Broadband transmission line transformer design ?

[Al (astral highway)]

Photo from 'Practical Electronics for Inventors', which is a neat reference work of mine. It discusses a broadband transmission line transformer in the context of overcoming RF impedance-matching challenges.

I'm still focussed on this as I optimise my home AM rebroadcast set-up and try to improve VSWR into a short antenna. The base-loaded antenna design mooted in another thread by Argus25 is a great start, but I'd like to pop such a broadband transmission line transformer into an RF power stage.

The concept is described:

'A broadband transmission line transformer is a simple device that consists of a few turn of miniature coax cable or twisted-pair cable wound about a ferrite core. Unlike conventional transformers, this device can readily handle high-frequency matching (it geometry eliminates capacitative and inductive resonance behaviour ). These devices can handle various impedance transformations and can do so with incredibly good broadband performance (less than 1dB loss from 0.1 to 500MHz.)' . P.270. Sherz.P, Monk, S. Practical Electromics for Inventors, McGraw Hill,

Only the design isn't elaborated. My immediate application is for a centre frequency around 1.4MHz ...

Any help in design concept and detail would be greatly appreciated.

Thank you

[Radio Wrangler]

The seminal article is C L Ruthroff "Some broadband transformers" Proc IERE August 1959.

Then Jerry Sevick wrote an awful lot (probably far too much) categorising baluns in vanishing detail.

David

[kalee20]

Just looking at your application (matching a short antenna into an AM rebroadcast set-up (pantry transmitter?), this actually looks like a comparatively narrow-band issue! Frequency range a fraction of an octave... or have I misunderstood?

[Al (astral highway)]

Originally Posted by kalee20

Just looking at your application (matching a short antenna into an AM rebroadcast set-up (pantry transmitter?), this actually looks like a comparatively narrow-band issue!

It's possibly me who has understood. I am interested in the impedance-matching virtues that are underscored in the text I quoted. I assumed that from this description, such a broadband transformer has the capacity to match disparate impedances between any of the frequencies described. I didn't take it to mean that it does so only when there is a varying output signal between any of these frequencies...

[G8HQP Dave]

I think the point is that such a transformer is good at matching impedances near those for which it was designed, over a wide range of frequencies. You perhaps have the opposite problem: a wide range of impedance (nowhere near a typical transmission line impedance) over a narrow range of frequencies. For your purpose a tuned matcher may be better. However, you have to be careful as too good a match could turn you into a broadcaster.

[Al (astral highway)]

Originally Posted by G8HQP Dave

You perhaps have ... a wide range of impedance (nowhere near a typical transmission line impedance) over a narrow range of frequencies. For your purpose a tuned matcher may be better.

Spot on, Dave. I've now got it exactly. Thank you for clearing it up. The thing is, it's hard to come across good designs for tuned matchers.

I do want to keep my 2mH inductor base load and build from there.
At the moment, it's the 2mH base load plus a short antenna (3m) terminated in a stainless steel ball with an estimated capacitance of 15 pF. I don't have a scientific way of evaluating VSWR but I suppose I could use a series bulb or similar for tuning? I don't enjoy working in a haphazard way.

I have seen many variants of pi networks to deal with harmonics, but at the moment, these are not my focus, it's just the impedance match.

Originally Posted by G8HQP Dave

However, you have to be careful as too good a match could turn you into a broadcaster.

Ahah, yes. Reasonable match is good enough for me, but I'm far from close to it at the moment.

[MrBungle]

Definitely a matching network job. Look up L or Pi to T match tuner. Only problem is you need to either know the antenna impedance or have an SWR meter as well. You can build the latter with little effort. There are some neat designs that use LEDs instead of meters etc as well.

Ignore Practical Electronics for Inventors for this sort of stuff. I really don't think it's a very good book if I'm honest. Experimental Methods in RF design by Wes Hayward has a section (3.6) on matching networks. If you can get hold of a copy of this it'll keep you up all night with the perpetual enlightenment on matters such as this. It's not impossible to obtain an electronic copy for review if you know where to look.

[Al (astral highway)]

Originally Posted by MrBungle

Only problem is you need to either know the antenna impedance or have an SWR meter as well. You can build the latter with little effort. There are some neat designs that use LEDs instead of meters etc as well.

...Yes, I do need to know the antenna impedance, so I guess that's the next stage, building an SWR meter. Thanks for the tip...

Originally Posted by MrBungle

. Experimental Methods in RF design by Wes Hayward has a section (3.6) on matching networks. If you can get hold of a copy of this it'll keep you up all night with the perpetual enlightenment on matters such as this.

Ahah! I'll have a scout around, thank you! £100 for the hardback, I'll have to dig deeper on the internet!

[MrBungle]

Annoyingly RSGB shop used to stock it for £35 which is more reasonable but alas no longer.

[Radio Wrangler]

Is there something about books I like and silly prices. Experimental methods in RF design is a really interesting read. £100.... wow! I also have a copy of C M Hargreaves The Philips Stirling Engine which was over £600 the last time one appeared up the Amazon. I must up my security. Uggrade the wolverines to rabid wolverines or something.

You design matching networks on Smith charts. I use LLSmith from www.rfdude.com Lance was a software guy made redundant who decided to take up RFery. I was one of the people who helped him a bit with his transition.

David

[Terry_VK5TM]

Experimental Methods in RF Design has just recently been re-released by the ARRL (American Radio Relay League) for $49.95.

An added benefit is that it comes with a CD containing electronic copies of two other very popular books of the time (1990's)

Terry

[MrBungle]

Unfortunately if you order it from arrl.org, there's $30 on top of that for delivery that is tracked and then the poor exchange rate.

There are copies on a certain auction site and abebooks.co.uk at around £55 which is slightly less.

[ms660]

I wonder if a noise bridge would measure your antenna's impedance?:

https://www.youtube.com/watch?v=fiMbK8pjfNU

Lawrence.

[G8HQP Dave]

The thing is, it's hard to come across good designs for tuned matchers.

That is because they need to be designed individually to meet a particular requirement. There are websites which do this, I believe. Or you can learn how to do series-parallel transformations and calculate it yourself.

An SWR meter will be almost useless for measuring a small loop antenna. This is because an SWR meter is designed for use in a 50ohm system (or some system) and basically it measures how far off this standard you are - but doesn't tell you which direction (i.e. impedance) you are off. It gets less accurate the further away you are from a match. Your loop is nowhere near 50ohms so an SWR meter will simply tell you that it is a very bad match.

Your best bet is to measure the inductance of the loop or calculate it. Then calculate the radiation resistance (proportional to loop area) and add the conductor resistance (calculate or measure) to get the total resistance. Then at a particular frequency you can calculate the series impedance. From that you can design a matcher.

[Radio Wrangler]

Originally Posted by astral highway

so I guess that's the next stage, building an SWR meter. Thanks for the tip...

I've got that one covered for you

Sprat number 61. "A dual directional power meter"

It just sits there with 2 meters (or LED bargraphs or frog's legs or whatever indicators) One shows you the forwards power, one shows you how much gets reflected from the load. Much easier to visualise then VSWR.

David

...Yes, I do need to know the antenna impedance, so I guess that's the next stage, building an SWR meter.

Then you need to measure the complex* impedance at the working frequency, a vector network analyser will do nicely, then you can calculate the required matching network if you know the (complex*) output impedance of the source.

This will only work at one frequency, or with a bit of luck a few.

*measuring the resistance and reactance separately, and SWR meter only gives the vector sum. Complex as in leisure complex, not complicated. An SWR meter will tell you if you have got it right, not how to do it.

[Al (astral highway)]

Thank you folks! I'm out and about so will do short-form replies for now as using the quoting format takes ages to set up on iPhone!

MrBungle: thank you, that looks reasonable from AbeBooks

David: thanks for that link, will investigate what looks like a very useful tool from your description. I will probably treat myself to Experimental Methods if I can get it for less than £50, or I might send a drone up to your place to borrow one, if that suits you? ;-) Sprat 61: ? Is your analyser idea a concept or does it exist in a replicable form?

G8HQPDave: great points, thank you. Only I'm not necessarily committed to the loop; I'm now open minded about either loop or antenna.

MM: thanks for your take on this, helpful insights there. Only I think I'll skip the. Vector Analyser. Saw a good one one EBay for just £130,000!

I'm off to find an air-conditioned space. It's nudging 35 centigrade here and my brain is melting.

Only I think I'll skip the. Vector Analyser. Saw a good one one EBay for just £130,000!

I have MiniVNA that covers 1MHz upwards, PM if you want to borrow it.

[jimmc101]

How about Motorola AN267 'Matching Network designs with Computer Solutions' for a brief outline of common matching networks.

For designing networks on the computer my favourite is 'RFsim99' (freeware, available from http://www.microwavers.org/sw.htm. amongst other places. I run it on a virtual XP machine.
There's a short video giving an idea of how it works on Youtube (in Spanish) https://www.youtube.com/watch?v=zcQyIby-K6Y (Filter design but it handles matching networks similarly.)

Both date from the 90s but I still find them useful.

Jim

[Al (astral highway)]

Originally Posted by merlinmaxwell

I have MiniVNA that covers 1MHz upwards, PM if you want to borrow it.

That's brilliant, thank you! PM sent!