Clifton Laboratories 7236 Clifton Road  Clifton VA 20124 tel: (703) 830 0368 fax: (703) 830 0711

E-mail: Jack.Smith@cliftonlaboratories.com
 

Wes Hayward, W7ZOI, is a prolific equipment designer and author, including a clever low power, wide range RF power meter described, along with his co-author, Bob Larkin, W7PUA, in Simple RF-Power Measurement, June 2001, QST, page 38.

W7ZOI's power meter uses an Analog Devices AD8307 log amp, which has a DC output proportional to the logarithm of the input signal amplitude. This means the AD8307's output is directly proportional to the input power stated in dBm, or dBw or dB with respect to any other reference. (My Z90 and Z91 panadapters use the AD8307 as the log detector. It's a marvelous little chip.) The AD8307 is usable to above 500 MHz and has a very wide range, with good accuracy between -70 dBm and +13 dBm, and is usable well below -70 dBm.

If you search the Internet with the title of the article, you will find full PDF copies. I'm not sure of the copyright status of those PDF files, so I won't provide a link. In addition, Wes has a follow up at his web site that should also be read. http://w7zoi.net/Power%20meter%20updates.pdf

Gary, W2GNN, built a power meter from the June 2001 QST article and recently asked me to calibrate it for him. Gary has kindly consented to making my report available to readers of my web site, with the hope that it will spur others to build the power meter.

You can read my report by clicking here or by finding it on the Documents page of this site.

My conclusion is that if you need a low power termination meter, it's an excellent device. Very good return loss up to 148 MHz (important if power is to be measured accurately) and, with the proper calibration it can read power to 0.25 dB or so. (The AD8307 is rated at ±0.5 dB but Gary's meter does a bit better than that.)

The photo below is one I took of Gary's power meter. I didn't open it up, but the outside certainly looks nicely done. (Gary plans to add a custom calibrated scale to the now blank meter.)

The plot below shows the relationship between input power in dBm and the meter's DC output voltage A perfect log-linear relationship will be a straight line, and as the plot demonstrates, the fit is extremely close to linear over the range -70 dBm to +13 dBm.

Log Power = y + mx

where y is the "intercept" and m is the "slope."

These coefficients were developed for each amateur band between 160 meters and 450 Mhz using an automated signal generation and data collection system I developed.

The plot below shows the "residuals" which represent the difference between the power meter's expected output predicted by the equation above (with  the coefficients for the particular band used) and the actual measured values. The smaller the residuals, the closer the power meter's output is to being perfectly linear, and the data shows the residuals are nicely clustered within a relatively small range up through 148 MHz. The 221 and 441 MHz data is more dispersed, most likely due to the degraded input SWR of the power meter.

This calibration, of course, references W2GNN's meter to the accuracy of my test equipment, and an appropriate error allowance must be included for that.

These matters and more are covered in the detailed report.