
As antenna?
Ah, catenaries, yes.
I intend to actually um put a wire between two catenary supports and try and load into the European rail network.
>> [laughter] >> Look, uh but last year where is Shawn here?
No, Shawn's not here, is he? Last year he complained OH&S that I was telling people to throw um wires across um extra high tension. Mind you, a wire wrapped wire across 400 kilovolts isn't going to be anything but a fuse.
Poof.
Okay. Any other questions?
>> Yeah, I'm still here.
>> Oh, you are, Shawn.
>> The reason that I objected is because what you're doing is prohibited by the electrical code, not because you're absolutely going to kill yourself.
>> it on the wire.
>> But you're probably going to die one day and that that catenary that catenary trick is really going to get you.
We'll see.
>> And if you think the rail if you think the utilities don't like RF, they didn't put their going into their system, Yeah.
the railroad is going to like it a whole bunch. They got 15 kilovolts, very high current.
>> Yeah, isn't that impressive and deadly?
Sounds good.
>> But I'm not touching the power lines.
I'm not touching the power lines.
>> that say you can't do that stuff.
I'm not telling you to do it. I'm just saying that I did it.
I'm not telling you to do it.
And I wouldn't recommend doing it because you need to be a very good shot with a slingshot to do it properly.
Yeah.
These um These circuits are are very very Q very high Q. They're very Well, 20 20 Q on a uh a 8900 How much does would your SWR change if you're moving maybe 50 or 100 hertz?
Would it be very small?
>> Good question. Yeah.
That's a very good question.
>> the US >> Let me investigate it.
The US Navy for quite a while has transmitted to their submarines at 20 20 kilohertz, and they had a a coil for mark and another coil for space. Yeah, and massive antennas. Uh a couple of acres suspended up at about 1200 feet.
>> I am doing QRP portable equipment.
>> I understand.
>> And um my my equipment is small.
And it's all because I've developed this ability to use Litz wire and um the printed toroids and printed variometers that actually work very well. My printed variometers that I showed the first year it's 124 mm spherical variometer on which the inside coil was 0.5 mm spacing from the outside coil, which means it had very very close coupling.
Its peak inductance of a single winding was 1,860 microhenries.
The minimum inductance was 11 microhenries, which is a 181 range on that. Why? Because of the coupling and the geometry.
The the Navy's uh variometer was about 4 m in diameter and a million watts. Yeah.
That's why Litz wire is good fun. Yeah, good. That's it for our questions. We do need to clear out so that staff can get tables broken down.
Videos Relacionados
U.S. Military Just Flexed The Most Dangerous Aircraft Ever Built The F-47
MaxAfterburnerusa
11K views•2026-05-29
Heating Staying On On The Hottest Day Of The Year
PlumbLikeTom
507 views•2026-05-29
발전 효율을 높이는 태양광 추적 시스템의 기술적 원리 #공학 #공정 #태양광 #알고리즘 #재생에너지
찐현장기술
2K views•2026-05-29
How Far Can A Tomahawk Missile Actually Travel?
WarCurious
13K views•2026-05-28
직관 및 곡관 배관 결합 고정 작업 #worker #process #fabrication #pipework #clamp
월드촌촌
2K views•2026-05-30
Wire To Wire Connection Trick | Strong And Secure Electrical Joint #shortvideo #wireworks
ElectricianTips-b1h
5K views•2026-06-02
Peterborough to Newark Northgate Driver's Eye View aboard an InterCity 225 - East Coast Main Line
TrainsTrainsTrains
822 views•2026-05-31
AI turbine design: hypersonic cooling leap #shorts #ai #hypersonic
bobbby_rn
671 views•2026-05-31











