GPS Tracking with a modified TCXO RTL-SDR
Michele from Michele’s GNSS blog has posted his results with using a modified R820T RTL-SDR with Temperature Controlled Oscillator (TCXO) for GPS reception and decoding. The RTL-SDR is capable of tracking GPS even without TCXO but improved performance can be expected with a more stable oscillator. He notes that the R820T with it’s 3.57 MHz IF is ideally suited for GPS reception when combined with an active GPS antenna. Using this setup he was able to track GPS satellites and the Galileo E1B/C GNSS satellites as well.
Michele modified his R820T RTL-SDR with a 28.8 MHz TCXO he obtained from a friend. It is however possible to purchase modified TCXO R820T dongles directly from the 1090mhz webstore.
Also found that a blue LED connected in reverse between “Hot” end and Gnd will work to some extent. Halfway house if you can’t obtain a modded dongle but it worked for a while at least.
You might have to try a few diodes as their reverse capacitance can be variable, for me the first one out of the strip worked but capacitance is quite critical. If you overdo it then the dongle will stop working completely.
Dont misunderstand me. The TCXO solution sounds really good. Capacitor compensation is the poor mans solution. My local 2 meter repeater and Heathrow Tower ATC came in so far off frequency I could not find them. I had a pack of trimmers so I tried one. The end result was better than nothing and drift was reduced.
A cheaper alternative to the TCXO is to use capacitor compensation. The XO normally runs high in frequency allowing a small ceramic 2 to 10pF trimmer to be fitted across the crystal to reduce its frequency. With careful adjustment the frequency error can be reduced. Fitting the capacitor also gives an improvement in stability. The improvement is definitely worthwhile for for putting telephony channels on frequency and would make GPS easier to handle as well.
Capacitor compensation will correct the frequency error, but not the thermal drift. Thermal stability is the whole point of a TCXO.
Correct. However the end result was more stable. The added trimmer was NPO so the added capacitance would swamp the effect of any circuit capacitance which was either negative or positive coefficient, I did not note which. To misquote Tennyson “Mine is not to question why. Mine is just to do and try”