# Determining values when designing with thermal fuses; calculating influence of location/proximity to heat source & properties of protective coverings

The root question: Are there any best practice rules or guidelines for spec'ing out a system with a non-resettable thermal fuse? What criteria is used to determine what rating fuse is needed?

I'm primarily talking about the fuses you find in most appliances that have heating elements or can fail in a way that leads to continuous heat generation, fuses that serve as a final fail-safe against an overheating device igniting a fire, the ones that self-destruct when the load + environmental temperature literally melts the internal workings and breaks continuity.

They come in a wide range of failure points, with sub-degree precision, but unlike other fuses, the physical location of the component in relation to both the primary heat source and the target of concern would have a huge influence on what rating should be used. So how does one choose?

For really temperature critical applications presumably you'd have a multidisciplinary team on it, but there must be some general concepts, ground rules, something for the EE to consider and take into account when designing the circuit?

Real world inspiration for the question:

The loop for the defroster in my freezer has a thermal fuse rated 77 C. It seems logical enough to assume that something has gone horribly wrong if your sub-zero freezer is now hot enough to burn your skin off - but that's not actually what the fuse is measuring. It's measuring a localized temperature from its spot nestled between the heating coil and the fins, behind a panel and insulation, and your ice cream could still be frozen rock solid at the same time that fuses trips.

It's not measuring the surface temperature of the heating element itself, nor is it in contact with it (or anything else) so it's clearly meant to trigger based on the wider environmental temperature - but what catastrophic failure in particular is it preventing? What happens above 78 C that this fuse is trying to prevent?

B/C physical failure of another part of the system isn't only influenced by the temperature at that one spot. A BIC lighter will melt if held over the flame of another BIC lighter, but it can't melt itself with it's own flame - you've got all sorts of thermal mass and surface area distributing that heat energy out. A fuse that fails at the melting point of a material it's trying to protect will be grossly over sensitive. It's like saying nothing in your furnace can get above 30 C because that's what you set your thermostat to. Heating elements get much hotter, else (even in a theoretical heat-loss-less world) it would take forever to warm your house up.

The true motivation of this question:

I've got a blown thermal fuse. I've fixed the problem that caused the overheating, but need to replace the fuse - and the replacement part is no longer available. It's literally the fuse on a wiring harness, crimped into some sort of protective tube housing. I can buy a replacement component easy - but I don't know what to do about the tube housing. The air gap alone would provide significant temperature buffering capacity, and the case material is not insignificant either. I want to ensure that whatever I cobble together to replace it will behave as originally designed, but to that, I need some inclination of how it was designed and/or identify what it is intending to protect.

Image is of a typical thermal fuse assembly and harness, shamelessly stolen from the internet.

The actual component itself is a SW-105T by Sungwoo. Rated for 77 C, 250 V, 10 A. (Mfr details here; US distributor spec sheet here)

I can buy a replacement component easy - but I don't know what to do about the tube housing.

It's a PVC tube hot-crimped on the wire. Get a piece of similar diameter PVC tubing, string onto the wire and thermistor, then crimp - if you want to be really close to the original.

In practice, grab some adhesive-lined heat shrink tube with a diameter just large enough to slide over the fuse, shrink the ends - or the whole length - to cover the fuse & wire just like the PVC pipe did. It'll be conservative, i.e. will trigger no later than the original fuse did, since the heat shrink+adhesive will be thicker than the tube most likely, or will even be in contact with the fuse. Replacing air with a solid lowers thermal resistance. So, as long as the fuse won't blow, everything will be all right.

The air gap alone would provide significant temperature buffering capacity, and the case material is not insignificant either.

The original design had very wide margins too. This is a catastrophic failure detector - it needs to trip sometime before things get too dangerous and/or likely to catch fire. It doesn't need to be ultra-precise.

They come in a wide range of failure points, with sub-degree specificity

If you mean sub-degree accuracy: not really. And anyway - the faster the thermal transient, the smaller the volume where you can measure a given temperature to any desired resolution.

The fuses are usually tested in a prescribed test scenario that may not have much to do with conditions it's applied in. The "resistor-like"-packaged fuses you show are IIRC tested in an oil bath. The designer of the fridge did enough for the fuse to trip soon enough, and that's probably the extent of thought that went into it. Sure, you can model such things, but the cost quickly outpaces the benefits in the fridge application you mention.

• I'm not sure heat shrinking is a viable option, given the low fuse threshold in this specific scenario. Most polyolefin tubes have a shrink temp above 90* C, and most with adhesive linings are also double walled - a non-negligible amount of thermal mass that exceeds the fuse rating. Heat sinking with pliers as you might when trying to solder a temp-sensitive component isn't an option, and pre-chilling would likely interfere with the adhesive. Commented Aug 17, 2023 at 18:53
• Sub-degree precision, not accuracy. The availability of components rated in small increments indicates that the components are highly precise - implying that there are use case scenarios that require that degree of precision. supporting the basis of the question that there must exist a methodology to determine the value necessary to achieve the desired results. Whether or not the value printed on the part when it comes in is accurate is irrelevant. Commented Aug 17, 2023 at 19:24