L.F. On-the-air

With a transmitter completed and your HF rig, or a converter for receive, you'll need to match the system into an antenna.
With UK amateurs having had access to 136kHz for several years,  a lot of experience has been gained matching transmitters to a variety of "LF" antennas.

Throughout this period small and large antennas have been used by the UK LF operators.
The most popular and most effective has been the Marconi T where a vertical wire is top loaded with a large (or small) top capacity hat. Vertical heights from 90 feet down to 25 feet have been used with success.   A variation on this theme is for an existing top band or 80m dipole to have the two elements and the coax strapped forming a Marconi T.
Also popular are the inverted L and single / multiturn loops.
During my time on LF I've tried a small 27 foot vertical from home and  made many QSOs from the Whitton ARG club station using the club's G0MIN callsign. The radio club has a  234 feet sloping long wire attached to the top of the adjacent church at 90 feet AGL. As much wire as possible is a good rule for LF, but an attempt with an indoor vertical and 50 Watts made it from my home in London over 150km to G4GVC in Leicester. I finally gave up this idea after seeing a high voltage arc pass straight through the 1kV antenna insulation and then through a ceramic wall tile.  20kV+  is not unusual on short LF antennas.

Matching Antennas.

Any practical antenna is going to be electrically short at 136kHz. With a wavelength of 2200m a quarter wave is 550m or 1700 feet long!  A typical wire antenna can be compared to a short mobile whip on HF, and matching techniques are similar - but on a grand scale!

You can resonate any "short" length of wire using a series loading coil. The feedpoint impedance at resonance will depend on your earth system, the length of the antenna and the "Quality" of the inductor. My 27 foot vertical has a feed point impedance of 140 Ohms, while  the G0MIN antenna is 68 Ohms at resonance and can be driven directly from the amplifier described earlier. The relatively high impedances are due primarily to earth losses. The antennas alone have radiation resistances of less than 1 Ohm- Efficiency is low!  Where an antenna system has a feed impedance that is much greater than 50 Ohms, or perhaps I should say, where the mismatch between the antenna and the amplifier is significant, then you can use a transformer to match between them. A suitable transformer can be made using an appropriately  large 3C85 toroid. I've used the same type as used in the output stage of the amplifier. Dave, G3YXM, has used these toroids up to 800 Watts. Another technique is to use a loading coil to resonate the antenna but to achieve an impedance match by tapping the power into the coil a few turns up from the earthed end.

The most difficult component to construct  is the loading coil which is used to bring the antenna to resonance. Ideally this should be robust, using a low resistance, large cross section wire - 2.5mm2 is  reasonable  for a high performance coil.  Ideally the coil should be variable to allow perfect matching. You can use a series of fine taps, or use taps for coarse adjustment in combination with a variometer - A smaller coil wired in series that can be rotated inside the larger one. You will need between 1 and 5 millihenrys for a typical antenna! These coils are big.....15 inches / 400mm diameter is not unusual. This size also fits well under a upturned plastic "dustbin". A possible alternative is to use a small coil with ferrite loading. Several portable radio antenna ferrite rods can work - but you'll need to locate the right type of ferrite. The inductance is varied by moving the ferrite in and out of the coil.

The  Whitton Amateur Radio Group ,G0MIN, uses an unusual design of two basket weave coils that can be moved relative to each other. The photo below shows the current version of this coil. The lower coil is fixed while the upper coil can be moved by turning the handle at  the top.
Inductance can be varied from 500 to 2500 micro Henrys. During a test with DJ8WL using slow CW, we noticed significant heating in the original 1.5mm windings.  The coil was then rebuilt  using 54 conductor Litz. The wire has three bundles of 18 conductors. The overall diameter of the 54 conductors is 5mm.
The loading coil was constructed by WARG club member Otto -  G0RKA.

LOADING COILS - 36k bytes

Also shown on the left of the photo is a conventional loading coil used for 136 and 73kHz. Each of the three sections contain 100 metres of 1mm2.  The wire came from a roll of "Twin and Earth" used in house wiring. The red and black were used in the coil, the bare copper earth wire is buried under the garden. Another 100m (Blue) was added later. LF is great business for wire suppliers. Maximum inductance is about 8mH.
In an attempt to measure the Q of the two coils,  we  borrowed  an old Advance Q meter from a local college. Unfortunately, we did not have a manual for calibrating the instrument, however we did manage to approximate a measurement of Q.    For the conventional coil, the Q was measured as 140.  A Q of  233 was obtained for the for the basket weave Litz coil. Although we can not say that these figures are totally accurate, it is safe to say that the Litz coil significantly out performs the other (by 68%)

The wooden former used for the conventional coil is not the best choice for a wet British Winter!  The small grey box at the base of the coil contains a 3C85 toroidal conventional transformer with a tapped secondary. This allows matching from 50 Ohms to a range of impedances from 50 to 400 Ohms.

In a further attempt to increase performance Otto constructed two coils from 10 strand Litz. One is 8 inches in diameter one is 6 inches. The two are mounted horizontally with the larger fixed and the smaller connected to a nylon threaded rod. This arrangement gives a maximum inductance of 10mH and a Q measured at a maximum of 400. We've noticed that the Q is maximum when the coils are not interacting. At full "mesh" the Q measured fell significantly.  This suggests that for maximum performance, any variometer type adjustment should only contribute a small percentage when compared to the major part of the inductor.
Loading coil

                                                                                    Photo  G4VBH
Amplifier limits + Protection

There have been many attempts to reach the maximum power output of the amplifier described in the LF Source book. With 4 x IRFP450s fitted, the total dissipation is 760 Watts. An initial thought is that if an amplifier in class B is 60 - 65% efficient then it should be possible, with a Pd of 760 Watts, to achieve an output power of 800 - 900W.   In fact 900W has been seen coming out of one of these amps, but...."briefly".   The reality is that 760W Pd is at 25 degrees C.    A typical FET will rise in temperature at a rate of 0.6 degrees for each Watt of power dissipated. So if your FETs are running 100W of dissipation each, then the devices will increase in temperature by 0.6  x  100. 
i.e.  ambient (say 25) plus 60 degrees = 85.

For the 60 degrees increase, the devices have to be derated. For a good FET a typical figure for derating is 1.5W for each degree.  So for 60 degrees, you derate the 760W   by: 
                      60 degrees x 1.5Watts x 4 devices.   = 90 x 4 = 360W.

So you see that your 760 W  dissipation amp is now actually 760 - 360 =400 W.
Even if you ignore all the other losses, between the transistor and the heatsink and the heatsink to air, the amp has a maximum practical limit of 400W diss' which at 60 - 65% gives an output of about 500 W   AT VERY BEST!


All amplifiers should be followed by a Low pass or band pass filter. Personally, I use a LPF based on G3YXM'S design using 2x  T-200-2 powdered iron toroids with capacitors to ground on the input, output and the junction of the two inductors.  Even a good linear will have harmonics at -20dB,and -12dB is not unusual.  A LPF will stop the harmonic energy being reflected from your narrow band antenna back to the amplifier. The gain or gm of the FETS is very high and if the amplifier is to be unconditionally stable then it should be terminated with a suitable load at all frequencies. If the amp is used "as is" then harmonics have no load and will cause instability or oscillation. You will not see this problem when testing into a dummy load.  Another "fix" is to install a capacitor from the junction of the drains to ground.  One 4n7  or 10nF on each side of the amp will solve it. This will of course reduce the HF response of the amplifier to something much lower than 900kHz.

If you need more power then go to class D or E with square waves. Class D amplifiers switch the FETs on and off. Efficiency is high but depends on the devices output capacitance. Class E amplifiers contain components to compensate for this capacitance and have even higher efficiency. 75-92% has been measured for D/E operation.
Contrary to popular belief, switching type amplifiers can be used for linear amplification. It's not easy, but amplitude variations can be digitally modulated onto the supply Voltage, while variations in phase can be modulated onto the drive signal. Examples of high efficiency linears can be found on some of the satellites operated by AMSAT.     HELAPS technology (High Efficiency Linear Amplification by Parametric Synthesis) is used to efficiently convert the limited energy from solar panels into energy radiated down to Earth as SSB or other form of "linear" signal.

High power?  Take a look at G3YXM's site for a class D circuit - but forget " A linear".










Friday night at Whitton ARG.                                              Photo   G4VBH

G0MRF tuning for resonance on 136k as visiting A45ZN looks on. Equipment from far end of table; Coil! 250 Watt amp and driver with keyer and LF synthesizer on top. IC756 in foreground with RX converter behind. At rear of table; Avometer and 5A RF ammeter. Finally the SMPSU on the chair.

Protecting your Amplifier

Setting up an LF station can be hard on amplifiers. High voltage arcs, accidental antenna shorts, and bandwidths that can put you way off-tune 500Hz down the band all cause major problems for your P.A.
One year on and my prototype amp is still on its original FETs. I believe the reason for this is that I always use a post amp LPF and a directly fed series resonant antenna.  The Low pass filter will prevent harmonics reflecting from the antenna back into the PA stage. Also a capacitive input LPF will present a low impedance to any high frequencies present in the amplifier's output. This effectively damps the stage at HF preventing high voltage spikes and enhancing stability.
By avoiding coupling loops in the antenna feed my antenna reaches its lowest impedance at resonance. If the wire shorts or arcs or even falls to the ground, then the impedance seen by the amplifier increases. At worst, this causes 1 or 200V to appear in the circuit. This is no-problem as the FETs are rated at 500V. What is important is that the current will reduce to a safe level saving the FETs from excess current or excess power dissipation.

              ============                 =============                ============

The WARG operation on 136kHz is limited to a just a few hours on a Friday evening - This is club night!. Although operating time is restricted, G0MIN has worked radio amateurs in  7 Countries:
England, Wales, Scotland, Ireland, Belgium Germany and Finland.

"Best DX" is a QSO between G0MIN and OH1TN on 1st Jan /99. The distance was 1800km with 569 reports both ways. OH1TN used 100 Watts to a 500m wire antenna.

The "DX record" for the 136kHz band is a QSO between G3AQC and VA3LK

Another event of note was a  crossband QSO between G3WSC and UB5WF. This contact took place in daylight in April 2000. UB5WF was operating on 7011kHz while G3WSC was on 137.250kHz. The distance was 2225km. G3WSC was running 50 Watts ERP from a 100m decommissioned Decca mast in Hertfordshire in Southern England.  This one weekend with a high power variation to the licence also provided the first crossband contacts into Jersey and Guernsey.
In September 2000 the first amateur signal from Europe was seen in Canada and resulted in a crossband QSO between G0MRF/P  (136kHz) and VE1ZJ (14 MHz) over 4332km

The 73kHz band is allocated until June 30th 2003. Access to the band is via a notice of variation to the UK class A licence. No portable operation is permitted. With the subsequent release of 136kHz no new N.O.V.'s are being issued.

Worked two way or cross band on 73kHz are:

G3YXM            G3KAU          G3LDO           G3XTZ          G0ONA        MI0AYZ        G3AQC
G4JNT            G3GRO
           G3WSC          G3XDV          M0BMU         G4GVC       GM3YXM
Heard but not worked are  G2AJV and G0AKN

A record for a two way QSO on 73kHz was set by Peter, G3LDO, and GD0MRF on 21/Nov/99 at 478km using QRSS. 

This 478km was later extended with a QSO between Peter and MI0AYZ
The distance was 570km. Ian MI0AYZ was operating from a decommissioned Decca mast in Northern Ireland. Both used CW

A sunny weekend in March 2003 saw 3 groups operating from temporary locations around the UK.
Dave G3YXM was operating from New Galloway in Scotland. Derek G3GRO and Jim M0BMU had driven to Porthcurno near lands end in Cornwall. Meanwhile G0MRF had driven to Ochiltree Ayrshire Scotland, to the QTH of Simon Lewis GM4PLM.

On the 31st of March 2003, Peter Dodd worked GM0MRF extending his own record for a 2 way QSO by 9km to 580km. The QSO was with GM0MRF located at the home of Simon Lewis GM4PLM which is 15km East of Ayr.  The next day the 'LF Dx-Pedition' mounted by Derek G3GRO and Jim M0BMU worked GM0MRF over a mainly sea path of 609km. Both stations had problems with high local noise levels and David and Simon transmitted for nearly 6 hours before the noise dropped and propagation peaked around sunset just long enough to make the QRSS contact.

Although 73kHz is unique to the UK, hard work and good planning between Peter, G3LDO, and Reino, OH1TN, finally resulted in a two way cross band QSO  73 / 136  using QRSS. The distance was around 1900km.

Countries with a 136kHz allocation include:

In the USA the FCC has issued experimental permits to 12  amateurs. AMRAD have an experimental beacon operating on 136.750 running 100W from Virginia.
Canada has also issued experimental permits to VE3LK and VE3OT. The first Canadian 2 way QSO was made on 22/July/2000 over a distance of 431km / 268 miles
ZL and VK amateurs are able to operate for limited periods with 'experimental licences'


Home                                                                                                     Amateur Radio Links