thermal imagers

Thermal imagers, from Fluke, Flir, and Wahl, have come down hugely in price in the last couple of years, with decent looking units around $2500 now. An imager is a fabulous tool to have around.

John

Reply to
John Larkin
Loading thread data ...

I keep looking for a good enough excuse... TEquipment.net has the FLIR I7 for about $2000 and it has just enough resolution / minimum focus distance to put one 0805 into one pixel, which seems to be about the threshold of usefulness.

--
Rich Webb     Norfolk, VA
Reply to
Rich Webb

Doesn't CenTec, out of China, make a thermal imager for REALLY cheap, as I recall $299 or such.

Reply to
Robert Macy

We have a FLIR E40 which are around $4K The E30 is cheaper and most likely would have sufficed but there were evidently none in stock at the time. The first thing we found that was absolutely necessary was the manual focus adjustment.

We looked at an imager over 10 years ago and it was around $50K and probably not as good as today's instruments.

It is truly a wonderful instrument for production and engineering.

boB K7IQ

Reply to
boB

We have an E45, which we paid about $12K for a few years ago. It has a close-up lens that will resolve the hot-spot temp on an 0603 resistor. I wish we'd paid less for it, but it was sure worth it.

Our test people use it a lot. If some power supply is being bogged down, just peep the board and see what chip is running hot.

It's great for stuff like this:

ftp://jjlarkin.lmi.net/UPA800_80mW_one-side.jpg

That's a dual transistor that, thermally, isn't much better than two separate transistors.

John

Reply to
John Larkin

In the dark ages, at least for chips, we used liquid crystals to find hot spots. If the device on the chip wasn't getting hot enough for a phase change, you would bring a soldering iron nearby to "prime the pump".

Reply to
miso

I also use a hot spot finder with audible enunciator...

If the component is very hot when the finger-probe is drug over the top of the hot part, it emits a loud "OUCH !"

boB

Reply to
boB

Wow! almost 90F background. Seems a little warm already, no wonder the transistor went up to 50+C

Did you calculate the chip temperature?

Everybody seems to forget that it's temperature rise ABOVE ambient! The concept was reinforced for me once when the office got to 90 and the rise above ambient for the monitor went beyond that to provide a 'burn' on my arm as I laid it across the top of the monitor

Temp all adds up.

Reply to
Robert Macy

That may be a little deceiving. The PCB itself is fairly reflective at thermal wavelengths; copper is an almost perfect mirror at thermal wavelengths, so some of what you see is the reflection of ambient, specifically the ceiling of my office on July 21. That's the big gotcha with thermal imaging or thermometry: if the object's emissivity is not 1.0, what you see is some mixture of actual object temperature and reflections from elsewhere.

Two tricks: dab the object with black whiteboard marker, or put a bit of Kapton tape on it. Both are almost black at thermal wavelengths.

The epoxy used to encapsulate transistors is pretty good as regares emissivity. The point of that photo was that one transistor of the pair can get pretty hot without much warming the other.

No, I was just interested in differentials.

John

Reply to
John Larkin

We use the Fluke TiS ($2500) which has a focus control. Allows you to get within 6 inches of the subject. This works nicely for electronics and is an invaluable troubleshooting tool as well as sussing out heat issues in new designs. The FLIR cheapo model has fixed focus which limits the closest useable distance.

Fluke claims the image is a 120x120 pixel image. If you process their image file, you'll find the image is really 160x120 pixels.

Reply to
qrk

Speaking of reflection and emissivity, do you have a good idea for what the emissivity adjustment of say, a TO-220 mounting tab should be ??

boB K7IQ

Reply to
boB

I was thinking for board level use, there should be some device to wave over components. It would have to be fast, not a thermal couple. Maybe PIR film.

Reply to
miso

The metal part? It's so reflective that an e correction probably isn't worth trying. Put a dab of black whiteboard marker on it, or a piece of electrical tape.

Measuring the epoxy body shouldn't be bad.

Or use a thermocouple!

John

Reply to
John Larkin

Back in the day, I was using a thermal imager to find board shorts. If you spray foot powder on the board, it becomes uniformly emissive and much easier to diagnose. I was elated until I discovered that I couldn't get the foot powder off the board. Bummer. But at least, the board never suffered from athelete's foot.

Reply to
mike

Snow has an emissivity of almost 1. Maybe sprinkling snow on a board is the way to go.

John

Reply to
John Larkin

Yeah, we use thermocouples when testing units for UL/ETL and that kind of stuff but for production that's not too practical. I do put black tape or mark shiny things like FET tabs for better accuracy, but it seems accurate enough at least for comparative measurements withtout any blacking. I just though you might have stumbled upon a setting that might work better.

I really don't like the IR reflective properties of things either. Sometimes it even makes is kinda hard just to point the camera at right part of the object. It's weird.

boB

Reply to
boB

One quick test for emissivity: image the object from an angle, not too close, and wave your hand around in the area where you'd expect an optical reflection, or just try to have your hand fill in as much space as possible around the object. If the temperature changes as you move your hand, or if you can see tha reflected image of your hand moving around, the imager is seeing the reflection of your hand, because of the reflectivity (low emissivity) of the object.

John

Reply to
John Larkin

Yes, Great idea !

boB

Reply to
boB

film.

I have used a near field IR thermometer with a hole in some cardboard for a lens. Holding it less than an inch away from components.

Greg

Reply to
gregz

On a sunny day (Sun, 15 Jan 2012 01:59:26 +0000 (UTC)) it happened gregz wrote in :

I like that :-)

Reply to
Jan Panteltje

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.