Andrew Cornwall
Home Page


The small magnetic loop antenna autotuner is the subject of an article in the May-June 2018 The Canadian Amateur (TCA) magazine published by the Radio Amateurs of Canada. The article describes the autotuner project and reports several sets of trials with various options - more options than are shown here. For the TCA article the trials were conducted in Nova Scotia during the spring of 2017, using a Yaesu FT-897 transceiver with a 5 watt tuning signal. The trials here took place in the Arizona desert where my wife and I were on a camping excursion during the winter 2018. The transceiver was a Yaesu FT-817, QRP rig, my radio traveling companion. The tuning signal was nominally 2.5 watts - as explained below. The trials here using Program 3b (quarter step reconnaissance sweep, quarter step active tuning sweep, and backwards steps) are the most similar in approach to those published in TCA. The results of numerous trials I've conducted in many locations demonstrate that the autotuner concept works very well.

(Use Link to Download Programs, Trials Data, and Sample Tuning Reports)

Note: There are two program files for each of three programs:
'BAS' is text that forms the basis of a microcontroller executable; PDF is for inspection.
Neither version is executable on a computer nor will it infect a computer.
These programs continue to be experimental and should help get your autotuner up and running. Don't be afraid to change timing, partial stepping, and to add or delete features. Reprogramming the PICAXE takes about a minute so you can try changes 'on the fly'.

* * * * * *

The programs here are the same except for partial stepping used with the reconnaissance sweep then the active tuning sweep(s), and for possible backwards stepping when tuning is nearly done. Backwards stepping was needed with two earlier models of the autotuner where the construction was not as rigid and tension in the assembly continued to turn the variable capacitor when autotuning stoped. By turning backwards for a partial step (or more) the tension is released.

For each program available here there are five sets of trials taken consecutively for the 40 and then the 20 meter bands, using a small magnetic loop antenna similar to the one shown on the referring web page. The transmitter loop is made of high quality RG8/U coax, with a circumference of 3.9 meters. This is similar to the small magnetic loop antenna tested in the TCA article. The antenna's variable capacitor is 260 pF which is twice the optimum size for the intended frequency range (doubling the change in capacitance per turning increment presents a tuning challenge). Trial power was limited to nominally 2.5 watts to protect the FT-817 transceiver from overheating due to repeated, protracted CW transmissions. Even so, there was an indication of overheating during the especially long tuning transmission by program 5c (one-half step for reconnaissance and one eighth steps for active tuning).

The SWR recorded for each trial is taken from an MFJ-813 QRP SWR Wattmeter that I use in the field. The meter's SWR reading is consistent with the FT-817's SWR bar graph. However, I believe the measurement of SWR is low, but not wildly low.

Each trial has from one to four active tuning sweeps needed to achieve satisfactory tuning. For most trials the number is 'one'. But it may be as many as as four before the autotuner gives up. In my experience the inability to tune means something is faulty in the antenna set up.

Most of the "Percentage of Highest Signal" tuning trial results are greater than 90%. A few are in the high eighty's. Even a percentage as low as 80% is only 1 dB below maximum possible radiated power. The array of trial results here (and in the TCA article) are typical of the autotuner's performance.

The computer tuning report shows the history of the tuning process for an individual trial. There is a sample tuning report here for each of the three programs. A tuning report presents useful diagnostic information. For example, the report provides data about the trials upon which the percentage of maximum power is calculated. To obtain a report a computer must be attached to the autotuner via a serial to USB PICAXE adapter cable. For somewhat remote operation I've extended the length of the PICAXE adapter cable using a 1/8" plug-jack stereo extension cable up to 10 meters long.

After tuning is completed the tuning report continues to measure radiated power for a few seconds to determine if there is any lingering torque turning the variable capacitor. Even if the mechanism is stable the post-tuning measurement sometimes "jiggles" around by several percent. I cannot explain this other than to think the transmitter's CW output is varying a relatively small amount, perhaps as much as 0.5 dB. This may be a characteristic of the FT-817 when transmitting a lengthy, continuous CW signal. I am planning to look into this some more.

The autotuner is content to operate without a computer attached. In the field a computer would not normally be used, and the small magnetic loop antenna can be set up anywhere a coax transmission cable and 12 volt aututuner power line can reach. (The autotuner is powered only when tuning.)

- Some browsers will automatically add the extension 'txt' when downloading the 'BAS' program file. Please rename the downloaded file to remove the 'txt'.

- The PDF files are not downloaded but displayed, where they may be printed.

The reconnaissance and active tuning sweeps run at stepper motor quarter steps. This is a resolution of 1,600 (partial) steps per 360 deg. turn (full circle). The quarter step resolution of the active tuning sweep(s) is the same. There are no final backward quarter steps.
This program version has the same step specification as above. The difference is backward (partial) steps occur at the conclusion of tuning. '5b' may be appropriate where the construction of the autotuner is not very rigid allowing a pent-up twist to develop between the stepper motor and variable capacitor - causing continued movement when tuning should be completed. By turning backward after tuning is finished, built-up torsion is relaxed.
The reconnaissance sweep occurs at half-stepping, and the active tuning sweep at eighth stepping which provides a tuning resolution of 3,200 eighth steps per 360 degree turn of the variable capacitor. Half-stepping for the reconnaissance sweep speeds up the entire tuning porcess, but it is still slow.

(* To view articles click on link)
The TCA Autotuner Article

This is the final draft article submitted to the Radio Amateurs of Canada for publication in The Canadian Amateur (TCA) of May-June 2018. Pictures, figures and tables are at the end of the draft. The version in the magazine is edited and composed in an appropriate format with pictures, figures, and tables included with the text. RAC members can download the magazine version from the RAC website.
Things you need to be a PICAXE programmer

PICAXE Programming Items.pdf
There are four things you need to obtain to be a PICAXE programmer:
- PICAXE microntroller chip. The Autotuner project uses an 18M2+
- AXE027 programming cable that plugs into a computer's USB port
- Programming software suitable for your computer
- A set of PICAXE programming manuals.
Estimating the Full Transmit Power Your Autotuner Can Withstand

The Autotuner receiver contains 1N34A diodes and a 4N35 optocoupler that have limits on the amount of current they can carry before being damaged. Current to these components is determined by the configuration of the sampling antenna and by the amount of power being radiated by the small magnetic loop antenna. During tuning, when power is low, overloading is not an issue. However, current continues to flow to the receiver during normal transmission when power to the antenna may be 100 watts or more. There is a simple way to find out how much full power operation the autotuner diodes and optocoupler can safely tolerate.

Note, this is different than the variable capacitor's high voltage limitation on the power that a small magnetic loop antenna can handle.
WSPR Trials of the Autotuned Small Magnetic Loop Antenna
Using a QRP TX Shield for WSPR on 20 Meters from TAPR and a Raspberry Pi

WSPR, 'Weak Signal Propagation Reporter' is a method for demonstrating the effectiveness of an antenna and propagation conditions. Reporting is done by means of on-line WSPRNet which collects, catalogues, and displays in real-time the continuous reports of amateur and SWL, WSPR signal monitoring stations world wide.

I conducted WSPR trials of the autotuned small magnetic loop antenna using a 'QRP TX Shield for WSPR on 20 Meters' module (commonly referred to as QRPi Shield) purchased from TAPR for $29US before shipping. The QRPi Shield plugs into a Raspberry Pi computer to comprise a 100 Mw, automatic WSPR transmitter. The antenna was set up in an open field where Internet WiFi was not available to deliver NTP server connectivity. Question: without NTP for timing regulation would the accuracy of the Pi's on-board clock and transmitting frequency be up to the WSPR standard? Another unknown, would a 100 Mw signal into the antenna be large enough for the aututuner to detect? The answers: Yes and Yes.