MSi001 dongles. Frequency vs temperature.
(Jan 24 2014)
This page shows (among other things) how to make a dongle frequency stable and should apply to rtlsdr and any other small radio.

It is a sub-page to The Mirics MSi001 with Linrad. The Mirics MSi001 chip is used in the FUNcube Dongle Pro+ and in the Logitec LDT-1S310U/J USB dongle. These two dongles are compared below.

Frequency stability vs temperature.

To avoid the influences of room temperature changes, the dongles were buried in a 10 litre bucket filled with sand. In that way a very long time constant is provided for the dongles without degradation of the cooling.

By placing a light bulb above the bucket and a temperature probe in the sand near the dongles, it is possible to evaluate the temperature dependence of the frequency. Figure 1 shows how the frequency changes when a 25 W light bulb on top of the sand is switched on causing the temperature to rise by about 13 degrees in 4 hours. Figure 2 shows what happened after the 25 W bulb was replaced with a 70W bulb that made the temperature rise to 53.6 degrees.


Fig. 1. FUNcube PRO+ (left)and Logitec (right) dongles on the D5400XS computer. The temperature is rising from 25 to 39 degrees Centigrade in about 4 hours.


Fig. 2. FUNcube PRO+ (left)and Logitec (right) dongles on the D5400XS computer. The temperature is rising from 39 to 54 degrees Centigrade in about 3 hours.

Figure 1 shows that the temperature coefficient goes through zero for both dongles at about 20.30 at which time the temperature was about 33 ºC. Table 1 gives temperatures and frequencies for the two dongles.



Time    Temp     FCDPP   Logitec
        (ºC)     (Hz)     (Hz)
18.28   25.96    101      861
18.52   25.97    101      861
19.27   28.17     95      887
19.48   29.89     93      902
20.18   32.09     91      917
21.11   35.03     92      923
22.20   37.85     92      909
23.22   41.71     94      868
00.09   46.65     92      722
01.08   51.38     80      477


Table 1Time, temperature and frequencies. The actual temperatures for the dongles may differ from each other and from the temperature at the temperature probe tip.
Table 1 which is extracted from the figures shows that the temperature compensated oscillator in the FUNcube PRO+ is about 10 times more temperature stable than the Logitec dongle.

Figure 3 shows the frequency stability of two different Logitec tuners buried in sand with a 25 W light bulb at some distance giving a temperature around 33 ºC. There is no thermostat and the temperature drifts slowly. It would be trivial to make the temperature stabilized at the point of zero temperature coefficient for one or the other dongle. That point differs by about 2 degrees between these two dongles.

It is quite clear that the stability is fully adequate for EME (moonbounce) with the JT65 protocol and that the close in sideband noise is low enough not to degrade detection. The dongle at the right hand side has about twice as much sideband noise as the one on the left hand side. The dongles differ a little, but both of them are perfectly adequate. Note however that instabilities on the 5V supply from the USB connectors easily degrades stability to the extent that JT65 detection is degraded even if the dongles are temperature stabilized


Fig. 3. Two different Logitec dongles on the D5400XS computer.