I asked this on the R-390 reflector and didn't get much of a response. It sounds like a dumb question to me but I'm really not certain about an answer.

When doing an AM sensitivity measurement, I feed unmodulated carrier to the receiver (AGC is off) to set the baseline "noise" and then turn on the 30% modulation. I fiddle with the generator level until the difference between the modulated and unmodulated is 10db. Fine.

So what am I listening to when I apply unmodulated carrier? Why does the output noise level rise with an unmodulated carrier? Where is this extra "noise" coming from? I can see some rectification of the carrier moving the bias around on tubes but why would that increase the noise level? My intuition says that the extra bias would reduce the gain and so the noise but my intuition has been known to be suspect.

Maybe some noise from the signal generator? Nah. That would call the whole method into question.

What do you think?

Tony

Hi Tony;
Just wondering if the noise is present at all levels? 10 uv? 10,000 uv? If its the same, perhaps it's coming from the signal generator? My Motorola 2008 has a pretty noisy carrier in AM mode for some reason and my old military (tube generated) signal gen is extremely quiet.
If you have the AVC off, the tubes are running at max gain and there is always some secondary emission that may contribute to the overall noise. Not all pentodes are created equal as are some pentagrid converters.
Some converters I've used also are quite noisy in comparison to others.
So, there ya go, I didn't answer your question. LOL. Got to ask a few myself, though.
Perhaps I'll pick up some good info from this thread. Great question.

Peter

Thanks Peter. I don't really know if this behavior is peculiar to my R-390a. I think it is pretty typical actually. My HQ-180 does the same thing (maybe more so). No carrier and it's quiet. Carrier gives a substantial rushing noise. Certainly some of it comes from the generator. There has to be random variation in the carrier amplitude and that would be detected as noise. Hey, I just had a thought. As the carrier goes through the stages of the radio, the noise in each stage "modulates" the carrier. Maybe without the carrier, the noise can't get through the tuned stages. Add the carrier and it rides all the way to the detector.

I like that idea. I am thinking "out loud" here... Noise has a spectrum. It isn't necessarily flat. Suppose the noise at a stage is large at audio frequencies but falls off at RF. The portion of the noise that is "in band" for the tuned circuits downstream would be all that gets through without a carrier. Add a carrier and the low frequency stuff might be able to hitch a ride as well. Hey, it sounds good .

Tony

If that idea has any merit, it would pick up the most noise at the RF amp because the carrier is the smallest there relative to the noise. As it moved through the stages the carrier gets bigger but the stage noise stays about the same.

Perhaps I have been doing it wrong, but, I interpret the signal + noise to noise ratio (sensitivity) as connecting an ac voltmeter to the audio output, take a reading then apply R.F. at 30% modulation, raising the generator output until the AC voltmeter reads 10db higher than before.

My SP 600 met its specifications using that method. am I doing it wrong or is this different than what you are trying to measure?

Charlie

Hi Charlie,

Thats another way and it tends to give much better sensitivity numbers but many argue that this is not representative of the actual AM performance. Look here for some more info: http://r-390a.us/R-390A%20Performance.htm . I measure both ways. I have found some interesting differences between the two. For example, when setting up an R-390a, your method benefits greatly from reducing the IF gain setting to a much lower level than the manual suggests. The modulation-on/off method not so much.

Tony

Edit: By the way, the R-390a manual specifies your method. It is pretty much the standard in BA manuals that discuss it.

Thanks for the answer Tony. FWIW I wound a torrid transformer to match the output impedance (50 ohms) to the 125 ohms of the SP600. The problem is: does the SP 600 maintain a 125 impedance and at all frequencies?

The reason I tried this was so I could read the dbm value of the generator attenuator and meter directly, assuming nil loss in the transformer.

I have to make something to measure the antenna input impedance of the receiver at any frequency.

Charlie

The 390 group often reinvents the wheel to their own liking but the long established SNR measurement was as mentioned above.

Set a baseline audio noise level using the dB scale of your meter with no signal, AGC off and RF gain on full. The IF stage gain should already be set and left alone. An AM 10dB SNR measurement includes the signal plus noise as well as the receivers ability at AM recovery which is not a constant among models.

Carl

Can't say about the SP600 input impedance but my guess would be that it is not constant across the coverage range. The R-390a input impedance is definitely not constant. I make the AM measurement and then crank up the step attenuator on the gen high enough to get a reading on the millivoltmeter at the RX antenna terminals. Reduce the millivoltmeter reading by the attenuator setting to get the actual signal level delivered to the radio.

Carl is right that there is no shortage of opinions on the R390 list. My take is that the traditional method gives you a number that you can readily compare with other receivers because most are measured the same way. The modulation-on/off method, in my opinion, correlates better to the ability of the receiver to produce clear audio on weak signals. A better number means that I will be able to hear better. If you use the traditional method, a carrier with no recovered audio will still produce 6 or 8 db of "signal to noise". The mod on/off method will yield 0 as you would expect.

It's just another measurement you can make that helps you figure out how well the radio is working.

Tony