Measuring an antenna matching unit's settings?

[Chris Wilson]

I have a borrowed Yaesu FC-902 antenna matching (tuning) unit that I have set to match a monoband 20 meter two element cubical quad to the 50 Ohm output / input impedances of my Kenwood TS-590S transceiver. I want to build either a fixed component LC network to match in the same way, or modify an unknown, bought used, home built ATU to match in a similar way on 20 metres.

I have an AIM 4170 antenna analyser and I have scanned the feedline at the radio end direct to the antenna, then via the borrowed Yaesu ATU set to "straight through" so I am really just adding a piece of 2 foot long RG=213 and connectors, plus the internal LC of the straight through wiring and switching of the ATU, then finally I scanned with the ATU set for best match, which it did actually set for the 17 metre, rather than the actual 20 metre band.

Do I have enough information here to make an LC network or know if the LC range of the bought used ATU are adequate to duplicate this matching network? If not what do i need to do save dismantling the Yaesu and measuring L and C values disconnected, which as it's not mine is a none starter ? Thanks

[iMo]

Do sweep the antenna "system" with your vna at the point you want insert your L-match.
Export the s1p data.
Import s1p file into SimSmith and you are done.. (you tune up there the LC for the best match).

Example:
https://www.eevblog.com/forum/rf-microwave/question-about-longwire-antenna-setup/msg3151928/#msg3151928

[DH1AKF_Wolfgang]

There are some matching calculators online, e.g.

http://home.sandiego.edu/~ekim/e194rfs01/jwmatcher/matcher2.html

Your values give the following results: (see attachement)

[A.Z.]

Since you have an analyzer, if it allows to, you may run a check of the antenna impedance at the antenna side (yes, not at coax end) and consider the idea of adding an impedance adaptation there, also, and since we're at it, doesn't that quad already have a tuning circuit (stub or the like) to allow tuning it to resonance ?

[Chris Wilson]

Thanks for the replies so far, much appreciated.


I can tune the loop diameter size for resonance but there is still a biggish mismatch of impedance. If I tune for minimum SWR then resonance is miles away. I am told not to worry too much about it being none resonant, but tests with the borrowed  Yaesu antenna matching unit showed a near perfect 1.02 to 1 SWR, something loop size adjustment and playing with 75 series stubs has failed to approach.

I knocked together a variable cap and a roller inductor, parts of the used home brew ATU I bought at a rally. (I now believe the home brewed ATU is an "Ultimate Transmatch" with a split stator variable of 250pF per section, a roller inductor and a separate single stage variable cap of 250pF, but the wiring is suspect and i think incorrect). If you look at the drawing I found on line, in theory image 2 should be what I need, with the load higher resistance than the source. But that would not match the antenna. I tried connecting the input and outputs in reverse, in effect creating image 1. This was able to generate a near perfect match of 1.03 to 1 at really close to 50 Ohms. I don't understand why this version should match when it's really for matching a load of lower impedance than the source...?



I think I will examine the home brew used transmatch and now I have the details for the "Ultimate Transmatch" I am pretty sure it was based on, rewire it properly and see what it does then.


http://www.arrl.org/files/file/Technology/tis/info/pdf/7007024.pdf


But any other suggestions very welcome. I have a scan of the antenna using the AIM "refer to antenna" option, which is similar to having the analyser connected at the actual antenna feedpoint, if it's of help?

[cdev]

There is a program lilled LinSmith on Linux that I have been fooling around, with, trying to get a feel for the effects of various impedance transforms.

Its based on a Windows program - I think its name is SimSmith.

 You have a directional coupler. )


I recently got a nanovna 2 for $60 - it works up to over 3 Gz which should keep me for a while. You obviously know more than I do. I just wanted to put a plug in for these tools which I am just beginning to explore as some things that might help.

[A.Z.]

I can tune the loop diameter size for resonance but there is still a biggish mismatch of impedance.

So, what's the impedance when the antenna is brought to resonance ? See, knowing that it would be easy to add a simple impedance transformation unit to the feedpoint so that the impedance of the antenna will then match the one of the feedline, although... in general cubical-quad antennas are designed so that element spacing gives an impedance around 50 Ohm, and then a gamma or omega match (or similar) is used to match the antenna to the feedline, it's strange that your quad doesn't have such a match at its feedpoint

[edit]

As for the matching, have a look at this (let alone the relay - it's to switch a multiband quad, not needed in your case)

[Chris Wilson]

OK, bear in mind I was told that I should ignore resonance and trim the driven element for minimum SWR. I did this and then added a 1/4 wave length of 75 Ohm co-ax at the balun. So I currently have the driven loop, a dual core ferrite choke balun as per G3TXQ's  balun pages for 20 metres direct to the ends of the driven loop, in a box, and then a 1/4 wave length of 75 Ohm co-ax cut for 14.2 MHz, then the (temporary) RG-58 feeder to the TS-590S.

The AIM has a "refer to antenna" function whereby you do the calibration with an open, short, and 50 Ohm load at the END of the co-ax and the software magically ignores the co-ax and the results are if one had the AIM4170 at the end of the coa-ax at the antenna feed point. this test currently shows this result for a sweep of 13 to 15 MHz, and for 10 to 20 MHz. Thanks.

http://www.chriswilson.tv/refer-to-antenna-sweep.jpg

http://www.chriswilson.tv/10-to-20mhz-refer-to-antenna.jpg

[A.Z.]

Hmmm... what's the balun ratio, please ?

[Chris Wilson]

Hmmm... what's the balun ratio, please ?

It's a choke, a ferrite toroid wound with whatever the recommended turns of co-ax was. Correct number of turns, correct mix of ferrite, would need to refer to my notes which are not here at the moment. It's not a transformer. The only impedance transformation is in the series 1/4 wave stub of 75 Ohm co-ax. Thanks A-Z

[A.Z.]

Ok, so I'm assuming it's a 1:1 choke, now, looking at the pics you posted (not there anymore ?!?) the impedance seems to be around 30 Ohms (if I saw that correctly), so you may place a 1:2 balun (or even better a 1:2.5 one) between the antenna feedpoint and the choke, that should give you an impedance of around 60 Ohms (75 Ohms with the 1:2.5) which will better match the coax impedance and give you lower SWR

[Chris Wilson]

[ Specified attachment is not available ]Right, thanks, I can certainly try that! In the meantime I have hastily, and without much if any attention to using copper strip connections, returned my "Ultimate Transmatch" to standard configuration. At first, injecting a very low level signal at 14.2 MHz from the AIM4170 into it, and with the antenna as it is, I thought I had done something wrong in the wiring, but eventually I got a good match. BUT it's INCREDIBLY touch, even with vernier drives on the two variable caps,  a minute change sees the SWR go off the scale. this is reflected in the graphs, too. When matched if I put my hand within 10 inches of the variable inductor, it goes to pot. My only input is maybe C2 is too small? C1A and C1B are 13 to 346 pF, it's a split stator variable, and those values are per side. C2 is also 13 to 346pF and needs the vanes fully meshed to get a match. Could it need a padding cap? I think the term is it's very high Q?

http://www.arrl.org/files/file/Technology/tis/info/pdf/7007024.pdf

Anyone understand transmatch tech better than I please step forward

[A.Z.]

as I already suggested, pick a pigtail, connect it directly to the antenna feedpoint and to the analyzer and run a sweep, then post the results here, all the remainder is a waste of time and effort, plus, the lack of a matching circuit at the antenna radiator makes me think that you either assembled it the wrong way or are lacking some bits and pieces, see usually quads have both a gamma/omega/... match at feedpoint and a tail, at the reflector, to adjust the f/b

[A.Z.]

As for feeding and matching the antenna, I believe you may find some useful informations about impedance here, and then here is the pdf version of the "All about cubical quad antennas" by William I. Orr (W6SAI)

The above being said, I've looked at previous discussions here and sounds like you have "bits and pieces" from a "Gem cubical quad" antenna, now, a bit of googling allowed me to find that the (commercial version of the) Gem quad is fed using a balun and that the reflector has a tuning stub to adjust the F/B ratio, so you may seek for informations about the original balun, plus, after you solve your matching issue, you may want to adjust the reflector for max attenuation when the back of the antenna is pointed toward a signal, that should allow you to squeeze out the best from your quad

[A.Z.]

I got a good match. BUT it's INCREDIBLY touch, even with vernier drives on the two variable caps,  a minute change...

My suggestion is to read this (and the following pages covering matching units), also and since you're at it, also check the links related to the antenna tuners found on the left sidebar and this link about symmetrical matching units, I think (hope) that the informations found there may be enough to get your started; just remember that an antenna matchbox isn't a "magical device", it doesn't "tune the antenna", it only allows your TX to see a good match and low SWR, but the antenna (and the feedline) will remain unmatched

[cdev]

I am trying to match a homemade version of a two band stub loaded jpole (the one by Ed Fong, thats famous here in the US) I have it matched to 1.06 on 146 MHz, but the match on 445 MHz is only around 1.65:1 So I have been chopping off bits here, adding bits there. Were it not for the nanoVNA I got recently it would be hopeless figuring out what is going on. Even with it it isnt easy because proximity to anything at all screws it up totally. This antenna to work, literally needs toi be hanging in free space with nothing around it for two wavelengths.

Try to figure out the theory of operation of the antenna and use that to figure out what to do.

[A.Z.]

Try to figure out the theory of operation of the antenna and use that to figure out what to do.

Isn't that similar to a "slim jim" antenna ? If so, here are some infos and a calculator, then I just wonder why you "hijacked" this thread instead of opening a separate one, not that it's a problem with me, but a separate discussion may help others fiddling with the same antenna to easily find informations 

[cdev]

Well, the one that I am thinking about is quite different because it uses a shorted coaxial stub for both impedance matching and decoupling.


Look up DBJ-1 and DBJ-2.  Its quite effective but makes the design very finicky about its immediate environment, and I think that quality is a fairly universal one for very high Q optimized antennas.

[A.Z.]

check out this

https://sites.google.com/view/sp3l-hf-antennas/cubical-quad

the quad there is a tad different, but the document may give you some ideas