Special Mixer Mess

This is a little tour of what you find. A series of pictures that come with a warning that it is uncomfortable viewing for some, meaning those who do not have it in them to do a disgusting bodged lashup on a older, but decently engineered spectrum analyser. This one had a 1.5GHz span with tracking generator, YIG local oscillator, more than 90dB dynamic range, +/- 1dB accuracy built to deliver a straight line across the screen without the benefit of software correction map to remove every little wiggle after a sweep through a indifferent piece of thin cable. I could not do it, even had I pockets so deep, and a life so fine that I had a newer model to spare!

In my self-employed, hence insecure and somewhat scary life, I take out nice new calibrated kit hired for the occasion when out at the customer's site. Back at the lab, I accumulate what suitable classy kit I can afford. For development, it is inevitably "older", kept calibrated by the passing hired equipment. I can wring 0.2dB accuracy out of a ugly lashup if I need to, but I try not to act like some sad reactionary, locked in the past. If I feel a bit outraged at insensitive treatment of measuring equipment, think of it as a minor personality feature!

Hopefully, this also becomes educational, because on the way, questions arise about how this device actually works, and also fun and games with a soldering iron. A sort of microwave forensic dissection, with questions and hard sums on the way. OK - weepy tissues and maybe also a stiff drink at the ready - here we go..

Hmm.. thats a nice enough 1.5GHz lowpass filter...Eh! Whats that on the left??

What might be this strange lump of plastic on the underside of the lid, with a matching hole cut into shielding? The display trace of 0-1GHz at 100MHz/DIV looks like a dogs hind leg, but it does not change hardly at all if you take the lid off the mixer and wave a finger about in it. The topside of the stripline filter part has of course to be put back, but the mixer can run open. Actually touching the active bits does affect the trace a little.

Looking back at the filter, the little rectangles are capacitors having a dielectric layer to both upper and lower groundplanes. The little tracks that join them are the stripline inductors. The route taken packs a long filter into a shorter space, and tries to keep those capacitors far apart. You know it's a low-pass filter because the ultimate low frequency (ie. dc.) would pass straight through all those gold-plated copper bits.




Page 1 Special Mixer Mess (description of the problem)

Page 2 - Spectrum Analyser Mixer Dissection - Taking things apart

Page 3 - Spectrum Analyser Mixer Dissection - A closer look

Page 4 - Spectrum Analyser Mixer Dissection - Oh Yes - it is still full of surprises!

Page 5 - Spectrum Analyser Mixer Dissection - How it ended up

How We Finally Got it Right (and all the secrets too!)

The Result - Performance Traces

Some (not so good) Traces


Author : Graham Seale