High Powered Amplifier Bias and monitor PCB

When I began as an engineer at UCAR, my first main project was the design and implementation of a fix for these custom amplifiers that we utilized in our 449[MHz] system. These HPAs blew out constantly without a culprit, losing us about 5k when one blew out every month. Investigating them, I first developed and documented a fell repair and validation procedure so as to make it that the technicians I worked alongside could replicate the repairs necessary to salvage an amplifier. This led me to two possible cultprits, the biasing circuitry on these amplifiers varying wildly due to being a separate PCB bridged with haphazard jumpers, or the system’s total power chain being implemented in a supply chain that contained 48V fans that drew immense current whenever they were enabled. Fixing the former I developed a new bias circuit that eliminated the variation to voltages that resulted from manufactured discrepancies as well as adding rudimentary current and temperature monitoring via a generic Atmel mcu that I chose for the reliability and ease of flashing. Separating it from the analog circuitry was a hefty discrete low-pass filtering network preceding biasing (as best as it could be isolated within the format), and powering via a separate supply LDO.

While I was only allowed the revision of the bias circuit, I took many issues with the outdated design of the amplifier sections as it was hard to assemble and hyper dependent on assembler’s proficiency with hot air rework tools, which none of our technicians were. I had only had it as I used to reflow the chips on ps3 and xbox360 systems while in high school. However they chose not to pursue this as between the new PCB and dedicated supply distribution changes in the powering and operation of the fans, we stopped blowing an amp every month. So far only a single amplifier burned out over the year after these changes were implemented, and it was most likely due to a mass of snow that had piled on the antenna.