Date: December 13, 2004 Subject : picomotor paddle The picomotor paddle does not have a good zero point. Here are the reverse engineering the paddle The slider is a 100K ohm linear pot, It is ballanced with a 200+ ohm resistive circuit The three channels are similar. We can re-do the control with rotating pots or with push button switches - Fast+ slow+ slow- fast- I don't know how big the dead band should be but this type of box should not be hard to make - if we want it. How slow and how fast do we want to go if we have buttons? Date: December 3, 2004 Subject: Operation note The long delays have a laser does not enter at exactly the right height. It appears that it should be a little higher. To fix this the laser, the pointing mirror, the interferometer, the receiver and the lens should all be raised slightly which would take the system offline for a couple of days. As a temporary fix we can reain the injection mirror. I found that, with an allen key in the vertical adjustment screw: 7 o'clock is correct for home 6 o'clock is good for 0 to 21 meters 5 o'clock is good for 20 to 25 meters 4 o'clock is good for 21 to 28 meters As an indicator - at home you should adjust the return spot down until the signal begins to drop. You should repeat this when the cart is between 20 and 25 meters if you wish to go beyond 25 meters. Date: October 25, 2004 From: Marc Lacasse Subject: operation notes A few comments on operation Cycling power A) The Lincoln Lab Star Tracker is currently operating at a lower temperature. In order to do this we need to circulate cold water through a fixture at the back of the cold finger. The switch above the control console controls power ti multiple items on the star tracker - The electronics, the TE Cooler, the Water chiller. The compressor on the water chiller does not like to have power cycled quickly so I suggest that the power Star tracker be turned off at the Power Supply above the Star Tracker CCD unit. B) It appears that the VME system does not always reinitialize properly when the power is cycled. I suspect that the reason may be that some power comes in through the Star Tracker and the Picnic camera if these items are left on. I propose that these electronics be turned off whenever the power to the VME is turned off. This got the system back operating at least one time when repeated cycling of the VME power would not clear the state of the electronics. If you do a reboot and have to correct the file system Answering yes to all questions about block is accomplished by starting with fsck -y /dev/dsk/c1t?d0s? With the appropriate numbers in for the "?" The disk partitions are (as of October 2004) /dev/dsk/c1t1d0s3 6193583 3102625 3029023 51% /usr/local /dev/dsk/c1t2d0s2 35009161 21672732 12986338 63% /home/iota16/d1 /dev/dsk/c1t1d0s4 10323610 8133763 2086611 80% /home/iota16/u1 /dev/dsk/c1t1d0s5 5883722 3393174 2431711 59% /home/iota16/u2 /dev/dsk/c1t1d0s6 5895278 34 5836292 1% /home/iota16/u3 Next error- 11/2/2004 Observers tried to find fringes before the delay line was tracking fringes. There was a crash and part of the tornado target server and registry were corrupted. doing launch & gives a window with only 2 PPC's vxworks-start or daq-start both generate a wtx error registry not found Note from Gary on the fix: To fix your tornado windshell problem (you could windsh to iota-ppc1 and iota-ppc2, but not iota-ppc0), I killed off the 3 target servers and the registry server, and then restarted them. To kill off the target servers, find the tornado processes running tgtsvr.ex (there should be 3, one for each ppc) and use kill -9 PID. to find the process type in ps -aef | grep svr For the registry server, find the tornado process running wtxregd.ex and use kill -9 PID. to find the process type in ps -aef | grep regd To then start the registry server and target servers, just run /etc/init.d/tornado.start. (Probably needs to be done as root)