Thermal IR detector

on

degree

I like it, John. ISTM that it should do the trick.

Getting Q1 to respond strongly to external temperature will be the trick. It should be thermally insulated -- give it long leads and back it with styrofoam or some other insulating material. It should get a lot of "view" of the hot area, but I'm not sure what'll work best aside from lenses made from expensive materials. (Possibly a tube that's shiny- polished inside -- but how do you achieve _that_ reliably?)

We talked about that in a thread above. It needs a small metal parabolic or spherical reflector to aim at the stove, Q1 at the focal point. I've used a flashlight reflector for stuff like this, worked great.

A tube won't add much gain or directionality.

Hi,

This one is pretty neat:

Superimposes a 64pixel (16x4 array) onto a visual image

cheers, Jamie

That's cool. It probably uses the Melexis 64-pixel imager.

Thermal imagers are getting a lot cheaper lately. An imager is a fabulous tool for testing electronics.

Which one do you recommend, John?

tool

We have a Flir, but it cost $12K a few years ago. There are others around for under $3K now, like the Fluke. I haven't tried one.

The Flir PC software is kind of awful.

Extech had a cheap imager, so Flir acquired them.

The super closeup germanium lens on our Flir cost more than that. It will image the hot spot on an 0603 resistor. I'd try one out before I bought it. If it has cheap optics, it could do fine imaging HVAC ducts but might not be good for electronics.

Maybe someone else has tried a lower-price imager for electronics.

tool

image

has

Thanks for the information. I'll look into it.

If you want directionality then a parabolic reflector. If you want maximum flux concentration it is a parabolic shape but with the focus arranged to be at the opposite edge of the cylinder of rotation.

Parabola with focus at centre dot.

| | \ / ' ' ' becomes a simple non focussing flux concentrator

| | \ / ' ' There is a proof that any ray entering the front aperture will hit the exit aperture after a finite number of reflections

Or rather nice warm thermal IR toy imager. For $200 well worth a punt.

Proper calibrated ones retail for in excess of £2k (and they have a few more pixels as well)

thanks John. Because my stove is so close, should work great, even without a perfect reflector. I'll put it in a box to control a relay for my fan.

What I want is a pen-like device which measures temperature a few inches away with a small diameter, say < 5mm, and shines a < 5mm beam of appropriately coloured light on the measuring point.

Then I can 'paint' colour onto PCBs to see the hot spots.

Is there such a thing?

Cheers

perfect reflector. I'll put it in a box to control a relay for my fan.

An IR-transparent window would keep convection off of the transistors. Black plastic garbage bags are nearly transparent at thermal IR, so you could use a piece of that as a window.

--

John Larkin Highland Technology Inc

Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom timing and laser controllers Photonics and fiberoptic TTL data links VME analog, thermocouple, LVDT, synchro, tachometer Multichannel arbitrary waveform generators

There are companies that specialize in IR transparent materials. I've used a IR filter over a B&W CDD camera and have found some optical devices whose black paint that is totally transparent to IR. [Not sure if that is good or bad, but I think it is good if you want the device to stay cool.]

Note that the traffic cameras (red light/speeding) use either extended red film or B&W CCDs to image through the tinted windshield glass.

Incidentally, it is these Sony Super HAD (hole accumulator didoe) sensors that have the good IR response.

tool

image

has

I had no way to try it, so I just ordered it.

It is a hassle because of government export regulations. I had to sign a document to show who I am and to testify that it will remain in the US.

However, that's okay. I simply want to see what my PCB is doing when I run max current through it and how the heat spreads. I've always wanted to know that ever since my technician days about 50 years ago.

Cheers, John

I'm cheap and just use a liquid crystal sheets. Resolution is admittedly lousy, but good enough to detect hot spots, especially on the flat circuit side of the PCB. If you want better resolution, buy one of the relatively cheap surplus germanium lenses on eBay, and project the far IR onto the liquid crystal screen. While not as sensitive as direct contact, the image is much sharper. If you're really really really cheap, you can make a passable pinhole camera lens that will pass IR, using a milk carton, and a liquid crystal sheet as the "film" plate.

lot

tool

image

has

The $2500 Fluke IR cameras work great for circuit boards. Our technicians love using this camera. No problem picking out a hot 0603 part. They focus down to about 12 inches. I like the Fluke units since they have a focus ring. There are other cameras out there in the same price range which are fixed focus - thus not suitable for electronics work.

Do you think that would work? I'd envision microkelvin contrast levels on the image plane.

Yes, it sorta works, but it's not like a photograph. With a pinhole lens, it takes forever to pass enough IR to get a visible image. I've only tried the pinhole trick once and found it useful.

The problem was I was trying to measure the heat distribution in an RF power amplifier assembly, which was a 3D problem. I was looking for hot spots along several PC transmission lines and for heat spreading on the heat sink. I couldn't lay the liquid crystal sheet flat on the board. I also couldn't use thermocouple probes because all the RF would wreck the readings. The hot spots (later measured with a thermistor probe) were in the 75C and up region which provided plenty of contrast between the approx 30C PCB. One was hot enough to melt the solder holding a porcelain capacitor.

The pinhole trick worked well because at f/100 (or something like that), the depth of field was impressive. The hot spots were in focus at almost any camera position, but quite dim due to the small aperture. I don't have any liquid crystal sheets or I would throw something together and post some photos. I don't think it will work with small SMT resistors and low power levels because of the lack of temperature contrast. However, if you can get a sheet in direct contact with the PCB, it should provide a useful thermal picture.

The liquid crystal sheets are sensitive over a range of about 5C but most of the color change occurs in about a 2C range. See the graph at: under the "Technical Images" tab. That means that I have to be careful as to which sheet I select. That proved rather awkward, so I simply moved the average sheet temperature by blowing hot or cold air over the back of the sheet to move it into the correct range. That produced some odd looking patterns, but the hot spots appears with reasonable contrast and clearly visible.

I once had a small germanium lens, which worked much better. I haven't seen it in years, but I'll see if I can find (or borrow) one.

tool

image

has

As I said, I ordered it. It came in yesterday, I charged it, and used it on a board I've been working on.

It was a revelation. I took a shot of the board (after it was hot) and was able to paste it as a transparent image over the board assembly drawing.

Now I see.

Thanks, John Larkin, for the lead.

John S