Starting with a microwave synthesizer

So. Electronics has been a hobby for me for two decades or so, and profession for a decade. Being a somewhat active hobbyist, there's a pretty large number of various projects I've worked on. Which is to say, there's not that many projects that have actually made it all the way to getting ready. I'll post some pics and descriptions of some of those older projects if/when I'll bother, mostly I'll be concentrating on the gadgets I'm constructing now.

Lately I've been mostly working on a microwave test system that's been suffering a feature creep of pretty epic proportions. I've done some early drafts of various measurement functionalities as early as 1998, with some not very successful attempts at partial functions once in a while, but the design really started in 2010 when I acquired a couple of ADF4350 chips which allowed the construction of a signal generator with quite little effort.

This was the first module that was finished, it implements a frequency synthesizer with a frequency range of 137.5 MHz to 4.4 GHz with a 5 kHz resolution, and a thermally compensated power control loop with a power adjustment range between about -35 and +5 dBm with a 12 bit DAC control. The synthesizer and DAC are controlled through an SPI bus. Below is a block diagram of the board.

Surprisingly the design worked right away, with just minor adjustments in the PLL loop filter and power control loop gain. The low cost regulators used were however pretty noisy, and switching a lower noise device and some extra supply filtering for significantly improved phase noise left the device rather uglier than pictured above. 

The phase noise is far from awful, but nothing that would compare with a proper test instrument. The power control feedback loop would optimally reduce frequency variation, but of course the detector chip does not have a flat response, resulting in an uncalibrated variation of around 10dB throughout the bandwidth.

The device required some control, of course, and after adding a graphical display, keyboard, an ARM7 card and some power supplies and constructing a pretty slick looking case, the above device was born. However, being a crappy programmer with not that much interest in coding, and with new ideas for the synthesizer drawing my interest elsewhere, the user interface coding slowed and stopped before the device was actually useful. The power reading on the screen is rather more than the device actually outputs, but I never got around to implementing power level settings.