Dual-Band Transmitter Board Version X2

[NZ0I]

There is some progress to report on the build out of the Rev X2 Transmitter board. All of the power supplies have been installed on the transmitter board and are working adequately. The only challenging power supply was the switching "buck" supply that will provide settable bias voltage to the transmitter final amplifiers (both 2m and 80m). The switching supply takes the 12V coming from the power supply board and efficiently drops it down to 2.5V to 10V, settable by the processor, for controlling the amount of transmit power being generated by the final amplifier.

The LTC3600 buck switching supply controller (U208 in the schematic) by Linear Technology/Analog Devices is being used for the first time in this project, so of course there was some tweaking required. It turned out that, as designed, the supply was unstable. The output from the supply would swing in an oscillatory manner with an amplitude of several volts. It was possible to stabilize the supply by changing the switching frequency (making it higher) - a solution that required changing a single resistor value. The LTC power-supply design software proved very helpful for determining the best solution. 

With no other parts changes, the buck supply will now provide a stable variable voltage between +2.5 and +9.2 Volts. It doesn't quite reach the design goal of +10V. Some additional tweaking of component values should get the voltage up to +10V. But that might not be necessary if it turns out that 9.2V is sufficient to achieve a reasonable transmit power level. So further power supply tweaking can wait until there is more data on transmitter performance.

I haven't attempted any efficiency measurements of the buck supply. But it puts 3.9 watts into a 22-ohm load resistor (that's over 400 mA) and all power supply components remain cool to the touch. So it is clearly operating more efficiently than a linear regulator. Linear Technology claims is can be over 90% efficient, but the efficiency will vary depending on the load and the output voltage... so a single efficiency measurement won't tell the whole story.

The 22-ohm 5W carbon load resistor I am using did get warm enough to melt the plastic O-scope clip that was attached to one resistor lead. Oops. The clip is still usable, but now has a little pit melted into it. The power supply components remained cool throughout the ordeal.