The Nanuc 500MHz Cold Probe
by Deryck Webb
Varian 500MHz Cold Probe #80 is now installed on the NANUC 500MHz system (Figure1). The expert installation was facilitated by Robert Short of Varian Inc. (Fort Collins, Colorado). The installation went very smoothly with two notable exceptions. The first of which illustrates the need to observe and understand all the systems involved with the cold probe, and the second, which, when found and rectified, gave the user a better understanding of the sensitivity of the components involved.
Figure 1 (Click to enlarge)
Robert Short was scheduled to arrive on Tuesday the 18th of May and the plan was to have the probe evacuated on its bike rack by the time he arrived. On the Monday the probe was unpacked and attached to the pump by 10am. The valves were closed up to the probe to begin evacuation of the vacuum line and everything seemed in order. The pressure gauge was not immediately activated due to its sensitivity to higher pressures that may damage the filament designed to measure extremely low pressures. After 3 hours of pumping the gauge was activated and an indicated pressure of 3.3*10-6 mbar was noted.
The valve to the probe was then opened and the pressure began to slowly increase. We assumed that the vacuum loss was due to pressure from the probe being added to the system. It was thought that the vacuum would plateau and then continue down evacuating the entire system including the probe. Therefore the system was allowed to evacuate overnight. The following day the gauge indicated a high-pressure error (E - 09).
The gauge was reset under the assumption that the error was induced from pressure introduced by connecting the probe. The gauge displayed a reading of 4.2*10-6 mbar and was again observed going up slowly until after about 4 minutes a high-pressure error (E - 09) was again observed. The probe vacuum valve was closed and the junction box valve was closed as well to allow the vacuum line only to evacuate again. There was a fear that the probe was not able to hold a vacuum, however when the gauge was reset again it was found that even the vacuum line itself could not hold a vacuum!
Figure 2 (Click to enlarge)
Robert Short arrived and confirmed that the turbo system on the pump was not engaged. He observed that the operating speed of the pump was much too low (12kRPM vs. ~56kRPM, nominal) and the operating current was much too high (I=1.62 Amps vs.0.86 Amps, nominal). His first thought was that the pump's bearings had dried out since the last time the pump was operated in January. The solution ended up being much simpler than a leaky probe or dried bearings. In the end, it was determined that a vent switch at the back of the pump was inadvertently left open (Figure2). The vent, which allows efficient releasing of the vacuum, was opened during the previous system disassembly and not closed after the system was disassembled or prior to engaging the pump. The vent switch is not identified in our current manual and was not referenced in the start up or shut down procedure for the cold probe. This vent was closed and the vacuum line was immediately drawn to a high vacuum. The complete system was opened to the vacuum and, when an appropriate vacuum (2*10-6 mbar) was achieved, the system was cooled.
After 24 hours the cold probe was stable at 25K and Robert began to test the probe. All of the probe specifications were being met and/or exceeded (Table 1). One anomaly we noted was that the pump pressure gauge indicated a vacuum of 1.3*10-5 mbar. This is much higher than the normal operating pressure of ~10-8 mbar. It was not clear yet why the vacuum reading was so high.
One procedure that was not previously done was the cryo_noisetest that ‘burns' impurities off of the RF coils by sending high power pulses over multiple channels. The effects of impurities on the RF coils was not apparent until the 13C 90° pulse width test. The pwx parameter was arrayed from 0 to 152.6 and produced the data array seen in Figure 3.
Figure 3 (Click to enlarge)
The test indicates that high pwx values cause the probe to arc. This behaviour is indicative of contaminants on the active coil. A second cryo_noisetest was performed in an effort to remove the contaminants however this and several subsequent efforts failed to correct the problem It was determined that the vacuum was insufficient to draw away any of the vaporized contaminants. The problem was localized to a leak in the connection between the probe vacuum valve and the aluminum box valve (Figure 4). The valves were closed and the coupling was disassembled revealing a small paint flake lying across the O-ring and likely causing the leak. The O-ring also appeared worn so it was replaced.
Figure 4 (Click to enlarge)
The fittings were reassembled and then opened. The vacuum pressure improved immediately. After running another cryo_noisetest the 13C pulse width array was perfect (Figure 5). This situation gives rise to an interesting question That is: with a leak occurring at the position we determined, would atmospheric contamination travel toward the pump, or toward the probe and the RF coils? We speculate that a pressure differential may be biased toward the probe due to the low temperature and the potential for cryo-pumping. If our speculation is borne out then this leak will introduce new contamination to the coils that will be turned to plasma during the noisetest and will simply re-adhere to the coils as soon as the test is over.
Once the high vacuum was restored the contaminants were drawn away from the coils thus eliminating the arcing.
Figure 5 (Click to enlarge)
The whole installation was very informative and we at NANUC would like to thank Robert Short for all of his help and insight.
Since installation was completed we have run 3 ,3D spectra on a 1mM ubiquitin sample in just 3 days. We are now excited to offer cold probe experiments to all our users. Please watch for a cold probe spectra in our spectra of the month.
|Table 1 - Calibration Tests and Results
|1H Signal to Noise
||0.1% Ethylbenzene in CDCL3 Wilmad 535-PP
|1H Signal to Noise
||2mM sucrose in 90% H2O
||Completed on July 23, 2003
|1% CHC13 in acetone-d6
|1H 90° pulse width
||tpwr = 57
|1H RF homogeneity
|Doped 1% H2O in 99% D2O
|13C 90° pulse width
||tpwr = 60
|13C RF homogeneity
|15N 90° pulse width
||tpwr = 58
|15N RF homogeneity
||Doped 1% H2O in 99% D2O
|gradient recovery @ 90%
||Doped 1% H2O in 99% D2O