back to index

Homemade thermal fuse


Thermal cutoffs are a good thing for electronics. They come in two flavors.

Thermal fuses are typically single-shot devices. They contain a spring-loaded fusible link, which melts when a specified temperature is reached and interrupts the circuit. They are inexpensive, but they cannot be reset after they trip and have to be replaced.

Thermal resets, in contrast, open when a temperature is achieved and reclose again after it falls below a threshold. They tend to be bigger and more expensive, and reclose on their own (which can be an advantage or a disadvantage).

A middle-ground device, cheap-to-free and small enough so it can be attached e.g. to the back of a TO220 transistor, with the ability to be user-reclosed, is missing on the market. Such device would be excellent for [?:protoboard] designs, where they could guard parts that are not supposed to get hot but could (e.g. resistors, diodes, transistors). The relatively low trip temperature gives an early warning and prevents the protected parts from getting so hot they could injure the operator.


A fusible link approach was selected for its simplicity. Wood's metal was chosen for the joint, on the basis of availability and a suitable melting point (71 °C).

For the substrate, a piece of universal circuitboard was made. A connector made from header pins was attached to one end. A piece of wire was soldered along the board to reach its other end. The end of the wire, coiled somewhat back, was not covered with solder. A piece of header socket was used as a holder for the fuse boards.

(A third version of the fuse used a small piece of copper as the substrate. This one has the advantage of better thermal conductivity and the disadvantage of not being insulated from the substrate.)

A spring was adapted from a phosphor-bronze wire, salvaged from a PCI connector from an old motherboard. Its "business end" was coated with an aggressive flux, then the middle of the spring was heated (to avoid direct contact with the soldering iron tip with the Wood's metal) and coated with molten Wood's metal.

The ends of the wires on the boards were smeared with aggressive flux. A couple shavings of Wood's metal were placed on the wire ends, and the wires were heated from the side with the soldering iron. The metal melted readily and wetted the wire and the underlying circuitboard.

The springs were soldered by normal means to the board or to the header pins (depending on the fuse version). Then they were bent back, to create a degree of tension.

To arm the new fuse, or to reset the tripped one, the spring was gently pressed down by its very end and the side of the fixed wire was heated with a soldering iron until the Wood's metal melted and fused with the spring. The heating was then stopped and the metal was allowed to solidify. Only then was the spring let go. (Sometimes the joint would not hold. Just redo it in such cases. Maybe melt the metal and wipe its surface to remove oxides if it gets really stubborn.) The spring's end was then gently pulled back to test the mechanical integrity of the joint; if it did not hold well, it was remade.

The heating for the fuse reset can be performed optimally with a soldering iron, but any heat source (e.g. a cigarette lighter or hot water) will do the job. This makes the design suitable for deployment even in field conditions outside of the lab.

The fuses were tested multiple times by heating on a soldering iron. This step is critical to get the right "feel" for the proper spring tension. Avoid pressing it down by the middle when arming, as that may bend it too much and lower the tension below what's sufficient for the fuse to trip.

Initial parts: mounting boards, connectors, springs, Wood's metal

Connectors attached to boards

Tinning the spring with the Wood's metal

Fixed contact

Fixed contact, Wood's metal added

Fixed contact, Wood's metal added

Spring pin attached

Spring attached and loaded

Spring attached and loaded

Joint detail

Joint detail

Joint detail

Finished fuses

Finished fuses

Finished fuses


Wood's metal contains cadmium. If subjected to arcing, it could vaporize and be somewhat poisonous. For such applications Field's metal may be preferred.

Prolonged higer temperature or repeated tripping may cause oxide buildup on the joint. Occassional wiping and refreshing of the Wood's metal bead when the fuse reset becomes difficult, or maybe encapsulation in e.g. a dab of silicone vaseline, may be useful.

Due to contamination of the Wood's metal with tin/lead or copper and other metals from the plating on the parts and substrate, the tripping temperature may become slightly higher.

A high temperature device can be made when solder is replaced with e.g. a silver braze, Wood's metal is replaced with a conventional solder, and the spring material is ensured to not lose elasticity at the maximum service temperature.


Other use of Wood's metal is a good thermal interface material. The metal wets copper well (when helped with rubbing and some aggressive flux), and also wets the plating on the backs of TO220 (and other) cases. While the thermal conductivity of the alloy is fairly low for metals, it is still pretty good in comparison with nonmetal thermal interface materials.


If you have any comments or questions about the topic, please let me know here:
Your name:
Your email:
Leave this empty!
Only spambots enter stuff here.