Question about longwire antenna setup

[mio83]

Hello,

I am currently experimenting with the following setup:



This is composed of a SDR (HackRF), whose output is connected to a cheap 1-2W amplifier (by a 50ohm coax), which is then connected (with a 50ohm coax) to an antenna tuner (T-match, MFJ-902) which is directly connected (end-fed) to a "long wire" (about 7 meters). Finally, the ground of the Antenna tuner is directly connected to the mains ground (as recommended in the MFJ-902 manual).

The goal is to transmit on 14Mhz. At point (B), the nanoVNA measures 1.3 SWR at 14Mhz.

So far my tests seem to indicate that the antenna does not radiate any significant power.
It works OK in receive mode (I guess): I can hear voice communications and this is the result of running WSJTX (FT8) for 24hours:




QUESTION: is there some kind of relatively accessible "reference mathematical model" that could predict the behaviour of a system of this type in terms of efficiency? I'd like to know how much the antenna is radiating.

(Disclaimer: I know that antenna/transmission theory is a very complex topic. I am just an amateur trying to learn a bit, just to have fun. So here I am looking for some educated Back-of-the-envelope calculations [1], if possible).

The energy wasted cannot (I think... ) be in the form of reflections to the amplifier, because the SWR at point (B) is quite low.
The antenna tuner (T-match) is made of capacitors and inductors. So I am not sure where the wasted energy is going. At the same time, I know that this antenna tuner can also "tune" (i.e., impedance match) with a good SWR an open circuit (a long wire antenna of 0cm), so I guess the signal is simply being reflected back and forth inside the tuner.

QUESTION 2: what could be a way to precisely measure the emitted power (for example with an oscilloscope)? Attaching the scope to the long-wire would change the impedance of the long-wire and the whole setup, I guess.


Thanks a lot for any answer and suggestion.

PS: of course I will try with other antenna setups: longer "long wire", center fed dipole, etc. But this post is about this specific (and surely not ideal) transmission setup.

 

[1] https://en.wikipedia.org/wiki/Back-of-the-envelope_calculation

[iMo]

The key here is the "Radiation resistance" of the antenna. A short antenna has a low Rr.

https://en.wikipedia.org/wiki/Radiation_resistance

The SWR itself is not an indication how efficient is your antenna - as you may tune a paper clip for 1:1 at 20m with a good matching circuit.

PS: You may build a simple field strength meter, ie. 1m antenna + 2 Ge diodes + 10n capacitor + small 60uA meter (or a dmm). With 1W PA the reading will be rather low, however.
You may also use the websdr pages (http://websdr.org/) to monitor yourself on-line.

[mio83]

The key here is the "Radiation resistance" of the antenna. A short antenna has a low Rr.

https://en.wikipedia.org/wiki/Radiation_resistance

The SWR itself is not an indication how efficient is your antenna - as you may tune a paper clip for 1:1 at 20m with a good matching circuit.

Thanks! Yes indeed with my tuner I can tune the open circuit to about 1.2 SWR...

PS: You may build a simple field strength meter, ie. 1m antenna + 2 Ge diodes + 10n capacitor + small 60uA meter (or a dmm). With 1W PA the reading will be rather low, however.
You may also use the websdr pages (http://websdr.org/) to monitor yourself on-line.

Thanks, I'll look into field strength meters.
I have indeed using websdr and FT8 to monitor my signal, but nobody can hear me

cheers 

[iMo]

With a 7-8m long wire mounted off your balcony, and 10-20W at 20/17/15/12/10m you certainly can make entire Europe+ with FT8 (stations placed around a circle with 2000km radius with your QTH in the middle) even in poor conditions..

[A.Z.]

As I see it, your antenna is too short to be efficient on the 20 meters band, my suggestion is to go here http://www.hamuniverse.com/randomwireantennalengths.html and pick a length (from the GREEN numbers list) which fits your available space and, at the same time, is longer than 1/4 lambda at the desired band (the longer the better), then instead of feeding it through an antenna tuner, just place a 9:1 current UnUn between the antenna and the coax feedline, add a 1:1 common mode choke to the feedline and ensure to connect the coax braid to a good ground/counterpoise system, that should give you low SWR and a better radiation efficiency; also, you didn't specify how the antenna wire is laid out, I hope it's suspended as high (and far from obstacles) as possible

[radiolistener]

Yes indeed with my tuner I can tune the open circuit to about 1.2 SWR...

Good SWR don't show efficiency of your antenna. It just means that power is consumed by some load and is not returned back to the transmitter. But it don't explain which component exactly consume power. It may be antenna for radiation. It may be antenna for wire heating. Or it may be coax cable heating. Or it may be tuner heating. Each component consume some part of power. And you don't know what is proportion.

The power may be consumed by your tuner, or by your cable and other components. If you use very big or very low antenna impedance, then significant part of the power will be consumed by your tuner or cable and transferred into heat.

[mio83]

Thanks for your useful answers.

I did my test with 1W amplifier, but nobody received any of my FT8 packets. So, even if my antenna is not perfect, I suspect it is not well set up, for some reasons.

Regarding the position: it should be pretty good. It's a straight horizontal wire (about 8m, out of my windows) and I live on the top of a hill (see picture, the wire goes outside the 4 windows).



I was expecting better results given the relatively good height.

Question: I have heard many talking about UNUN for long wires, but how is this approach better (theoretically speaking) than a T-match antenna tuner? My (naive) line of reasoning is that inductors/transformers are more lossy than capacitors, so a T-match might be a better solution.

thank again to everybody. 

[iMo]

A naked T-match is a little bit better than T-match+unun, provided the T-match can tune the antenna properly (the tuning ranges of the tuners are usually limited, like 10-500ohm or similar). The ununs (like 1:4 or 1:9) introduce some losses (it depends on the design), but lowers the high antenna impedance to the region the tuner can match.
Btw, your MFJ-902 has a built-in unun/balun (1:4) you may use as well, afaik..

PS: also double-check there is a signal coming into your tuner. Load the end of the coax with a dummy load (ie 50ohm/1W resistor, or two 100ohm resistors in parallel)) and look at the signal at the resistor with oscilloscope. You should see a sine wave when keying a none modulated carrier in. The voltage should be around 10Vpeak (20Vpeak-peak) with 1W, afaik.

Also mind you have to have a low-pass filter (for 14MHz in your case) wired at the output of your power amplifier, otherwise you could be out of the spec for harmonics levels (the T-match does not provide low-pass filtering, it is a high pass)..

And finally you have to be a licensed HAM..

[A.Z.]

an atu (be it a pi, l,t ... match) will have losses, a good antenna with a proper impedance transformer (and proper grounding/choking) will have lower losses

[ZigmundRat]

Do you have someone local that can listen for you and successfully decode your packets? You might also try something simpler than FT8 with its specific synchronization requirements. If you can operate CW, just calling CQ for awhile, you can check the reverse beacon network (RBN) to see if you’ve been heard. Might be worth a try. You might also try WSPR. We know HackRF with no amp has been used in this mode successfully.

I should think a 7M wire with tuner would radiate enough, it’s not that different to what many QRPers use for SOTA or when hiking. Maybe not optimal, but a good enough.

With regard to your amp: In your diagram your amp is unspecified and appears untuned. This would mean that it’s power is being used for the desired fundamental (14MHz) as well as harmonics generated by the HackRF. Between that and amp generated products, your power at 14MHz is reduced. In addition, the HackRF generates a max of 15dBm@10MHz. To get to 1W would require about 18dB gain, which is reasonable, but again we don’t know anything about the amplifier (1W input or output? Class of operation? Gain? etc.) Pretty sure you’re not getting anywhere close to 1W into the tuner, much less the antenna.

[mio83]

Thank you everybody.

I am particularly interested in understanding this properly:

an atu (be it a pi, l,t ... match) will have losses, a good antenna with a proper impedance transformer (and proper grounding/choking) will have lower losses

what is the reason? Can you point to some general argument/calculations that would help me reach the same conclusion regarding T-match VS transformer?  I haven't been able to find much.

thanks again!

[radiolistener]

what is the reason? Can you point to some general argument/calculations that would help me reach the same conclusion regarding T-match VS transformer?  I haven't been able to find much.

The reason is that impedance transformation leads to losses in the impedance match circuit and in the cable between impedance match circuit and antenna. Higher impedance transformation ratio leads to higher loss.

So, it's better to use antenna with impedance close to the required. It minimizes losses in impedance match circuit and simplifying impedance match.

This is why any unbalanced antenna (such as long wire) with high or low impedance has very low efficiency. Just because a lot of power is wasted on impedance match and feeder line.

But unbalanced antenna (such as long wire) also has another bad issue. It's hard to prevent using your feeder line coax as a part of antenna, so it leads to high interference to your home electronics for TX and high noise from home electronics at RX. You will need a lot of very good RF choke on your coax cable in order to fight with these unwanted effects.

Technically you will put a lot of Watts into heating of your home walls and make interference for your home electronics. This is why any unbalanced antenna (such as long wire) requires good RF choke.

It's more easy to build proper antenna than fighting with these unwanted effects of common mode currents and losses on impedance ratio transformation.

[Circlotron]

Good SWR don't show efficiency of your antenna. It just means that power is consumed by some load and is not returned back to the transmitter.

Is SWR more or less the same thing as power factor in an AC circuit that has some amount of capacitance or inductance?

[radiolistener]

Is SWR more or less the same thing as power factor in an AC circuit that has some amount of capacitance or inductance?

They are essentially the same, but have different meaning.

VSWR shows relation between forward wave (from transmitter to the load) and reflected wave (reflected back from the load due to impedance mismatch). This reflection depends on the reflection coefficient Γ (gamma) at the border between two transmission lines with different impedances:

Γ = (Z - Z0) / (Z + Z0)

where
Z0 is the source impedance (transmission line which provide power)
Z is the load impedance (followed transmission line which consume power)

VSWR = (1 + |Г|) / (1 - |Г|)


Power factor PF is a measure of instantaneous power used by a device versus the average power. For the mains, the resistance of the load is always higher than the source impedance. But frequency is too low, so reflected wave is not an issue here.

For pure resistive load:

P = U * I

But for reactive load there is a phase difference between voltage and current and it leads to:

P = U * I * cos( θ )

where
θ is a phase angle between voltage and current due to reactive load

PF = cos( θ )

For pure resistive load PF = 1. For reactive load PF is less than 1, so it leads to higher currents in the transmission line and it leads to higher loss on transmission line resistance.

When VSWR > 1, it means that there is reflected wave, so the part of energy is not consumed by the load and reflected back to the source. When you use impedance match circuit (antenna tuner for example) you catch reflected wave from the load and put it again into the load, so the part of reflected wave will be consumed by load again and again and a part of it will be delivered to the load on each cycle.

In such way there are a lot of back and forth energy transfer through transmission line between impedance match circuit (antenna tuner) and antenna. On each forth and back energy transfer there are transmission losses in the coax cable and in your antenna tuner. So, high VSWR between feeding cable and antenna leads to high energy loss in the cable between antenna and antenna tuner.

But when you measure VSWR between transmitter and antenna tuner you will not see that your antenna has bad VSWR. Just because antenna tuner catch reflected wave and put it back into the antenna.

For example VSWR between your feeding coax cable and antenna is 2. But VSWR between transmitter and antenna tuner is 1. In such case there is a high loss in the coax line between antenna tuner and antenna. But if you measure VSWR before antenna tuner you will not see this problem and don't know that you have problem.

In order to catch this issue, you're needs to measure VSWR at the feeding point of antenna.

And this is why it's much better to place impedance match or antenna tuner as close to antenna as possible, because energy loss in the cable between antenna tuner and antenna will be extremely high. So, when you place antenna tuner just at the antenna, you will eliminate power loss in the cable.

But even if you place antenna tuner at the antenna side, you're still have power loss on each reflection, but now it depends on heat loss in your antenna tuner and antenna wires. This is why it's better to have antenna with input impedance as close to required (usually 50 Ohm) as possible.

When your antenna is good balanced and has input impedance the same as your feeding cable (usually 50 Ohms), there is no energy loss in the antenna tuner. In such case total efficiency will depends directly on your antenna efficiency.

This is why it's better to strongly avoid unbalanced antennas (such as long wire) and avoid antennas which impedance is very different from your feeding coax cable.

[iMo]

FYI - there is a great tool called SimSmith.
You may measure your antenna with nanovna (for example), import the s1p file (as the "load Z" on the left hand side) and do a simulation of entire signal chain.
Below is an example:
1. imported .s1p file from nano-vna (8m wire + 1:4unun + choke) - the load Z
2. added coax and T-match
3. G is your transceiver.
You may interactively "tune up with the T-match" such you get the best SWR. You also get some information on losses through the signal chain. There are models for coax cables, etc.

[Wallace Gasiewicz]

The relative low SWR means that you signal is going into the antenna tuner and not being reflected back. Does not say anything about how much is getting into the antenna.
I think we can look at the tuner and antenna (not a long wire for any practical purpose) as a resonant circuit. It however does not have very good radiation. If you want better radiation you need a better antenna. I have built a small loop and I think from you pictures that it would work for you. For your low power operation, it is easy and cheap to build one.
One of the important things in a small antenna is the radiating surface, this is quite a factor in loop antennas and I think a very big factor in the radiation of your signal. I think most of your signal is lost in the tuner.

Wally

[mio83]

FYI - there is a great tool called SimSmith.
You may measure your antenna with nanovna (for example), import the s1p file (as the "load Z" on the left hand side) and do a simulation of entire signal chain.
Below is an example:
1. imported .s1p file from nano-vna (8m wire + 1:4unun + choke) - the load Z
2. added coax and T-match
3. G is your transceiver.
You may interactively "tune up with the T-match" such you get the best SWR. You also get some information on losses through the signal chain. There are models for coax cables, etc.
(Attachment Link)

wow this looks like an amazing software! I downloaded it and will try to learn how to use it. thanks!

[iMo]

FYI - there is a great tool called SimSmith.
You may measure your antenna with nanovna (for example), import the s1p file (as the "load Z" on the left hand side) and do a simulation of entire signal chain.
Below is an example:
1. imported .s1p file from nano-vna (8m wire + 1:4unun + choke) - the load Z
2. added coax and T-match
3. G is your transceiver.
You may interactively "tune up with the T-match" such you get the best SWR. You also get some information on losses through the signal chain. There are models for coax cables, etc.
(Attachment Link)

wow this looks like an amazing software! I downloaded it and will try to learn how to use it. thanks!

FYI - there are many dozens of videos on how to use the SimSmith and vna in ham practice made by W0QE:

[tkamiya]

Long wire (aka random wire) is not a great choice for QRP.  (low power operations)  I've experimented with this in USA.  With 10 watts, no QSO.  100 watts, great fun!

If you can tune into nothing and get a decent SWR, something is very lossy.  You might want to look into L match which has wider impedance transition and more suitable for long wire type antenna.  Ground becomes prime importance as well.  I'd try adding "counter poise" to improve the situation.  It is extremely difficult to get a good RF ground connecting to the actual earth.

[iMo]

Long wire (aka random wire) is not a great choice for QRP.  (low power operations)  I've experimented with this in USA.  With 10 watts, no QSO.  100 watts, great fun!

If you can tune into nothing and get a decent SWR, something is very lossy.  You might want to look into L match which has wider impedance transition and more suitable for long wire type antenna.  Ground becomes prime importance as well.  I'd try adding "counter poise" to improve the situation.  It is extremely difficult to get a good RF ground connecting to the actual earth.

There are basically two kinds of "long wires" or "end fed" antennas:

1. End fed random wire
2. End fed half wave (EFHW).

The EFHW antennas work pretty good in QRP area.

PS: you may try with EFHW one.
In your 14MHz case you need a wire of aprox. 10m (you have to cut it into the resonance, it could be of any reasonable shape which fits, ie. an "L") and a 1:49 impedance ratio (1:56, 1:64) unun transformer (a special winding on a 43 material ferrite toroid, for the 1W even the ~2cm diameter is ok).
When the wire is cut into the resonance your SWR will be close to 1:1 and you neither need the T-tuner, nor a counterpoise.

[A.Z.]

Thank you everybody.

I am particularly interested in understanding this properly:

an atu (be it a pi, l,t ... match) will have losses, a good antenna with a proper impedance transformer (and proper grounding/choking) will have lower losses

what is the reason? Can you point to some general argument/calculations that would help me reach the same conclusion regarding T-match VS transformer?  I haven't been able to find much.

In your case the antenna (wire) is directly connected to the tuner, so either a transformer or a matching unit will make little difference from the loss/efficiency standpoint; the only problem (let me call it so) is that with that config you will have "RF in the shack" since part of the ratiating element (the antenna) is inside; that's why it's usually better to either use a remote antenna matching unit sitting at the (external) antenna feedpoint or a transformer (again, at the feedpoint) and feeding the antenna using some coax (with proper choking/grounding).

As for simulating the antenna, you may have a look at the free 4NEC2 program which you can download here it will allow you to create a model of your antenna and then perform a number of simulations; for a primer on 4NEC2, see this and this (you can find more guides/manuals/primers on the net)

in your case, you have two main issues, first of all your 7 meters antenna is too short for the 20 meters band (14MHz), second, you lack a counterpoise, the latter may even be a piece of wire hanging down from the window and connected to the gnd of your matching unit (or transformer); as I already wrote, my suggestion is to start by going here and selecting from the GREEN numbers list the longest length of wire which will fit your available space, next build a 9:1 UnUn like this one (the toroid may be smaller given that you'll use low power); done so, pick up some length of coax to connect it between the UnUn and the wire and connect a second run of wire hanging down from the UnUn "gnd" connector; use a piece of rope to hold the UnUn and the wires as far from the building as possible and a second one to hold it at the far end; that will give you a pretty decent antenna, sure it won't be a Yagi or whatever, but it will at least be better than the current one

Willing to go for a resonant antenna, you may consider putting up a resonant dipole, in such a case the calculation is pretty simple/straightforward

Lambda = 300000 / KHz

HalfWave = Lambda / 2

WireLength = HalfWave * 0.954

ArmLength = WireLength / 2

with the above infos in your hands, cut two wires at ArmLength and connect their center point to a 1:1 BalUn (yes, you want a BalUn there since it's a balanced antenna) like this and connect the balun to the coax, hang the center as far as possible from the building (using a piece of rope) and hang the two arm ends (through ropes) to a couple supports outside, the dipole will then have an "inverted V" shape but will work w/o problems; to bring it to resonance you may just fold back (a bit at a time) the wire ends over themselves until you'll find the sweet spot

Ok, enough I think, hope it will be of some help

[edit]

The 0.954 factor I used is the velocity factor for "phosphor/bronze" wire, I prefer using that since, even if it will give you a (slightly) longer value, it's easier to shorten the wire than to lengthen it

[edit 2]

As for the "random wire" lengths, the table below shows them in meters, the first colum shows the "lowest" band which a given length will cover, the second one the length of the wire in meters, knowing the band of interest, pick the longest length which will cover the desired band (or a lower one) and fits your available space; notice that while laying out the antenna in a straight line would be the best setup, it's not mandatory, the antenna wire may be sloping or zig-zagging (horizontally and/or vertically) as long as the angles described are above 100 degrees, sure that will alter the radiation pattern, but it will allow to fit the wire into your available space

-----------  -------------
lowest band  wire (meters)
-----------  -------------
    20            8.8392
    40           10.8204
    40           12.4968
    40           17.6784
    80           21.6408
    80           25.6032
    80           32.6136
    80           36.2712
   160           45.1104
   160           61.8744
   160          105.7660
   160          124.0540
   160          128.9300
-----------  -------------

[cdev]

A picture of the antenna in its context would be helpful. Maybe some obvious issue exists thats not in your description. Have you tried a quarter wave wire? (or a quarter wave "fan monopole" multiple quarter wave wires parallel with each other.)

That would likely be easier to match.

If you could take a photo or photos that shows where the antenna is and what its relationship to ground and counterpoise. If its hanging down out from a window it may benefit from being stood off from your building a meter or two if you can.  Because buildings contain steel and steel is conductive and grounded.

[cdev]

If you do this for best performance try to keep the far (high impedance) area of the wire away from other metal and ideally as far as you can away from other objects like a building. The wire could be thin and almost invisible.

You should consider getting a NanoVNA or similar. It would be very helpful to you in getting this working.

Long wire (aka random wire) is not a great choice for QRP.  (low power operations)  I've experimented with this in USA.  With 10 watts, no QSO.  100 watts, great fun!

If you can tune into nothing and get a decent SWR, something is very lossy.  You might want to look into L match which has wider impedance transition and more suitable for long wire type antenna.  Ground becomes prime importance as well.  I'd try adding "counter poise" to improve the situation.  It is extremely difficult to get a good RF ground connecting to the actual earth.

There are basically two kinds of "long wires" or "end fed" antennas:

1. End fed random wire
2. End fed half wave (EFHW).

The EFHW antennas work pretty good in QRP area.

PS: you may try with EFHW one.
In your 14MHz case you need a wire of aprox. 10m (you have to cut it into the resonance, it could be of any reasonable shape which fits, ie. an "L") and a 1:49 impedance ratio (1:56, 1:64) unun transformer (a special winding on a 43 material ferrite toroid, for the 1W even the ~2cm diameter is ok).
When the wire is cut into the resonance your SWR will be close to 1:1 and you neither need the T-tuner, nor a counterpoise.

[A.Z.]

If you do this for best performance try to keep the far (high impedance) area of the wire away from other metal and ideally as far as you can away from other objects like a building. The wire could be thin and almost invisible.

Yes, it can be thin, but at some point a too thin wire will have an impact on radiation efficiency and then, depending from the location, a too thin wire may break due to wind or ice load, and while a pulley and weight may help, I believe that a (say) 2 mm wire would be a good start, just ensure to keep both ends of the wire as far from obstacles and other lines as possible and fed it using coax and an appropriate UnUn, alternatively one may consider putting up a rybakov antenna or a small transmitting loop

You should consider getting a NanoVNA or similar. It would be very helpful to you in getting this working.

A VNA would be nice, but a directional wattmeter would be cheaper 

[A.Z.]

There are basically two kinds of "long wires" or "end fed" antennas:

1. End fed random wire
2. End fed half wave (EFHW).

Three 

1. End fed random wire

2. End fed half wave

3. End fed longwire

the third is, by definition some multiples of a wavelenght, while the first should be longer than 1/4 wavelenght at the longest desired wavelenght and never a multiple of 1/2 wavelenght for any of the bands you want to use the antenna on; #1 and #3 are often "interchangeably" called "random wire" but they are different beasts