You might like this document that reverse engineers how the four tones are used to transmit a wspr beacon of 162 symbols of information at 1.46Hz shift. Each tone lasting for 0.683 seconds. It takes 110.6 seconds to send out your position.
However convenient, not necessary to use a GPS reference clock if you have a good TCXO (or the NTP calibration as James Peroulas did in RPi). I done a few things in WSPR and used for a few years in various ways.
And to go into “black magic details” not only the long term stability what matters but also the very short term fluctuation too due to the extreme narrow band modulation. I had very interesting conversation once with Hans Summers from QRP Labs and he described his experiments with his OCXO (oven compensated) for 2m and above WSPR bands.
Yeah, I know 2m and 70cm is quite exotic now in WSPR. However building a good transmitter not necessary helps making it more popular as still the receivers needs to be compliant (hopefully general ham radios can do the decoding with their TCXO references but I never dig deeper into this topic).
I guess the trick is to know how good is good enough. How rock-solid does your frequency reference have to be.
By the way, a scrap HP 435B wattmeter (without the expensive RF sensor) has in it a very stable OCXO running at 50 MHz. Not only is it a rock-solid 50 MHz, but it’s precisely set to 0 dBm. It’s totally worth the $20 for that part alone.
I restored a little fancier HP438 by fixing its power supply and fan. The HP8482H sensor was not cheap.
I had thought of using the L-band dreamcatcher for the gps part of a GPSDO since it had all the saw filters near the gps signal. (maybe I should have posted this in the forum on the DC V2.03). But the purpose was to generate a reference 10MHz with 1 pps accuracy. I have access to some test equipment that can accept a 10Mhz input.
I kinda put it on the back table. But it brings up the question of entering a correction into the moRFeus. (or using it as the reference signal). I didn’t see anything in the menus for building a table of correction values… or to do it dynamically.
I guess in mixer mode I could determine a correction when I applied a calibrated reference signal. Then at least I would know the correction factor at that frequency. Since the moRFeus has several internal Freq bands based on the desired generated frequency, it would imply a correction factor that followed the step changes in the internal oscillator.
So the real question is, Are the various step bands within the moRfeus known and are they fixed ?
UPDATE: Do clarify my question — I thought that as the vco requested frequency was changed, that it had to recalcuate it’s internal dividers and known temperature compensations. I thought it might be in fixed steps.
Yes, we talked about re-purposing a DC2.03 into a GPS receiver for frequency reference. I had found the reference, but then never made one. I was too busy with getting my DC3.0 going.
Are there step bands in moRFeus? I got the impression is it was a continuous broadband tune.
Got it to work this morning on Linux Lubuntu 32 bits applying last README.
Compilation was very long but successful.
On the Raspberry, I tried again but still unlucky. Compilation of hidapi took one hour (not a joke), and succedeed, but still stuck at launch :
~/moRFeus_Qt $ sudo python moRFeus.py
Traceback (most recent call last):
File “moRFeus.py”, line 179, in
device = window.initMoRFeus()
File “moRFeus.py”, line 51, in initMoRFeus
device = hid.device()
AttributeError: ‘module’ object has no attribute ‘device’
“try.py” python script from hidapi sources says it’s OK :
python try.pyPreformatted text
It a pleasure, this is my first solo project(5th day with Python) and I find it quite fun to play with this device, I have some pretty awesome things I am incubating atm…
You are more than welcome to ask for some help, I would gladly provide it
This is the current version that I would push soon.
Yes we know for the article we know we are famous also
In the meantime I pushed two hours ago the latest release on my github repository .
It’s well tested on Linux, RPi platforms with automatic install.
No plan to add more feature at this moment, reached my goal.
Edit 26/04 : once again we have an article about moRFeus on rtl-sdr.com blog . And as usual Ohan and myself are sharing the same post
For the next 24hrs, you can access the live demo of moRFeus_GUI in real situation : installed on Raspberry Pi3 , moRFeus device running, and link established to GQRX running in background (but not playing).
From a linux computer (running X-session, not from ssh console) run from terminal console :
(from the above line you have to replace \aat\ by @ and /d0t/ by you know what )
answer ‘yes’ to the SSH security question and password is Outernet (with O = upper case).
Note 1 : running from remote location (with my poor upload b/w) will bring latency, and it works faster in real situation.
Note 2: I don’t know if 2 or more people can connect at the same time ! Take a ticket if busy, thanks.
Note 3: Windows users I’m sorry but you have to install third-party software I remember Xming can do it.