Akademisk Radioklubb


Kuhne amplifiers and 1 to 10 GHz stage three

ARK is currently working on a project that will allow us to work Earth-Moon-Earth, satellites and various scattering modes on the amateur bands between 1 and 10 GHz. Our solution uses a 3 m parabolic dish together with a set of discrete amplifiers, the entire system is excited by a USRP SDR.

We have split the 1 to 10 GHz project into four sections. Up until now we have completed stage one and two, while stage three and four are still remaining. The stages are roughly:

Stage 1: Literature study, ordering of components – Further details in “An update on the 1 to 10 GHz project”

Stage 2: Construction of the parabolic dish and mast – Further details in “3m parabole dish ready”

Stage 3: PCB development and integration onto parabolic dish – Further details in this blogpost 😀

Stage 4: Long term projects with the dish, software, amplifiers, new antenna feed – Further details in the future.

This is a good time to elaborate more on our plans for stage 3 as the amplifiers that will be used for the project just arrived! We have purchased:

144/432 MHz IF to 10.5 GHz mixer – MKU 10 G4, 3 cm Transverter

To keep the work more organised we have split stage three into four sections. The sections are three PCB-design projects and one final assembly of all the components at the back of the parabolic dish.

1: Wideband driver amplifier
The Kuhne amplifiers and transverter seen in the previous section, will bring the output power to the level that is required to achieve Earth-Moon-Earth communications in the amateur radio bands. In order to be able to excite the Kuhne amplifiers and transverter from a USRP SDR with 10 mW max output power, an intermediate stage is required. The next step is to design and construct this amplified.

Very simplified schematic of 1 W driver.

We have conducted a study of available parts, and concluded that it is indeed possible to create an amplifier that will deliver 1 W across the frequency band from 0.1 GHz to 6 GHz.

In the figure above an example schematic using Guerilla RFs GRF4001 together with Analog Devices HMC637LP5 is shown. The device will deliver 1 W across 0.1 GHz to 6 GHz. This will allow exitation of all amplifiers, as well as the transverter that enables 10 GHz coverage. The gain should be on the order of 30 dB in order to avoid operating the USRP at its saturation power, where it is known to be quite noisy.

2: Wideband low noise amplifier
Another component that we want to develop ourselves is the low noise amplifier (LNA). There are not many good and cheap LNAs available for the amateur radio market, despite there being integrated circuits that boast very good performance for this application. If we are able to make an LNA and provide the design notes as open source, it will likely be beneficial for many people.

The LNA is one of the more challenging circuits. It needs to work using relatively cheap equipment while being largely immune to electromagnetic noise. A lot of work will likely be spent on making the supply-lines that power the HMC753 LNA circuit noise free, as well as ensuring that the metallic shielding is sufficiently tight. Another consideration is that the amplifier must be able to sustain relatively high input powers that will leak through the coaxial relays during transmit.

Outline of LNA.

The figure above shows a draft for a test assembly for performance testing of HMC753, which is a device that could be used in our LNA.

3: Controller board
Interfacing with the amplifiers will be handled through a controller board that communicates with either a computer or the USRP directly over the serial protocol RS232. The interface board is responsible for managing power supply states for all amplifiers as well as startup sequencing.

A set of RF relays are used to select which of the discrete amplifiers should be connected to the different points along the circuit. These are available as surplus devices on auction sites such as eBay.

Essentially, the interface board is responsible for ensuring that all connections and devices in the figure below are connected and powered correctly for a given configuration. It should also be able to alter the configuration in a rapid way.

Relays, amplifiers, transverters and SDR connection diagram.


4: Mechanical integration

After the three sections above are complete, the mechanical integration of the RF system onto the dish can start. This is an extensive effort as there are many concerns to deal with. Thermal management and waterproofing are two likely issues. So far we have an idea revolving around a gutter heater solution to keep the system from freezing during the winter. To keep the system cool enough we are experimenting with different heatsinks and weatherproof fans (IP68).

We hope to have the first three stages finished by the end of november, and the mechanical work started and delivered some time early next year. The time it takes to develop the PCBs gives us a good chance to secure the final funds that are required for the mechanical work (cabinets, fans, heatsinks). Overall we are really excited to see the project taking shape.

Before we finish all the sub-projects in stage three we might try to work some contacts using the 23cm module on our IC-9100, a MKU 131 AH 23 cm LNA, coaxial relays and the 200 W 23 cm PA we just bought. More on that in a later post.

New licencees H17

The results are in from the license exam last Tuesday. We are proud to announce five new amateurs:

Vegard Josvanger: LB0LH
Eirik Veløy Nadim: LB0MH
Håvard Gautneb: LB0NH
Jordi Frances: LB0OH
Sivert Bakken:  LB0PH

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

Microham Double Ten performance outside the specified frequency range

ARK has been expanding our antenna park with DK7ZB yagis for 4 m (70 MHz) and 6 m (50 MHz). In order to make the setup more practical we would like to have the antennas selectable via our Microham Double Ten antenna switch. The Double Ten is only specified up to a frequency of 30 MHz, so we wanted to investigate if it would be usable beyond this.

6m and 4m beam

The DK7ZB 6 m (bottom) and 4 m (top) yagis we want to use with the Double Ten switch.

To perform these measurements we used a HP 8753E network analyzer calibrated with a 85052D calibration kit.

Since we are intending to use this device outside its specified frequency range we decided to carefully measure the S-parameters between each output and input port for a sweep from 0 MHz to 200 MHz, focusing on the amateur radio bands covered.

The S-parameters (or scattering-parameters) are a way to measure the response of a RF system by applying a test signal to one port, and assessing the levels that result on the other ports. The results are categorised as S_xy, where S_xy denotes the power that appears at port x from an input on port y. S_xx is also a valid measurement, and gives an indication of how well matched a device is to the target impedance, in this case 50 ohm.

Schematic for a similiar switch, 6 to 2 switch by OK2ZAW. Source: Remoteqth.com

To better understand the internals of the device we wanted to check the circuit schematic, unfortunately none was available from Microham. Luckily OK2ZAW has made an open source 6 to 2 switch, that turns out to be very similar on the inside. Due to substrate and component losses we expect that the antenna ports that are placed further from the radio ports should be have more loss than the ones that are placed closer.

The inside of the Double Ten switch. Wires are attached to a lab power supply to manually select which ports are active/inactive.

To avoid noise interfering with our measurements we took care to re-mount the metal chassis between each change of ports.

Fewer measurements are required if the response between a given port and radio A/B is the same, so this was the first thing we checked.

As seen in the figures above the response from port 10 to radio A and radio B are nearly identical. We assume that this is also the case for the other ports.

This slideshow requires JavaScript.

Insertion loss is a measure of how much signal is lost when translating through a device. For the switch the insertion loss varies quite a bit from port to port. Lowest insertion loss is observed at ports 1 and 10, which are on the edge of the switch. One possible explanation for this is that the edge ports have more series inductance, which could help match the relay better at high frequencies. At port 10 the insertion loss at 70 MHz is only 0.35 dB. This is a promising result, and could indicate that it might be useable for our purposes.

To ensure that the device can operate at 50 MHz and 70 MHz we also need to check that the reflectivity of the input port is not excessively high. Low reflectivity means that the device will translate most of the applied power to the output, and high reflectivity means that a significant portion of the input power is reflected back at the transmitter.

Return loss of port B when output is coupled to port 6

The return loss varies between -8 and -13 dB at 70 MHz, depending on which output is selected. The outputs with lower insertion loss have better match. A return loss of -13 dB corresponds to a SWR of ~1.6. At the antenna port we are not so concerned about the reflectivity, as the power levels are not a cause of concern when receiving.

The final characteristic that is investigated is the isolation between Radio A and Radio B ports. If two transceivers are operating on the switch simultaneously, the isolation between the ports must be sufficiently high to ensure that either radio is not saturated when the other transmits.

Isolation between Radio A and Radio B ports.

The isolation between ports is not very high. When transmitting with 100 W at 70 MHz, 2 mW will be coupled to the other radio. This can cause the receiver to saturate if the filtering is not sufficient. Typically the antenna will receive signals that are on the order of pWs, so the receiver would need tremendous dynamic range in order to cope with 2 mW coupled.

Practically this is not a large problem, as we don’t intend to use two radios at 4 m simultaneously. Furthermore, 70 MHz is heavily attenuated by the preselectors filters that are common in HF transceivers, so the intermediate frequency stages are likely to be sufficiently shielded from saturation.

At low frequencies (>30 MHz) the measurement is seen to be noise floor limited.

In conclusion, given that port 1 or 10 is used, the Double Ten will work well at 6 m, and decently at 4 m. Care should be taken to not run excessive power at these frequencies, atleast not without extra filtering on the other radio ports.

Thank you to LA1BFA for providing a Double Ten switch for testing! It was very convenient to not have to disassemble the one we already have installed.


Samfundet is currently closed down in preparation for UKA. In between the sound checks, rehearsals and other activities that occur in this hectic period, we managed to sneak in some contacts as LM100UKA in the CQ WW RTTY contest. We decided to run as Multi-Two to get the most out of the few timeslots when we could operate.

Henrik LB5DH working 20 m on Saturday.

One of the challenges we face when operating in contests from Samfundet is Electro Magnetic Interference (EMI) to the other groups at Samfundet. Particularly when operating legal limit on the low bands, audio and lighting equipment tend to get unwanted interference. Normally we try to locate the weak spots and seal them with ferrites and filters, but during UKA there is a lot of new equipment so it is hard to get to all the problems in time. The solution is often to cooperate with the groups that are responsible for audio and visuals and coordinate so that we do not transmit high power levels on the low bands while they are working.

Conditions on 20 m was also good on Sunday.

The contest was also our first good test of the Flex 6500 for RTTY. Thanks to the panadapter, finding an open frequency to work was a breeze. It also allowed us to quickly diagnose that we had set the wrong tone spacing (oops). We had a bit of trouble setting up SmartSDR to talk to MMTTY through N1MM+, but in the end we got it working through the help from some good resources online. What was not so great was setting RIT/XIT using mouse and keyboard. Hopefully our Maestro will arrive soon to address this problem.

QSOs per hour. Orange: 15 m, Red: 20 m and Blue: 40 m.

The QSO-rate was largely influenced by when we had a chance to operate in between different sound checks. As the graph shows, this was mostly during the afternoon on Sunday. Conditions were ok, but only one band was significantly open at a given time, typically either 20 m or 40 m.

Score and multipliers per band

We got 657 QSOs and 107 mults over the weekend. A good result considering the operating conditions, both on the air and at Samfundet. Thanks to LB5QG, LB1MH, LB5DH, LB7JG and LA3WUA for their operating efforts.

Repairing broken receiver on a USRP N210 WBX40 daughterboard

Kimmo Kansanen at NTNU recently donated some USRP N210 units they no longer had a use for. We have started to use them for various communications experiments, for example estimating the antenna patterns of our VHF/UHF antennas against the LA2VHF and LA2UHF beacons.

While operating one of the N210s we were sudddenly unable to receive the LA2VHF beacon that had previously been easily decodeable. We also saw that the noise floor had increased by 20 dB. Jens, LB6RH, decided to investigate matters further.

Inside the N210 a WBX-40 daughtercard provides the RF-frontend. The device functions as normal, except for in receive mode. The receiver section of the WBX-40 should be a good place to start looking.

The WBX-40 daughterboard after removing coaxial patches that attach to the front panel of the N210.

To further investigate what may be wrong we started investigating the board for any obvious short circuited connections. We were unable to find any such sources. The next step was to probe around the board with a multimeter to check the voltages being generated from the different voltage regulators on the board. Since Ettus Research provides the schematic for the WBX-40 online this process was greatly simplified.

LB6RH found that the 3.3 V rail on the output of voltage regulator U308 was only at 1.2 V. Suspecting that something was wrong with U308, he removed the component and attached a laboratory power-supply to the 3.3 V output pin. After turning on the supply (with the current limit set low, to avoid frying the circuit) he noticed that there was still a short circuit. To identify which component caused the short we borrowed a FLIR thermal imaging camera to check what components got hot when we turned up the current limit on the power-supply.

Unfortunately we did not get a good picture of the thermal test, but the chip that got hottest was the Low Noise Amplifier (LNA) U313. Since our problems are related to poor reception we thought this might be a likely candidate. After removing U313 we turned the power-supply back on, and saw that there was no longer a short, hurray!

A closeup after removing components U313 and U308.

We ordered new components for U308 and U313 from Digikey, and soldered them back in place.

U308 – Analog Devices Inc. ADP3336ARMZ linear regulator

U313 – Broadcom Limited MGA-82563-TR1G MMIC broadband LNA 

Replacements for U308 and U313 have arrived.

The soldering battlestation. A microscope helps when soldering small parts.














After replacing the two parts we connected everything back and tried powering on the device again.

We are now able to receive LA2VHF again!

« Older posts