Viewing page 71 of 170

This transcription has been completed. Contact us with corrections.

[left margin]RMS[/left margin]
consistency check problems: constants were selected & verified based on SIMFAC/ASAD results. SIMFAC/ASAD model not completely correct (2nd order terms left out... noticed after comparison to systems tests). With improved SIMFAC model, present parameters produce false alarms [particularly in SINGLE/COARSE, though also in MAN AVG... apparently not AUTO (...?...)]. ~80% of false alarms are produced by the "encoder" part of the check (could be fixed by changing 6 constants); ~20% produced by envelope part of the check (an additional 18 constants — including the filter term 2N->4N for each joint) => all are level c, except filtering, which are design constants.

DR exists against 18.11 consistency check: IBM has to re-code and re-verify anyway. 6 constants same amount of work as 24 constants. Aldridge supports constant change; SPAR knows constants => proceed with sim

ORB PCL - MPS buried?
Sound level meter