Buy the print and get FREE access to our online edition!

Tech Forum

May/June 2018

Measuring Heat With Crystal Diodes

I have a bunch of crystal diodes in my junkbox that I want to use to measure heat. I seem to remember a circuit for it some years back that I’m trying to reconstruct from memory. It’s not working very well, so I must be missing something. Anyone have a simple circuit or explanation of how it should work?

John Marion
Bend, OR


Texas instruments always encouraged its engineers to use a diode in the emmiter of thier transistors. Why, because the reverse temp curve, cancelled the xistor temp curve. So we built several little circuits to monitor the temp. So, a series circuit that lets a low current through your diode, monitored with a sensitive meter, will give a value for different temps.

Caution, some diodes are sensitive to light, so take that into account. The 1n914 and 4148 were a common device. Never messed with the gallium devices, just be cautious they are easy to let the smoke out. Luck in your endeavors.

Tom Sides
Phoenix, AZ

I’m not certain what you mean by “crystal diodes” but most modern semiconductor diodes can be used as temperature sensors. A standard silicon diode will have a forward voltage drop, usually stated as “0.6V” or “0.7V.” However, it varies with temperature. Typically it will drop about 2 millivolts per degree C. By measuring that drop and calibrating at two extreme temperatures, the current temp can be determined.

The diode should have its cathode connected to ground and the anode should have a resistor to the positive supply. The voltage measured between the junction of the resistor /anode and ground will be the forward voltage to measure.

If by crystal diode you mean the old “cat’s whisker” crystals (e.g. galena) I suspect they will have much lower voltage drop. However, it will probably still show a temperature sensitivity, so may work. Silicon diodes work pretty well.

For an in depth explanation, with lots of math and theory, look here:

William Cooke
Clarksville, TN

Theoretically, the forward-biased voltage drop is about -2.2 mv/deg C at reasonable currents like a few milliamps. This works pretty well for silicon diodes like the 1N400x and the 1N914. It also works well for the base-emitter junction of a silicon planar transistor (almost anything in a TO-92), especially if you short out the base-collector junction. I have not tried it, but I think it does not work as well for point-contact germanium diodes like a 1N34 (which might include your "crystal" diodes).

Joseph Feng
San Jose, CA