SDN0007.06 – GNIRS Cryocooler Selection
Revision A

1. Introduction

 This document summarizes the criteria used to the select the cryocoolers used for GNIRS, namely four Leybold RG5/100 heads.

 2. Selection Criteria

 The most important selection criterion for the GNIRS coldheads was that they be capable of providing enough cooling for the instrument. Studies (SDN007.02) showed that even with a liquid nitrogen pre-cool, substantial cooling power is needed to get from 80K down to ~60K, and that even with high capacity heads four will be needed.

 Three cryocooler models were identified as having sufficient capacity, all from Leybold. These are the RG5/100 heads and the Coolpower 130 and 150 heads. Note that the Coolpower 150 heads are single-stage heads, and thus one must use at least one of the others to provide detector cooling. Subsequently, Leybold has come out with another single-stage model, the 120T.

 Although the RG5/100 heads have the smallest capacity, the impact on overall cool-down time is not very large, and any four of these heads would perform satisfactorily.

 Therefore, we considered additional criteria:

 ·        Cost. The RG5/100 heads are the cheapest, and the Coolpower 130 heads are the most expensive; the single stage Coolpower 150 is in between.

 ·        Vibration. The RG5/100 (and 120T) heads use a drive with lower vibration. The vendor can add a low-vibration option for further improvements. This is inexpensive; it is surely cheaper to do this than to prove that isn’t required.

 ·        Reliability. There are several aspects to this. First of all, the RG5/100 heads have a higher lower limit on the helium gas input temperature, compared with the Coolpower heads. The difference is 50F vs. 40F, but the evidence suggests that in practice the Coolpower heads will run under ambient conditions, while the RG5/100 heads will require keeping the gas above 50F. It should be noted, though, that the experience with the RG5/100 heads has been with Phoenix, whose structural temperature is not controlled, and which therefore is quite sensitive to cryocooler efficiency. An instrument with temperature control might be more forgiving. In any case, the gas temperature problem must be solved for Phoenix, so at least one reasonable solution will exist when needed.

          A second aspect of reliability is time between failures or maintenance intervals. There is not enough experience at the telescope with any of the heads to provide a clear preference. The people at ESO have developed maintenance procedures for the RG5/100 heads, which are claimed to eliminate failures, and NOAO will (probably) implement these for Phoenix, which will use the RG5/100 heads.

 ·        Future availability. The RG5/100 heads are widely used in medical imaging equipment,
 whereas the Coolpower heads are much less extensively used, and the manufacturer does
 not seem particularly committed to maintain two parallel lines of heads.

 ·        Drive electronics. The RG5/100 and 120T heads use much simpler drive electronics than the Coolpower 130 and 150 heads. They cannot be phased, but there is no requirement to do this. It should be noted that there should be some dependence on the power frequency for the RG5/100 heads (50 vs 60 Hz) which may be relevant if GNIRS goes to Chile. (One would want to run the heads at 50 Hz in the instrument, in the lab, before shipping.) The simpler electronics have lower cost, heat generation, and weight.

 ·        Spares. Of the Gemini instruments, NIRI is using the Coolpower 130 heads, Phoenix is using the RG5/100 heads, and the mid-ir instruments are using something completely different. If GNIRS goes south, it will need its own spares on-site, unless it shares with Phoenix.

 3. Selection

 Of the secondary criteria listed, four favor the RG5/100 heads, one (reliable low temperature operation) somewhat favors the Coolpower 130 (and 150) heads, and one (spares) is neutral. None of these considerations absolutely excludes either head type. We therefore have selected the head favored by the preponderance of criteria (and consistent with “build to cost”), which is the RG5/100.

4. Spares

 Some means of dealing with cryocooler failures or preventive maintenance is required. What we have done up until now is to replace heads in pairs if one fails or after suitable intervals of use; the heads removed are then reconditioned and become the new spares. If the ESO procedures are tried and shown to work (be transferable), then one might get away with having the appropriate maintenance kit rather than complete heads.

 Gemini might, if they wish, purchase two spare heads now, which could then be used as spares during GNIRS testing at NOAO and then follow the instrument to the telescope. (NOAO would deal with reconditioning if they were used while at NOAO.) This is clearly preferable from NOAO’s point of view.

 Alternatively, one could wait until NOAO gains more experience with maintaining the RG5/100 heads and then decide what level of spares (complete heads vs. parts) is needed. This would require a decision in 6-9 months in order to ensure that GNIRS has needed spares during testing at NOAO, since NOAO’s own spares will be supporting Phoenix for individual targets.