Usually i do not post links to foreign content but here is a really interesting talk by Bruce K6BP on the state of digital voice and open source usage within amateur radio in general.
Thanks a lot to Hamradio Now for making this talk available !
Since some time i own a DF9NP GPSDO. Currently i only use it to lock my signal generator but i also want to use it for my microwave transverter OCXOs. When connecting a SMIQ signal generator and checking the signal output at 6GHz i noticed some +-50Hz jitter under the poor reception conditions at thelocation of this signal generator. So i want to look a bit closer into that topic. In China i ordered a Neo-M8N module that can be configured to provide a 10MHz clock at its timepulse output. I tried to compare it with the GPSDO and the OCXO of my SMY-02 signal generator.
Of cause the digital clock has significant jitter because it is generated by a CPU (specified with +-10ns).
Neo-M8N clock output at 10MHz
The picture of the timepulse at 10MHz shows, that it seems that comparing the clock to a stable oscillator over a relatively short period could be sufficient for adjusting the frequency of this oscillator to the GPS.
GPSDO, M8N, OCXO
The screenshot shows the oscilloscope triggered to the GPSDO (green). The output of the M8N module (yellow) shows the jitter and the adjustment range within 60 seconds (yellow shadow). The third (blue) trace comes from my SMY-02 which was locked to the GPSDO. The SMY signal shows some slight jitter compared to the reference. To me it is not clear if the cause is the reference or the locking in the generator. The clock from the M8N module shows significant adjustment of the clock frequency within the 60 seconds shown compared to the GPSDO which has a TCXO that is slowly compensated by the GPS inside the reference.
Looking to the signal in frequency domain shows this picture:
Neo M8N spectrum
There are rather close sidebands that require narrow band adjustment of a oscillator eventually locked to this GPS clock. Wideband the spectrum is noisy as well.
M8N phasenoise at 10MHz
Finally i took a short video showing the the scope triggered to the 10MHz OCXO reference of the SMY and comparing the GPSDO and the M8N output. The OCXO is slightly off 10MHz and therefore the picture is moving all the time. You can also see that the M8N is slowly adjusting compared to the GPSDO output.
The above shows my 9cm transverter. It gives about 5W in 3400MHz. The boxes are rather old but still work fine after adjustment of all currents. The PA has 46dB gain which makes a attenuator necessary.
transverter component side
This photo shows the assembly side with the RF transistors. The other side carries the filter cups aß Well as some other larger components.
transverter top side
The small box ist the oscillator/multiplier and the other one the transverter.
103.5MHz oscillator briefly measured with R&S FSP30..
PN Plot G8ACE OCXO
I think above 1kHz the PN figure is determined by the FSP oscillator.
Does someone have a better measurement of such an oscillator, especially above 1kHz up to 200kHz ? Unfortunately i do not have a better PN testset.
The following picture shows the cold-start behavior of the OCXO from room temperature to 60 degree celcius. The X-axis shows time in seconds and the y-axis frequency in Hz.
I shortened the probe pins of my experimental pipecap filter to 5mm in order to get rid of the unwanted response around 7GHz. As expected the filter is rather narrow now and the attenuation increases a lot.
Marcel made some new measurements up to 14GHz in order to see how the suppression behaves. It looks a lot better now but you can also see that at the upper end of the measurement range the attenuation is very low (keep in mind that the probes are nice quarter wavelength antennas there).
The following picture shows the filter tuned to the 6cm band:
pipecap 5mm probes 5760MHz
The passband attenuation is now always somewhere in the range of 2..3dB.
Tuned to the upper end of the possible range you see that it behaves more than a lowpass than a bandpass ;) The passband gets a bit wider.
pipecap 5mm probes 10610MHz
pipecap 5mm probes 11815MHz
I would assume that it makes most sense to design the probes beeing quarter lambda for the frequency were the notch of the filter appears (or slightly above). Since this depends on the frequency you want to tune the filter to you need to consider that before you make the filter.
Triggered by some forum discussion about interference risk of operating repeaters in neighbor channels in 6.125kHz channels I was curious about the TX spectrum of my IC-E92. The measurement was done radiated. I checked for neighbor channel interference. You can see the delta values bottom right.
IC-E92 FM wide 1750Hz
IC-E92 FM narrow 1750Hz
D-Star (only half span screenshot):
You can see that D-Star is really the most narrow band mode. Using neighbor frequencies at close QTH still cannot be suggested but at least the interference will be less than for narrow band FM.
The measurements show about 0.4dB insertion loss. Do not expect to much suppression. As a simple RX image reject filter or for multipliers its certainly a good option.
S-Parameter for filter tuned to 5720MHz
However for my construction it seems that the rejection above the tuned frequency is not very good. According to the paper of Paul it seems i have choosen for too long probes.
S-Parameter for filter tuned to 4375MHz
[Update1 – wideband measurement]
DL2MRE provided some extended measurement of the filter up to 14GHz.
I have to say, that it looks better than i thought. I was expecting worse suppression at high frequencies due to radiation but it is ok. The resonance at about 13GHz might come from the 90degree SMA connectors. However the bandwith is rather large and has an unexpected shape. Might be some parasitic effect ? Thanks to Marcel !
pipe cap 6cm wide measurement 4-14GHz
[Update2 – unwanted response and tuning range]
Marcel made some more pictures with the filter tuned to different frequencies.
It can be seen that there is some unwanted respone of -15dB around 7GHz. This stays more or less static with respect to the frequency the filter is tuned to.
From above you see that the coupling probes have 11mm length. 11mm x 4 in wavelength equals about 6.8GHz in free space. So direct coupling between the probes is the most probable reason.
Filter tuned to 4380MHz
Filter tuned to 10368MHz
Another outcome was that the tuning range of the filter is very large from about 3.4GHz up to 12 or 13GHz.
Because of the static unwanted filter response i decided to open the cap again and shorten the probes to about 5mm (so about 15GHz). As described at W1GHz this has significant impact to bandwith and insertion loss of the filter. Now it is rather hard to tune it to the correct frequency. I will post some measurement results later.
A evening of TX. The picture was made by DK1RS in a distance of 80km and shows my signal (horizontal line) during about half an hour. The crossing lines are reflections from aircraft +- Dopplershift. So far signal raises about 10dB compared to the direct path between Rainer and myself. WSPR cannot decode if the Aircraft Reflection is present because it cannot deal with the Doppler drift.