Sensitivity measurements on 200 ohm comms receiver

[Alan Bain]

I now have a working AR88D on my bench and have run through the alignment procedure for IF and RF and I would like to check how it matches the sensitivity specs in case there are any latent faults. I'm used to testing receivers with a 50 ohm antenna input.

In the alignment instructions mention is made of a "dummy aerial" of 200pF on lowest range (1) and 200 ohm on (2-5) in series between the generator and receiver. Elsewhere it states that the receiver is designed to match to a 200ohm line on bands 2-5. This would seem to suggest that the output impedance of the sig gen is expected to be 0 rather than the 50 ohms I would have expected. I feel I must be missing something obvious here...

The most convenient generator here is an HP8662A and the output is specified as PD at output terminals into a 50 ohm load (checked with the manual and an HP power meter!) i.e. half of the internal source EMF. So if connected in series with a 150ohm resistor and I assume the receiver provides a 200ohm matched load, then voltage at receiver terminals will be the sig gen output voltage. Is this a reasonable approach, or should I be measuring at the antenna terminals with an SA?

Alan

[G6Tanuki]

The way I have done similar things in the past is to use a 10dB attenuator pad between the source and the receiver. This will present a predictable source impedance to the receiver pretty much whatever impedance is at the input of the pad.

[trh01uk]

The AR-88D comes from the days well before the standard 50 ohm RF impedance became widespread. Most WWII era equipment has all sorts of weird and wonderful requirements for dummy aerials - just check a few military manpacks and see.

So you are right that you need to put a 150 ohm (non-inductive) in series with your nice modern 50 ohm source generator. Without it, any results will be optimistic, because the correct voltage divider for the generator's EMF will not be present. If you simply connect your 50 ohm generator, I make the error around 4dB. Not huge - unless you receiver is marginal.

By the way, your 150 ohm resistor should be non-reactive, so probably a carbon composition type, to avoid frequency dependent effects.


Richard

[Radio Wrangler]

EMF is helpful insomuch as EMF is the voltage you get when you don't load the output and consequently it stays the same no matter what impedance you choose to not load the output with! If you load the output with a matched impedance, say 50 Ohms, then you get EMF/2 volts across the load.

Load with a mismatch like 50 Ohms in the sig gen, a 150 Ohm resistor and then a 200 Ohm AR88 and then you hve to do a simple sum to find the signal voltage at the AR88. The AR88 will be as per its specified operating condition, thinking it's fed from 200 OHms.

The sig gen will be less happy as it sees 200 + 150 = 350 Ohms load.

The other thing not happy will be any cable needed to get some distance between sig gen and AR88. The idea of an AR88 and an 8662 screwed directly together will flabber one's gast. Both are rather hefty.

The way out is to use nice standard 50 Ohm cable from the sig gen and then either a minimum loss 50 -> 200 Ohm pad, or a 2:1 turns ratio broadband transformer.

If you do the pad, you need to do the sums to get the level scaling factor.
If you do the transformer, then the sig gen power in dBm is the power reaching the receiver, less any known cable and transformer losses.

Now to let some reality in:

Receiver input impedances are usually a long way from quoted ideal values. In fact the best gain and least noise source impedences are dead different.

Receivers are often far from ideal in terms of Zin, and on HF, it hardly matters.

David

[Alan Bain]

The AR88D is the only thing I have ever found to test with the 8662A which weighs more than the signal generator The bench won't float away.

Some very helpful comments here - the 10db pad is a very simple and effective solution which probably takes less time than finding a suitable wideband transformer.

Reality remarks are also helpful -- I guess I'm too used to higher frequencies where input impedances (or S11) are tightly specified.

[ms660]

Back in the day the impedance of signal generators was generally very low, 10 Ohms or thereabout, that's why the standard dummy antenna was to become a dummy antenna fed from a 10 Ohm source.

Lawrence.

[Jez1234]

Nice signal generator!

[Radio Wrangler]

I'm rather fond of the 8662 and had a pair at work set up with amplifiers, pads and combiners for intermod testing. I'd have liked 8663s with the low phase noise options, but needs must

David

[Cruisin Marine]

The sig gen will be less happy as it sees 200 + 150 = 350 Ohms load.

But surely the sig gen should be loaded with a 50 ohm across the RF o/p, then a 150 ohm in series to the 200 ohm Z receiver.

So, wouldn't it be approx. 50 ohm in parallel with 350 ohms as far as the genny is concerned?

Of course the genny will be 6dB down at the o/p terminals, and there will be more loss as well, but that can be measured or calculated.

That is the way I would approach it anyway.

[Radio Wrangler]

Originally Posted by Cruisin Marine

But surely the sig gen should be loaded with a 50 ohm across the RF o/p, then a 150 ohm in series to the 200 ohm Z receiver.
.

You need to offset the 50 Ohm resistor to accommodate the loading of the rest of the circuit and the receiver to get the load on the sig gen/cable right, and what you wind up with is a resistive L-network called a minimum-loss matching pad. For any impedance transformation, there is a minimum loss needed to make it work. What's more, they are flat and broadband and go all the way down to DC. They are commonly used when you have a 50 Ohm network analyser and want to look at some 75 Ohm networks.

david