Notes on In-band IMD
In the most common form of IMD testing, a receiver is subjected to two strong signals, both of which are outside the receiver's passband. In the case of third order IMD testing, the signals are close to the receiver's tuned frequency. With second order IMD testing, they are far removed.

In-band IMD testing is a different kettle of fish. Here, both test signals are placed inside the receiver's passband. In the test we use, the signals are a mere 100 Hz apart. The test signals are -76 dBm, which registers "S-9" on an FT-1000MP, and is a little less that the old Collins definition of an S-9 signal (-70 dBm).

In out-of-passband testing, the receiver's response to the IMD products is monitored with a true RMS AF voltmeter connected to the receiver's AF output. With in-band testing, we monitor the response with an AF spectrum analyzer. We use the same computer-based SpectraPLUS system you've encountered with our testing of IF and AF response.

In-band testing is not usually a big deal with commercially manufactured transceivers. Generally, their in-band responses are thought to be "well behaved," if their IMD products 30 or more dB below the test tones.

However, we suspect that in-band IMD might be more important than commonly believed. Our intuition comes from our experience with the famous "presence" of a high quality DC receiver. For example, take a look at the article by Rick Campbell, KK7B, on his R1 direct conversion receiver. (QST, August, 1992, page 19). He shows two AF spectra side by side (reproduced below). The left spectrum is from the R-1, while the one on the right is from a NE-602/LM383 receiver. Even though the receiver on the right satisfies the criterion of IMD products being more than 30 dB below the test tones, we have no doubt that the R-1 sounded better.