Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.arcetri.astro.it/~salinari/MMT336images/thermo.ps
Дата изменения: Wed Apr 5 14:32:31 2000
Дата индексирования: Sun Apr 10 03:28:09 2016
Кодировка:

MMT 336 ADAPTIVE SECONDARY
DESIGN REPORT
Doc. : A1RPAD98001
Issue : DRAFT
Page 34 of 42
3.13 Power dissipation
The nominal power dissipated through each coil is:
RMS: 1 Watt
PEAK: 9 Watt
capacitive sensors
air damping
The nominal power dissipated by each crate is:
RMS: 475 Watt
PEAK: 570 Watt
These figures are given by the following contributions:
capsens boards ??? Watt
control board ??? Watt
com board ??? Watt
3.14 Mirror thermal control design
The cooling system get input refrigerant some C below ambient temperature from the
MMT plant.
The main pipe passes through spider shadow and ends at the adaptive secondary
distribution stage, located ???.
Then, from the distribution box seven parallel circuits feed the cold plate cooling lines,
each one cooling an average of 60 actuators.
Then, after passing through the cold plate the refrigerant pipes input the crates.
Six pipes cool the three crates and the last line is available to cool the housekeeping
computer stuff.

MMT 336 ADAPTIVE SECONDARY
DESIGN REPORT
Doc. : A1RPAD98001
Issue : DRAFT
Page 35 of 42
At the end all the seven pipes converge into a single box to exit the M2 hub.
3.14.1 Cold finger
The cold plate can be kept at a temperature such as the RMS coil power is removed, while
the PEAK power dissipation can not be controlled.
Aluminum cold finger section is: AF = 0.0075 2 x 3.14 x 0.6 = 106x10 6 m 2
where the circular section has been reduced by 60% to account for the capacitive sensor
board pod.
Cold finger length is: LF = 0.080 m
The thermal conductance is then: C = k x AF / LF = 0.265 (Watt/C)
assuming k = 200 (W / m C) for the AL conductivity.
With these values, the temperature gradient needed between the coil and the cold plate to
remove 1 W is: DTF = 1 / 0.265 = 3.8 C .
This result drives the refrigerant input temperature from the MMT plant.
3.14.2 Thermal test on P30 Results
In order to evaluate the thermal parameter of the cooling system a test was performed on
the 30 Actuators Prototype -- P30 (ref ???).
The P30 crate is designed to be fully compliant with the actual one.
During the test the crate was filled by dummy control boards, designed to dissipate the
nominal power of the actual boards.
The measured power introduced by the boards is: 265 Watt
The measured temperature of the water is: INPUT T IN = 24.5 C
OUTPUT TOUT = 27.3 C.

MMT 336 ADAPTIVE SECONDARY
DESIGN REPORT
Doc. : A1RPAD98001
Issue : DRAFT
Page 36 of 42
The exchange surface between the refrigerant and the heat sinks is:
Di = 0.004 (m) pipe internal diameter
LC = 2 x 17 x 0.09 = 3.06 (m) pipe length (both sides of the crate)
AC = 6.28 x Di/2 x LC = 38.4x10 3 (m 2 )
From these data the heat transfer coefficient is computed:
hC = q / ( AC x (TOUT T IN ) ) = 2463 ( Watt / m 2 C )
The heat transfer coefficient depends on the refrigerant properties and speed.
The fluid velocity inside the pipes (2 in parallel for one crate) is computed from the output
flow measured ( Q = 0.03 liters / second ).
The average speed results: v c = 1.2 (m/s)
The heat transfer coefficient ``h'' depends on 1/3 power of the fluid speed.
3.14.3 Crate cooling circuit
Given the 570 Watt PEAK power to be dissipated for each crate and assuming water as
refrigerant, the required DT can be computed based on the P30 measurement:
0
1000
2000
3000
4000
5000
6000
0 5 10 15
(m/s)
(W
/
m2
C)

MMT 336 ADAPTIVE SECONDARY
DESIGN REPORT
Doc. : A1RPAD98001
Issue : DRAFT
Page 37 of 42
DT = 2.8 x 570 / 265 = 6 (C) , while the fluid velocity is kept equal to 1,2 (m/s).
On the other hand, by imposing 5 C temperature difference, the required fluid speed
becomes:
h c ' = 570 / (38.4x10 3 x 5 ) = 2969 ( Watt / m 2 C )
v c ' = 1.2 x (2969 / 2463 ) 3 = 2.1 (m/s) .
By using Glycole instead of water, the previous result is scaled according to the different
properties of the fluid:
Cp = ??? ( J / Kg C)
conductivity = ??? ( W / m C)
density = ??? ( Kg / m 3 )
viscosity = ??? (Kg / m s)
[...]
3.14.4 Cold Plate cooling circuit
Before entering the crates, the refrigerant warms up while passing through the cold plate.
The DT between the coil and the cold plate must be at least 3.8 C in order to remove the 1
Watt RMS power.
We assume the fluid enter the cold plate with DT = 6 C.
The fluid temperature increase is computed by assuming an average of 60 actuators
cooled by each one of the seven lines.
P = 0.03 (m) pipe internal perimeter
LP = 0.3 (m) pipe average length
AP = 0.3 x 0.03 = 9.x10 3 (m 2 ) contact surface

MMT 336 ADAPTIVE SECONDARY
DESIGN REPORT
Doc. : A1RPAD98001
Issue : DRAFT
Page 38 of 42
The fluid velocity is computed from the velocity in the crate's pipes:
vP = v c ' x SC / SP = 2.1 x 12.6 / 36 = 0.74 (m/s) .
Then, the heat transfer coefficient is estimated from the crate measured one.
Beside the dependence on the (1/3) power of the velocity, ``h'' is function of the length by
the (1/3) power.
Thus:
hPLATE = h c ' x ( v c ' / vP ) 1/3 x ( L c / LP ) 1/3 = 2969 x ( 2.1 / 0.74 ) 1/3 x ( 3.1 / 0.3 ) 1/3
= 2969 x 1.41 x 2.16 = 9048 ( Watt / m 2 C )
The temperature increase results:
DTP = q / ( AP x hPLATE ) = 60 / (9.x10 3 x 9048) = 0.74 C
Concluding, this configuration wants the MMT plant supply refrigerant at --6 C.
The flow rate results:
Q = 7 x ( vC ' x SC ) = 7 x 2.1 x 3.14 x 0.002 2 = 0.18x10 3 (m 3 /sec)
Then, after removing the 1x336 W RMS actuator power from the cold plate, the refrigerant
warms up by 0.74 C and has still more than the 5 C DT required to dissipate the 570 PEAK
power per crate.
3.14.5 Magnet -- mirror interface
[...]