Updated  11 Aug'11

--> home/projects/1W_10GHz_PA/



1W - 10GHz SSB
Power Amplifier



This project by IW9ARO was made to upgrade his homebuilt 10GHz transverter to 1W output power. It is a 2-stages design, and even if it is a straightforward, standard approach, some good informations could be passed to hams starting to build similar blocks.


Thanks to the help of a good italian hamfriend, we got the main component needed to build this project: the teflon PCB, with etched traces to accomodate a 2-stages amplifier.  The choice of suitable transistors has permitted to reach the targeted 19dB gain and 30dBm output power in a small sized box, that was fitted inside the transverter in the same place where a single stage 20dBm PA was. As always happens in microwave projects, a large part of the time was spent for the mechanical arrangement.

-IMPORTANT- I did not believe, at the beginning, but it is absolutely true; GaAsFETS are veeeeery delicate. Anything metallic that gets in touch with their leads (no matter if they are on the PCB or not) can damage them. Use only non-metallic tools, hot-air soldering station and ESD precautions (the simplest one is not to wear shoes). They should be mounted on the PCB after everything else is already in place, and after that negative bias on the gate traces has been verified to be present. Power GaAsFETs are high-priced, and great care will save you great money.    



The two stages lineup is straightforward, and transistor selection is discussed in the next section. The bias point can be set individually for each transistor, with the aid of the two sensing resistors on each drain. Being 1ohm resistors, each mV measured with the multimeter corresponds to 1mA Id. The output power is sampled with a simple directional coupler to monitor the correct operation during transmission.
The supply section is designed to give a different Drain voltage for each stage: 5V to the driver and 10V to the final. Drain supplies are enabled only during transmission to limit power consumption and heating. Negative bias for the gates is always present, and is generated by a charge-pump inverter supplied by a regulated 5V. A protection circuit disables both Drain supplies in case of failure of the negative bias generator, and true shut-down regulator type has been chosen on the 10V line because even a 1V residual Drain voltage on the final GaAsFET is enough for a couple of Amps current to flow. 
Among the 3 linear regulators, the one dissipating significant power during normal operation is U1 (an LD29150 or MIC29150), therefore is should be mounted on a dissipating surface. Being a very low-drop regulator, its 10V output is still regulated down to 10.3V minimum supply. A power-schottky diode protects the whole circuit against reverse polarity. 

  Back to top



Basically, the choice of the final transistor depends on the target power, its commercial availability and of course, price. Do not expect high gains from power GaAsFETs: most common GaAsFET in the 1W range is the Mitsubishi MGF2430, capable of delivering 1.2W with 7dB gain at 10GHz. Then, considering the amount of total gain needed for my PA (19dB), the driver stage should have a minimum gain of 12dB while capable to deliver a minimum of 23dBm (200mW). There could be none of such a GaAsFET on the market: for example, the popular MGF1801 delivers 23dBm, but its gain is just 9dB... 
Doing some research I found the EPA060B-70  by Excelics: 25dBm and 11dB gain at 12GHz (so easily going to 12dB at 10GHz). A strange 400mW microwave HEMT, probably designed for high-dynamic front-end, because it has NF= 0.5dB and is packed in an ordinary package for small signal GaAsFET (not flange mount, then).  The price was fair so I bought two of them to keep one spare, that was lately used to upgrade my beacon transmitter. I also bought one MGF2430 for the final stage and my experiments begun...

  Back to top



The enclosure box for the PA has been tailored around the PCB dimensions. Since the final placement inside the transverter should be vertical, the only possible way for heat dissipation was through one of the sidewalls. Not having the possibility to mill a custom aluminum case, my choice was to use al "L" shaped frame, cut from an aluminum bar, that would constitute the bottom and one of the long sidewalls. The heat would easily spread from the bottom where the GaAsFETs are mounted to the thick sidewall and then to the front panel of the transverter. The other three sides of the box were made of tinned iron sheet, secured to the edges of the L-shaped frame by miniature self-tapping screws. SMA connectors are put on the bottom part of the frame, and not as usual on the edges of the short sides. This was done so because of the mechanical constrains inside the transverter.  The shielding cover for the completed box was built by cutting and bending a piece of tinned  iron sheet. After the PCB preparation, accurate holes marking was done simultaneously on the PCB and on the Al frame. Two holes are for the MGF2430 flange,  two holes are for the two SMA center pins, and all others are for the many miniature screws that hold the PCB to the bottom of the frame. 

PCB preparation is an important step. Making mistakes here means that the PCB could not be used any longer...  After cutting the PCB edges to the desired final size, a very precise slot should be cut for the MGF2430 mounting flange. To restore the Ground copper layer continuity, a thin copper foil has been soldered on the PCB backside, using solder sparingly to not alter too much the planarity.  

To make the vias from the Ground layer to the top I used 1.5mm wide copper stripes, passing through narrow slots punched through the teflon laminate. After soldering, intense use of desoldering braid will remove excess solder. After drilling the holes on the PCB, cut some Ground copper around the two holes where SMA center pins will pass through.
The two GaAsFETs must be soldered after all components are mounted and all voltages generated by the supply section are checked. Both bias trimmers should be initially set to give -3V to the gates. The MGF2430 is a very precious component, so strictly observe all ESD precautions. To avoid mechanical damages, screw it in its definitive position before soldering its gate and drain. By measuring the voltage drop on the sensing resistors, set the Drain idle bias current of each GaAsFET with both input and outputs of the PA terminated with 50ohm, possibly applying power with a current clamped power supply.  Adjust driver and final idle current, respectively, at 50mA and 300mA. Keep monitoring output power so ensure there is no self-oscillation once the two transistor are brought into the linear region. 
After bias adjust, apply the input power and make the matching adjustments with small pieces of copper foil. I prefer to do this with a swept frequency input and a SNA, because I can see if there is more gain in the neighborhoods of the 10368MHz band. An important step is to glue a foil of microwave absorber on the inner side of the shielding cover, to minimize the detuning effect and instability when the box is closed.


Final tests, after accurate (and time spending) matching optimization,  confirmed the project targets of 19dB gain and 1W output power. The small size has allowed a smooth retrofit of the existing 100mW PA, bringing my 10Ghz homebuilt transverter to a very respectable transmitting performance. Special thanks to Armando I3OPW for the big help given.  


  Back to top



  Back to Ham Radio Pages summary  or Back to Our Projects Pages summary


The projects presented in these pages are our own design and have been tested and verified by ourselves at the best we can. However, they might be inspired by concepts, ideas, solutions coming from known-art or free resources on the Web. We provide them as  reference designs to skilled hobbyists and technicians  who are willing to reproduce them for non-commercial use. Your results might be different from ours and we cannot be considered responsible for that. Similarly, we are not responsible for any damage or injury you might incur while building, assembling or using the equipments, projects or ideas presented in these pages. The firmware embedded in our projects is our property unless differently stated and, when available in the Download Area, it is license-free only for non-commercial purposes.  

Hotwater 2007-2011