The linearity of HxET allowed the back-calibration of all SLR measurements from 1994 recorded using Stanford counters.  A comparison of the timer measured range is shown below for three Stanford timers.  Two timers show significant error characteristics within the 60 ms range, yet the third timer shows close to linear measurements. 

Effect on LAGEOS data 1994-2006

Using the appropriate counter along with the actual calibration range given in the ILRS normal point header of SGF LAGEOS data for the period 1994-2006 and estimated correction in mm to calibrations is calculated.  From the Figure right, it is apparent that errors of between 5 and 8mm have been made to the calibration values. However, given our estimate of the uncertainty of these average values, we finally derive an average calibration error of 7±2 mm, and in the sense that the calibration correction is too large by that amount. 

During this re-assessment we also discovered that no account had been taken for the effect on total delay of a glass neutral density filter that is placed in the optical path during calibration but not during satellite ranging. This correction amounts to 1.5mm, again in the sense that the calibration correction derived from target-board ranging is too long. 

Therefore our calibration corrections in the period 1994-date are too long by 8.5±2 mm and thus calibrated satellite ranges short by the same amount. This correction, which affects all satellite data equally, is of course in addition to the range-dependent correction discussed under ‘previous calibrations’ above and announced for the period 1994 October to 2002 January in SLRMail 0891 in 2002 January.