Viewing page 174 of 250

This transcription has been completed. Contact us with corrections.

• 134B is not picked up (apparently) because it can result in a runaway when a joint is backdriving. Tests at Montreal will be performed next week on the hardware. If this malf isn't picked up, they'll pull the SPA's to fix it. [what else could they fix while they're in there?]
•  check is (1) entirely independent of encoder checks, and (2) picks up all encoder failures except encoder failed to last value (constant) - it picks that up when cmd is removed (check can have "easy" fix = to accumulate integrated tach data as well as encoder data - then it would pick up all encoder failures... but it would set brakes as well.
• deceleration = backdrive
• if joint is not moving = forward drive
• check is apparently not sensitive to nominal / off-nominal parameters ( check this in runs)
• different parameters are used for loaded / unloaded cases (we'll use 32K PL)