Why no scanning non-contact thermometer?

I'll assume you want a thermal imager (far infrared) rather than a near infrared imager, which can be done with a modified CCD camera.

The magic buzzwords are: "infrared pyrometer" (a detector) "thermopile" (the most common detector) "thermal imager" (the name of most products) "thermal imaging" (the name of the technology) "flying spot scanner" (the name of the rotating mirror scanner) "infrared line Scanner" (flying spot scanner and imager)

For example:

which has a nifty description of how theirs works.

More than you ever wanted to know about thermal imaging and measurement:

I built one about 10 years ago from of butchered supermarket bar code reader, by substituting a do it myself thermocouple for the laser. I was initially able to get a decent image on an oscilloscope using an LED in place of the laser, but when I tried the thermocouple, nothing worked. It turned out I had multiple problems. The thermocouple had very slow response. My mirror was spinning far to fast. Checking the frequency response, it was less than 1Hz or a 1 second response time. The front surface mirrors worked at near IR, but not far IR. Using a single thermocouple was also problematic because of the enormous gain required to obtain a usable signal. The picture was mostly noise. I gave up.

Later, I had the bright idea of using either a commercial SMD thermopile detector, or perhaps the detector from an ear thermometer. Something like this device:

However, I never tried it.

More reading on thermopile thermometers:

from white papers at:

Looks like response time is going to be a big problem.

Problem is, you need a germanium lens and/or mirrors, the detector is easier, get the thinnest platinum wire you can get, make a very small(0.2 mm diameter and length) for a fast response time at the focuspoint of the scanner, and put a constant current through the wire, measure/amplify the voltage across it, and voila, a (slow) IR scanner.

The problem is that the OP didn't bother to mention what he's looking at and for what application. Thermal imaging is very different from FLIR imaging, which work at different wavelengths.

I assumed thermal imaging (which is the most fun) from the thermometer term in the subject line.

You need a long wave IR (8-20 um) pass filter or germanium window, which can be made or purchased. I found mine at Edmund Scientific when they were selling seconds and rejects. Today, the prices are much higher:

While germanium is the best, there are some cheaper filters that will work:

There are some on eBay for tolerable prices:

If you're totally cheap, a piece of exposed 35mm color film will make a marginal substitute.

The front surface gold mirrors were salvaged from a supermarket scanner. They weren't the best but they were cheap. As long as the anti-reflective coating on the optics doesn't totally block IR, it will work. The problem with the supermarket scanner is that it's anything but portable. It was roughly a cube, about 10" on a side. Not exactly portable.

That's exactly what I did. The problem is that a single thermocouple just didn't generate enough voltage to be useful. The signal was buried under useless noise. I used the best low noise amplifier I could design, but it wasn't good enough. The solution was an array of thermocouples, called a thermopile. If I did that with my home made thermocouples, it would be too large to focus properly. In addition, it was important to add thermal compensation to the thermocouple, so that I would be measuring IR, and not heat from the electronics and packaging. Most thermopile detectors have this built in.

Hint: The lack of do-it-thyself scanning thermal imaging projects on the web should offer some clue as to the difficulty.