Continued from the Low Cost Wireless Network How-To, Appendix E
Select a picture for larger image.
Picture of the transmit/receive control logic boards and the voltage regulators. Once again, housed in an old California Amplifier MMDS downconverter case. +15 VDC comes in through the top via a 1000 pF feed-through capacitor. The transmit/receive control logic from the WaveLAN card enters via one of the F connectors on the right side. The +5 VDC on transmit and +9 VDC on receive signals leave the case through standard feed-throughs (non-capacitor).
WaveLAN side PIN diode switch module. This is a Mini-Circuits ZMWS-1211, and was purchased from Fair Radio for $20. The RF connectors are SMA, and the switch logic is +5 VDC @ 2 mA. They are good up to 2.5 GHz.
Control line #1 (RF port #1 control) shown.. Along with the 100 pF mica bypass capacitor, a type 43 ferrite bead, and a 2 k resistor used to drop the +9 VDC signal down to the +5 VDC used by the switch. Math lesson! R = (9V - 5V) / 2 mA) = 2000 ohms
The beast. The PF0031 based power amplifier. I say it's only 6 watts, but in reality, it'll do 10 watts easy. Note, however, in the following pictures you'll need to perform some serious RF engineering skillz, otherwise the amplifier will oscillate, probably destructively (AKA, blow up). I salvaged the heatsink and copper heat spreader from and old Nokia cellular phone. I suggest you try to do the same. The PC board the amplifier is mounted to was bought surplus from Down East Microwave, #CL11. This isn't the right PC board, so it will need some modifications to fit the module. The PF0031 module is available from RF Parts.
Closeup of the PF0031 module and its support componets. RF input is on the left side, through a SMA bulkhead and a 20 dB resistive pi attenuator using carbon composition resistors. The Vapc line is next (going left to right) and is air supported on a 56 pF ATC porcelain capacitor. The Vdd line is next, and is also air supported on a 56 pF ATC porcelain capacitor. Finally, on the right side is the RF output, also going through a SMA bulkhead. You can sorta see the copper heat spreader under the PF0031 module. DO NOT use heatsink grease on this module because the module's case is the ground. Heatsink grease would isolate the module from ground, and it will freak out and crash airplanes and stuff.
Closeup of the RF input side and the 20 dB attenuator. Should probably use surface mount resistors in the attenuator pad, but they won't handle the fairly high RF input power. You can also see how the circuit board was cut to isolate the Vapc line.
Alternate view to show how the Vapc and Vdd lines are supported in the air by the 56 pF ATC capacitors and the little jumper wires to connect them back to the rest of the circuit.
To reduce power amplifier intermodulation distortion and to protect the amplifier in case of high SWR, you should use an isolator. This is a surplus cellular phone isolator from Down East Microwave, #SE031. The magnet helps to "move it up" to the 900 MHz band. You'll need to sweep it with a signal generator and a diode detector, while moving the magnet around, to tune it. Glue the magnet in place when you're done. The 5 watt, 50 ohm load is also avaiable from Down East Microwave, #SC170. You'll need a N lesbian (N female to N female connector) to connect it to the isolator.
Receive pre-amplifier. Don't even bother with homebrew here, order the 33 cm LNA kit from Down East Microwave, #33LNAK ($40). It include all the parts, including the case (Hammond 1590A), PC board and SMA connectors. It will need to be modified slightly for use in this project. Most notable is the fact that it is feed directly with +9 VDC instead of going through a voltage regulator, and some of the bypass capacitors are removed. Also added, was a 915 MHz, +/- 15 MHz, bandpass filter (Down East Microwave #SC167, $5) in the pre-amplifier's output stage. There is also a small, 0.6 inch, #22 hairpin inductor, shunt to ground, on the pre-amplifier's input. This acts as a highpass filter, filtering out anything below about 400 MHz. This should help reduce intermodulation problems from the mixing with VHF pagers, FM broadcast, etc. Ohh, I forgot, you'll need to remove 2 turns from the series gate inductor if you add the hairpin inductor.
Closeup of the receive pre-amplifier's RF output, showing the big 'ole bandpass filter. You'll need to cut the output trace (after the series capacitor) in order to insert the bandpass filter. I used tiny RG-178 coax jumpers to insert the filter. The bandpass filter was added after the pre-amplifier in order to preserve the amplifier's low noise figure (<0.7 dB).
I lined the insides of all the cases with that anti-static foam you can buy at Radio Shack, part number 276-2400. Theoretically, this will absorb stray RF inside the case and help improve the circuit's input/output isolation. Or maybe I just smoke $2 crack....
Here is the antenna side RF PIN diode module. This is where I had the original difficulty in designing high power amplifiers. How do you switch a high powered 900 MHz signal back and forth in microseconds? You get ideas from someone else, that's how! A million thanks to Glenn Elmore, N6GN for this section of the amplifier. The RF PIN diode module is a Mitsubishi MD003. 0 VDC on the bias line connects the antenna line to the receive line, a +5 VDC @ 50 mA bias connects it to the transmit line. Mounted on a homebrewed PC board inside a Hammond 1590A case. Be sure all the ground pins on the module are connected to a low inductance ground. Also be sure to use high quality RF bypass capacitors, 56 pF ATC porcelain, in this case.
You can see the bias resistor and ferrite bead floating in the air. The bias line passes through a regular feed-through (non-capacitor). RF connectors are SMA, with the teflon covered pin (ala the 33LNACK), the connectors are also available from Down East Microwave.
Inside case shot.