MF Operation at DF0WD

by Wolfgang Buescher, DL4YHF

last updated: January 2013.


Initial Reception Tests (at DF0WD)
Transmission tests and antenna matching
Using an Icom IC706 as 'exciter' on medium wave
A small 'Linear' power amplifier for MF
Initial activity on MF (starting in June 2012)
2014-03: Various reception tests, VO1NA in slow CW (QRSS)
Links (medium wave related)

Initial reception tests

During initial reception tests on the UK's former 500 kHz band, it became obvious that the 'station ground' (connected to mains ground) carried a lot of noise, most likely from switching mode power supplies, compact fluorescent lamps, and similar. The receiver made a buzzing sound, and weak amateur radio signals were completely masked.
The following RX antenna configuration gave a significant improvement:

An important feature of this LC parallel configuration is its ability to act as a transformer. The antenna, and its counterpoise ("Earth") are completely isolated from the noisy "mains" ground.
The first coil was wound on a plastic tube, 15 cm diameter, 20 cm length, using 60 turns of 1.8 mm enamelled copper wire. The spacing between two turns is ensured by winding a thin plastic rope (greenish colour in the photo below).

(antenna loading coil, still without variometer for 'receive-only')

Primary coils (60 turns) and secondary coil (2..3 turns, with red plastic coated wire) are separated by about 1 cm. For simplicity, foam plastic from an iso-mattress (sic!) was used.
The spacing between primary windings is important to achieve the largest possible Q factor, because this coil will be used for a transmit antenna one fine day .. when German radio amateurs may operate on MF, which was not the case in April 2012. For reception, a much smaller coil (thinner wire) will do the job just as well.

The original 'receive-only' setup consisted of a short wire, which was connected to the topmost tap of the primary coil (largest impedance). To connect longer wires without degrading the 'Q' (which helps to keep the strong MF broadcast signals away from the receiver), different taps were made on the primary coil. They will later be used to find the best impedance match for a longer transmit antenna.

(closeup of the antenna loading coil with multiple taps)

With the above receive-setup, crossband contacts with various stations from the UK and Ireland on the 'old' band around 500 kHz were made.

Transmission tests and antenna matching

In June 2012, German radio amateurs were positively surprised by the authorities (Bundesnetzagentur, BNetzA) when the frequency range 472 to 479 kHz was allocated on a secondary, non-interference basis with a maximum ERP of 1 watt. On this occasion, a big thank you to everyone involved !
The loading coil shown above was turned into a variometer (to make it suitable for transmission with a low power level), and a few homebrew accessories were added:

(very first transmit matching test, with too many alligator clips)

In the foreground, from left to right:
SWR meter (measures forward and reflected power) combined with an impedance matching unit (RF transformer with various output taps, from 12.5 to 50 Ohms);
RF current meter ("feed through"-type with another ferrite toroid acting as current transformer).

The variometer consists of the old loading coil wound on a plastic tube with 15 cm OD, and a smaller inner tube with 11 cm diameter. The wire for the inner (rotatable) coil is old 'Tensolite' wire (most likely Teflon covered, silver plated Litz wire).
The outer coil is tapped to achieve resonance with the inner coil rotated for maximum inductance (about 350 uH in this case), which ensures the lowest loss in the coil.

(photo with antenna matching accessories)

The loading coil compensates the capacitive reactive part of the antenna impedance (-j * 1 kOhm here), by connecting +j * 1 kOhm in series. The result is purely resistive, and consists of the ground loss, environmental loss (trees), coil loss, and a tiny bit of radiation resistance (in this case, way below 1 Ohm). The output tap on the 'impedance transformer' (integrated in the SWR meter housing) was set for 26 Ohms to achieve the lowest possible SWR, thus with a 30 watt transmitter it should be possible to push 1 Ampere into the antenna wire. It was.

(schematic of antenna matching transformer and loading coil)

The transmitter used for the initial tests was built in a rush (before discovering that the old IC706 can be used as an exciter for 472 kHz):

(photo of the 472 kHz transmitter prototype, with variable crystal oscillator)

An Icom IC706 as 'exciter' (in german: "Steuersender") on MF

By accident, it was found that the author's old IC706 would transmit around 472 kHz. It's not sure which of the modifications in this radio's former life did the trick, but quite certain it is just a matter of configuration (these "modifications" used to be on or other sites).
When trying this on an IC706 or similar shortwave transceiver, be sure to use the lowest power setting. A radio designed to operate between 1.8 and 29 MHz will not have the necessary 'large inductors' in the final and the driver stage ! Trying to transmit with such a radio outside the specified range will void your device's warranty, and of course you will have to try this at your own risk and expense !
At the lowest power setting ("L" in the IC706's "Q1 RF POWER" menu), the radio produced about 2 watts radio frequency on 475 kHz on a 50 Ohm dummy load. But the slighly clipped waveform indicates that the rig doesn't really like to transmit on this frequency:

(waveform of IC706 loaded with 50 Ohm on MF)

Anyway, the external MF "power" amplifier will remove harmonics, so this less-than-ideal waveform is not a problem. 10 Vrms on 50 Ohms is approximately 2 Watts of RF, but voltage and 50 Ohm impedance are too large to drive the power MOSFET's gates directly.
Thus low-impedance matching is required to drive the MOSFETs:

("linear" MOSFET gate driving circuit, adapted for 2 watts from IC706)

The complementary push/pull driver was left unchanged. Note that the 22 nF capacitors parallel to the MOSFET gates are not a typo: Together with the rather low inductivity of the step-down transformer (trifillar wound ferrite toroid), they form a resonant tank (with low Q) near 475 kHz.
With 2 watts of driving power, and a sufficiently dimensioned step-down transformer (from 50 to a few ohms), it may be possible to drive the MOSFET gates directly.
Due to the "linear" operation, the efficiency of this PA is not spectacular (compared to a switching mode PA with rectangular drive) but who cares if the required RF output power (to reach approximately 1 Watt ERP) is only 50 watts.
The MOSFETs are HUF75343, slightly oversized (75 A max. drain current, 55 V max drain voltage, up to 270 W power dissipation but that's not realistic for a TO-220 housing without liquid nitrogen cooling). But the decision was easy because from an older project (LF PA), a bunch of transistors was available, and some safety margin helps in cases of maltreatment (bad SWR, transmitting with no antenna connected at all, etc..).

("linear" MOSFET output circuit)

The output power could be increased with more secondary turns on the output transformer, if a larger ferrite toroid was used (in the author's prototype, an FT114A-61 was used because nothing else was found in the junk box). Through the secondary taps, the output power can be selected without sacrificing the PA's efficiency. When tested with 14 turns secondary, the PA delivered 30 watts RF, and consumed 3 amperes DC input current. This was at the amplifier's clipping point, i.e. more input drive didn't significantly increase the output power anymore. For CW, this is acceptable, and SSB isn't an option on MF (even though it would be technically possible with this amplifier, when "moderately driven").

An RF vox circuit, a timer, and two relays for RX / TX switching were added to the PA board to complete the 'MF station':

(IC706 driving the experimental PA; with automatic RX / TX switching)

With a 'linear' power amplifier (in addition to a 'linear' exciter, such at the IC-706 in this case) all kinds of soundcard-based modes are possible, including those with a non-constant RF envelope.
Here for example, a Chirped Hell (aka Fourier Hell) transmission by DF0WD, with approx. 20 Watts transmitter output power, received at ON7YD. You can see some unwanted sidebands near the figure '3' and the letter 'G', when the transmitter was driven into compression:

(Chirped Hell transmission "73 GN" by DF0WD, received at ON7YD's MF grabber)

Sidenote: Similar effects can be seen on shortwave, when operators overdrive their transmitters in PSK31, trying to squeeze 100 watts average power out of a 100-watt radio ;o)

Next plan: Install a preselector for reception on the PA board, since there is already a 'receive-only' path on it (between the two relays).

Initial activity on MF (starting in June 2012)

With the IC706 (HF transceiver), using 30 .. 40 W from the amplifier shown above, and 1.2 A antenna current, reception was reported through the WSPR network from a number of receivers in western europe (DL, F, I, PA, ON, G, GM, EI). Even during daylight conditions, the MF signal with an estimated ERP of 500 mW (which was later found to be much less (*) ) was copied at GM4SLV in Shetland, over a 1000-kilometer distance.
Contacts in normal CW in the late summertime evenings were difficult due to QRN but possible - the first successful two-way QSO from DF0WD was DL2HRE.
Nice CW QSOs with 'armchair'-copy on both sides were made with DJ9IE and DK8KW. These stations are well inside groundwave range, and there was no QSB (fading) at all.
After midnight (in mid summer), the band "opens up wide", and with some luck the QRN goes down significantly. Under such conditions, two-way contacts on MF in normal CW with 'armchair copy' were possible over several hundred kilometers distance (like DF6NM and DK7FC). A first 'highlight' was EI0CF - thanks Finbar ! - crossing the 1000-km limit with an estimated ERP of 0.8 Watts (*) on this side.

(*) ERP (Effective Radiated Power) much lower than expected ?
Comparing the ERP with other active stations on MF, and by judgment from 'calibrated' receivers (and their estimation of DF0WD's ERP on MF), it was later found that the trees surrounding the antenna had quite an impact on the 'true' radiated power. Moderately increasing the TX power from 30..40 Watts to something around 100 W seemed easier than improving the antenna efficiency, or felling a couple of trees around the club station ;-)
... using a high-efficiency linear PA with switching mode modulator - details further below)

In winter time, the ionospheric absorption drops faster, and 'DX' signals appear from the east before sunset. With a moderate ERP, and place not plagued by local QRM, contacts were made in relatively slow CW (but not QRSS) with Malta (9H) and Romania (YO).

In January 2013, radio amateurs in other European countries got access to the 630 meter band, like the UK (requires an NoV) and the Netherlands. This will hopefully boost activity beyond 2013-01-01, when the first stations in PA, and one hour later G, GW, GM, and EI got 'on air' just a few minutes after their local midnight.

2014: Improved linear amplifier, with improved efficiency

When it became obvious that the ERP was 'much lower than expected', and some stations had difficulties to copy DF0WD on MF, a new final amplifier was built. To use the existing '13.8 V, 20 A' power supply, a more efficient design was required. The choice was that, instead of operating the MOSFETs in the 'linear' region (which resulted in a poor efficiency), the RF power output stage should run in switching mode, and produce a maximum power of approx. 150 Watts; ideally from a DC supply of not more than 15 A (at 13.8 V) to have some margin for the IC-706 (exciter) from the station's '20 Ampere' power supply. With a DC input power of approx. 200 W, and an RF output of 150 W, the design goal was "at least 75 %".
To keep the CW modulation sidebands as low as possible, and to be prepared for modes which require envelop shaping (like PSK31), the new power amplifier also had to be 'linear'.
As a proof of concept, the switching-mode 'linear' PA was built from various 'modules' on a large copper-clad board. The entire board is at ground level, and there are no traces etched (or cut) into it. All modules (like sync rectifier, auxiliary negative voltage supply, PWM, SWR sense, current sense, protection) are built on smaller, individual PCBs.. in fact, the PCBs are neither 'printed' nor etched, but cut with a sharp tool.

Photograph of the PA in an 'early stage', running in a two-hour stress test (during which much water was boiled to cool the dummy load resistor on the left side) :

Click on the image for a full-screen view
( isn't it strange how fast a workbench can turn into a mess, with cut-off wires floating all over the place ? :o)

Details about the new PA (with switching-mode 'power modulator', put into service in January 2014) are in another document; see 'Switch-Mode Linear'.

2014-03: Various reception tests, VO1NA in slow CW (QRSS)

VO1NA received via DF0WD on 477.7 kHz in QRSS10, 2014-03-22 .
Times are in UTC. Local sunrise (in Germany, RX side) at 05:22 .
Click on the image for a full-screen view

... to be continued ...

Links (medium wave related)

DX cluster with filter for the 472 kHz band (OH8X DX Summit)
DM4TR MF Grabber: 472-479 kHz live spectrogram by Thomas, DM4TR, in JO61DE
DK7FC MF Grabber: by Stefan, DK7FC, in JN49IK
Grabbers in Birmingham: A long-lasting service for LF and MF by Dave, G3YXM
TF3HZ live spectrograms (covers a part of the 630 meter band)
The Shetland Grabber: 472-479 kHz live spectrogram by John, GM4SLV (temporarily offline?)
WSPR Spot Database for MF and LF ("old" interface but imo easier to use)
The RSGB LF Group (at Yahoo), also used by MF operators
GW3UEP's site devoted to 500 kHz CW with a variety of homebrew CW transmitters
Operating Portable on Medium Wave by Finbar, EI0CF - Amateur Radio at it's best !
Homepage of the Montenegro LF / MF Group(seems to have disappeared; or redirects to a far-east provider)
G4WGT Multi-Grabber page: now includes the new 630 meter band.
Temporary 500 kHz (or 475 kHz) grabber by Rik, ON7YD
WebSDR at the Universtiy of Twente; with gapless coverage from 0 to 29 MHz .

back to DL4YHF's main ham radio homepage