Receiver aerial input impedance: questions.

[Skywave]

These Qs. are simply to tie-up a few loose ends in my understanding as per this Thread's name: not applicable to any specific comms. radio. I have consulted many standard reference books in my technical library and the 'Net, but have failed to find a clear, concise answer. So I thought I'd ask here . . . .

The aerial input impedance is frequently found to be stated as "75 ohms". This can be for a balanced or unbalanced aerial connection. However, should this not really be stated as aerial input resistance is 75 ohms ? And presumably a nominal figure at that and also, presumably, a figure that will deviate from that figure as the receiver is tuned over a particular bandwidth? And will that figure remain resistive, or will it swing from an impedance that is of the form R + jX to R - jX over that tuning range, the average value thus being approx. 75 ohms resistive? Or . . . . ?

All replies appreciated; thank you.

Al.

[Bazz4CQJ]

As an RF signal is an alternating current, and the input to any tuning circuits will have elements of inductive reactance, with some capacitive reactance and some resistance, then it is appropriate to designate it as an an impedance.

I believe that 75ohms became a common standard mainly because it was used for TV's. Most professional equipment is now 50ohm, but especially older communication receivers were often 600 ohm.

It is really the aerial and the feeder which determines what the impedance of the system is and that is why an aerial tuning unit is often used to match the receiver to the aerial.

Barrie

[Skywave]

O.K. I can understand and appreciate that. So do we have a situation then whereby for the aerial input impedance, Z = R ± jX, the quadrature component of Z changes sign from one end of the tuning range to the other, and that the quantities R and X also vary in that range such that |Z| remains approximately constant at some figure - like 75 ohms?

And for the reason for 75 ohms becoming a standard figure for aerial input impedance, surely that was caused by the (approx.) input resistance of a half-wave dipole?

Al.

[Bazz4CQJ]

IIRC, the impedance of long wire antennas for HF work varies with kinds of factors including height above ground but I don't recall 75ohm being a specific value for dipoles.

The one receiver on which I can comment in some detail is the HRO on which the aerial input is the primary winding (inductively) coupled to the the the tuned circuit input of the 1st RF AMP. That primary can be connected either as a balanced or unbalanced input but within the limited tuning range of one coil pack, there is no "front-panel" means of adjustment when in use. I guess that might imply that having a broad un-tuned aerial input is more that compensated for by having 2 tuned RF amps, but of course, that was just what the guys at National came up with and no doubt other designers went with alternative approaches.

[G8HQP Dave]

The feed resistance of a half wave dipole in free space is 73 ohms. 75 ohms was chosen as one of the standard coax impedances because it gives the lowest attenuation for air dielectric. See Wikipedia.

When people talk about the 'input impedance' of a receiver, they may mean one of two quite different things:
1. the input impedance i.e. what you would measure if you attached an impedance measuring device to the input.
2. the source impedance which the receiver was designed to work with i.e. the impedance which gives the stated gain/lowest noise etc.
These two impedances are rarely equal, because receiver circuitry does not give the best noise figure when there is an impedance match at the input. For MF and HF receivers the actual input impedance is typically quite a bit higher than the optimum source impedance, and it varies across the band. By now I bet you wish you hadn't asked!

There is no clash between calling it a resistance or an impedance. An impedance with a zero reactive part is just a resistance.

So 'a receiver with 75 ohm input impedance' probably means a receiver designed to give good performance when fed from a 75 ohm source (e.g. antenna plus cable). It almost certainly does not mean a receiver which presents an input impedance of 75 ohms. The actual input impedance is likely to be higher than 75 ohms and somewhat reactive.

[Robert Darwent]

You are correct in your assumption Al regarding 75 ohms (actually 73 ohms) being the theoretical feed resistance for a half-wave dipole. For a folded dipole it's 300 ohms.

(NB: Posted at the same time as Dave above)

[Skywave]

Thank you Dave, (G8HQP) for your comprehensive reply. There are a few points in that reply which I would like to discuss further . . .

You said:
The feed resistance of a half-wave dipole in free space is 73 ohms. 75 ohms was chosen as one of the standard coax impedances because it gives the lowest attenuation for air dielectric. See Wikipedia.

My response: A rather convenient co-incidence then!

You said:
When people talk about the 'input impedance' of a receiver, they may mean one of two quite different things . . . . (etc.)
By now, I bet you wish you hadn't asked!

My response: Not at all! I understand all of that.

You said:
There is no clash between calling it a resistance or an impedance. An impedance with a zero reactive part is just a resistance.

My response: Of course; agreed.

You said:
So 'a receiver with 75 ohm input impedance' probably means a receiver designed to give good performance when fed from a 75 ohm source (e.g. antenna plus cable). It almost certainly does not mean a receiver which presents an input impedance of 75 ohms.

My response: Quite so: it is exactly that 'probably' that caused me to produce this very Thread! However, operator's manuals for many comms. receivers, e.g. typically Eddystones', clearly state: "the aerial input impedance is 75 ohms". Moreover, if the receiver does not present an input impedance that matches the source impedance - which in this case is purely resistive - does that then not suggest that the designer failed in his attempt to make an 'impedance' match that will produce the maximum power transfer from source to load?

My further thoughts . . .
For classic valve comms. receivers (which is what I have in mind here) typically these have an input transformer with a tuned secondary coupled to the grid of the R.F. amplifier. The primary is usually untuned and is connected to the aerial input terminals. The impedance at the primary will produce a transformed impedance to the secondary. Since the secondary is tuned to resonance as the receiver is tuned, it will reflect the dynamic resistance (L/CR) of that tuned secondary into the primary: hence the primary presents a resistance plus some small reactance of the primary to the aerial. Presumably, this secondary reactance will be small at centre frequency of the band coverage, but will increase (mainly on account of the changes at the secondary by variation of the tuning capacitor; there are other factors) as the set is tuned over the band selected. Hence, at approx. mid-band, an impedance of R ± jX is presented to the aerial, where X = 0 at mid-band, and for the cases under consideration, R = 75 ohms.

Al.

[G8HQP Dave]

It is unlikely that an Eddystone receiver actually has an input impedance of 75 ohms, although I suppose that is possible. An HF receiver has to balance two things: voltage stepup to overcome valve noise, and resistive loading of the first tuned circuit to attenuate thermal noise from the dynamic impedance of that tuned circuit. These two things pull in opposite directions. Perhaps surprisingly, it is thermal noise which usually dominates so a lowish stepup is used. This means that the antenna socket sees a relatively high impedance. See my website for the details.

Tuning the input coil means that roughly the same impedance can be presented across the band, set by the Q. It won't be exact because tracking is never exact, and Q can vary with frequency. It does not follow that X=0 at band centre. If correctly tuned X=0 everywhere, but this is an ideal not fully achieved.

At higher frequencies where valve grid noise becomes significant it may be necessary to mistune the input coil as best signal-noise ratio corresponds to the valve seeing a slightly capacitive input source. You can sometimes see this either if an HF receiver has a separate antenna trimmer or if you are adjusting a VHF set: maximum signal does not quite coincide with best signal-noise although it won't be far off.

[Craig Sawyers]

I run one of these http://www.wellbrook.uk.com/pdf/ALA1530.pdf into my RA17. Nominally this is an impedance mismatch since the Welbrook is 50 ohms and the RA17 is 75 ohms. But a least the Welbrook provides a real 50 ohms from its head amplifier, and not something that varies with frequency. I run 50 ohm coax from the Welbrook.

If I was moved to do so, I could change the resistor values in the RA17 input attenuator, and R15A to 50 ohms. But the VSWR loss is not too bad with a 75/50 mismatch.

For a long wire, you need a decent balun to match the varying impedance to 75 ohms, or use an antenna tuner.

Craig

[Skywave]

Originally Posted by G8HQP Dave

It is unlikely that an Eddystone receiver actually has an input impedance of 75 ohms, although I suppose that is possible.

Thanks for your reply: all points noted and I will visit your Web Site.
Plus that reply, in turn, does raise the question of what design procedure was taken to determine that stated figure of input impedance in the first place. I presume that 'educated guesswork' was not used!

So let me tackle this question from a different angle - albeit a conceptual one, for the purposes of this discussion.

Suppose I have a receiver whose nominal aerial input impedance is stated as 75 ohms. Suppose that I also have a signal generator of a known output impedance, (with a pad inserted in its output, if necessary, to convert that output impedance to 75 ohms). I connect a length of 75-ohm co-ax cable from that pad to a test box comprising a variable resistor and a variable reactance (i.e. a coil or capacitor). I set the generator and receiver to the frequency of measurement, then measure the level of the signal output from the receiver. Then I adjust the values of those two components in that test box for (a) maximum output and then (b) for maximum SNR. The corresponding values of the R and the X in the test box will tell me what the input R and X of the receiver is for that frequency and each test condition, yes? And, if so, am I going to find that the input impedance, |Z| = sq. rt.(R² + X²) = approx. 75 ohms for each test condition?

Now don't take that method too literally please - it neglects various sources of error and the practicalities of performing an accurate measurement: as I said, I am only talking conceptually at the moment.

Al.

[G8HQP Dave]

Maximum output will correspond to a conjugate match at the input, so your R will be the receiver input impedance and -X will be input reactance. If the receiver has a genuine 75 ohm input impedance then R=75 and X=0. What is more likely is that R=75 will corrrespond to maximum SNR, and X will be small.

[kalee20]

If anything, I would have thought that a receiver specified as 75Ω input impedance would present a resistance of 75Ω (pure resistive) at the aerial terminal. Or at least, this would be the design aim - in practice, some variation from 75Ω, and some variation of phase angle not being 0° would be expected.

The reason? When fed from 75Ω coax, if the receiver presents anything other than 75Ω, there will be a reflection generated, which will travel back to the aerial. At the aerial, there will almost certainly be another reflection back towards the receiver, because a half wave dipole is only 75Ω (or 73Ω) resistive at the one frequency. So standing waves would be set up of varying degrees across most of the receiver's tuning band.

At other frequencies, the dipole not being of 75Ω source impedance wouldn't matter much even if connected to 75Ω coax - unless it got hit by a reflection from the improperly terminated receiver-end of the cable.

This is quite independent of noise performance. But, I'd regard myself as of very limited expertise in designing RF coupling circuitry, so please, somebody, comment!

[MichaelR]

Originally Posted by G8HQP Dave

So 'a receiver with 75 ohm input impedance' probably means a receiver designed to give good performance when fed from a 75 ohm source (e.g. antenna plus cable). It almost certainly does not mean a receiver which presents an input impedance of 75 ohms. The actual input impedance is likely to be higher than 75 ohms and somewhat reactive.

Hi Dave,
Many comms receivers and definitely some Eddystone receivers such as the EA12 have bandpass filtering (front panel tuneable preselection circuits) preceded by highpass or broad bandpass filters. that in itself surely would reduce noise and allow broad 75 ohm input impedance.

Mike

[Herald1360]

Input filters will only reduce out of (filter) band noise. They won't improve the noise performance of the receiver as such. Unless of course their own output impedance provides exactly what the receiver input needs for this from a perfectly resistive antenna impedance.

[Skywave]

Dave, G8HQP - that is exactly my understanding. (And I did visit your Web Site: interesting reading there).
If I had the necessary lab. equipment, I suppose a better method would be to use a spectrum analyzer (with its tracking oscillator) and a directional coupler. That way I could determine the return loss at the aerial input (with the receiver powered up, of course!) and thus calculate the input impedance. Nevertheless, that conceptual idea of mine has provided me with food for thought . . . .

Kalee20: what you have said is my understanding too; thanks for your contribution.

And thank you to everyone else who has contributed to this Thread: I am somewhat more enlightened now - but please don't let that stop anyone from adding to this fascinating topic. For example, Kalee20 did say "please, somebody, comment!"

Al.

[MichaelR]

Originally Posted by Herald1360

Input filters will only reduce out of (filter) band noise. They won't improve the noise performance of the receiver as such. Unless of course their own output impedance provides exactly what the receiver input needs for this from a perfectly resistive antenna impedance.

Chris,
My comment re noise was meant to referrr to received noise not generated noise of the receiver itself altough I do not see at the moment how you can seperate received noise from generated noise.

Mike

[G8HQP Dave]

Any filter will increase noise, because all filters attenuate the signal and add thermal noise. Their benefit comes from reducing interference, so there is a balance to be struck. At HF the balance is/was shifted towards reducing interference, so highish Q (in valve receivers) and accept some noise. At VHF and above low noise is usually more important so tuned circuits are more heavily loaded to get less attenuation at the cost of broader fractional bandwidth bandwidth.

Mismatch causes signal reflections, but if the antenna is a reasonable match to the feeder then most of the reflected signal will be re-radiated by the antenna. Very little will end up bouncing up and down the coax. This a loss, but we already decided that we were designing for best SNR rather than highest signal. Note that unless the feeder is long when compared with the signal modulation rate, what we see is not bouncing pulses but a net impedance which should keep the receiver happy.

Maybe I can clarify something. When I speak of receivers I automatically think of valve receivers, as that is my interest. Modern solid-state receivers are different, because the active devices are inherently quieter (no hot cathode!) and have a lower input impedance. That means that an impedance matched input can still have lowish noise. Assume you have a perfect noiseless amplifying device, but you still need to put a filter in front of it. Then provided that the filter has low loss you can have good matching and lowish noise. Valves are different, because at lower frequencies they have very high input impedance to the grid. The incoming signal doesn't 'see' the valve grid, just the dynamic impedance of the input tuned circuit. Sorry if I have caused confusion.

I think it can be shown that a perfectly matched receiver, operating at room temperature, can't have a noise figure better than 3dB. Fine (too good?) for HF, marginal for VHF, and too low for small signal work at UHF and beyond.

[ppppenguin]

Originally Posted by G8HQP Dave

Any filter will increase noise, because all filters attenuate the signal and add thermal noise.

Not sure that's quite correct. Only resistive elements have thermal noise. L and C do not. So except for any losses in the filter due to finite resistance an LC filter will not add noise. Since a good filter will have low resistive losses any noise should be negligible.

[jimmc101]

The 3dB minimum noise figure is true if the match is achieved by normal means. It arises because the thermal noise of the input resistance is equal to that of the source (must be equal resistance and assuming equal temperature.)

However there are ways of achieving a match without generating additional noise and without cooling the receiver input circuit.

For instance this can done by using feedback via a lossless element, either a transformer or a reactance together with a 90degree phase shift in the amplifier. (see http://www.ece.vt.edu/swe/lwa/memo/lwa0071.pdf for an example.)


Input filters using perfect components have no loss, however filters using practical components have losses.
In general, using similar components, the narrower the filter the higher the loss.

Jim

[G8HQP Dave]

That lossless feedback link does not work. Could you check it?

Filters add noise due to losses in non-ideal L and C (mainly L). A good filter won't add much, but my point was that no filter can reduce noise (apart from an image filter, which we usually take for granted anyway).