Once upon a time there was a contest ....!

For many years the OZ9KY contestgroup has been participating in the VHF/UHF contesting. Having on of the most exposed - and best - QTH in this country, and running high power, big antenna systems and sensitive receivers/preamplifiers we have experienced our share of interference problems.
We have received reports from nearby staions complaining over our broad signals occupying much of the band - as well as being disturbed ourselves, when trying to receive weak signals in the presense of broad and noisy signals from others. Interference is definitely a two way street!

To improve the situation, we must understand the problems involved and we must be able to do some sort of measuring of signals and equipment. In this note I'll describe some of my observations along with measurement on a number of transmitters and receivers.

How to interpret interference.

When we experience interference or "splatter" we tend to blame the other guy - after all the problem wasn't there until he started. But of course we have to realize that both the receiving as well as the transmitting side may be blamed and often both are. No transmitter is perfect. Besides the wanted signal it may put out unwanted signals like: spurious (discret frequencies with or without modulation), sideband noise associated with and adjacent to the wanted signal and wideband noise.

The receiver on its side may suffer from overload from simply not being able to handle one or more of the signals, the selectivity may be insufficient to keep faraway >20 kHz stations out of the IF amplifier and the phase noise from the local oscillator may end up in the passband through reciprocal mixing.

These are all well known and described in the litterature.

Limits for transmitter broadband noise.

When you are located on a hilltop you will have line-of-sight to a lot of other contesting stations. A drawing will show the actual exampel.
The difference between a very strong local signal - mayby 20 kHz away - and the weak signals we want to hear is around 100 dB. Or to put it into another perspective: - both of us must have transmitters with noise and spurious level of -100 dB in order not to interfere with each other. (This is just one criteria. There are similar "severe" requirements to the receivers.)

- measurements of real life signals. It's often assumed in calculations, that stations with line-of-sight have free space attenuation between them. To check this hypothetis a number of real life measurements are being made. More later.

Direct measurement of transmitter broadband noise.

You need a spectrum analyzer with more than 100 dB dynamic range in order to directly measure transmitter broadband noise. This is much more than the 60 to 70 dB dynamic range of an "ordinary" amateur surplus spectrum analyzer.

But with SSB transmitter's there is an alternative - measure without modulation.
With a suitable amount of attenuation between transmitter and analyzer adjust for full screen with a modulated signal. Now remove the modulation but keep on keying the transmitter. The signal drops down some 40 to 50 dB depending on carrier suppression and microphone noise. Now the attenuation can be reduced and therefore analyzer sensitivity. In this fashion it's not difficult to measure 100 to 110 dB below peak level.
- measurements of different transmitters. (This may come as a surprise to some ..)

Limits of transmitter IMD and spurious signals.

Close to the transmitter frequency the IMD level is nowhere near -100 dB. Typical 3'rd and 5'th order products may be 26 to 36 dB down from PEP. But at >20 kHz we ideally need the -100 dB, and this is what we should aim for!
One thing to keep in mind is the difference between technical IMD measurement and real life speech modulation. With a standard two-tone measurement the tone spacing is normally 1 - 1,5 kHz. As an example this means that the 41'st order sidebands are 20 to 30 kHz removed from the transmitting frequency.
However in the human speech no such frequency difference exists between individual voiced frequency components. In the male speech the individual components are harmonics of 100 - 120 Hz (up to abt 5'th harmonic depending on actual wowel), making the 41'st order sideband very much closer to the wanted signals. The higher frequencies in the human speech are unvoiced and have a much lower energy content.
So the IMD spread of speech modulated signals are (fortunately) less than the spread you get with a two-tone measurement.
With spurious it's normally not so critical. They will be on isolated frequencies, not occupying the whole band and usually outside the contest band - although there are exemptions - like the FT 847 which has -70 to -80 dB sidebands +- 15 kHz away.

Direct measurement of transmitter IMD and spurious signals.

Measuring IMD and spurious in a 60 dB range is relative straight forward on a good spectrum analyzer. It's possible to "overload" the analyzer by 10 dB - which works in most cases - to get 70 to 80 dB range. Any more "overloading" may saturate the IF section, and should be approached with caution
- see the measurements here.

But measuring in a 100 dB range during modulation is TOUGH. However by using an x-tal notch filter to reduce the SSB signal while passing the sidebands it's possibel to get down to -100 dB, at least for some of the side band frequencies.
- read more about the technique.

Limits of receiver single-signal selectivity.

As shown in the example a situations with an interfering signal of -40dBm at the antenna is not uncommon. (If you want to speculate in worst cases there can be more of them - and stronger) As contest stations often uses a preamplifier at the antenna to increase sensitivity, the signal down at the receiver is even higher. With a "well tempered" preamplifier (ie not too much) we end up with -30dBm at the input of the transceiver/transverter.
Ideally the receiver should cope with this and - at 20 kHz distance - be able to receive a signal some 100 dB weaker.

Indirect measurement of receiver selectivity and phase noise.

A way of checking receiver performance - both blocking and selectivity - is by taking an x-tal oscillator with an output of 5 to 10 mV (-33dBm to -27 dBm), and using this as a signal generator with very low phase noise.

The simple way to check a receiver:
Connect the x-tal oscillator to the receiver and tune around the x-tal frequency. If your receiver can be tuned to within +- 10 kHz or closer from this signal without the S-meter responding or strange noises coming from the loudspeaker, you have very good single signal selectivity.
On the other hand, if the noise in your receiver starts increasing when you're still 50 - 100 kHz away, thats an indication of either too much phase noise from the LO or a poor selectivity in the IF. Only very few stations can pass this test!

The simple way to measure the selectivity:
This methode uses the x-tal oscillator and the receivers own S-meter as indicator. First start by calibrating the S-meter using a signal generator. You will probably need to know the values for S3 to S9. Then connect the x-tal oscillator through some fixed and an adjustable attenuator to the receiver. It's important that the attenuators and cables are well screened so the signal doesn't "jump over". Tune for max. signal and set the attenuators to give an S-meter reading of f.ex. S5.
Now tune to either side of the signal keeping the S-meter in the S5 range by changing the attenuation. You have to tune very carefully close to the filter edges, as things happends very quickly here.
By noting the S-meter reading and attenuation setting you can now draw up a curve of the receiver response to a single signal.
See my measurements of different receivers.

Specifications for transmitter and receiver.

For the "big gun" contest station: (1kW, preamp., big antennas)
Transmitter broadband noise <-100 dB relative PEP.
IMD <-100 dB at 20 kHz distance.
Spurious <100 dB inside contest band.
Receiver: able to handle a single signal at 10 mV and have full sensitivity 20 kHz away

For the "small pistol" station: (25W, no preamp, single yagi)
Transmitter broadband noise <-80 dB relative PEP.
IMD <-80 dB at 20 kHz distance.
Spurious <80 dB inside contest band.
Receiver: able to handle a single signal at 3 mV and have full sensitivity 20 kHz away