This shows you the differences between two versions of the page.
Next revision | Previous revision | ||
hi:cooling:cooling:blmrackcooling [2023/02/01 11:49] p.boutachkov created |
hi:cooling:cooling:blmrackcooling [2023/02/01 21:10] (current) p.boutachkov |
||
---|---|---|---|
Line 1: | Line 1: | ||
====== Improving the cooling of the rack with the BLM electronics ====== | ====== Improving the cooling of the rack with the BLM electronics ====== | ||
- | Task: | + | Task: In the last few summers were hot leading to higher temperatures at the ESR roof, which led to temperatures in the BLM electronic rack above **35 C**. |
+ | Temperatures close to ** 40 C ** may lead to electronics Failure. For instance all {{ : | ||
+ | are rated to operate up to ** 40 C **. | ||
+ | |||
+ | ====== Electronics Rack cooling ====== | ||
+ | There is a air-condition unit installed on the top of the electronics rack. The unit model number is: TEXA ~DEK15BT0B, manual | ||
+ | {{ : | ||
+ | The manufacturer recommends to set it to 30 C. If the temperature goes 2 K above the set point the unit will be engaged. | ||
+ | One user is expected to set the temperature, | ||
+ | |||
+ | ====== Optimizing the Electronic rack cooling ====== | ||
+ | In the experiments described bellow the following tasks were accomplished: | ||
+ | * Determine the optimum configuration to reach the lowest rack temperature | ||
+ | * Determine the heat sources in order of significance | ||
+ | * Determine the cooling power needed to control the rack temperature | ||
+ | |||
+ | A RPi with a temperature sensor is used to monitor the temperature in the rack bellow the NIM crate positioned at the bottom. | ||
+ | Here is the starting configuration: | ||
+ | |||
+ | {{: | ||
+ | * All electronics off, AC set to 25 C. Temperature in the rack __28.3 C__. The switch and RPi are heating the rack. | ||
+ | * Switched on the NIM Crate. Now one can see the AC turning ON and OFF. | ||
+ | |||
+ | {{: | ||
+ | |||
+ | The period of oscillation changes. It is most likely related to outside temperature. | ||
+ | * Switched the Tek MSO44 scope. The temperature in the rack is __28.3 C__ but there are no more oscillation. Seems the AC is ON all the time | ||
+ | * Switched the CAEN HV, no HV ON. The temperature in the rack is __31 C__ | ||
+ | * Switched the DAQ electronics ON. The temperature in the rack got up to __ 34.5 C __. During the night it went down to __ 33.6 C__ | ||
+ | * Added a triangle shape " | ||
+ | |||
+ | {{: | ||
+ | |||
+ | * Removed the paper " | ||
+ | * Removed the HV main frame, the temperature dropped from __31 C__ to __26 C __ | ||
+ | |||
+ | {{: | ||
+ | |||
+ | * Installed the HV under the scope, above the DAQ. Turned it on. The temperature stabilized at __31.5 C __. | ||
+ | * There is a possibility to measure the HVCardsTemperature. The temperature is __ 30 C__ | ||
+ | |||
+ | {{: | ||
+ | |||
+ | * Installed the HV under the DAQ in place of the NIM bin. The temperature stabilized at __ 29.5 C__. The HV card temperature is __ 29 C__. | ||
+ | |||
+ | {{: | ||
+ | |||
+ | * Installed the NIM bin above the HV and the DAQ above the NIM bin. everything is on now. The temperature stabilized at __ 29.1 C __. | ||
+ | |||
+ | {{: | ||
+ | * Open the front and back door. The temperature now is __ 24 C __. The HV card temperature is 29 C. As this does not change I am not sure if the HV card reading is reliable. | ||
+ | * Christian set the AC set point to __ 23 C__, hysteresis to __ +3 C__. The temperature is now __ 27 C __. The second sensor above the DAQ shows __ 27 C__ too. | ||
+ | * Used the CAEN Java tool. The fan speed was set to Med. The HV card temperature was __ 28 C__ for the 24 Ch card and __ 27 C __ for the 12 Ch cards. | ||
+ | I have measured the temperature above the DAQ too. Here is a photo of the second sensor read via an Arduino. | ||
+ | |||
+ | {{: | ||
+ | |||
+ | The temperature was the same for both sensors. Hence there is sufficient air mixing. | ||
+ | |||
+ | When increased the fan speed to High the temperature of the 24 ch card dropped to __ 27 C __. Note that the temperature is available via the ssh tool too. | ||
+ | * Set the CAEN HV fan speed to High. The rack temperature set to __ 27.6 C __. The HV card temperature is __ 28 C__. | ||
+ | * Set the CAEN HV fan to Low speed. The temperature is now __ 28.5 C __. The HV card temperature is __ 28 - 29 C __. | ||
+ | |||
+ | Moved the second sensor outside of the rack. It is now measuring the hall temperature. | ||
+ | |||
+ | {{: | ||
+ | |||
+ | The outside temperature is __ 23 C__. | ||
+ | |||
+ | * Leave the CAEN HV on, turn off the DAQ, scope and NIM crate. The temperature stabilized at __ 28 C __ | ||
+ | |||
+ | Observations: | ||
+ | * The maximum temperature in a close rack was reached with the paper roof on to of the HV: __ 35 C__. | ||
+ | * The minimum temperature in a close rack was __ 26 C__. Reached with the HV removed from the rack. Note that the AC set points is __25 C __. | ||
+ | |||
+ | Therefore the AC has sufficient cooling power to take care of the system without the HV. The HV is the main heat load. | ||
+ | Here is an estimate of the power generated from the electronics. | ||
+ | |||
+ | {{: | ||
+ | |||
+ | These are the parameters of the {{ : | ||
+ | {{ : | ||
+ | |||
+ | Based on the estimate we need of order of 1600-1700 W cooling power. So lets use 1750 W for the search. | ||
+ | A DEK20 at __ 35 C __ ambient temperature should be able to cool down to __ 30 C __. This is a drop in replacement of a DEK15. It is 6 kg heavier than DEK15. | ||
+ | |||
+ | A DEK30 can cool bellow __ 20 C __ when operated at __ 35 C __ room temperature. But it requires more space and is 80 kg | ||
+ | |||
+ | As extreme case lest take __ 40 C __ hall temperature. The the DEK20 should stabilize the rack temperature just bellow __ 35 C __. A DEK30 unit will no have problems keeping the temperature bellow __ 20 C __ even at __ 40 C __ ambient temperature. | ||
+ | |||
+ | Therefore a DEK20 is the minimum suitable choice for the rack cooling. A DEK30 is to be considered if additional BLM will be operated from the same rack. | ||
+ | |||
+ | ====== Summary ====== | ||
+ | * The main source of heat is the CAEN HV. The heat output of the rest is much lower. | ||
+ | * Placing the HV frame under the scope has the largest effect on the cooling efficiency in the rack. | ||
+ | * The optimum configuration is with the CAEN HV at the bottom. Speculation: | ||
+ | * The cooling power needed to control the temperature in the rack is of order of 1700 W. I expect to add 4 more units, 2 LED drivers and 2 switching matrices. The will add of order of 40 W. Hence, I use 1750 W for the estimates | ||
+ | * A DEK15 unit operating in __ 35 C __ hall temperature will stabilize the rack temperature at around __ 40 C __. | ||
+ | * A DEK20 unit operating in __ 35 C __ hall temperature will stabilize the rack temperature at around __ 30 C __. (Note: this is a drop in replacement of DEK15, it adds 6 kg to the roof load) | ||
+ | * A DEK30 unit can keep the rack temperature bellow __ 20 C __ even in __ 35 C __ hall. |