Cooling Box

A cooling box for up to seven detector modules was developed at HEPHY [60] to provide conditions as close as possible to the CMS tracker. An ambient temperature of $-10^{\circ}\,\rm C$ is even mandatory when testing irradiated sensors (see section , p. ). The cooling box is based on direct-to-liquid cooling elements employing the thermoelectric effect discovered by J.C.A.PELTIER^5.1. This principle was soon adopted by the CMS collaboration for other test setups after successful operation in our beam tests.

Two highly efficient Peltier elements with an electrical power consumption of up to $350\,\rm W$ each are used to cool the interior of an isolated box. The heat at the warm side of the elements is carried away by water flow. This method outperforms air cooling, because the Peltier performance degrades with the temperature difference between warm and cold sides. While the warm side is tied to the tap water temperature, which is usually between $10$ and $20^{\circ}\,\rm C$, air cooling would inevitably cause a warm side considerably above room temperature.

The cold side of the Peltier elements is connected to two massive aluminum plates at top and bottom ends inside the box. Small aluminum boxes, which carry the silicon detector modules, are inserted into rails in the aluminum plates. Thus, the small boxes are thermally connected to the Peltier elements. To prevent condensation of water on sensitive elements, the cooling box is flushed with a dry gas (e.g. nitrogen) prior to and during cooling. Inside the cooling box, the gas is first guided through caverns in the aluminum plates to cool it down. Then, it flushes the small boxes and exhausts into the large volume.

Not only the mechanics, but also the power supplies for the Peltier elements and their control and readout interface have been developed at HEPHY. Moreover, a ten-channel temperature monitoring system (MultiTherm) was built to watch the system temperature in several spots and automatically control the power supplies of the Peltier elements for a set temperature.

In terms of control theory, the conversion of electric power at the Peltier elements to the temperature decrease inside the box reveals $\rm PT_1$ behavior with a short dead time. At typical internal heat dissipation, a time constant of approximately $45\,\rm min$ and a dead time of about $10\,\rm min$ were measured.