Trouble-Shooting Inclinometry

The document Taking Inclinometry Data describes how to obtain sufficient data from which to create a new track model. The process is fairly automated and normally reliable, and I assume that the steps in that document are being followed. This document attempts to aid in troubleshooting known problems. In the event of new problems you may be best advised to contact appropriate technical staff.

The expected sequence of starting up and testing the system is :

• icl> incline_load
• icl> send incline obey log test 1 -- record the output from all channels each second and put the results into test.log .
• icl> send incline cancel log -- stop the recording
• icl> $ type test.log -- display the results

SOLUTIONS

1. Interrogating individual channels .
   icl> send incline obey command "#1RD" ~ or ... "#2RD" etc.
   • The returned value will take the form, eg, #1RD+00072.10A5, where the last two characters are the checksum, and +00072.10 is in mV. A failure will be more obvious and indicates a particular problem with that channel - maybe go see if the inclinometer is actually plugged in and that the gain setting is correct. Data channels after number 9 are in hexadecimal and should be entered into this command capitalized : eg . . . "#ERD" to interrogate channel 14.

2. Resetting the DGH units .
   
   • The DGH units are housed in the 3 small gray boxes near the inclinometers, i.e. on the A-frames and below the TMU, and are fed power via the breaker box mounted near the center of the antenna floor. This box also hold breakers for the antenna drives. Cycle breaker switch number 14, marked Inclinometry.

     NB : power-cycling the RS485/232 converter box (in the computer room, located on the right side of the aisle at about thigh-level) may well cause the DGH units to 'hang'. There is NO documented reason for power-cycling the RS485/232 converter at this time.

3. Restarting the terminal server port :
   1. Log onto mwttel
   2. Type "mc ncp conn node jcmt2 via isa-0" at the $ prompt.
   3. Hit enter
   4. At the # prompt type "boojum"
   5. Type anything for the username ("x" is fine)
   6. at the local> prompt type "set priv"
   7. enter the password (ask the telescope support specialist)
   8. Type "show port 2 status" to see the port state - (port 8 as of Apr 2000)
   9. Type "log port 2" to log the port out - (port 8 as of Apr 2000)
   10. Type "lo" to log out of the terminal server
   11. Type control-d to get out of ncp
   12. Type "lo" to log out of MWTTEL

4. (more seriously) rebooting the DECserver .
   
   • This may be necessary if general communications have at some stage be cycled or, eg, hit by power failure, but the reboot will also affect other subsystems such as the weather station - so beware of subsequent TEL problems.

If each channel contains sensible data and scatters are < 5mV or so in each column then all is OK, and you may then proceed to take data explicitly, as described in "Taking Inclinometry Data" .

Problems in data quality alone (noted by comparison with previous datasets) may imply

• loose mountings
• electrical or mechanical jolting
• mechanical failure (or imminent failure - as indicated, for example, in Feb 1998 by a ripple in the A-frame difference data which was diagnosed as due to a failing bearing.)