Akademisk Radioklubb

LA1K / LA1ARK / LA1UKA

Author: LA3WUA (page 1 of 2)

3m parabole dish ready

We have finally finished one of the major goals of the 1 to 10 GHz project. The mast is installed, the rotor is mounted and the 3m parabole is built.

Below is a narrated selection of pictures from the build. You can find the full album here.

Espen Molven, LA2USA, advised that the mast would be easily liftable by two persons.

We barely made it with 4 people.

The mast was easily fitted to the lamp post.

First hoisting of the mast.

The mast hoisting mechanism is really strong. Hydraylic lifting allows us to easily lift a grown man.

The rotor (SPID BIG-RAS/HR) mounted on the mast.

To get the rotor running some outdoors soldering was needed.

The center hub of the parababole fully assembeled.

Two of twelve spokes mounted. The spokes came ready assembeled.

Progress on spoke mounting.

A view of the attachment point of spokes to the center hub.

All spokes in place.

Mounting circular bracing rings around the dish. In the background, you can see our 5.8 GHz wifi link and the four bay array for 144 MHz.

The twelve pieces of mesh that the dish is made of.

One of twelve mesh pieces mounted.

Clamps were very helpful when mounting the mesh.

Half way there. In the background are our homebrewed DK7ZB antennas for 4m and 6m.

We had time to do a re-enactment of the photo from the original 5 m parabole.

Twelve hours after starting mounting the mesh the parabole is complete!

 

A big thanks to everyone who helped make this possible. Stay tuned as we attempt to make our first contacts with the dish, and progress on the RF hardware.

Measuring LNA characteristics

As mentioned in the post on our future ADS-B setup we have conducted a series of measurements on amateur radio low noise amplifiers (LNAs). I had the opportunity to spend some time at NTNUs microwave laboratory while working on my masters thesis. Some of these results and figures are also presented in the thesis (“Ground station considerations for the AMOS satellite programme”, Øyvind Karlsen, NTNU, 2017). This post shows how we measured the Noise Figure (NF), Gain, Third-order intercept (TOI/IP3), spurious free dynamic range (DR_f) and insertion loss of the LNAs in the figure below.

LNAs under test. From top left to bottom: SSB-Electronic SP-7000, SP-2000, SP-200, SP-70, VHFDesign.com LNA 04-2m-v.03, Kuhne MKU LNA 131 AH and LNA4ALL

To start I would like to elaborate which characteristics are important for a LNA and how to measure them. If you are more interested in the results, feel free to skip ahead by clicking this link. The LNAs under test are:

SSB-Electronic:
– SP-7000 (432 MHz)
– SP-2000 (144 MHz)
– SP-70 (432 MHz)
– SP-200 (144 MHz)
VHFDesign.com:
– LNA 04-0m-v.03 (144 MHz)
Kuhne:
– MKU LNA 131 AH (1296 MHz)
LNA4ALL:
– LNA4ALL @ 5 V (30 MHz to 2000 MHz)

A simple way to check if LNAs are working, is to turn on the power supply with a multimeter inline. The multimeter should then be able to read that the LNA is drawing the amount of current specified in the datasheet.
By toggling on and off the power to the LNA there should also be an observable increase of the noise floor at the receiver compared to when the LNA is off. To get a better idea of how they contribute to receive performance more sophisticated measurements must be made.

For all measurements an Aim-TTi EL302Tv precision power supply is used to power the LNAs. All devices are operated at 12 V, except the LNA4ALL which is operated at 5 V.

Noise figure and gain analysis

The noise figure of an active component describes how much noise it adds to an incoming signal. The noise and gain performance of the first stage in a receiver is particularly important as it contributes the most to the system noise figure. The background for this lies in Friis formula for noise, for a more thorough explanation see this link. Using the FSV-K30 option on a R&S FSQ signal analyser together with a HP 346B noise source, the gain and noise figure of the LNAs can be measured. The measurement setup is seen the figure below.

Gain and NF measurement setup

The measurement works by having a wideband noise source, which is a device with a very precise noise contribution over a large frequency range, connected and calibrated to a spectrum analyzer (in this case called a signal analyzer). The noise contribution of this noise source is listed in a table at the back of the device, as seen below.

Signal analyzer calibration procedure. Noise source table shown on the back of the noise source to the right.

Once the spectrum analyzer is calibrated to the noise source, the noise source is connected to the input of the device under test (DUT). The output of the DUT is then connected to the spectrum analyzer. The resulting signal measured at the spectrum analyzer will now be the amplified noise from the noise source. Since the power level and noise contribution of the source is precisely known it is possible to calculate both noise and gain of the DUT.

This is where option FSV-K30 makes the measurement very simple. The signal analyzer sweeps  and calculates automatically, the output is a plot of both noise figure and gain vs frequency.

These measurements are very temperature sensitive. They are done at room temperature, and for thermal stability the equipment is powered on for an hour before starting measurements. Furthermore, the signal analyser is regularly re-calibrated to the noise source.

The SSB-Electronic LNAs feature adjustable gain by tuning a potentiometer. Each LNA is tested for three cases: maximum position, middle position and minimum position.

Linearity analysis

The linearity of a LNA will limit what signals it can reliably amplify. Any active device will generate spurious emissions, for a LNA the close-by intermodulation products that occur between two or more tones in close proximity are the most detrimental. We will show two linearity measures that originate from the third order intermodulation product (IM3) of a two tone test, third order intercept point (TOI or IP3) and Spurious Free Dynamic Range (SFDR or DR_f).

The n-th order intercept point (IPn) is where the n-th order intermodulation-product (IMn) crosses the linear growth curve. This is commonly used as a measure of linearity for non-linear devices, and a higher IPn value is better. Since our devices are highly linear, the measurement is noise floor limited beyond third order intercept (IP3). The relation between the two tone test and IP3 is shown in the figure below.

Relationship between two tone test and IP3 measurement

The two tone test is conducted by setting up two equally strong carriers with a given spacing. By setting up a two-tone test, intermodulation products will be generated in any non-linear device. We are interested in the third order intermodulation products, these can be found at F_high + F_spacing and F_low – F_spacing.
The measured level of these intermodulation products may be used to calculacte output referred IP3 (OIP3) and input referred IP3 (IIP3), as illustrated in the figure above.
The third order intercept point is located at the intersection between the linear line and the third order line. In the logarithmic domain this may be mathematically expressed as:

Solving for X gives output referred location of the third order intercept point

Where P_fundamental is the power of one of the fundamental tones and P_IM3 is the power of the resulting intermodulation. Since the power difference between the output and the input of a LNA is gain, the input referred third order intercept point is found by

There is no guarantee that the low and high intermodulation power is equal, so both low and high side powers must be recorded.

Two-tone measurement is performed with 100 kHz spacing generated from a R&S SMU 200A signal generator that is connected to the device under test. The resulting distortion is measured by a R&S FSQ 40 signal analyzer. Inherent test setup distortion is measured for all frequency ranges, but intermodulation products were buried in the noise floor for generator output power of -37 dBm per carrier.

For the SSB-Electronic devices the IP3 measurement for mid-level of potentiometer is done right after NF and gain measurement to ensure that these occur at the same level.

Saturation

In presence of a strong transmitter the LNA might saturate. If this happens it may be necessary to add filtering to attenuate the saturation source.

The measure of how much input power an amplifier can take before spurious emissions are generated is the Spurious Free Dynamic Range (DR_f). DR_f is given as the magnitude relation between the fundamental power and the third order intermodulation product.

If the goal is to measure signals that are at the noise floor, the signal that will induce detrimental spurious behaviour is located at DR_f dB over the noise floor.

It is important to note that DR_f is measured at a specific frequency spacing, and that DR_f is typically lower for close spacing. For HF amateur radio it is more common to use smaller spacing for two tone tests, this way interference from nearby stations during contests may be assessed. A common spacing is 2 kHz, where an adjacent station may be located during contests. The intention with these LNAs is to work weak signals. Local interference is more likely than interference from the intended mode of propagation so wide band blocking is

Finally the insertion loss of the LNAs with internal bypass relays are measured. This is measured by a R&S ZNB 8 Vector Network Analyser (VNA) . Calibration is done with a HP 85052B calibration kit.

Results

In the following figures the gain of the devices is shown with a black trace with values in dB on the right hand axis. Similarly the noise figure is shown with a blue trace with values in dB on the left hand axis.

Click each text section to expand the results for the different LNAs.

SSB-Electronics SP-7000

SSB-Electronics SP-7000

SP7000 High gain setting

SP7000 High gain setting:

Gain: 22.6 dB

Noise figure: 1.252 dB

Low fundamental tone power (434.95 MHz): -21.22 dBm

High fundamental tone power (435.05 MHz): -22.6 dBm

Low IM3 power (434.85 MHz): – 96.8 dBm

High IM3 power (435.15 MHz): -98.9 dBm

Low side OIP3: 37.73 dBm

High side OIP3: 38.84 dBm

Low side IIP3: 16.24 dBm

High side IIP3: 15.13 dBm

Low side DR_f: 77.68 dB

High side DR_f: 75.46 dB

Insertion loss: 0.28 dB

SP7000 Medium gain setting

SP7000 Medium gain setting:Gain: 16.96 dB

Noise figure: 1.514 dB

Low fundamental tone power (434.95 MHz): -26.93 dBm

High fundamental tone power (435.05 MHz): -26.79 dBm

Low IM3 power (434.85 MHz): -102.2 dBm

High IM3 power (435.15 MHz): -104.3 dBm

Low side OIP3: 37.635 dBm

High side OIP3: 38.755 dBm

Low side IIP3: 20.675 dBm

High side IIP3: 21.795 dBm

Low side DR_f: 75.27 dB

High side DR_f: 77.51 dB

Insertion loss: 0.28 dB

SP7000 low gain setting

SP7000 Low gain setting:Gain: 12.23 dB

Noise figure: 1.841 dB

Low fundamental tone power (434.95 MHz) : -31.35 dBm

High fundamental tone power (435.05 MHz): -31.3 dBm

Low IM3 power (434.85 MHz): -106.8 dBm

High IM3 power (435.15 MHz): -109.1 dBm

Low side OIP3: 37.73 dBm

High side OIP3: 38.9 dBm

Low side IIP3: 25.5 dBm

High side IIP3: 26.67 dBm

Low side DR_f: 75.45 dB

High side DR_f: 77.8 dB

Insertion loss: 0.28 dB

SSB-Electronics SP-70

SSB-Electronics SP-7000

SP70 high gain setting

SP70 High gain setting:

Gain: 22.13 dB

Noise figure: 0.68 dB

Low fundamental tone power (434.95 MHz): -21.64 dBm

High fundamental tone power (435.05 MHz): -21.6 dBm

Low IM3 power (434.85 MHz): – 95.6 dBm

High IM3 power (435.15 MHz): -99.4 dBm

Low side OIP3: 36.98 dB

High side OIP3: 38.9 dB

Low side IIP3: 14.85 dB

High side IIP3: 16.77 dB

Low side DR_f: 73.96 dB

High side DR_f: 77.8 dB

Insertion loss: 0.17 dB

 

SP70 medium gain setting

SP70 Medium gain setting:

Gain:  19.26 dB

Noise figure: 0.78 dB

Low fundamental tone power (434.95 MHz): -24.1 dBm

High fundamental tone power (435.05 MHz): -24.04 dBm

Low IM3 power (434.85 MHz): -98.2 dBm

High IM3 power (435.15 MHz): -102.1 dBm

Low side OIP3: 37.05 dBm

High side OIP3: 39.03 dBm

Low side IIP3: 17.46 dBm

High side IIP3: 19.44 dBm

Low side DR_f: 74.1 dB

High side DR_f: 78.06 dB

Insertion loss: 0.17 dB

SP70 low gain setting

SP70 Low gain setting:

Gain: 12.72 dB

Noise figure: 1.22 dB

Low fundamental tone power (434.95 MHz) : -31 dBm

High fundamental tone power (435.05 MHz): -30.95 dBm

Low IM3 power (434.85 MHz): -105.3 dBm

High IM3 power (435.15 MHz): -108.5 dBm

Low side OIP3: 37.15 dBm

High side OIP3: 38.78 dBm

Low side IIP3: 24.43 dBm

High side IIP3: 26.06 dBm

Low side DR_f: 74.3 dB

High side DR_f: 77.6 dB

Insertion loss: 0.17 dB

SSB-Electronics SP-2000

SSB-Electronics SP-2000

SP2000 high gain setting

SP2000 High gain setting:

Gain: 23.08 dB

Noise figure: 1.9 dB

Low fundamental tone power (144.95 MHz) : -20.95 dBm

High fundamental tone power (145.05 MHz): -20.79 dBm

Low IM3 power (144.85 MHz): -84.5 dBm

High IM3 power (145.15 MHz): -87.8 dBm

Low side OIP3: 31.78 dB

High side OIP3: 33.51 dB

Low side IIP3: 17.86 dB

High side IIP3: 19.60 dB

Low side DR_f: 63.55 dB

High side DR_f:  67.01 dB

Insertion loss: 0.09 dB

SP2000 medium gain setting

SP2000 Medium gain setting:

Gain: 18.58 dB

Noise figure: 1.92 dB

Low fundamental tone power (144.95 MHz) : -25.23 dBm

High fundamental tone power (145.05 MHz): -25.07 dBm

Low IM3 power (144.85 MHz): -89.1 dBm

High IM3 power (145.15 MHz): -92.5 dBm

Low side OIP3: 31.94 dB

High side OIP3: 33.72 dB

Low side IIP3: 8.86 dB

High side IIP3: 10.64 dB

Low side DR_f: 63.87 dB

High side DR_f:  67.43 dB

Insertion loss: 0.09 dB

SP2000 low gain setting

SP2000 Low gain setting:

Gain: 13.91 dB

Noise figure: 2.03 dB

Low fundamental tone power (144.95 MHz) : -29.95 dBm

High fundamental tone power (145.05 MHz): -29.78 dBm

Low IM3 power (144.85 MHz): -93.8 dBm

High IM3 power (145.15 MHz): -96.6 dBm

Low side OIP3: 31.93 dB

High side OIP3: 33.41 dB

Low side IIP3: 13.34 dB

High side IIP3: 14.83 dB

Low side DR_f: 63.85 dB

High side DR_f:  66.82 dB

Insertion loss: 0.09 dB

SSB-Electronics SP-200

SSB-Electronics SP-200

SP200 high gain setting

SP 200 High gain setting:

Gain: 20.5 dB

Noise figure: 0.4 dB

Low fundamental tone power (144.95 MHz) : -22.79 dBm

High fundamental tone power (145.05 MHz): -22.64 dBm

Low IM3 power (144.85 MHz): -95.5 dBm

High IM3 power (145.15 MHz): -95.6 dBm

Low side OIP3: 36.36 dB

High side OIP3: 36.48 dB

Low side IIP3: 23.63 dB

High side IIP3: 23.75 dB

Low side DR_f: 72.71 dB

High side DR_f:  72.96 dB

Insertion loss: 0.06 dB

SP200 medium gain setting

SP 200 Medium gain setting:

Gain: 16.88 dB

Noise figure: 0.36 dB

Low fundamental tone power (144.95 MHz) : -26.67 dBm

High fundamental tone power (145.05 MHz): -26.53 dBm

Low IM3 power (144.85 MHz): -98.6 dBm

High IM3 power (145.15 MHz): -98.4 dBm

Low side OIP3: 35.97 dB

High side OIP3: 35.94 dB

Low side IIP3: 15.47 dB

High side IIP3: 15.44 dB

Low side DR_f: 71.83 dB

High side DR_f:  71.87 dB

Insertion loss: 0.06 dB

SP200 low gain setting

SP200 Low gain setting:

Gain: 12.73 dB

Noise figure: 0.57 dB

Low fundamental tone power (144.95 MHz) : -30.91 dBm

High fundamental tone power (145.05 MHz): -30.75 dBm

Low IM3 power (144.85 MHz): -102.6 dBm

High IM3 power (145.15 MHz): -102.2 dBm

Low side OIP3: 35.85 dB

High side OIP3: 35.73 dB

Low side IIP3: 18.97 dB

High side IIP3: 18.85 dB

Low side DR_f: 71.69 dB

High side DR_f:  71.45 dB

Insertion loss: 0.06 dB

VHFDesign.com LNA 04-0m-v.03

VHFDesign.com LNA 04-0m-v.03

VHFDesign.com 2m LNA

Gain: 23.58 dB

Noise figure: 0.52 dB

Low fundamental tone power (144.95 MHz) : -20.3 dBm

High fundamental tone power (145.05 MHz): -20.1 dBm

Low IM3 power (144.85 MHz): -97.3 dBm

High IM3 power (145.15 MHz): -96.6 dBm

Low side OIP3: 38.5 dB

High side OIP3: 38.25 dB

Low side IIP3: 14.92 dB

High side IIP3: 14.67 dB

Low side DR_f: 77 dB

High side DR_f:  76.5 dB

Kuhne MKU LNA 131 AH

Kuhne MKU LNA 131 AH

Kuhne 23cm LNA

Gain: 21.42 dB

Noise figure: 0.43 dB

Low fundamental tone power (1294.95 MHz) : -23.18 dBm

High fundamental tone power (1295.05 MHz): -23.21 dBm

Low IM3 power (1294.85 MHz): -94.5 dBm

High IM3 power (1295.15 MHz): -102.2 dBm

Low side OIP3: 35.66 dB

High side OIP3: 39.5 dB

Low side IIP3: 14.24 dB

High side IIP3: 39.5 dB

Low side DR_f: 71.32 dB

High side DR_f:  78.99 dB

LNA4ALL

LNA4ALL

The IP3 is found to be on average 34.75 dB by measurements done by DG1TRF, these can be found on the LNA4ALL product page. Below is the noise and gain measurements from 30 MHz to 2 GHz measured with 5 V supply.

LNA4ALL gain and noise figure

 

 

 

 

 

 

 

Installing an AIS receiver at Vassfjellet

This weekend we put up a Automatic Identification System (AIS) receiver while making some improvements on our internet link to Vassfjellet.

Improving the link

Last year we installed a 5.8 GHz radio link between Samfundet and Vassfjellet, where we have our radio beacons. The radio link will allow us to remotely check status of the beacons, and allow for several exciting monitoring applications.

We found out that we had done a mistake in choosing the feedline between the Ubiquity rocket M5 and the antenna. The mistake has nagged us over the entire winter, so when the snow on the road finally melted, we bolted up the mountain.

LB0VG terminating RJ45 plugs for the link. LA1BFA inspecting important assets in the background.

The feedline was around 6-7 meters of a RG-58 type. At 5.8 GHz this turns out to have a massive attenuation, approximately 13 dB. By simply replacing with a shorter and better cable, we could get a huge improvement in link quality.

To get as low cable loss as possible we decided to mount the Rocket right behind the antenna. We bought some 15 cm RP-SMA to SMA pigtails that use RG174 cable, which should give a cable loss of only 0.56 dB. The resulting improvement is seen below.

When we took the link down at 12.00 UTC the link margin was 10 dB. We had it up again an hour later, and the link margin is now 22 dB. This is very much in line with the cable loss improvement mentioned above.
This improvement in link margin will be very nice when we start adding more services up there.

Marinetraffic AIS receiver

Boats over a certain size are required to report their position using AIS. This makes for very interesting listening, as you can effectively stalk the movements of large boats.

Marinetraffic is a website where reports from a network of AIS receiving stations are gathered.  Marinetraffic are also interested in unique sites that will allow them to expand their coverage, see their application form here. We got in touch with them, and they were interested enough to send us:

We tuned the antenna to 162 MHz using our AA-170 antenna analyzer, and got it to resonate with about 1.2 VSWR. The antenna was plugged to the SLR350Ni, which surprisingly is based on a Raspberry Pi 3 with a radio daughterboard. After a small power struggle with the software and trying to set it to a static IP, we started receiving ships.

We were a little worried about the receiver getting a lot of interference from LA2VHF, as they are in the same band, and very close. But it looks like everything is working smoothly.

LB0VG handing LA3WUA the Rocket modem. Behind LB0VGs head is the newly installed AIS antenna.

Below is a display of the ships that we have received. On average we get 200 AIS messages a minute from about 100 ships, with a maixmum reception distance of 463 km. I’m confident that by adding some filtering, an LNA and maybe a small yagi antenna, we can get more than double of this.
You can also find live information on our Marinetraffic station page.

AIS messages from boats near the Trondheim coastline. Vassfjellet receiving station in lower right.

It’s very nice to finally get some traffic over the link. Over summer we’re hoping to expand with more monitoring services, but that’s a story for another blogpost.

LA3WUA demonstrating a patented LA1K antenna hoisting method.

An update on the 1 to 10 GHz project

ARK is making an open source ground station that is compatible with GENSO and SatNOGS. We want to focus on documentation, making it easy for others to adopt.

We are attempting to approach this problem from a radio system perspective. We want to give a guideline to set up a modular distributed network node. The node should support common ham activities in the bands covered, such as EME and satellite work.

The ground station will cover 1 to 10 GHz and be RX and TX capable.

Block schematic of the project

The reference ground station will support up- and downlink covering 1 to 10 GHz using a cross-polarised dish feed. A dual channel software defined radio is used as the central transceiver, allowing for a wide range of applications to be programmed on the fly.
By using two radio channels and two polarisations we want to control polarisation schemes dynamically in software. This will allow for optimization of a large range of satellites. Using this setup we aim for access to linear vertical and horizontal, right and left handed circular as well as some elliptical polarisations.
This ground station (and network) could be interesting for amateur radio groups, CubeSat initiatives and research projects.

The Antenna(s)

We are going to use a parabolic dish antenna with around 30 to 45 dBi gain, increasing with frequency. The parabolic dish will be a 3 m kit by rfhamdesign. In the long run we might attempt a custom parabole design that aims to lower costs even further, utilizing machined or 3D-printed parts and metallic mesh.

At the core of the project at this stage is the choice of feed antenna for the dish. We want a feed that has the following attributes:

Continuous coverage over 1 to 10 GHz, S11 < -10 dB: Covering the ham bands from 23 cm to 3 cm. There are also a lot of interesting satellites in these frequencies.

Stable phase center over the frequency range: The phase center is the apparent center of radiation, for a feed antenna this is particularly important as it decides where the antenna should be placed in relation to the dish. If the phase center shifts with frequency the main lobe of the antenna will also change with frequency, making calibration and gain optimization hard.

Constant antenna pattern over the frequency range: Similarly to the previous requirement it is desireable to have a stable main lobe. Distance to the dish, as well as a backwards ground plane will have to be considered.

100 W power handling: In Norway this is the legal limit output power for the bands 23 cm to 3 cm.

Dual linear or dual circular polarisation:  By having either of these combinations it is possible to combine the remaining polarisations as well as a set of elliptical polarisations.

We are looking into several antenna topologies. Luckily there’s a lot of research available for ultra wideband (UWB) dish feeds, as they are very common in radio astronomy. Our design differs a bit from these however, since we also intend to use the antenna for transmit applications. The main difference is the power handling, as radio astronomy applications usually are RX only.

So far we have found two promising designs, and are running simulations to get the exact dimensions.

Dual linear spiral antenna

The first of the two is the dual linear spiral antenna. This is a planar antenna that is realizable on printed circuit boards. For additional power handling machining the elements may be a better choice. For high frequencies the antenna has to be scaled in a way that leads to lots of tight gaps between metallic elements, for high power levels this will cause spark-gaps, and the antenna will not function as intended.

Dual linear spiral antenna

Dual ridge vivaldi antenna, alternately called quad ridge horn antenna:

The more promising of the two designs is the horn antenna, as several commercial designs that fit our specification already exist. One such design by Schwarzbeck Mess-Elekronik is shown in a picture below. This antenna covers 0.4 to 10.5 GHz with an efficiency of 90% or better. It also handles 200 W.

We are looking forward to learn more about the design as we progress with our own.

A dual polarised horn by Schwarzbeck Mess-Elektronik

Getting funding

The other thing that we have been working with at this stage is getting funding for development. Transistor costs alone are so high that we couldn’t do this on our own.

We would like to give a big thanks to NTNU IES, KSAT, Marlink, Sit, Radionor, Jotron and Kongsberg for making this happen!

What’s next?

We have prepared a  lamp post mast that will serve as an attachment point for a tiltable mast, which LA2USA is designing. The lamp post previously housed ARKs 5.5m spoil in the early 90s, seen in the photo below.

The story of the old spoil ended with a winter storm, which is why we’re making some changes to the mast. Hopefully by tilting it down during windy periods it will catch less of the harsh winds. This also comes with the benefit of easier maintenance since the feed will be more accessible.

We have also ordered a rotor and 3 m mesh parabole kit from rfhamdesign. Building will commence when parabole kit and the LA2USA mast arrive here mid June.

As the design on LNA, PA, antennas and software continues there will be more updates.

New hams

After the test last Wednesday there are 11 new amateur radio operators. We are proud to announce:

Martin Hergot Festøy: LB7AH
Ken Are Meisler: LB7CH
Ole Christian Tvedt: LB7DH
Håkon Eide: LB7EH
Haavard Knibe Fiskaa: LB7FH
Anders Liland: LB7GH
Anders Selfjord Eriksen: LB7HH
Dennis Skulbru Eriksen:  LB7IH
Einar Uvsløkk: LB7JH
Svein Ove Undal: LB7KH
Ragni Helene Halvorsen:  LB7RH

You will recieve a letter from NKOM with the final details.

Congratulations, we look forward to hearing you on the air.

Older posts