Karl-Heinz Sulanke (Desy Zeuthen)  provided five samples of a RALTRON model RTXT 16.80000 MHz oscillator.
The oscillators were measured for at least an hour, with frequency samples taken every five seconds.
The frequency samples were analyzed in blocks of 100 samples, and blocks of 500 samples,  using our standard Allan Variance code.
 
 Sample Total of samples in data file Best Allan Variance for blocks of 100 samples Worst Allan Variance for blocks of 100 samples Allan Variance for blocks of 500 samples
1 of 5 1000 4.82377e-11 5.71838e-11 5.29407e-11
2 of 5 1117 
640
28844		 6.68643e-11
9.29478e-11
5.12798e-11 		 7.84872e-11
4.02818e-10
1.54877e-10		 7.06435e-11 
2.00688e-10
5.70214e-11
3 of 5 1000  1.57479e-10 3.85815e-10 1.81410e-10 
4 of 5 1000 
10,000		 5.76400e-10 
3.64539e-10		 3.01120e-09
2.15924e-08		 1.16787e-09 
4.37509e-10
5 of 5 1000 5.08528e-11 2.35990e-10  6.83699e-11

In any particular table row, the upper entry is earlier, the lower entry is later.

Results of #1, #2 and #5 suggest that some percentage of  the RTXT part could be good enough for IceCube.

Short term stability appears to improves (toward an asymptote) as the parts accumulate more time under power..
For these samples, the integrated operating time was not recorded, unfortunately...

Statistics are poor.
Note that #4 didn't improve to the level of acceptability.
Note that #2 doesn't appear to settle down.
When time permits, both #2 and #4 will be burned in longer and retested.

Data files are available on request.  Email me...

Measurements and analysis by Gerald Przybylski, Lawrence Berkeley National Laboratory, for IceCube.