User's Guide to the Gemini Twin-Arrays Infrared Camera

Table of Contents

Quick Reference
What is Gemini?
Summary Table
Quick start - for the expert
Not so quick start - set-up
Graphical User Interface
More About Gemini
Signal-to-Noise Estimates
Sampling Modes
Writing Scripts
Observing Recipes
Computer Setup
Testing the Arrays
Instrument Maintenance and Trouble-Shooting

Mt. Hamilton Homepage

Maintenance and Trouble-Shooting

Full Data Disk & Clearing Data off Gemini PC
Detector Temperatures
Liquid nitrogen level
Cables to the closed-cycle refrigerator(CCR)
Electrical connections to the detectors
Baking the Getter
Powering up the Motion Control System
Trouble shooting the motion control system
Dichroic slide won't move or initialize
Trouble shooting the telescope control system(TCS)
PCT link or PCT error
SBRC Array (B Channel) all 0.0
Vertical bars in SBRC images
Undesirable Software Characteristics

Full Data Disk & Clearing Data of Gemini PC

Gemini's D: disk only has 500 MB of space for data. When the disk is full Gemini will continue to try to take data and will display a message with a blue background saying insufficient disk space. Unfortunately this message gets partially overwritten by the Status message and may be missed by observers, resulting in lost data.

The best thing to do is try to prevent data loss from lack of disk space by clearing data every other day (it is unlikely, though not impossible, that the D: drive will fill up with only two nights of data). To do this safely, do the following:

  1. On the Gemini PC start windows (if not already in windows) by typing wi.
  2. Click on the "My Computer" icon and navigate to the D: drive and Gemini folder.
  3. Confirm that all data on drive has been copied to gouda /data/gemini/.
  4. If all data on the Gemini PC is not on gouda, carefully follow the directions on the Quick Reference page for FTPing data to gouda's /data/gemini/ directory.
  5. Use archclear on gouda to make sure all the data are in the archive. Please do not delete the data on gouda until after the Gemini run is complete because the observers will probably still want to copy it to their own computer(s).
  6. Use the "My Computer" window on Gemini to select all the .FTS files and put them in the Recycle Bin.
  7. Right click on the "Recycle Bin" icon and select "Empty Recycle Bin."

Detector temperatures:

6 temperatures are routinely monitored using Lakeshore Cryotronics temperature diodes and displays. The small blue monitor cycles through 4 diodes, displaying each temperature reading for a few seconds. A temperature Log Sheet is provided. The four points monitored are

  1. a filter wheel (80 K)
  2. the LN2 can (78 K)
  3. the NICMOS array (70 K)
  4. the CCR first stage (70 K)

The other Lakeshore monitor shows the second stage (lowest temperature) of the CCR head (about 11 K) and the temperature of the SBRC array (28 K). This unit is a servo system set to keep the SBRC array at 28 K.

The most important temperature to check during the night is the LN2 can. If the temperature begins to rise then the LN2 level is too low. It should not be allowed to warm up more than 10 K, otherwise the detector temperatures will begin to change and thermal gradients will be set up in the internal optical bench.

Liquid nitrogen level:

Gauging the LN2 level is best done by the temperature meter, but the lack of a white boil-off plume clearly indicates a low level.

The LN2 should be filled at least twice a day, once in the very late afternoon just before dinner and again at the end of the night.

Prolonged observations low in the south (below the Celestial Equator) will result in quite a large spillage of LN2. It is occasionally necessary to do a refill before dawn in this case. The LN2 can is "pressure-filled" from a large reservoir using a metal transfer tube which can be fed into the neck of the can and down to its base. To reach the instrument, the LN2 reservoir must be set on the rising/mobile platform.

Extreme care is required when the platform is rising. It is very easy to hit something or trap a cable.

Extreme care is also required when inserting or withdrawing the tube from the Gemini LN2 can, since the metal fill tube can damage the baffle in the neck of the LN2 can.

To execute the fill once the tube is inserted, simply open the valve on the big canister. Close the valve again when liquid nitrogen starts to stream from the Gemini fill tube.

Cables to the closed-cycle refrigerator (CCR):

The air-cooled compressor which drives the CCR is attached to the telescope by means of a poor-man's gimble - a shackle and some bungy cords - and the pair of 12 ft long high-pressure helium lines from the compressor to the CCR head are supported by a cord to minimize twisting.

Experience has shown that almost all parts of the sky can be reached with this arrangement. Nevertheless, some twisting can occur in the extreme west and the observer should check the condition of the cables every 15 minutes ore so when the Hour Angle exceeds 3 hrs west. Excessive twisting can cause small helium leaks, which can be monitored using the pressure gauge on the compressor. The gauge should read about 250 psi. If it falls below this level a bottle of pure helium is available for replenishment.

Electrical connections to the detectors:

The electrical connections to the detector outputs emerge from the dewar and fan out into four coax cables which go into the long, shiny preamplifier box. These cables should be considered fragile and sensitive and should not be touched, especially when the arrays are powered up and cold. The output sides of the preamps, from which four white coax cables go to the Gemini Transputer System, are more robust. These outputs can be connected to an oscilloscope if required.

Baking the Getter:

The getter (desiccant) needs to be baked periodically (approximately yearly or after the vacuum dewar has been opened). The getter should only be baked when Gemini is warm and connected to the vacuum pump.

There should be a cable that connects to the heater hermetic Military connector on Gemini's wall (this cable is stored in the wooden Gemini Cable box). The other end of the cable terminates with two BNC connectors. There is usually an adaptor on the BNC end to convert the BNC to a two-prong straight connector. Only one of the BNC's is connected to a 25 Ohm, 1 Amp heater resistor. The other is open. Use a meter to verify which one, if necessary. Put 15 Volts DC across the resistor for 15-20 minutes. If it has been a while since the last bake, you can go a little longer (e.g. 30 minutes).

Powering up the Motion Control System:

After switching on the Transputer crate, the Motor Controller and the Motor Power Supply unit, it is very important to run up the Gemini software. Unfortunately, switching on the power to the motor control system results in all motors being energized (though they do not actually move). They will remain this way, generating heat inside the camera, until a command from the software is received. Running up the GEMINI program initializes the Motor Control System and sends a "power-down" command to all motors. Power outages can cause the Drivers to power-up without the software restarting, so it is very important to turn off the Motor Drivers at the end of each night.

Trouble shooting the motion control system:

In most situations, the user should not have to interact with the motion control system except through menu selections. In a few rare cases one may desire to slightly reposition a filter wheel, or one may be forced to trouble-shoot a mechanical failure. Before considering specific cases of possible motor trouble-shooting, there are some sage words of advice that should be remembered at all times.

Assuming motors are being run at temperature (i.e. 48 hours after start of cooldown), they should be reliable. If there are problems, they are not likely to be in the form of a motor "almost" getting where it should go; problems will be either motors not starting or motors failing in some "catastrophic" way (for example jamming while ramping up to full speed). The switches are considered dependable, and may thus be believed. An "aesthetic" problem that might be encountered would be a slit or grism needing to be rotated through a few degrees.

The first check should always be the switches. The switches can be read via the computer using the "read switches" selection in the engineering menu. The switches may also be observed by looking at the LEDs on the motor transputer box. The green lights indicate that a primary switch is on; the orange lights indicate that a secondary switch is on. The order of the LEDs is the same as the order of the switches in the software. Make sure they are on when the motor is at that position. Watch them being activated as the motor moves past them. Below is a list of which switch corresponds to what:

MechanismPrimary SwitchSecondary Switch
FPA0 degrees -- open 270 degrees = 72000 steps
AUX0 degrees -- open 120 degrees = 32000 steps
LWF0 degrees -- blank 180 degrees =
SWF0 degrees -- blank 180 degrees =
DichroicK @ LWK @ SW
Waveplate Slide------
Waveplate Rotator0 degrees---

If wheels are failing to initialize or to fully reach the position requested, one can move the motor controlling that wheel by a few hundred steps at a time to try to get the wheel to the appropriate switch. Motors may be moved by steps rather than position using the "move motor" selection under the "Engineering" menu. If problems persist, notify the IR Lab team.

Finally, aesthetic problems can be solved by rotating the wheels through a small angle. The angle of rotation can usually be calculated from the image on the screen (for example (delta y)/(delta x) = tan(alpha)). Once the angle has been identified, simply go to the engineering menu and select "move motors". It will prompt for a number of steps. Consult the chart below to calculate the number to steps needed to move through a given angle. The same procedure can be used to check that motors are moving from one position to another (see below for the number of steps between positions).

WheelTotal StepsSteps/degree# Positions Steps/Positions
FPA96000~2674 24000
AUX96000~2673 32000
LWF7200020010 7200
SWF7200020010 7200
Waveplate36000100--- ---

SlideSteps between Positions

Dichroic slide won't move or initialize:

First make sure motor driver power is ON. If power is on then dichroic has gotten stuck (most likely in K@LW position). There is no fix for this at the telescope. The vacuum dewar must be opened up and the dichroic slide manually fixed. To prevent this problem, ALWAYS initialize the dichroic motor from the Engineering - Motor Init menu before moving the dichroic from the Setup - Motor Setup menu.

Trouble shooting the telescope control system (TCS):

Once in a while there have been problems communicating with the TCS. Normally there will be an error message at the Gemini prompt saying TCS link not up. Some things to try are:

PCT link:

If there is a problem with the PCT link, try the PCT reset under the "Engineering Menu" in the software. If this does not work, it will be necessary to cycle the power on the front of the transputer crate. There is a lighted green button that needs to be turned off for a few seconds and then back on. Reboot Gemini program and start setup procedure again. If problem still persists then also cycle power on the back of the motor driver's controller box and the transputer box (because there is one more transputer in the motor controller).

If the Gemini software will not start and gives a PCT link error, use the check program to identify where the problem is. At the DOS C:\ prompt type check. You should get a table that lists the transputers and the connection between them. There should be 16 of them, numbered from 0 to 15. Typical output of the program is

check 2.2
#  Part rate  Mb  Bt [ Link0 Link1 Link2 Link3 ]
0  T805b-25  0.33  0 [ HOST   1:1   2:1   3:0  ]
1  T800c-20  0.68  1 [  4:3   0:1   ...   5:0  ]
2  T800c-20  0.66  1 [  6:3   0:2   ...   7:0  ]
3  T800c-17  0.66  0 [  0:3   ...   ...   ...  ]
4  T805d-25  0.90  3 [  8:3   ...   ...   1:0  ]
5  T805d-25  0.90  0 [  1:3   ...   ...   9:0  ]
6  T805d-25  0.89  3 [ 10:3   ...   ...   2:0  ]
7  T805d-25  0.89  0 [  2:3   ...   ...  11:0  ]
8  T805d-25  0.90  3 [  9:3   ...   ...   4:0  ]
9  T805d-25  0.90  0 [  5:3   ...   ...   8:0  ]
10 T805d-25  0.90  3 [ 12:3   ...   ...   6:0  ]
11 T805d-25  0.90  0 [  7:3   ...   ...  13:0  ]
12 T805d-25  0.90  3 [ 14:3   ...   ...  10:0  ]
13 T805d-25  0.90  0 [ 11:3   ...   ...  15:0  ]
14 T805d-25  0.90  3 [ 15:3   ...   ...  12:0  ]
15 T805d-25  0.90  0 [ 13:3   ...   ...  14:0  ]

If you are seeing less than this number that is the cause of the problem. Make sure both the transputer box and the motor driver controller box are powered on.

If you only see number 0, then the problem is between the PC and the transputer crate - number 0 is in a plug-in card in the PC. The likely suspect in this case is the DB25 cable into the back of the PC was not properly connected.

If the check command gives a Time Out Error, reconnect the DB25 cable to the back of the PC.

If you see no transputers, then the PC should be rebooted.

With the current 486 PC used to run Gemini (as of 1 Feb 2002), there will occasionally be trouble communicating with all the transputers, even though check shows communication to all the transputers. If this is the case and starting the Gemini software gives a PCT Boot error, reboot the PC. If you still get the error, then you may also have to cycle power to the transputer crate and motor controller.

SBRC Array (B Channel) all 0.0:

This is a PCT link problem and is fixed by cycling the power on the transputer (see above).

Vertical bars in SBRC images:

Vertical bars (each 4 pixels wide) in SBRC chip images indicates that the chip is warming up. The chip will start warming up if the temperature controller somehow gets set to the wrong temperature (i.e. is any temperature other than 28 K) or if the CCR stops working. Check to make sure the CCR is on, all cables are properly connected, and that the cold head is working.

Undesirable Software Characteristics:

When running a script double check that the integration time is sufficiently long enough for the number of co-adds and multireads chosen. The program will not inform you if the integration time is too short and will run the script anyway.

To abort a script during an exposure press CTRL F2 and then CTRL F3. Otherwise, the script will be aborted only after the current exposure is finished.

If the computer crashes after 12:00am during a run, it will save the remaining Log Files in a directory on the hard drive named for the following day.

Last modified: Fri Oct 1 01:20:51 PDT 2010 by Elinor Gates