User`s guide

FLAMINGOS at the MMT
VI.Troubleshooting
1) Problem: The image looks funny. You can see all 32 amplifiers, and the signal level is ~58,000 ADU.
Solution: This is a hardware problem related to electrical noise causing an early readout of the detector. In
imaging mode, MCE4 reads the array once, and then lets it integrate before reading it again. The final image
written to disk is the difference of these two images. This problem occurs when MCE4 spits out the image
before making the second read and difference. There is nothing you can do to resolve the problem. Typically,
we get bad reads for about 2-5% of the images taken. If you are getting a much higher failure rate you should
call for help You should be cautious with the image immediately following a bad read frame, because the ADU
level in the image is frequently lower than it should be.
2) Problem: Excess Background
Possible solutions:
a) Image a very bright star (V ~ 2) at Ks-band, with the telescope way out of focus, say 2000 units lower.
You should see a donut with a central obscuration and spider vanes. If it's not circular, then you could have
something obscuring the beam---e.g., the dome, the mirror covers, or a mispositioned filter, lyot stop, or grism
wheel.
b) Check that the ion gauge on the Granville-Phillips pressure monitor is off. When left on it casts a
large amount of warm illumination on the lower right hand corner of the array.
c) Check that the Mos dewar temperature is lower than 200 K. If the temperature is not between 80 – 85
K, did you forget to fill the dewar with LN2? What about the camera dewar?
3) Various Problems related to image readout, wheel motion and iocomm communication.
The three daemons started by initflam.pl ( temperature, motor, and MCE4) work by issuing commands to
particular hardware addresses on a networked device called the iocomm. This device is like a multiplexer. We
have three devices (the temperature monitor, the motor controller, and the array controller, MCE4) which accept
commands via serial ports. The iocomm has one serial port connection for each device. Commands from the
software daemons running on flamingos1b are received by the iocomm via an ethernet connection, and then
passed to correct device.
Troubleshooting several types of problems involves deciding if the problem is intrinsic to the device, or to
the communication via the iocomm itself.
a) Problem: The image has not read out. The exposure time counter never appears.
This is our most common failure mode. Sometimes the array controller crashes before the image
acquisition script issues the START command to MCE4. For example, a successfully started image
acquistion should show:
/usr/local/uf/bin/ufdc -host flamingos1b -port 52008 -acq "start -
timeout 9 -cnt 1 -lut /usr/
local/flamingos/flamingos.headers.lut/Flamingos.lut -fits
16/23