2 Nov 2001
W. A. Traub

SHORT DELAY PRESSURE SETTINGS
-----------------------------                   

I measured optimum pressure settings for SD-1/2 in early July.  The units are 
pounds per square inch (psi), which can be read directly on the gauges near the 
control knobs.  The measurements were made with air in the tank, as follows:

         Location     Pressure range (air)   Pressure range (vac)
            
         SD-1 top     63 - 100 (best = 70)    est. best = 70  
         SD-1 side    55 - 100 (best = 60)    est. best = 60  
         SD-2 top     45 - 46  (best = 45)    est. best = 45, maybe 36 
         SD-2 side    42 - 46  (best = 43)    est. best = 43, maybe 34  

The minimum pressure was found by noting at what point the force needed to pull 
the cars started to increase, as pressure was reduced.  The minimum force, at 
pressures abobe the minimum, leveled out at about 0.1-0.4 kg for SD-1, and 
0.1-0.2 kg for SD-2. The measurement was made with a spring scale, over a range 
of about 50 cm, near home, and the worst case force was noted.  It may in fact 
be worse farther down the track, but I could not measure there.

The carts should *not* be run at low pressures which might allow dragging of the 
air pads to occur, because this will cause deep grooves to be cut in the granite 
surface, and will increase the oscillation dramatically.  We know this from 
experience.  The way to be sure that we are not rubbing on the granite is, 
ideally, to measure the current needed to drive the cart, and to be sure that 
the current is never greater than some value, which we still need to determine 
empirically.  If someone can set up these tests, I would be glad to discuss this 
with them; we have an ADC card which sould be used to monitor this current, or 
perhaps there is a better way.  Ultimately it should be logged as a housekeeping 
parameter.

The maximum pressure is that at which the cart starts to vibrate, at about 60 
Hz.  At pressures above this point the vibration increases linearly.  The 
vibration was measured by looking at the oscilloscope display of the error 
signal, so it is a measure of the OPD noise, which is the aspect that we care 
most about.  The noise was measured when slewing the SD at 1 cm/s.  Note that an 
error beyond the range of +/-5 volts (I believe) on the oscilloscope is bad, 
because at this point the error rolls over and we pick up a permanent path error 
of several (10 ?) microns (MGL please fill in here).  The error signal can be 
digitized with the ADC card and logged.  Both this and the drive current should 
appear in a little graph that the operator can see on the VxW monitor, and a 
flag set if either value goes out of bounds.

If vibration is a problem at all pressures, possibly especially on SD-2, then 
the magnetic pull-down force should be adjusted to minimize vibration.  This is 
done by turning one of the 3 jack screws on the top (usually) of the SD 
platform.  The dihedral will have to be removed to do this (4 screws from base 
plate to cart top).  About 1/4 turn, either CW or CCW, should help.  Try to mark 
where you start, so you don't get lost.  A full turn is probably far too much.
The reason for this problem is that the original adjustment decays, due to a 
poor locking mechanism on the jack screw; if this worked better, we might not 
lose the adjustment.
               
In vacuum, the atmospheric pressure of about 670 torr or 9 psi is removed,
so the proper flotation pressure should be less than in air by this amount.
However I have not tested this in the present set-up, so the safe way to proceed 
would be to start in vacuum with the same pressure as in air, and reduce the 
pressure cautiously watching both the drag and the error measurements.