Need help with selecting a heatsink for a TRIAC

[rick_1976]

Greetings,

The math 'aint usually the hard part, so much as properly interpreting the documentation so as to have proper numbers to math on… Pre-cooked discussion on the topic is available in the posts linked below.

A thermal management example (part one)

The purpose of this page is to illustrate the process of making basic thermal analysis calculations common in electronics applications, and finding info in product literature that's relevant to that process. For sake of example, we’ll assume that we’ve got an IRL3713 transistor in a TO-220 package attached to an HS278-ND heat sink via a BER183-ND thermal interface pad, looking similar to the model in figure 1. [image] Figure 1. Model of the thermal assembly under consideration. In the conte…

A Thermal Management Example (Part 2: When Datasheets Attack)

In part 1, the process of making basic thermal management calculations was described. This page builds on that discussion, and demonstrates the importance of such analyses as an inflation gauge for certain component specifications. The previous example assumed use of an IRL3713PBF transistor (TO-220 package), HS278-ND heatsink, and BER183-ND thermal pad, and showed the calculation of device temperatures using (for sake of simplicity) an assumed 4 watts of electrical power dissipation in the d…

Some relevant snippets from the datasheets are below.

The FET figure 1 indicates that at your specified 10A maximum, the triac can be expected to burn ~10watts.

That amount of thermal power (modeled as a current flow) has to go from the semiconductor device to the outside of the semiconductor package (modeled as a resistance, R_ΘJA) then laterally through the copper mounting pad to the heat sink’s mouting point, then from the heat sink’s own thermal resistance into the ambient air.

Unfortunately there’s some missing information here:

1. How much thermal resistance to expect from that lateral transfer through the copper pad
2. What sort of air velocity the fans mentioned are going to offer
3. What you specify as the maximum permitted ambient temperature for your system

Part 1’s kinda tough. For sake of discussion, let’s spitball that figure as 2°C/W
Part 2: Let’s assume an air velocity of 500 FPM. Per the heatsink’s chart, that should make the heatsink’s effective thermal resistance about 2.5°C/W.
Part 3: Let’s assume you call it 50°C.

Total thermal resistance is then expected to be about 1.4+2+2.5=5.9°C/W

Multiply that by your thermal load: 10W*5.9°C/W=59°C. This is how much you would expect the semiconductor junction inside the chip to increase in temperature above what you’re cooling it with. Since we stipulated that to be 50°C, under these conditions and assumptions we’d expect the silicon in the chip to max out at about 109°C, which is a comfortable distance from the 150°C max mentioned in the datasheet.

Now, let’s contextualize that a bit:

• That thermal resistance figure across the copper pad is an educated WAG. Reality will differ.
• The reality of your airspeed situation will also differ.
• We’re ignoring effects of direct cooling of the package from this forced airflow. That should end up working in one’s favor.
• If reliability/longevity are concerns, one probably doesn’t want to visit that 150°C mark on the regular, and certainly not play peek-a-boo with it. Differences in thermal expansion coefficients put stress on the bond wires within a part and degrade the thermal contact of the silicon to the outside of the package. If your application is of a sort that’s going to be cycling between full-on and off every few minutes, targeting a lower maximum temperature is probably wise.

In summary, we could probably say the proposed approach is plausible, but needful of validation.

Do be mindful of how/where you plant those temp sensors mentioned. Under your max conditions the junction’s already going to be 14° hotter than the copper pad, so if you were to have a bit of copper between your triac and the sensor with a bunch of cool air flowing over the works… Well, it’s not hard to imagine getting measurements that are on the order of ~30° low.