In the lattice lab we use three Thermocube 600L chillers to cool the fiber amplifiers, Feshbach supply, and transport system. The chiller for the fiber amplifiers uses a centrifugal pump (recommended by SS Cooling for low noise), and the chiller for the transport system uses a Gear pump tuned to ~2LPM flow rate (necessary to generate more pressure at the expense of more noise). The Feshbach chiller also uses a gear pump but it is tuned to ~1LPM making it quieter. These chillers are water cooled themselves, and so there are two separate water lines, House (cooling the chiller) and Process (cooling the experiment). Heat exchange occurs at a stack of three water cooled plates with thermoelectric devices inbetween. The House water feeds the top and bottom plates in parallel, and the Process water circulates through the middle plate.
We fill the circulating water line for the chiller with a ~10% mix of Propylene Glycol in water. The Propylene Glycol acts to reduce growth of algae in the water lines, and also acts as a lubricant for the gear pump in the transport system chiller.
Low Flow Rate From the Wall
Since Fall 2013, we have noticed that the large filter for the house water needs to be changed very frequently. The flow rate of water from the wall (measured without going through the chillers) should be about 3Lpm (fills a 500mL beaker in ~5s). About every two weeks, this flow is seen to drop to about 1.5Lpm or worse, but quickly improves when the filter is replaced. If the chillers are having trouble, this should be the first thing to check. We must keep a supply of the Rainfresh CF2 filter cartridges (Home Hardware).
Tank Level Low
Inside the tank of the chiller is a styrofoam 'cork' that floats to the top of the water level. If the floater sits too low, it will trigger the 'Tank Level Low' error. From time to time a little bit of water may need to be added to the tank if this error occurs. It has also happened a few times that the cork simply gets stuck in a lower position by the flow of water in the tank, and needs to be jostled around a bit to get it floating properly again.
The standard model chillers from SS Cooling are fan-cooled rather than water cooled, and so this error indicates that the chiller is unable to effectively cool the process water. This will obvoiusly happen if the water from the wall which cools the chillers is not flowing, but can also happen due to poor flow inside the chillers.
There are 6 temperature sensors in the chiller, two on each of the three heat exchange plates. These sensors are visible at the front of the chiller if the cover is removed. Once the temperature on any one of the sensors becomes too high it will trip, resulting in an open connection and leading to the fan failure error (thus the fan failure error could also be caused by bad electrical connections to these sensors, but this hasn't been observed in the lab yet). The connection through the sensor will not close until the temperature comes back down, and so turning the chiller off and on will result in getting the same error again quickly.
Since a sensor in either heat exchange plate can cause this error, we have had trouble with uneven flow rate of the house water through the two plates, since the flow splits in parallel but sometimes unevenly due to blockage in one plate. It may be necessary from time to time to clear out the plates of any dirt by using compressed air and pipe cleaners.
The flow rate of the circulating water to cool the transport system is monitored by a flowmeter (Proteus). This flowmeter will trip an interlock in the CATS if the flow rate becomes too low. If the flow rate seems to be dropping too much, the liquid filter in the chiller may need to be cleaned or replaced (McMaster Carr part # 45175K15).
Previously when the filter seemed to be causing the problem, we would see a good flow rate initially when turning the chiller on (~0.4 on the interlock panel of the CATS - corresponds to about 1.3 Litres per minute) which then decreased over a period of ~15 minutes to a poor flow rate (<0.3 on the interlock panel - about 1 Litre per minute).
After replacing the liquid filter in the chiller, the flow reading on the CATS interlock panel was stable at nearly 0.4 V.
Backfilling of the Tank
The circulating water lines of the chillers both run up to the ceiling and then down to the experiment table. When filling or re-filling the lines by pouring coolant into the chiller tank while the pump is running, pressure levels above atmospheric pressure can build up in the lines near the chiller, forcing the circulating liquid up and out of the tank. In order to refill the lines with coolant, it is helpful to pour only a little bit of coolant into the tank, and then close the tank before running the pump to force the liquid into the lines.