M1 / Primary Mirror Systems

There are 91 mirrors, each with its own positional feedback system and means of mechanical control. Each mirror has 6 active sensors, 6 passive sensors, and 3 linear actuators that work together with a “good stack” to maintain an aligned sphere, despite changes in mirror positions caused by thermal expansion, contraction, wind and instrumentation. The systems used for this are SAMS and SCS.

SAMS (Segment Alignment Maintenance System) provides information about the relative positions of mirror segments with respect to each other through the use of edge sensors. It calculates the corrections needed to maintain or restore an alignment that was saved in a previous reference, preferably a stacked-mirror reference.

SAMS is made up of 480 edge sensor pairs (an active and a passive) located on the segment edges. Each sensor pair measures mirror (relative) positions, in terms of vertical shear heights and gap widths. There are up to six active sensors on each of the 91 segments, and these (up to) six values are carried to an ADC / DSP board package located on the back of the mirror segment. The 91 segments are grouped into three “hubs”, each containing 30 or 31 segments. A hub collects all of the data from the ADC/DSP packages in its group. This information is then carried to “the Brick” and the SAMS computer which are both located in the UER (Upper Electrical Room). SAMS is then able to calculate mirror corrections required based on current sensor positions.

It should be said that with all of these components, there are many points of possible failure. By understanding the system architecture, you can better react to a problem should it occur and better relay the problem to a mirror team member should they need to help troubleshoot. For example, if all six active sensors go out on a given segment, it most likely indicates a problem with the ADC/DSP boards; however, if all twelve sensors (active and passive) go out on a given segment, it would mean that segment is truly mechanically positioned very far out of range.

IMPORTANT: Anytime you close the SAMS Server, you should use the STOP button. Not the x in the right hand corner, LabVIEW does not like it when programs are closed incorrectly.

Also IMPORTANT: Always set a reference before going into Operate Mode. When you first open the server, it is in Standby mode-- seen in the topmost button. Click on Set Reference to choose from recent references the best file, making sure to note the temperature at which the reference was saved. After the reference is loaded, you can toggle to Operate Mode where SAMS will begin actively taking data every second.

Note: This server interface is liable to change relatively soon (by end of 2018?), as Randy Bryant has designed a new and improved SAMS Server Gui.

SCS and Sensors lights - These lights are shown in the upper-right, just beneath the Errors Encountered log. They are intuitive, Green is “good”, Red is “BAD”, and yellow is “be cautious”. SCS light only turns green when SAMS is talking directly to SCS, which happens once every 34 seconds or so when the two programs are in closed loop. See Common SAMS problems for more information about what to do when Sensors light turns yellow or red.

Commands Received Log - This log shows various communications SAMS has with other software (e.g. getttp is request from SCS for tip, tilt, piston corrections) and changes the user makes to SAMS (e.g. if you remove a sensor, it will log it here; if you set/save a new reference, it will log it here). While stacking, there will be a setreloff command logged during every array correction.

Errors Encountered Log - This log shows… errors encountered. If there is a sensor with a relatively high TSE (while there is a global TRSE, target residual square error, plotted in the history tab, each sensor has its own TSE, target sensor error), it will be logged here. There are a variety of rare errors that can appear here as well. While the TO cannot monitor this log all night, it is a good place to start investigating if the sensor light turns yellow or red, or SAMS behaves oddly.

RSE / TRSE - Watching the relationship between RSE and Target RSE is important when operating the SAMS Server. The global RSE/TRSE values are plotted under the History tab. RSE (Residual Square Error) is the current real error across the array; successful SAMS corrections will bring this value down closer to the Target RSE (TRSE) which is the targeted or lowest possible error the system can achieve. Low RSE and TRSE (<200nm) are ideal, but acceptable TRSE up to 500nm is possible depending on the science and time available in the night. High TRSE indicates either a significant temperature change since the last reference or a problem.

Depending on the TO, they will use Tip / Tilt Error, Tip / Tilt Scale, or both as their metric for how the primary mirror is maintaining the stack.

Tip / Tilt Error - This value is shown directly above the plot for RSE/TRSE in the history tab and as noted is the FWHM (full-width half max) of the tip/tilt errors of all of the mirrors. This is a statistical value, of which the definition will not be explained here. Suffice to say, less the 0.1 is ideal for the mirror while in closed loop, maintaining a stack.

Tip / Tilt Scale - This value is shown in the upper-right of the TTP’s tab, showing the color-coded array. This graphic is normally left in auto-scale mode, so the T/T scale is the length of a vector from the center of a segment to its edge. It can more easily be thought of as a maximum T/T error of a single segment. The ideal value for this metric is <0.2.

Segment Data - This tab shows detailed information about individual sensors and the tip, tilt and piston corrections needed for a single segment.

Configuration - This tab is where the user can disable or enable sensors and segments in the array. The selection of changes can be made while the Server is in Operate mode, but must be put into Standby before the changes will actually occur. If done quickly enough between SCS moves, these changes can be made while SAMS is in closed loop with SCS.

Operation Control - In general, this tab should be ignored. The only change the TOs should make in this tab is to change the max TRSE value (but no higher than 1500).

Shear / Gap / and Temperature Tabs - These tabs show colored array graphics of the corresponding values; e.g. unser shear, the shear values for all sensors are scaled relative to each other. Similar, to the TTPs tab, these are normally left in auto-scale mode.

Common SAMS Problems

High Target RSE:

High TRSE is anything above 1000nm in the SAMS Server. Large change in temperature since the last SAMS reference can cause high TRSE, but if temperature change has not occurred since the previous night, high TRSE alerts the TO there could be a hardware problem. Changing the Operation Control above 1500nm should only be done if the cause is understood. If not understood by the TO, contact a member of the mirror team otherwise rather than lowering the error, it will most likely only be distributed across the array. High TRSE could be caused by a failing sensor, bad segments or segment positions, a bad SAMS reference, or other system failures. Because the cause could be a number of things, the solution depends on the problem. Removing sensors or segments, loading a SAMS reference closer to the current temp, re-loading the last SCS position file, or applying GRoC to the mirror are all potential solutions to HIGH TRSE. Please don’t guess at a solution and call another TO or mirror team member to help instead.

Bad sensor (poor sensor alignment, faulty electronics, broken cables, broken sensor mount, etc):

When a sensor fails, it will report in the SAMS Server with either a warning (yellow light) or a failure (red light). SAMS will continue to operate through warnings though they can increase TRSE; the offending sensor may need to be removed. SAMS will not operate through a failure; the sensor must be removed. Typically, reported LTSEs (Large Target Sensor Error) are associated with yellow light warnings and +/- 384s are associated with red light failures. If sensor failure extends across SAMS “Hub” groups, it is indicative of a larger comm problem. Note: For how to remove a sensor for any of the above cases, please refer to ‘Removing sensors from SAMS’ on page 23.

Bad segment (no active sensor values, comm problems, ADC/DSP connection failure, etc):

When a segment fails to report information, all active sensors on the segment will read +/- 384. The red failure light will be displayed. The segment must be removed. This problem is typically associated with poor hardware connections. If all twelve sensors read +/- 384, a very rare event (very hopefully avoided) could be due to the segment being pistoned above or below sphere.

Bad segment(s) (not moving in one axis, not moving at all, not applying SAMS corrections):When a segment stops moving, the array will not be able to correct to the reference. The RSE and TRSE will increase and the SAMS Server will display the problem as a segment(s) developing larger vectors in Tip / Tilt and error in Piston. You will see the Tip/Tilt? Scale increase. The segment must be removed and quickly because the rest of the array will move to accommodate its position. If this failure extends across many segments in SAMS or SCS control groups, it could be indicative of a larger comm or power failure. Restart may resolve the issue, but a call to the mirror team is likely necessary.

Some familiarity with the SAMS hardware will help you in troubleshooting or describing problems with M1. There are diagrams located above the TO console that explain the difference between: SAMS “Hub” groups, SAMS power groups, and SCS communication and power banks. At times, segment failures will be related to communication or power and the failed section of the array will clearly correspond to the hardware. Referring to these diagrams may help you describe a problem or segments failed to the Mirror Team over the phone if needed.

SCS (Segment Control System) is the system that moves the 91 segments. SCS is made up 273 linear actuators (3 actuators per mirror segment). While the 3 actuators can be moved individually, normally they are used to tip (rotate about the horizontal axis), tilt (rotate about the vertical axis) or piston (move up or down) the mirrors.

Mirrors can only be moved by setting a “move authority” in SCS. If the move authority is set to SAMS, SCS will begin querying SAMS for tip, tilt, and piston corrections that need to be made. If the move authority is set to manual, the user can move individual segments via the config menus. While stacking, you may observe the move authority switching from SAMS to AOA/MARS, as the AOA/MARS system will send its corrective moves to SCS.

The actuators are actuated by stepper motors. The stepper motors are driven by electronics that are currently stowed right beneath the mirrors, but will soon (hopefully by end of 2018) be stored in the Igloo. Currently, these controllers are part of SCS serial communication banks, of which there are 10 banks. There are 5 power banks, each of which powers two communication banks. SCS comm or power failures can sometimes be identified by issues that correspond to the comm banks; the diagrams above the TO console can be used to help identify or describe the problem.