13cm Power Amplifier
by Ian Bennett G6TVJ
Having built a simple synthesised 13cm ATV exciter my thoughts turned to what could be used as a power amplifier. The exciter produces about 10mW perhaps enough to drive some sort of MMIC or hybrid amplifier. The problem with 13cm is that at 2.3 GHz and beyond devices either discreet transistors or amplifiers get very expensive. After researching a number of manufacturers and suppliers I eventually came up with an IC amplifier costing about ú40 and producing almost a watt, at these frequencies not a bad power to price ratio.
The PM2104 manufactured by Pacific Monolithics is a GaAs MMIC amplifier device intended for ISM applications centred on 2.440 GHz, the device is relatively wide band so covers the whole of the 13cm amateur band. The IC has a gain of 24 dB and runs on a supply of +5V, it also requires two negative bias supplies similar to other gas fet amplifiers. The 2104 is a surface mount device housed in a SOT8 package and heatsinked via its metalised base, its pritty tidly so is not recommended for newcomers to surface mount construction techniques.
A complete amplifier can be made by mounting the device on an enclosed PCB and supplying it with negative bias and a positive supply.
All the clever stuff is done inside the PM2104 IC, fig 1 shows the circuit diagram of the amplifier. The bias supplies and positive supplies are decoupled, some 50R matching resistors and a 33p cap are also required. The device is nominally pre-matched to 50R but a 33p matching cap is needed at the input pin.
The device is powered on pins 4 and 8 with 5V. Negative bias is supplied to pins 2 and 5. An ICL7660 DC to DC converter chip is used to generate the negative supply in a similar manner to what is used in some of the Microwave Committee 3cms units. A simple comparator circuit is used to detect the presence of the bias supplies and then switch on the +5V VDD supply to the IC. If the bias voltage fails the IC may be destroyed so the comparator is needed to protect it. As a second measure the L200 5V regulator incorporates a current limit set to about 600 mA, I have not proved the action of this circuit but after the touch and go action of soldering the IC in and the ú40 whole in my pocket I thought I would put it in.
The bias generator, comparator, and regulators can be built on stripboard as convenient, only mild heatsinking is required. A standard fibreglass double sided PCB is etched to the pattern shown, the layout comes from the manufacturers data sheet. The base of the IC is metalised and must be soldered down to the board to form a heatsink. First tin the underside, solder down the pins and finally solder down the base by applying the iron to each side. The operation of mounting the IC is very tricky so not recommended for people without previous experience of working with surface mount devices. The other components can be added including several via pins which help earth the device. The PCB is then mounted in a tinplate box and fitted with SMA connectors in a similar way to how the RSGB microwave committee 3cm amplifiers are constructed.
The data sheet suggests that pin 2 should be set to -1.2V and pin 5 set to -1.5V, it is best to start with more negative bias and reduce them carefully while monitoring the supply current and power out. It is worth checking the action of the L200 current limit first with a dummy load. The protection circuit can be set up by adjusting the protect pot until the VDD supply just switches on (Best done before connecting the VDD supply to the 2104). I found that I could get the best efficiency with slightly different values of bias. By comparing the DC power in and RF power out the bias can be tweaked for minimum power dissipation of the PM2104. The values I ended up with were -1V and -2.2V for pins 2 and 5 respectively. The attainable output power should be up to about 800mW, +29 dBm with 5 mW of drive. The PM2104 runs warm to the touch in operation.
In the absence of a better solution for attaining medium levels of power at these frequencies the PM2104 does OK. The amplifier has provided good service facilitating P5 pictures over a few miles and some 50 miles on last years ATV contest. At ú40 not a bad mW per ú ratio at these frequencies. It is a tad awkward to use and mount so if anyone has a simpler solution please share it with us.
[ About me
| Acronyms | CW
| Data Sheets | Docs
| Download | E-mail
| HOME | Ham
projects | Hobby circuits
| Photo galery | PIC
Sign in my guestbook | View my guestbook ]
ę 2001 - YO5OFH, Csaba Gajdos