微波EDA网,见证研发工程师的成长!
首页 > 研发问答 > 微波和射频技术 > 天线设计和射频技术 > Is the impedance of the coax in a 4:1 lambda/2 balun important?

Is the impedance of the coax in a 4:1 lambda/2 balun important?

时间:04-06 整理:3721RD 点击:
A half-wave 4:1 balun can be made with coax. See for example

VK5AJL - Why buy baluns - make your own.
4:1 coax balun design.
http://www.creative-science.org.uk/balun.htm

is a pretty common device. Does anyone know what (if any) difference it makes if the impedance of the line is changed?

If converting 50 Ohms to 200 Ohms, is it best to use 50 Ohm coax, 200 Ohm coax (probably unobtainable). Perhaps something in between (75 Ohms is easy to get).

I must admit, I tend to think the impedance of the coax does not matter, but somehow that does not seem very logical.

Dave

It is the delay, that causes impedance transformation. Use any cable impedance as long as its length is correct relative its velocity factor, which can differ between different types of coaxial cables.

Thank you.

BTW, if you know an answer to a related, but more difficult problem I'd be interested. What happens if the length is not lambda/2? I'd like to know what the impedance looking into the low impedance side would be, if the length is an arbitrary length L, and the load is an arbitrary impedance R + j X.

I've seen designs for commercial antennas where the balun length is much shorter than I'd calculate, yet the impedance of the combination a better match to 50 Ohms than one would expect given the characteristics of the folded dipole. This leads me to think it might be possible to adjust the length of the balun somewhat to partially compensate for the changing impedance of an antenna with frequency. This might give a wider frequency range over which the SWR is reasonable. I'm guessing the ones I've seen have a balun about 80% of the length I'd calculate, even though I have taken into account the VF of the cable.


Dave

A very important aspect for the reason to use a balun is that it reduces problem with RF current that else is reflected back along coaxial cable. If the balun not is designed with each leg 180 degree apart will that function be less effective.
A commercial antenna with a shorter coaxial balun then 180 degree, does not really seem serious to me. Maybe they reached better impedance match, by adding the main coaxial cable as a part of the total antenna structure, as that is what is happening.
Such design in not at a professional level, but maybe commercial.
It is neither a tool to reduce antenna size. As best can it be a part in a resistive impedance match for for example a folded dipole but correcting reactive loss that way is a very poor method, if you not really want that the coaxial cable also should radiate. There are a few exceptions that is designed with coaxial cable as a part of the radiating element, but never after a balun.

Yes, I agree, but I'd still like to know if this does work out in theory, as the 1.5:1 SWR bandwidth of the antenna as measured on a VNA exceeds what I would expect given the characteristics of a folded dipole. I'm wondering if a short balun is a mistake, or done intentionally to improve the impedance matching at the expense of radiating from the cable.

I've personally been unable to theoretically predict the effect of a balun that is not lambda/2.

BTW, it's also clear that any losses in the balun must cause an inbalance in the currents, so radiation from the coax

Dave

The cable impedance matters for wideband VSWR. The calculation is difficult, because it needs to consider the common mode dipole input impedance. For a pure differential load, a cable impedance around 100 ohm gives best bandwidth.

It's a basic rule that I follow when I build this kind of λ/2 baluns.
The λ/2 segment must be the same type as the feeder coax cable, and for its length calculation I use the simple formula:

Length(meters) = (λ/2) x Velocity_Factor

For velocity factor this are the values of common materials used in coax cables.
Polyethylene = 0.659
Foam polyethylene = 0.88 - 0.79
Solid PTFE = 0.695

But why must the balun be the same as the feeder? Or even why it is theoretically.

I can't find any theoretical reason that the impedance of the coax must be that of the feeder. The basic idea is the half-wave generates a voltage the same as the main feeder, then the two are in series. That doubles the voltage. For a constant power, if the voltage is doubled, the current is halved, so the impedance is increased by a factor of 4.It would appear the cable impedance is immaterial - at least at a spot frequency, although I'm not sure about over a bandwidth.



Dave

A too short balun can improve impedance matching, most likely at cost of reduced antenna efficiency and bandwidth compared to doing things correct.
If the antenna structure gains in radiation efficiency and bandwidth due to radiation from the cable, why use an antenna that is less effective then the cable as radiator?
As parenthesis, some very good antenna structures can be built using the coaxial cable alone, such as stacked quarter-waves or end-feed dipoles.

Do you have anything to substantiate that claim? It's not clear to me why a shorter balun will improve impedance matching. I've got no idea if it makes the impedance matching better or worst. All I know is that I've seen it done in a commerical design. Whether that was based on the fact the designer just fiddled around with lengths, whether he mis-calculted them, or did not have a clue what he was doing, I do not know.

I'm not sure what you are defining as bandwidth here.

I don't disagree with you - doing this would not be my choice of method. But I'm interested in finding out some facts about what the effect of the shortened balun are to impedance matching, and the bandwidth over which the SWR is below a certain figure. As I say, I have seen it in commerical designs, but I've no idea on what basis such a decision was made.

I've seen stacked half-waves (coaxial colinear first published by Wheeler), but the characteristics of that are hard to predict. For example, the antenna here

Build A 9dB, 70cm Collinear Antenna

has an exagerated gain figure (9 dB). Although the author never states if it's dBi or dBd, computer simulations with both HFSS and a time-domain simulator show the gain to be be around 7 dBi.

I don't know the designs using stacked quarter waves. If you know of a decent reference on this, please let me know.

Dave

One small point; I would hesitate to use references such as this:

VK5AJL - Why buy baluns - make your own.


to base much theory or understanding of all BALUNs (or terminology) on, better to find peer-reviewed IEEE papers and pubs for items you are discussing.

References such as above are subject to too much 'pop' (culture) theory and interpretation ... for instance, 'current baluns' (chokes) seem to be in vogue, yet 30 years ago were called 'chokes' ... BALUNs balance, CHOKES choke. BALUNS can achieve near theoretical 'balance' whereas chokes only still 'choke' (present some finite series impedance with NO balance implied!) such as in the common-mode current-suppression transformers or chokes.

And also the term TLT (for transmission line transformer) seems to have been replaced with that same nebulous term in some cases.


Jim

It is not complicated, a balun, it adds a time delay which is as biggest at lambda/2. For its extreme variations in a 50 Ohm system with balun length 0 or 360 degree is it not hard guess that this is equal to no balun at all for this particular frequency. Variations for the feeding cable goes from not being a part of antenna to a pole in a dipole antenna (at zero length balun).

If antenna structure have a impedance of 50 Ohm resistive, a balun with zero length gives better impedance match?
A bit longer balun -> more then 50 Ohm impedance and a increasing phase delay.
It is possible to design antennas with impedance that fits any balun length but my customers do not want to pay for a balun with lousy or misused function due to improper antenna design. In most cases is RF current outside of coaxial cable assumed to be due to something faulty, that in worst case can destroy electronic circuits along its way traveling to ground.

The frequency range were an antenna have efficiency above a certain level. VSWR bandwidth can be perfect for a resistor also, but for an antenna is efficiency a vital function. Low VSWR is not a guarantee for high efficiency. To achieve high efficiency must also VSWR be reasonable low but that can be solved in a numerous ways, not related to the antenna it self, with a balun, stub or discrete components.

Yes,
I was not taking much notice of it. I was just Googling for a picture.

I'm well aware of a lot of stuff written in the amateur radio press / web sites needs to be taken with a large pinch of salt - I give one example of the 9 dB 70 cm collinear. A vastly exaggerated specification, but many people have followed it.

Dave

Strictly spoken, it isn't. For antennas without resistive elements, it's a good guess, though.

As the term suggests, the purpose of balun is balancing, impedance transformation is an important side effect. The effect of a non-ideal balun respectively criteria for optimal transmission line balun setup can't be discussed without a distiction between differential and common mode antenna impedance.

But that does not say anything about the effect of a balun of less than lambda/2.


Well, assuming we are talking about a system with an impedance of 50 Ohms, if the antenna has an impedance of 50 Ohms, then clearly a lossless balun which is not a non-negative integer of a half-wavelength will have change the impedance, and so make the match poorer. But that's saying nothing about how a short balun improves the match for an antenna of some arbitrary impedance R + j X.

I know precisely what the role of the balun is. I am still trying to get to the bottom of whether a mis-guided technician has designed a balun that's about 80 % of the expected length to improve the impedance match. You, nor anyone else has managed to convince me that this "trick" improves the impedance match.


OK, you have defined bandwidth. You need to do that, as otherwise the term is meaningless.
Yes, agreed.
I don't see what efficiency has to do with SWR. An antenna can have a very high SWR but still be efficient. Efficiency is defined as

(Power radiated by antenna) / (Power accepted by the antenna)

Obviously, if the SWR is high, the antenna will only accept a small fraction of the power it receives - the rest is reflected. But that does change the efficiency of the antenna.

Dave

In many cases yes, but suppose you measure with a 50 Ohm calibrated system, a folded dipole would not impress with low VSWR.
For me, working with embedded multiband antennas, I would be rather limited if trying designing antennas with low VSWR as primary goal.
No that idea have we thrown overboard in this discussion.
Technically should we maybe use another name then balun when it not not is used as a such.

---------- Post added at 13:17 ---------- Previous post was at 11:58 ----------

Coaxial length zero -> lamba/2
Impedance R -> 4R Ohm
Phase 0 -> 180 degree
Is this better understandable?
Select anything in between R and 4R and phase will also be something in between. You can plot these points in a polar impedance diagram. If it exist a reactive component in the impedance system can it be seen as rotating around its normalized system impedance with actual selected phase added.
Do not try to put word on me, I have NOT said so.

1. If an antenna have an impedance that is EQUAL to what the antenna can see in a shortened balun, then it it is a improved impedance match for that antenna at first sight, that can seem to be better solution then trying to match against a correct functional balun.

2. I have said that it improves nothing if the antenna designer knows what he is doing, Knowing what he is doing excludes use of shortened balun. Situation (1) is therefore excluded.

If you try with "but if it is something of value in it" again, I will not comment it.

No, since I knew all that.
That's an interesting point. It means that a cable with an electrical length of less than λ/2 would only make impedance matching worst to a folded dipole, not better. A λ/2 dipole has an impedance of around 70 Ohms (depends somewhat on radius of element). Folding increases that by a factor of 4 to around 280 Ohms. So to get the best match to 50 Ohms one would want an impedance transformation ratio of around 280/50 = 5.6:1. If the impedance transformation ratio achieved with a cable less than λ/2 long is under 4:1, then not only does the shorted cable cause the coax to radiate (as it not longer such a good balun), but it would also make the return loss worst, not better.

I'm suspecting the reason the return loss on a commercial antenna is better than I would have expected has nothing to do with the fact the cable used in the balun is electrically less than λ/2 long. If anything, that's just making the return loss worst.

Dave

Normally, when designing a folded dipole, besides that we want high efficiency more or less narrow-band must it also be given a fitting impedance,
Impedance can be varied by selecting material thickness, shape/distance between folded elements, micro structures, feeding point offset. Compare this antenna with a full wave loop antenna which regarding antenna element have about same length, just another shape.
When I am doing such antennas, when satisfying antenna Q have been found (narrow band peaked efficiency or wider bandwidth with a bit more even spread efficiency) do I try to minimize reactive components. Correct impedance is then often adjusted with some type of gamma-match. Mainly is this due to that it must be cheap PCB printable antennas (typical dual band 2.4/5 GHz). If it is dual band can impedance tuning be a bit tricky but it have so far been doable.
Close to the TX chip do I have a synthetic balun as RX/TX chip in this frequency range have often differential output with complex impedance and a impedance matching network consisting of discrete caps and inductors. These discrete components are also used to find optimized TIS and TRP settings if the customer requires that.

Designing baluns with RG-coax is propably more common at frequencys below 100 MHz, but I have in my toolbox a lot coax-baluns, also at 2.4 MHz. These are used when measuring in unknown balanced active systems.

O.K. I got your point. i was assuming, that the reference impedance has been choosen according to the antenna.

But I suppose it is used as a balun in the present case, because the dipole is assumed to be fed differentially. The suggestion to use a TL length different from lambda/2 implies an possibly unwanted common mode feed of the antenna, at least changing the direction characteristic. The same thing happens of course when sweeping the frequency.

Dont be too harsh on John !

He is well known world wide for his work on antennas and assoc items
you would do well to at least take notes from his findings and reports They will be helpful :)
and you will find he is well backed up by ARRL and other publications

Dave

Copyright © 2017-2020 微波EDA网 版权所有

网站地图

Top