[Radio Wrangler]

At HP we were involved with selective level measuring sets which had to winkle a single channel out of thousands, all thundering away at max level loaded with phone calls. We ran into problems trying to get crystal filters with adequate dynamic range to not be the limiting factor. 2N3866 and custom arrays of >5GHz bipolar power devices were used. The mixers were designed by Hugh Walker (Author of "Sources of intermodulation in diode ring mixers" Proc IRE)

Later, we made an NPR test set which used a superhet receiver with a next generation diode ring mixer (Very high level!) and the channel selectivity at IF. This worked with an FDM system where normal loading was simulated by white noise at max level across full system bandwidth. We could reliably get a slot depth of comfortably over 70dB, but for best NPR floor, we needed to have crystal preselector filters made. In the end, it looked like only Thomson-C.E.P.E. could make them linear enough.

After that sort of development, pushing off-air receivers to better large-signal handling and dynamic range wasn't exactly easy, but it was easier. The tale of the NPR measuring set is in the HP Journal for April 1982. The instrument was too late for the market. No more FDM was being deployed, PCM was taking over. What was learned got used, though when later we got involved in surveillance receivers for spectrum management agencies, and some other people.

Great fun!

David