This blog is about ham radio activities from PA1EJO, my previous call sign was PD1EJO between 16-mar-2016 and 25-may-2016, since that date it became PA1EJO which is a full license in the Netherlands, maidenhead locator JO21fx. More details about the station can be found on QRZ dot com.
After the most recent fox hunt I thought that a smaller antenna would be helpful, so I went for the moxon for 2 meter, material used: 32 mm PVC tube etc, coax cable, glue, tie wraps, 4 mm brass bar, propane burner (you can’t bend brass when it is cold). I used one of the moxon calculators, long end: 740,7 mm, sort end: 274,9 mm, short end top: 98,3 mm, short end bottom: 142,3 mm, spacer 34,3 mm.
Leo Bodnar’s webshop sells two GPS disciplined oscillators, and I tested the miniature version as shown in the image below connected to the airspy. It required 60dB attenuation to get a SNR on 60 dB, as a result I estimate the output of the GPS DO to be above 0 dBm when the drive current is set to 8 mA (you can increase it to 32 mA). With the rfmeter program on the VNA I verified that the output is at best 10 dBm over a 50 Ohm load.
The signal looks very clean on the waterfall plot, I could not detect spurious signals around the carrier that originate from the GPS DO. You program the frequency with the provided software for which you need a windows PC, I could not find anything that runs on a Mac or on Linux. But once it is programmed at a preset frequency the settings are retained. This makes the GPS DO ideal as a frequency reference that you can carry around, a USB charger is sufficient to keep it alive.
I verified the accuracy of the airspy SDR local oscillator which is temperature compensated (I believe), but its frequency is not disciplined to the internet or GPS. Here are the PPM offsets that I found with the sdrsharp software:
30 MHz: -1 ppm,
50 MHz: -4 ppm
145 MHz: -15 ppm
250 MHz: -13 ppm
430 MHz: -37 ppm
500 MHz: -41 ppm
750 MHz: -87 ppm
Very likely these values are temperature dependent, for the airspy HF+ I did a similar test for HF frequencies, in this case the calibration values were obtained by the SDR console software:
500 kHz : 3.40 ppm
1 MHz : 3.40 ppm
5 MHz : 2.70 ppm
9996 kHz : 2.65 ppm (it matches the result I got from the RWM time reference)
10 MHz : 2.64 ppm
20 MHz : 2.59 ppm
30 MHz : 2.58 ppm
Below 500 kHz I saw control loop oscillations in the GPS DO output, I did not attempt to estimate any ppm offset. There is an initialization time of about a second for the GPS DO to change to a new frequency, the manufacturer claims that its tuning range is between 400 Hz and 810 MHz.
Why do you need a GPS DO? If you go into the microwave domain (23cm and up) then usually any frequency comes from a lower value via frequency multiplying in which case you may want to use the GPSDO or a Rubidium or Cesium standard. For Doppler measurements to cubesats on the 2m and 70cm we want the observed frequency to be within a Hertz or better, and probably there are a dozen more applications that I forgot. Leo Bodnar’s webshop is here.
In this demonstration we measure the standing wave ratio (SWR) of a load, in this case an UHF antenna. A constraint is that we want to do this in a cost-effective way up to 1.7 GHz. My earlier discussed VNA is not capable of doing this, you can get VNAs that can but they rapidly become rather expensive.
In the demonstration video we generate noise with a tool that essentially amplifies the noise of a zener diode, you could make it yourself, but if you google for BG7TBL noise generator then you will soon find out it is easier to order it than to build it yourself. Noise from the BG7TBL tool has the property that it is random phase by frequency and that the amplitude is somewhat constant over the entire spectrum. Put it on the scope and you get what you expect: something that resembles white noise. The tool needs up to 12 Volt to operate, it will get hot, but it also works on a 9 Volt battery.
If you insert noise in the SDR (I used my airspy) then it will fill the entire HF VHF and UHF spectrum, and you may wonder why you would like to do that at all. Please insert attenuators because the output of the BG7TBL tool is rather high, so I added 30 dB attenuation before I did the measurement with the SDR. When I measure the reflection properties of a load then 10dB goes before the bridge, and 20 dB goed after the bridge to prevent that the bridge is influenced too much.
There is a very good reason why you would like to insert artificial noise into your SDR. The reason is that it allows you to measure the performance of filters or loads connected in between the source and the receiver. In the video we inserted a RF bridge between the source and the receiver.
The RF bridge comes from the transverter store on ebay, see the movie, also there the conclusion is, it is easier to buy one than to make one. The bridge has an input, an output, there is a reference port and there is a device under test port. There are only SMA connectors on the PCB, the schematic is as follows:
On the input (left in the schematic) you insert a signal into the RF bridge, and on the reference port (top port in the schematic) you insert a nominal dummy load of 50 Ohm. On the output of the RF bridge (right port in the schematic) you measure the throughput with your SDR (an airspy). There are now several possibilities depending on what you connected to the DUT port (bottom port in the schematic):
Nothing is connected to the DUT port, in this case the signal reflects back into the RF bridge and this is picked up by the receiver. The throughput is called the directivity of the RF bridge. (for details see the discussion on the transverter-store web page).
The DUT port has a dummy load of 50 Ohm (or the same amount of Ohm as the reference port). In this case you will notice that no signal will pass the bridge and that everything is dissipated in the dummy load at the DUT port.
Connect an unknown load to the DUT port, and now you will see that some of the signal is reflected back to the receiver, but that the rest of the signal is dissipated into the load on the DUT port.
With the first and the second measurement you calibrate the set-up, with the third measurement you can see who much dB is dissipated in the unknown load. You can measure this in the waterfall plot, and hence you can measure the so-called reflection loss RL in decibel caused by the unknown load. Next you need to convert the RL into a SWR value, the equation is as folows:
RL = 20 log10( (SWR-1)/(SWR+1) )
I measured on the waterfall a reflection loss of 20 dB near 440 MHz, this is compatible with a SWR of 1.2, the connected antenna is a miniature version of a 70 cm amateur band antenna.
Hybrid coupler as RF bridge
An alternative for the RF bridge discussed before is to use a 90 hybrid coupler, I found one at a ham market in Beetsterzwaag Friesland. This is what it looks like:
The purpose of the 90 degree hybrid coupler is to split an input signal into two paths, one is essentially forward with 0 degree -3 dB and the other is also forward but then 90 degree phase shifted. There is also an isolated port, anything that is reflected in the system ends up at P4.
To turn this into an RF bridge for measuring SWRs you connect the noise source to P1, the antenna to P2 and a 50 Ohm dummy load to P3. You could interchange P2 and P3, the device is symmetric, it has no effect on the SWR measurement. I use SMA 50 Ohm coaxial dummy load, you find them on ebay where they are a lot more affordable compared to buying them new. The SDR (preceeded by a 20 dB attenuator) is connected to port P4. For 70cm antennas this approach works, but don’t try this for frequencies that are far off the 500 MHz because the bridge 90 degree hybrid coupler was not designed for this.
A more extensive discussion about 90 degree hybrid couplers can be found here.