RE: Improvements to cooling system make it worse

[Tom Droege <tdroege@veriomail.com>]

One more thing.  The thermometers used, AD590, are rated at +/- 2 C.  They 
are usually much better than that, but since they put out 1 ua per degree, 
(i.e 273 ua at 0 C) one is taking difference of big numbers at our 
operating temperature.  You will note that I subtract 273 in the 
temperature computation.  This is all done through a 1% resistor.  So even 
if the AD590 was perfect, one could read 270 and another 276 at zero C just 
from the resistor.  Add to this the 2 degree spec of the AD590 and one 
could get 268 with one device and 278 with another, or +/- 5 C in absolute 
measurement.

The moral of all this is that you have to calibrate the thermometers if you 
want to know what the real temperature of everything is.  One way to do 
this is to let the whole mess settle to ambient temperature.  Turn off all 
the TEC drives and circulate room temperature water.  Turn off the dew 
heaters.  Now the two CCD sensors, and the water temperature sensor should 
all come to the same temperature.  Be sure you are not putting heat into 
the cooling water with the pump.  Now you can put a calibrated thermometer 
into the coolant, and put entries into the read out table to make the 
readings correct at that one temperature.   Now on another day, with 
another temperature, one can get another set of readings.  This should let 
you adjust the slope of each calibration to make all the thermometers track.

As you all can see, this is a big project, and not of much use.  In fact, 
the temperature readings tend to be much better than the possible errors 
indicate.  As I say, the emphasis in the design was on holding the 
temperature constant as the water temperature changes.  This it does with a 
significant gain.  It should have a gain of at least 100.  That is a 1 
degree change in the water should change the CCD temperature less than 0.01 
C if everything is working.  I did serious testing on the Mark III.  On the 
bench I was able to hold to 0.001 C with a gain of about 1000.  I have not 
done this for the Mark IV.  It is the same general design with the same 
amplifiers.  So it should be similar except the camera head has more 
thermal mass for the Mark IV.

Tom Droege



At 10:04 PM 1/8/02 -0500, you wrote:
>I was trying to test my mount pointing tonight but just when the Pleides
>came into view so did the clouds.  It's going to be cloudy for a while so
>I'll provide some comparison numbers for TEC current, temps, etc.
>
>Value       TOM       Mike
>TEC Power   11.954    11.35
>+15         14.798    14.66
>+5           4.88      4.86
>Ground      -0.02      0.00
>-5          -5.048    -5.10
>-15        -15.338   -15.01
>VCO Ref.     5.020     5.03
>ADC Ref      2.514     2.49
>DAC 14 set  -2        -2
>DAC 15 set  -2        -2
>DAC 14 read -1.992    -1.98
>DAC 15 read -2.036    -2.03
>
>Temp CCD0  -19.58    -23.64
>Curr CCD0    0.875     0.78
>Temp CCD1  -19.37    -23.06
>Curr CCD1    0.813     0.91
>Water Temp   7.27      4.80
>
>Tom's system and mine show almost identical values.  With nearly the same
>currents we get nearly the same cooling below ambient, in this case about
>28C below the water temp.
>
>
>Mike G.