Thermal controllers are available from numerous manufactures at various price points in various styles. They usually have a few things in common, however. Most thermal controllers mount in an empty drive bay and monitor temps and rotational speeds that probes and system fans report to the controller. You can use controllers to complement the internal sensors most CPUs and GPUs have or to augment the thermal monitoring of components that aren’t normally monitored, such as RAM, hard drives, or a northbridge.

Thermal controllers may seem like complicated hardware to some, but they are relatively basic and uncomplicated. Controllers typically report temps and fan rotational speeds, although more advanced models monitor various voltages, as well. Thermistors usually are used to monitor temperatures. Essentially, a thermistor is a special type of resistor in which its resistance changes depending on its temperature. As the thermistor’s temperature rises, the thermistor’s resistance increases, and as its temperature drops, its resistance decreases. By checking the thermistor’s resistance against a known scale, temperatures are deduced.

Monitoring fan speeds is also relatively simple, provided you use proper fans. Thermal controllers monitor the feedback signal output transmitted on the yellow wire that is incorporated into most fans using a 3-pin connector. The red wire is the positive lead, the black wire the ground, and the yellow wire the feedback signal output. The feedback signal reports the fan’s rotational speed.


Hook It Up

Installing a thermal controller is fairly straightforward. It typically involves mounting the thermistors/thermal probes, connecting fans to components, supplying power, and mounting the controller in an empty drive bay. There may also be some creative cable management involved if you’re anal about keeping your system’s internal wiring neat or hidden.

The Logisys FP206 (about $25) that we used for this project fits in any standard 3.5-inch external drive bay. The device has three thermal probes and can monitor rotational speeds of three system fans. The FP206 also has a clock and an IDE indicator, but we’ll focus here on the probes and fan speeds.

The Logisys FP206’s three thermal probes consist of basic two-conductor leads with thermistors at the ends. The thermistors monitor the temperature of whatever they are mounted to, displaying the temperature on a digital readout. You can mount the leads to nearly anything using nothing but adhesive tape, but we took a slightly more creative and reliable approach. We mounted our thermal probes to our processor, system RAM, and northbridge chip without any tape and neatly routed the wires to hide them.

For this project we used an Asus 975X Express chipset-based motherboard, an Intel Pentium Extreme Edition 840 processor, and 1GB of Kingston DDR2-5400 RAM. Because this CPU is known to generate quite a bit of heat, we wanted to monitor its temperature with one thermal probe and dedicate the remaining two probes to the northbridge and the RAM, as they’re both passively cooled and get relatively warm in a closed system.

To mount each probe, we used a component’s existing mounting mechanism or its preinstalled heat spreader to hold the probes securely in place. Because our CPU uses Intel’s Socket T (LGA775), there’s a metal retention bracket that surrounds the processor holding it in place. We slid one thermal probe under this bracket when installing the CPU. This locked the probe in place off to one side. This is an ideal location because it doesn’t interfere with the cooling fan, and it makes direct contact with the processor’s IHS (integrated heat spreader). If you’re not working with a similar socket and processor, be aware that it isn’t a good idea to mount a thermal probe directly on your processor’s IHS or die. If you mount the probe on top, the cooler won’t mate properly with the processor, which could cause it to overheat.

Because our RAM was equipped with heat spreaders with openings at each end, we threaded a thermal probe into one side until the probe’s end was roughly in the middle of the DIMM. Situating the probe this way worked well, as we were able to get a reading from the heart of the DIMM where it gets the hottest and keep the probe completely hidden. When we fully assembled the system, the only hint that a thermal probe was present in the DIMM was a pair of small wires protruding from one side.

The passive aluminum heatsink mounted to our motherboard’s northbridge also has a retention clip, making it an excellent location to mount a thermal probe. The northbridge heatsink’s retention clip is a simple, angled piece of steel that acts like a spring. To mount this third thermal probe, we disengaged the clip and lifted it slightly. Then we slipped the thermal probe underneath and reattached the clip. The clip provided enough pressure to securely hold the probe in place, and we could hardly notice the leads once we routed them to the edge and underneath the motherboard.

Numerous other system components are prime candidates for adding a thermal probe to. Today’s high-end hard drives tend to generate a lot of heat, making them very good possibilities. A motherboard’s VRM (voltage regulator module) is another component that gets quite hot. Parts of some VRMs are now even mounted on the motherboard’s underside where little air circulates, so keeping an eye on VRM temps is a good idea.


Make The Connection

Once our thermal probes were in place, we needed to connect our system fans to the thermal controller, supply the controller with power, and mount it. Connecting the fans was as simple as it sounds. We just had to slide the appropriate fan connectors over the headers coming from the controller. Again, our only requirement was using fans equipped with three-pin connectors and a feedback signal output. The Logisys FP206 has a common 4-pin Molex header that accepts any available connector coming from the system’s PSU. We made the connection and installed a couple screws to secure the controller in place in a 3.5-inch drive bay.


Knowledge Is Power

With the controller installed, we can monitor the temps of three vital components and check that three system fans are functional and rotating as they should. Using the temperature data while we’re overclocking will be helpful to determine if heat is the root of system instability. For example, if our system crashes or generates an error every time the northbridge or RAM hits a certain temperature, we’ll know we’ll have to back down on our overclock or more aggressively cool these components. Additionally, if a fan fails, we’ll know immediately and be able to resolve the problem quickly. We can also program the thermal controller to sound an alarm should any probes hit a certain temperature.

The data a thermal controller such as the one we used can prove invaluable in certain settings. Installing one is definitely a worthwhile mod.

by Marco Chiappetta