Test of the first sample of hte production splitter made by Nicolas
using the slow monitoring on the old TSFi at DCH-DCT teststand:

(A) = TSFi # 3545 in Slot 5, logical slot 1 [the prototype]
(B) = TSFi # 3543 in Slot 7, logical slot 6

Opt_Pow :  Total (V)    Short end (V)   Long end (V)    Efficiency (%)
TSFi:   (A)     (B)     (A)     (B)     (A)     (B)     (A)     (B)
Fiber:
  1     0.98    1.06    0.33    0.37    0.33    0.35    67 %    68 %
  2     0.31    0.35    0.12    0.12    0.10    0.10    71 %    63 %
  3     1.18    1.35    0.41    0.47    0.43    0.45    71 %    68 %

Note: 1) Same signal seen by different TSFi can vary 5-10%. 
      2) Long/short leg yield after splitter very close to 50:50.
      3) The total efficiency is rather poor. The test splitter borrowed
         from Ray is more like 90-95%. 

Comparing the earlier measurements on fibers by Sang-joon and Xuedong
using optical power meter: 
Fiber   Meter   mW/volt    
  1   0.66 mW    0.67 
  2   0.33 mW    1.00
  3   0.76 mW    0.64
Channels light yield may have deteorated. Only taking fibers 1 & 3,
a rough calibration of 1 volt TSF slow mon = 0.67 mW optical power.
 
For production system: 
   Oct/02           TSF Mon 
   Meter test (mW)  Volt    mW/volt
   Y0   0.692       0.90    0.77
   Y1   0.724       1.04    0.70
   Y2     ?         1.33
   Y3   0.398*      0.75    0.53
   Y4   0.646*      0.96    0.67
   Y5   0.676       1.08    0.63
   Y6   0.724       1.14    0.64
   Y7   0.776       0.35    0.22#
* Sangjoon's Oct/02  missed Y4. I suspect that Y2 was actually missed
  while Y3,Y4 measurements were mislabelled as Y2,Y3. 
# Old short fiber replaced once. TSFi receiving conversion suspicious