This small, light weight digital SWR and power meter uses an Atmel AVR "Butterfly" demo board and a "Stockton" type directional coupler. It is built into a "Whitmans Sampler" candy tin. The display board is mounted to the lid of the tin, so the cover can be closed to protect the display when the unit is not in use or is in transport. This makes it an idea meter to use when traveling or in the field.
BNC connectors are mounted to the back of the tin, which support the lid when it is open. The directional coupler is made using a BN-43-202 binocular core, rather than two individual cores, which is normally done. Using the binocular core results in a lower profile and eliminates the need to cut slots in the board for round cores. W7EL diode compensation amps are used to linearize the output of the bridge diode detectors and results in much more accurate readings at low power levels than would be possible with out the compensation amp. A micro power CMOS rail to rail op amp is used. This allows the amplifier to run off the same 3 V button cell which powers the AVR board and draws so little current, there is no need to remove power from it when the unit is not is use.
The meter has two power ranges and calculates SWR. A 5 way "joystick" switch is used to control the AVR board. Pushing the switch "UP" turns the display on. Pushing the switch "IN" steps through the meter functions. 15 watt scale (00.00 watts) 9 watt scale (0.000 watts) and SWR (01.00) . There is an automatic 5 minute shut off (turns off the display and puts the processor to sleep). The control switch can also be pushed "down" to put the meter to sleep.
Accuracy: Just because the display has 10 and 1 mw resolution, doesn't mean you should believe it! There appears to be some what of a "bow" to the response curve, with readings which are as much as 200 or 300 mw high at mid scale. I believe this is caused by slight over compensation by the diode compensation amp. Even so, this only a few percent error, so no big deal. Also be aware that a meter of this type is only accurate when driven by a true sine wave. A distorted or asymmetrical signal will not measure properly.
Round off error: The A/D converter is a 10 bit, 1024 count converter. In order to measure up to 15 watts, we need to measure the equivilent of 40 volts. Therefore, the A/D count is multiplied by 4 to scale it. The over all scaling factor of the coupler is 40:1, so things work out properly. Because of the X4 scaling of the A/D count and the relatively small resolution of the A/D, the display will not count in single digit incerments all the way to full scale. At some point, the display will start to count in 2's and then 3's as the numbers get larger. (or 20's and 30's if on the 1 mw scale)
SWR calculations are pretty accurate, as this is simply the ratio of the forward and reverse voltages. The main thing which will affect the accuracy of the SWR calculation is how well matched the forward and reverse voltage detectors are matched.
The Atmel AVR Butterfly demo board is available from Digi-Key for $19.95 (part # ATAVRBFLY). In order to use the demo board for this project, you will have to reprogram it. In order to do this, you will need to down load AVR Studio 4 from the Atmel web site (http://www.atmel.com) This is a 20 meg file, so if your on dial up, it will take a good hour to down load. While your at the Atmel site, also get the Mega-169 data sheet and the Butterfly user guide. The user guide will show you how to reprogram the Butterfly board. Reprogramming is a simple matter of wiring a DB-9 serial connector to the Butterfly board and using the serial port of your PC. The processor on the demo board has a "boot loader" program which makes this possible.
You will also need the power meter software which I wrote. Down load it here: Butterfly.zip
This zip file also contains a program to use the Butterfly board as a DDS controller for the AD9850 DDS daughter board by NJQRP. An html document explains this program.
Idealy, the meter would be calibrated by comparing it to a watt meter which is known to be accurate. Reasonably good calibration can be achived by applying a 2.21 DC voltage to the bridge side of the detector diode (D1 or D2), then setting the trimmer to display 10.00 watts. (actually, you can only get 10.01 or 9.98). If this method is used, be sure to disconnect the transformer first, or it will short out the DC supply.
Reverse power is not displayed, so to calibrate the reverse power side, move the forward power input lead to the Butterfly board to the reverse power output. If your using a watt meter for calibration, reverse the antenna and transmitter cables. Calibration of the reverse power detector is required to get accurate SWR measurements.
I built the bridge and diode compensation amp with primarily SMT parts and etched a circuit board. If you have a pcb layout program and can make your own boards, this is the best way to go. You could use leaded, through hole parts if you wish and build it "dead bug" style. The circuit is simple enough, it won't take too long to do this. In some places, like mounting the leads from the balun transformer, you may want to cut "Islands" out of the copper foil to make connections. I prefer this method over "Manhattan" style construction, as I can never get the little pads to stick, hi.
What might not be entirely clear from the diagram above is your making two transformers with one balun core. Therefore, there are two secondary windings, each wound through one hole and around the outside edge of the core. Use a small enough wire so most of the turns stay on opposite sides of the core. This will minimize any cross coupling between the windings. The primary is a simple wire which passes through the hole, one on each side.
The BN-43-202 core might be a little hard to find. W8DIZ of <http://www.kitsandparts.com> carries BN-43-302 cores which will work just as well ( the '302 cores are a little shorter than the '202's) Diz is a good source for all your toriod needs.
The op amp used is a STmicrodevices TS942. It has a very low operating current of 1.2 uA per amplifier, so it won't drain the battery much! These are available from Mouser in both SMT and DIP. Germanium 1N34A diodes are shown used on the schematic. These can be a little hard to find. High speed Schottky diodes (not rectifiers) can be used instead. A type such as Digi-Key # SD103CDICT would do. Ideally, you should get enough of either diode to be able to match the forward drops between the diode used as the detector and it's corresponding compensation amp.
There are a few modifications required to the Butterfly board. These are shown below. The photocell resistor needs to be removed, along with two little resistors. A third needs to jumped over, which is a little easier than removing it and then jumpering the pads.
The exposed side of the battery holder is "hot", so putting some electrical tape behind the lid of the tin is a good idea, so when you mount the board, it doesn't short out.