Thanks for your response. I do have one more question. We monitor these ammonia screw compressore 4 times a year for a certain company. While it is pretty easy to determine the lobe pass frequency in these units, we have no access to the bearing model or manufacturer. Our customers do not know the bearing numbers and the compressor manufactures will not give them out (they only give part numbers) so we cannot identify the possible fault frequencies. One month we collected spectral data out to about five times inner race frequency of some of most common bearings that may be in the unit. There was quite a bit of nonsynchronous energy on the spectrum in between the lobe pass frequencies -(we covered three times this frequency). We told the customer that due to this much nonsynchronous energy on the spectrum, there may be a problem with a bearing. Three months later we checked the unit again and all that appeared on the spectrum were the lobe pass frequencies 1, 2 and 3 times. All nonsychronous energy was gone! The customer had done nothing to the compressor since last check. It was the same unit. Can you explain what might be going on? Is there something we can watch for to be able to tell our customer if there is truly a problem? We have also had compressors fail that showed no indication of a problem. Thanks for your help.