[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Improvements to cooling system make it worse



Michael and all,

"  Way, way back in time, last summer, I ran some tests in a warm
room indoors.  The water temperature was around 30 C, and I was
able to reach CCD temperatures down to -34 C lower (CCD1),
or -44 C lower (CCD0).  Fine."

OK, this is about what I expect.  If you recorded the peak currents you 
were able to achieve, they should be something like 1.6 amps.  In the tests 
above, CCD1 was probably seeing hotter cooling water than CCD0.  Note that 
the thermometer is between the two cameras, so it is measuring the 
temperature of the water input to the second camera.

In theory, the TECs can be damaged if they get too hot.  In practice, there 
is enough metal in the camera heads to keep them below the temperature at 
which they are damaged, about 80 C.  The maximum current that can be put 
into the TECs from the circuit is about 1.7 amps at 11 volts or so out of 
the power amplifier.  I think the TECs are rated to draw 2 amps at 12 
volts.  So the circuit can simply not exceed the ratings of the TEC.

So the way the system is designed, it should be hard to damage the TECs in 
operation.  Still, I have had a number of them die.  The symptom is just 
what you are seeing, they don't draw enough current and so don't get very 
cool.

I have sat at the bench and tried to break them by driving them with 
various programs of voltage/current to see if they were damaged by 
transients.  They are not.  One way I have found for sure to break them is 
to drop a camera.  Just a couple of inches on to a hard surface will do 
it.  I have not seen any damage from vibration, as in moving a long 
distance by automobile.  Still, I always pack the camera heads in foam when 
I move them.

So the first question I would ask, is have you dropped the cameras on a 
hard surface since the lab tests?  Not likely, but I must ask it.  It is 
the shock that does it.  Just a couple of inches onto a hard surface will 
do it.

>   In order to modify the temperature of the chips, I am changing
>DACs 14 and 15.  The reported "TEC Current" values are at most
>1.35 Amps (for CCD0, which is currently making 14 degrees C below
>water), and 0.80 Amps (for CCD1, which is currently making 20
>degrees C below water).

OK, these values of current are low.  The first test to make is to turn the 
command full on, -9 volts.  (Be wary of setting DACs to their limits as 
they fold back at one end.  Now look at the current.  Be sure to check that 
the power supply is OK and that it is reading near 12 volts.  Let me know 
the camera numbers and the currents that they achieve.  I record the 
currents here when I assemble the cameras, and so should be able to tell if 
they have changed.

The reason I record the currents is that I have had a lot of TECs go 
bad.  Almost in every case it is because I have damaged them 
mechanically.  They are mounted with a thermal grease joint, with a very 
mild pressure.  Just the pressure produced by deforming a 1/32" printed 
circuit board abut 1/16".  This is just enough to hold the grease joint in 
place.  The other side of the TEC is glued to the heat sink.  The grease 
joint is at the back of the CCD.  Until I learned better, I would break the 
TEC trying to get the grease joint apart to change a CCD.  They do not like 
tension.  So if you pull the CCD off the TEC, the tension of the grease 
joint is enough to damage the TEC.  They do not look broken, but they quit 
cooling   I have also broken them by just dropping a camera a few inches 
onto a concrete floor.  (I try not to do that).

You do not just break one stage.  There are a few hundred junctions on each 
stage soldered between two plates.  The way they break is by very fine 
fractures of individual junctions.  The material is glassy, but since it is 
a lead alloy (Lead titenate zirconate I recall) it is a very weak 
material  You can almost crush a piece of it between your fingers.  I have 
lots of dead ones to play with.  ;^(

Now on the good side, unless you have done something nasty to the cameras 
mechanically, it is hard to imagine that they have died.  So the first test 
is to drive them full on with the cooling water on.  See how much current 
is drawn by each unit.  If you are up to it, measure the voltage between 
test point 2 and 3 on the driver printed circuit board.   These are marked 
x2 and x3 and are on each side of the connector that goes to the 
camera.  If you get 10 volts or so here, then the driver is probably 
OK.  If much less, then the power transistor may be sick, though I have 
never had one fail when operated properly.

Another thing to know is that TECs work better the hotter they are.  So you 
might get -40 C delta t with +20 C water, but only get -30 C delta t with 0 
C water.  These numbers are just guesses.  That is the direction that 
things work.  This is not so bad if one's goal is to hold the same 
temperature summer and winter.  If driven by greed to get the ultimate cold 
temperature, then this works against you.   ;^)  However, I suggest that 
you do not grasp too hard at this straw.

For comparison, here are some readings from TOM!:
TEC Power       11.954
+15             14.798
+5              4.88
Ground  -0.02
-5              -5.048
-15             -15.338
VCO Ref.        5.020
ADC Ref 2.514
DAC 14 set        -2
DAC 15 set        -2
DAC 14 read     -1.992
DAC 15 read     -2.036

Run TemPlt

Temp CCD0       -19.58
Curr CCD0       0.875
Temp CCD1       -19.37
Curr CCD1       0.813
Water Temp      7.27

Note that I did not wait long enough for the drive to settle.  If I had, 
they would have settled at -17 or so and somewhat less current.

Now set commands to -9.

Temp CCD0       -26.66
Curr CCD0       1.604
Temp CCD1       -28.12
Curr CCD1       1.646
Water Temp      9.34

Again I have not waited long enough, but the currents are at their limit as 
I can tell by the plot and things will not get much colder.

A check of the voltages shows TEC Power at 11.652, so it has dropped 
somewhat under load as expected.  The command voltages are also as 
expected, near -9 volts.

It is quite possible that you have sick TECs.  Again, I doubt that it could 
be caused by any water flow problem.  I can rebuild the cameras if that is 
needed.  I will just build more and trade you.  So don't fear too 
much.  Best to understand what has happened.  I would plan to get other 
things working.  In cold weather, just the cold water circulating will 
probably get you cold enough to be below the sky noise, but I know you want 
to do the best you can.  Possibly you can get a little colder if you unplug 
the heater ring that is screwed to the front of the camera.  This dumps 
about 10 watts into the face of the camera in an attempt to keep the dew 
off the filter/window.  If the humidity is low this is probably not 
needed.  Just disconnect the white connectors at the driver board.  But get 
the right ones.  ;^)  One disconnects the +12 TEC power.  You might also 
check that you do not have a loose connection, but that is improbable.  If 
you get nice curves with templot, then power is probably OK.

This is all I can think of for now.  A little more information will help.

Tom Droege

PS.  Tomorrow my ISP is selling me to Earthlink.  I may be out of 
communication for a few days.

I can be reached at:

630-879 7609
2S942 Thornecrest La.
Batavia, IL 60510









At 06:02 PM 1/8/02 -0500, you wrote:

>   I have been running tests of the cooling system, and am greatly
>puzzled: the more flow I induce in the coolant lines, the less cool
>the chips become.  Let me explain.
>
>   Way, way back in time, last summer, I ran some tests in a warm
>room indoors.  The water temperature was around 30 C, and I was
>able to reach CCD temperatures down to -34 C lower (CCD1),
>or -44 C lower (CCD0).  Fine.
>
>   Back on Jan 2, I ran tests with the same ol' pump I've been
>using for years.  As you can see from my status report on that day:
>
>       http://a188-l009.rit.edu/tass/status/status_mwr_03jan2002.html
>
>the lowest temperature I could achieve was about -20 C, which was
>23 degrees Celsius below the coolant temperature.
>
>   Over the next week, I've made some changes to the cooling
>system:
>               a. a new, more powerful pump, which increases the
>                     flow rate by about a factor of 2
>
>               b. replacing the quick-release connectors with
>                     straight-through connectors, again to increase
>                     flow rate
>
>               c. adding insulation to the section of coolant tubing
>                     which is exposed to the outdoors (between
>                     the coolant reservoir and the Mark IV shed)
>
>These changes have definitely increased the flow rate through
>the system.  They may have slightly increased the average daytime
>temperature of the coolant, but that's not important if I can
>really get 40 degrees C below the liquid.
>
>   Cooling tests I have run yesterday and today show that the
>performance is now WORSE:  the chips can reach at most 20 degrees
>below the coolant temperature.
>
>   I am wondering: has something bad happened?  Is it possible that
>I've damaged the TEC units, so that only one stage is operating?
>That might explain why the temperature difference between
>coolant and chip is now about half of what it was last summer.
>
>   In order to modify the temperature of the chips, I am changing
>DACs 14 and 15.  The reported "TEC Current" values are at most
>1.35 Amps (for CCD0, which is currently making 14 degrees C below
>water), and 0.80 Amps (for CCD1, which is currently making 20
>degrees C below water).
>
>   Perhaps there is some very simple explanation: maybe there's
>a voltage controlling the TECs which I'm setting incorrectly?
>
>   Advice, suggestions, help?
>
>                                         Michael Richmond