Hits:  since 20050614

The 145 MHz repeater's antenna was radiating enough energy to make the PIM-causing metallic junction to go unlinear in the mast structure, probably located in one of the antenna mounting clamps (Aluminium tubing and S.S. or galv. steel clamps and always on mid winter cold days). The PIM QRG moved with QRGs of 28, 50 and 432 MHz beacons, but of those, the 28 MHz beacon's harmonic, which was one of the factors in the formula of the PIM QRG, was the second highest harmonic (4th) and thus, a smaller QRG change was required to move the PIM QRG down and outside 145 MHz repeater FM RX IF filter's bandpass by about 10 kHz, not to mention 28 MHz beacon band was the easiest to re-allocate a new QRG compared to 6 m.  Fpim[MHz] = (4*28.2315 MHz) - (8*50 MHz) + 432.435 MHz.. When the new 28 MHz beacon QRG was implemented, the PIM resides now outside repeater's input bandwidth, on 145.086 MHz  +- 1 kHz (wobbled by beacons' QRG's annual frequency drift).

To cure the problem without creating new problems by too radical changes, the Radio Club of Kouvola, OH5AG kindly bought a new crystal for the 28 MHz beacon's TX for 28.2285 MHz. The PIM  has not been heard since on the repeater's input, after the QRG change to 28.2285 MHz was implemented in Aug. 2008. The antennas of six beacon's and 145 MHz FM repeater are all stacked on the same 6 m  mast.
 
 

Beaming SW towards Central Europe from 65 mAAT

OH5SHF's 28 MHz  beacon's ground wave coverage. Computed signal strengths (S-units) with 6 dBd RX antenna at 25 m AGL, S 1 = 0.2 uV (-121 dBm). Note:  the noise level is higher on 28 MHz and HF receivers are not  typically built to be as sensitive as 50 MHz radios, and unfortunately are most insensitive in particular on 10-meter's band.; expect signal/noise ratios much lower on 28 MHz than on 50 MHz

---

OH5SHF's 50 MHz  beacon's ground wave coverage. Computed signal strengths (S-units) with 8 dBd RX antenna at 25 m AGL, S 1 = 0.2 uV (-121 dBm)

---

"Corrected" Tropospheric Path Loss = TX ERP [in dBm] + RX Ant. Gain [dB]  -  RX feed line loss [in dB, often negligible] - antenna height loss  [in dB, see graph below] -  -145 dBm [CW RX sensitivity level] - band noise above 290 K [in dB, from upper graph per  your location's man-made noise level]

It is expected your RX system Noise Figure is substantially below 3 dB! (RX k curve)

Using the calculated Corrected Tropospheric Path Loss value, you are able to determine the approximate maximum range (CW mode) from upper graph (signal received 50% of time).

Note: the "Corrected" Path loss here is used include effects that reduce RX S/N ratio specially on 50 and 28 MHz due to higher noise levels on those bands to get the true range and not just pure path loss. If you need only the path loss for shorter range to estimate signal level when S/N is not an issue, do not correct for band noise!

Note: Range can be reduced by lower TX and/or RX antenna's height,  terrain and other obstructions and by factors such as poor RX system Noise Figure, higher local noise level which to some degree can be corrected using the graphs.
 
 

The CAD model used 1.4 mm enameled wire as elements and 4 mm brass welding rods as phasing line. Prototype antenna's element wires were stuffed inside four pcs of 3 m long telescopic fiber-glass fishing poles (2.50 € a piece), but in spite of being a prototype, it has survived the elements since September 2003 at the beacon site.

Computed 50 MHz gain was 1.5 dBd and F/B ratio 2.5 dB and on 28 MHz, the gain was 0.1 dBd and F/B- ratio 0.4 dB in free space, however antenna is mounted just 6 m above the roof.

Mechanical construction for such antenna includes insulator-supports for the 200-ohm open-wire phasing line, long aluminum element brackets to alleviate torque of fragile (center reinforced with short fit-inside PVC tube sections) fishing poles and finally, shielding the feed-connections from moisture and making sure the poles can not get filled with water - open ends - waterhole at center. The feed uses a ferite sleeve balun, many small toroids over RG-58 inside a heat shrinking tube installed inside the boom. This would call for a DC-shorting arrangement (perhaps a shorted "quarter-wave" stub after the ferrite sleeve, 1.24 m of RG-58 - but I have not verified how it affects SWR) so I recommed using a proper balun-transformer that works for 28 and 50 MHz both, instead. Don't forget to connect the feed line's shield to the boom, or mast for grounding.

The excess voids on fiber-glass poles' tips were cut off to reduce overall antenna size, wind and snow loads. The elements could also be made of aluminum tubing - then the element lengths should be shortened a little. Making it even more light and small it could also be mounted pointing upwards (at 90 degree tilt angle) by bolting the element brackets directly to the side of a mast (with no boom) - price: about 2 dB of ERP lost towards horizon on 50 MHz and 95 cm reduced height for either of  the elements.

Unless traps and loading coils are use, this may be the smallest "full-size" antenna that works both on 28 and 50 MHz since only 8.6 m (incl. boom) of tubing material is used. In comparison, a GP with trap and proper (3) radials for both bands needs some 14...15 m of element material and a GP can only deliver vertical polarization. This dual-band antenna can be used as vertically, or horizontally polarized. If set for vertical polarization, it is practically omnidirectional. Two such crossed antennas could be phased for circular polarisation. The same basic LPDA-style idea could be adopted for a set of any multiband HF, VHF or even UHF ham beacons fed with just single coaxial.
 
 

Few things to pay attention to: the stub on 28 MHz TX branch is mostly to keep 50 MHz matching good, 28 MHz TX has always LPF, which attenuates maybe 20 dB the 50 MHz signal, thus reducing transmitter created IMD products - one stub should be enough. However, typical  50 MHz  TX's LPF does not attenuate 28 MHz TX power and broad band power amplifier output transformers pass 28 MHz all the way in to the semiconductors and the 50-28=22 MHz transmitter created IMD product gets also passed out un-attenuated. The sum product at 78 MHz is not attenuated by the 50 MHz LPF very much either. The coupler has two stubs to attenuate 28 MHz signal by 66 dB from entering the 50 MHz TX. Quarter wavelength jumpers and stubs could be added on branches, if seen necessary to further increase attenuation, but even one stub / branch could be enough.

The construction can be made with T- connectors of preference (N, Amphenol, TNC, BNC....) or from  plumbing parts, but they may be as expensive as coaxial adapters and need a lot of soldering work.
Neither can any coaxial sections be changed with ease.

It is worth while to note that higher impedance cable has less capacitance and should be used in the stubs, even 150 ohm cable, if available.
The coupler I built, uses RG-62 in the stubs and RG-213 on 1/4 wave length jumpers, RG-58 (note the cable power limits) could be used as 28 MHz jumper, but not preferred on 50 MHz branch due to higher losses.
It you have T&M gear, or antenna analyzer, use it in cutting the stubs and jumpers, though for known cable types, simple math (((300 / MHz) / 4) * Vf) and metric tape scale works too.
It would have been too easy, if no extra trick would not be required: the 28 MHz port impedance is too low (because of the adjacent 1/4 wave length stub for 50 MHz), so the 28 MHz port needs LC- match to make SWR and losses good, 50 MHz port is fair with out any additional match network.

The values given are with 50 ohm dummy load as antenna. If the antenna has mismatch, it will reflect on TX port(s) SWR and a little bit on losses. The 28 MHz port LC- match gives some possibilities to compensate for antenna SWR by compressing - stretching the coil. Another form of match could be used if one likes to use variable capacitor there, from 28 MHz TX side: nominally 87 pF to ground and 7 turn coil in series (untested).

 The same scheme works scaled to other frequency pairs, if they are separated proportionally by the same amount. Quarter wave open-end stubs behave as "shorts" also on uneven harmonic frequencies, so it may not always be possible to use the stub coupler and if the difference in frequency is big enough, T- structure LC high pass - low pass filter is a working method. Too closely spaced frequencies call for cavities and anti resonance reactive components, as used in repeater duplexers. There are also compact commercial TX couplers manufactured for VHF and up.

Initial data from 1st summer Es season wth FET PAs comparing to 2013:
- 28 MHz: 6 dB power increase changed the number of DX spots by -50%
- 50 MHz:  6 dB power increase changed the number of DX spots by +18%

OH5RBG's 70 MHz  beacon's ground wave coverage. Computed signal strengths (S-units) with 8.5 dBd RX antenna at 25 m above average ground level with unobstructed take-off, S 1 = 0.2 uV (-121 dBm) @ RX input.
 

The 70 MHz beacon was built during fall of 2009 with latest technology: USB-controlled synthesizer, PIC-based PC-configurable keyer and MMIC driving the FET PA. The basic scheme works for all bands from 28 to 70 MHz with just modifications to driver LPF and PA stage coils and capacitors - the single-ended FET PA is narrow-band type. Using the 175 MHz application sheet for RD06HVF1 FET, the beacon can be built for 144 MHz operation also. 432 MHz model would need different version for the synthesizer chip - the faster LVDS reaching up to 910 MHz.

The first built beacon module was for 40 MHz, but uncertain and troublesome licencing caused a rebuild to 70 MHz (took some 2 h) which has been a ham band in Finland since 4th Nov. 2009. The enclosure holds the 70 MHz beacon TX module with G4JNT keyer and a 4.5 A SMPSU. FET PA feeds some 5 W to a horizontally polarized low-gain (7 dBi) antenna (Aerial AV1462-70) beaming Az. 232° (SW for sporadic-E propagation) mounted on a cellular tower on a hilltop 131 m + 8 m ASL and delivering 15 W of ERP.

In detail, the synthesizer, which is configurable for any QRG from 3.5 to 260 MHz in 1 Hz steps (and for exact FSK shift) via the USB connection, delivers a 0 dBm (square wave) fed via 2- section LPF to ERA-5SM MMIC amplifier. G4JNT RS-232-configurable PIC-keyer keys the SDR-Kits USB-synthesizer. The +16 dBm drive feeds a low-cost SWR tolerant Mitsubishi PA FET delivering 5 W RF-output. The output is filtered with several LPF stages, one for UHF for added attenuation on cellular frequencies (harmonics are attenuated by 119 dBc on 900 MHz UMTS band). Total currrent is monitored with a 2 A analog panel meter. The antenna output circuit has a 230 V gas-disharge tube, 2 kV series capacitor and antenna connector & case bonded to site ground bar.

OH5SHF 2320 MHz  beacon's ground-wave coverage. Computed signal strengths (S-units) with 24 dBd RX antenna at 25 m above average ground level and take-off clear of trees, S 1 = 0.2 uV (-121 dBm) @ RX input (needs a pre-amp.).
 
 

OH5SHF 10 GHz  beacon's ground-wave coverage. Computed signal strengths (S-units) with 30 dBd RX antenna (0.8 m dish) at 23 m above average ground level and take-off clear of trees, S 1 = 0.2 uV (-121 dBm) @ RX input.
 
 

10-2010->4-2014	28 MHz	28.8 MHz	50 MHz	70 MHz	432 MHz	1296 MHz	2320 MHz	5760 MHz (U.C.)	10368 MHz
TX Pout	+48 dBm, 62 W	+33 dBm, 2 W	+49 dBm, 75 W	+37 dBm, 5 W	+40 dBm, 10 W	+40 dBm, 10 W	+29.3 dBm, 0.83 W	+31.2 dBm 1.3 W	+27.4 dBm, 0.55 W
Refl. coeff, VSWR	-14 dB, 1/1.5	pi-match	-26.5 dB, 1/1.1	-24,3 1/1.13	-14.6 dB, 1/1.5	-18.4 dB, 1/1.27	-15.6 dB, 1/1.4	<-16 dB, 1/1.37 or better	N.A., tuned for best ERP
Pant, ERP	45 W, 45 WERP	2 W, 1.5 WERP	46 W, 73 WERP	4 W, 12 WERP	8.4 W,  26.5 WERP	8.3 W, 26 WERP	0.68 W,  30.4 WERP	1.3 W, 44 WERP	0.5 W, 4 WERP

 

Aerial's AV1525, two of such are used in OH2UHF, is basically a directional antennal with gain of 7.3 dBd and high F/B-ratio. We wanted to get more omnidirectional-style coverage but little ERP towards 130..180 degree zone. The solution was to abandon the two-wavelength ominidirectional cavity slot aerial and use a rugged commercially made second-hand (abt. 35 years in-use, but in pristine condition) dipole array and point the main lobe with most ERP towards WSW (SM7-/SM1-/SM0-/SM5- area), where most of the long-haul cluster spots come from and which iluminates the path of Baltic Sea tropo openings to DL and OZ etc. The illumination of southern OH5 and UA1-/ES border is reduced and St. Petersburg is just on the edge of the lobe as are OH6- and SM2-, but that is unavoidable without using more hardware in the form of extra gap-filler aerial.

The omission of the hefty bolt-on refector-grid reduced antenna's wind area (0.45 to 0.2 m²) and decreased the F/B from 20 dB down to 5 dB which gives enough ERP (about 5 W) towards NE sector for adequate illumination of OH7-/OH4- and eastern OH5-. The AV1525 horizontal support boom acts as a long reflector element and produces the 5 dB F/B and gives about 2 dB of more gain than the closely spaced dipoles alone. There are no stations beyond 200 km towards NE, so we do not spend more of the limited ERP for that area. The SWR of AV1525 ("P") is better than 1:1.5 on 432 MHz without the reflector and this antenna is insensitive to sleet & snow. The AV1525 is fed with 12 m of Ecoflex 15 cable.

1296 MHz beacon radiates from the controversial Alford Slot antenna, which delivers horizontal polarization and is practically omnidirectional (F/B: <3 dB), with slot facing towards azimuth 240°. If there is something good in it, this antenna fits to a 50 mm diameter radome where no quad loops or ordinary dipoles fit inside. The cavity slot's radome is now a black PVC tele cable protection pipe with 2.5 mm wall thickness. The antenna is fed with 7 m of Heliax LDF-4-50A through a 1/4-wavelength inside-the-cavity UT141 spliced balun. The SWR was improved during a service break in fall of 2010 by change to a less thick walled radome of same OD and widening the slot to 11 mm, adding two sub-pF flapper capacitors 7 cm away from the feed point for good SWR with radome on and reducing slot's length by 20 mm on both ends. Now the VSWR is 1/1.27 and the antenna appears no longer to be resonating >70 MHz too low. The slot's cylinder is made of rigid copper tube of 1.5 mm wall thickness with 35 mm OD.  The announced gain of Alford long slot antenna is given as  3...7 dBd, with the vertical pattern developing strong side lobes at upper end of operating frequency range and loosing gain. My estimate of OH5SHF's slot antenna's gain is now around 3 to 5 dBd and yes, it would still work slightly better on lower part of the 23 cm band. Must be related to cavity's tube diameter - no gimmick changes that.

The 2320 MHz radome is a 1 m long 110 mm Uponal PP pipe with end-caps and galvanized clamps bolted with stainless steel side-mounting supports. The antenna inside the radome is a re-scaled YU1AW-designed H-polarized sector antenna called Quados 8 originally for WLAN consisting of 8 stacked spaced quad loops fed with open-line. SWR is 1:1.4. The feeder coaxial is 1.3 m of RG-223. The front-side alumininum foil laminated reflector plane consists of  760 mm by 103 mm 12 mm thick plywood film veneer. Adhesives used were Poylurethane glue (Alum. foil laminating on plywood and coax through hole) and hotmelt-glue (element support spacers).

The 2320 MHz beacon's ODU built inside a ABS enclosure, consist of a bias-T with 96 MHz BPF, the *24 multiplier, 3-section 2320 MHz interdigital BPF and a 1 W WiMo ATV PA (modified for CW BCN use) added with home-made 2-section low-pass filter and a harmonic stub. The antenna illuminates a  +-35°-wide (-3 dB points) (+-57° at -10 dB) sector with about 25 W ERP from S to W to NW centered at QTF 240° towards OH2- and further out - propagation condtions permitting. Reception in the back-side of the antenna panel is limited by the low ERP of approximately 1 W to sectors N, E to SE and even less by immediate structural obstacle blocking NW to N sector. Reception is possible at shorter ranges by reflections from nearby structures located in the main-lobe.
 
 

The OH5SHF 10368.935 MHz beacon uses A1A keying with RF-power output of 0.5 W to horizontally polarized omnidirectional waveguide slot antenna yielding 4 W of ERP.

Mechanical difficulties in adjusting and temperature-aging  issues with ordinary XO oscillators caused the need to find an easier-to-adjust oscillator. The Si-570 LVDS 20 ppm version was the only one available, but with serious doubts about the temperature stabilty, while it's QRG is easily adjusted with USB cable and a laptop PC. After aging for 2 months the SDR-Kits unit and increasing Si-570 chip's temperature to +40 °C, the drift from changes of ambient temperature reduced to a fraction of what it was unheated.

The DG8SAQ Si-570- based USB synthesizer generates 108 MHz. This unit is located indoors in a partially insulated enclosure having inside a constant +35 °C(+- 1 °C) temperature for stability, while the Si-570 is heated to +40 °C with QH40A heater chip. This "double ovening" procedure achieves +-2.5 kHz stability at 10 GHz with IDU room temperature varying between +10 to +20 °C. Without heater and oven, the drift would be tens of kHz +- on 10 GHz. The equipment room's ventilation was fitted with a thermostat controlled fan in June 2011 and this should prevent thermal frequency run-away instabilities in the future during extremely hot summmer weeks. The synthesizer's output is low-pass filtered, boosted to +18 dBm with a keyed ERA5SM MMIC amplifier stage, filterd with 108 MHz LPF and BPF. The +14 dBm 108 MHz drive is fed to ODU via 35 m of CATV75 coaxial cable. The keying with message: "OH5SHF  KP30HV and a long carrier is cycled every 1 minute, is generated by a G4JNT RS232- programmable PIC-keyer installed by the Si-570 synthesizer.

The 10 GHz outdoor unit, partially sponsored by OH5AG, includes 108 MHz input BPF, DB6NT's MKU10G beacon unit with multiplier stages from 108 MHz to 10.368 GHz. The 10 GHz 200 mW output from a flanged  MGF1801-special drives RATS sponsored  FLX202MH GaAs-FET power amplifier producing about 0.5 W RF-output to a RATS sponsored 9dBd omni waveguide slot antenna at 144 m ASL with free take-off. Relative RF output power level is indicated with a LED-bar display located on the bottom of outer ODU enclosure. However, the indication depends strongly on ambient temperature since the PA RF couplers's diode is not temperature compensated. The outer polycarbonate enclosure is EMC screened with spray and EMC-gasket material (cavity RF adsorber foam sheet and ferrite sleeves on cables). It has a top-side mounted antenna radome pipe of Polypropylen sealed with rubber O-ring, MS Polymer and G.P. type hot-melt glue, sprayed over with rim silver spray paint for UV protection. While the Polypropylen material is easy to find and has very good electrical properties at 10 GHz, it is difficult to glue. Only adhesives with Polyolefins (w. PP surface yellow flame treatment) really work with polypropylen. The radome also unfortunately increases the 4 dips on the WG antenna's Az-pattern quite severely, which however is unavoidable and typical.

 The PIC keyer is G4JNT design which keys (A1A)  the ERA-5 SM  MMIC in the TX amplifier chain. Mitsubishi RD06HHF1 FET in the PA feeds via 2 dB cable lossses and mid-line LC-match (50 to 71-j20 ohms), some 1.9 W RF to a slanted-dipole with horizontal polarisation at 9 m AGL. The ID "OH5TEN" is send  every 90 seconds. Description of the unit was presented at OH5AG's summer meeting in June 2009 and in July 2009 issue of Radioamatööri-lehti (available at SRAL website in PDF format for SRAL members).

The frequency stability is typically better than 2*10^-12 on one minute basis and better for longer periods of time and can be only affected by temporary LOS of adequate number of GPS satellites, which should not occur, or during short periods of high electron content of the ionopshere - typical during strong auroral displays. The other notable benefit is the absolute correctness of the center frequency, which does not rely on any local frequency source prone to unknown offset.

28.8 MHz is in harmonic relationship with 144, 432, 1296 and 10368 MHz ham bands. Within the range of reception (of about 30...50 km), anyone having a 28 MHz horizontal antenna and a receiver, can zero-beat his own signal source (HF rig's TX or signal generator) against OH5TEN 28.8 MHz beacon and by using their own TX's signal's harmonics, get accurate frequency signal for 2 m, 70 cm, 23 cm and 3 cm bands. Since HF rig's harmonics are non-existent above 500 MHz, for 23 cm and 3 cm bands' a comb generator is usually needed - a simple device which creates harmonic signals from low-power RF (ie. 0.5 mW), which can be fed to U/SHF- receiver directly with a cable or RF-coupler. This comb generator can be fed on 28.8 MHz with signal generator or from HF TX via ~ 45 dB power attenuator of sufficient power handling capability.

OH5TEN's 28.8 MHz  beacon's ground wave coverage. Computed signal strengths (S-units) as received with a 6 dBd antenna at 25 m AGL.

I would like to thank Ari, OH5KFP and Aki, OH5MWZ for their help re-installing the OH5SHF beacons, it's antennas and OH5RAC's antennas and our radio club OH5AG for sponsoring crystals&synthesizers for all of the beacons from 10 meters to 3 cm and of course RATS ry for sponsoring two 1 W microwave PA's and the 10 GHz slotted WG antenna.
 
 

OH5RAC 145 MHz FM repeater's mobile coverage with different mobile (green 10 W, blue 50 W) & portable (yellow) equipment.
The repeater has been accessed 7000 times / year. TX and RX frequencies were adjusted in April 2006: deviation  <200 Hz from nominal.

50 and 432 MHz beacons were 7 dB above the noise level with 2.4 kHz B.W. and 2*6-el.  (CC 6176B) & 2*19-el. (CC 719B) Yagis 30 m AGL w. mast-head SP-7000 preamp in KP31PU.
OH5ADB received in KP30HV on 144.455 827 MHz with 19.7 dBd antenna and 1.4 dB feed losses at -117 dBm.