Thermal cameras can be a huge help around the electronics lab. With the ability to instantly visualize the many different temperatures of an area, there are applications beyond the obvious heat sink performance characterization. An engineer can use them when troubleshooting a board, to understand sensitive circuits, calculate junction temperatures, and work through power consumption issues. In the past these tools have been very expensive, on the order of $20 – 30k, such as the instrument-grade Mikron TH7515. Recently manufacturers have introduced a new low-cost line for these handy tools with Fluke's VT02 base model going for under $900.
Thermal cameras are tons of fun to play with, and I had the pleasure of reviewing Fluke's VT04 Visual IR Thermometer. It sits at the high end of their visual IR Thermometer line, giving a 4x sharper image with a wider field of view than it's VT02 cousin. The VT04 also boasts plenty of niceties not available in the VT02 like a rechargeable battery, time-lapse function, and temperature limit alarms which puts the price up to $1,200. For Fluke's complete listing of specs check out their datasheet, although it is disappointing that they don't spec resolution of the thermal image.
Fluke contends with consumer giants like Apple with their packaging. Not with some floaty, useless cardboard container but with a fitted foam-padded case; exactly what I want to see from a nice piece of kit. The unit comes with everything needed: the VT04, charger, USB cable, SD card and adapter, software CD and a quick start guide. I found getting my first measurement to be just as easy as using a cell phone camera. Another 2 minutes with the well-written and well-illustrated quick start guide and I was running the system like a pro. Measuring the temperature of points, changing views, blending the visual spectrum, and saving images are all intuitive and easy. The only negative I noticed was the unit audibly screeched and had a rattle to it, which does not meet the expectations one would have of Fluke.
With a good handle of the system's operation, I started to get down to business. The most basic use for a thermal camera is to help troubleshoot poorly performing boards. Examining a board with the VT04 makes it easy see if there is a part that is running hotter or cooler than usual. It won't tell you everything, but it certainly can be a time-saving hint! Here is an image of a board that has a fried voltage regulator:
The camera may also be helpful in determining the part on a board that was consuming more energy than expected. If a board is consuming 50mA when it should be consuming 1mA, finding the problem component with traditional methods can be tricky and involve lots of soldering and measurements. Using the VT04, finding the part that is consuming the most power is easy. Here is an example of an Arduino Due that is consuming 20mA but should be consuming much less thanks to the system being asleep. The image shows that most of the power is consumed by the Atmel chip, suggesting that the 'sleep' routine may not be implemented correctly.
Another general use of the thermal data is to aid sensitive layouts. Some analog components are most accurate with stable temperatures, so placing them near hot components can introduce unnecessary errors. Returning to the Due board above, the image shows which components on the board are warm, and how far the heat spreads on the board. This data suggests that I can place a temperature-sensitive component near the microcontroller and not worry about local temperature variations.
With the images saved in the camera, the next step was to install Fluke's 'SmartView' visualization software. Sadly, the pleasant user interface and plentiful features I enjoyed when starting with the unit hit a hard stop here. Test and measurement companies are some of the worst when it comes to computer interfaces, and the VT04 is no exception. Connecting the camera's USB port to my computer did not show the contents of the SD card, forcing the use of a micro SD adapter and card reader to transfer files. Upon reading the manual, I found that the micro USB cable is only mentioned as a charging cable, so USB data does not appear to be supported. Then, the most important feature of being able to read a temperature for each pixel on a thermal image was missing. Though the software was capable of displaying exact temperatures for the center point, coldest point, and hottest point. I assume this is because the non-specified thermal image resolution is poor and would show too few points to be useful.
Looking at more precise uses for the camera is limited because the resolution is not specified. One cannot be sure what size area is being averaged in the “point” measurement. With a known 'point area', confidently recording the case temperature of an individual IC without including the surrounding area can be a powerful tool. Take the instance where a camera measures the case temperature of a TO-92 transistor having a thermal conductivity spec from junction to case of 85*C/W. If the FET is consuming 1W of power when the case temperature is measured to be 50*C, one can calculate the junction at approximately 135*C; too hot! In some cases, this data can also be used backwards to calculate rough power consumption. If that same TO-92 transistor has a junction to ambient spec of 200*C/W (no heatsinking or fan), one can assume a thermal conductivity of case to ambient to be 115*C/W (200* J-A minus 85* J-C). Thermal data showing the case to be 40*C above ambient suggests power consumed by that part can be approximated at 0.34W (40*C / 115 *C/W = 0.34W).
All in all, I can say that bringing the cost down on visualizing temperatures is great news for anyone lucky enough to work in electronics. It will deliver a time saving and remarkably fun tool to benches that would otherwise go without. The ability of the VT04 to confidently measure a small point on a board, however, would be best left to the high-end cameras holding the confidence of complete resolution specifications.