Temperature tracking versus time under software control

[JohnG]

I have a problem that I would like to solve, and was hoping I could find some help here.

I'd like to be able to use a thermal camera to measure the temperature of several (2 - 10) points on a printed circuit board, and read these temperatures from a Python program in something approximating real time, i.e. a second or two delay would be ok. In a perfect world, I would also be able to command the camera to take a picture from the same program as well, and store it.

Right now, I have a FLIR T420, and ideally I would like to use this camera to do this, but I would consider one of the cheap solutions with an added lens, if it is more likely to be successful.

The purpose of this is to monitor the temperature of several components in a power converter over time. My program already runs a power supply, electronic load, and an Agilent data acquisition system, and takes data. I would like to add the capability to measure several temperatures in addition to voltages and currents.

Please note that I am not a good programmer and will never be one, but I can do a little bit of Python and Octave/MATLAB. I can follow directions pretty well, though...

Thanks,
John

[Fraser]

Hi,

Sorry I cannot help on the programming side either as I am no coder.

I know you want to use a thermal camera, but have you considered something like the FLUKE 2635 HYDRA DATA BUCKET ? It is designed for this sort of analysis. I have one and it provides many thermocouple channels that I can use to monitor the temperature at many points within a piece of equipment. The thermocouples do need to be taped down onto the IC's with Kapton tape but this approach has the advantage of allowing closed case testing so air flow is accurate for the normal operating conditions.

http://us.flukecal.com/products/obsolete-products/hydra-2635ac-portable-data-acquisition

http://www.bellnw.com/manufacturer/FlukeRef/2635A.htm

I no longer need my Hydra Data Bucket so might even sell my unit if you decide to go that route. Its a clever bit of kit that saves data to a flash card or interfaces to a PC. I have the very expensive FLUKE PC software for it as well 

Fraser

Just a thought.

[JohnG]

Hi Fraser,

Thanks for your reply. I have already considered using a data acquisition system, since I have an Agilent 34970A that I use to collect voltage, current, ambient temp, etc. The problem I have with using thermocouples is that the main parts of interest are chip-scale packaged parts, i.e. die with bumps, so it is really difficult to get even very small thermocouples to stay attached.

The bigger problem, though, is that at least one of the die surfaces that I want to measure is electrically active with a rectangular voltage pulse of 100-200V peak. I have yet to find a thermcouple meter that will accurately read the temp, and even a thin layer of electrical insulation is not enough due to the capacitive coupling, not to mention that it degrades the temperature measurement. I've also tried thermistors and RTDs, but they are only slightly better.

John

[Fraser]

Hi JohnG,

You have the perfect scenario for using a non contact method of temperature measurement....totally agree with your proposed approach of using a multi point thermal image for temperature tracking. Some of FLIR's advanced software packages designed for scientific use may offer such capabilities but sadly my experience of these software packages is that they are very expensive. FLIR ThermaCAM Researcher Pro was the software I used to use.

There is a free 30 day trial here:

http://www.flir.co.uk/cs/display/?id=42404


I think the new version is called ResearchIR and may be found here:

http://www.flir.com/science/display/?id=51371

http://www.flir.com/science/display/?id=66484

Best Wishes

Fraser

[hs3]

For E4 and some other models this software could maybe be of some use for automatic data collection:

http://joe-c.de/pages/posts/programm_ex_tools_en_162.php

[tomas123]

if you have an E4 and some program skills look in this thread

Flir Ex: Realtime raw radiometric data streaming via UVC
https://www.eevblog.com/forum/thermal-imaging/flir-ex-realtime-raw-radiometric-data-streaming-via-uvc/msg623251/#msg623251

you can stream the RAW frames to PC and calculate the temperature of every pixel



with the FLIR T420 WiFi Connectivity (or with USB) you can

• take an image over telnet
• load the radiometric image over FTP (flir/3vlig) to PC
• with exiftool and some mathematics get the temperature from some pixels

The purpose of this is to monitor the temperature of several components in a power converter over time.
....
Please note that I am not a good programmer and will never be one

my suggest:
use Flir Tools+ with RAW video:

[JohnG]

Hi Tomas,

My programming skills are such that I can follow instructions, but am not very creative . However, I think I can try the following, and control it from Python. If it works, it will be sufficient for my needs.

if you have an E4 and some program skills look in this thread


with the FLIR T420 WiFi Connectivity (or with USB) you can

• take an image over telnet
• load the radiometric image over FTP (flir/3vlig) to PC
• with exiftool and some mathematics get the temperature from some pixels

I do not know if I can extract the data output from FLIR Tools+ in real time. I also don't trust FLIR tools, as I find the basic version buggy. I also find the camera itself to be buggy and crash-prone, but I'm stuck with it for now. It works great most of the time, but if you leave it on, eventually it will crash. This has happened at some VERY inconvenient times, and FLIR support was not very helpful.

Thanks,
John

[tomas123]

I know two variants to take an image over telnet:

Code: [Select]

>store                       
usage: store [mode] [options] <filename> [2nd filename] [form template ID]
image mode:
x (no flag)  Radiometric FFF
x       -j   Radiometric JPEG
       -e   JPEG only (non-radiometric)
       -v   Large visual JPEG                                           >store -v \storagecard\40.jpg
       -w   Small visual JPEG
       -y   Visual FFF                                                        >store -y \storagecard\33.fff
       -z   Visual JPEG with preview (large visual inside)    >store -z \storagecard\32.jpg
       -s   Sequence FFF (SEQ)
x       -m   Combo IR + visual (must give 2 filenames)     >store -m \storagecard\29a.jpg \storagecard\29b.jpg
       -f   Fusioned IR/visual radiometric JPEG                   >ERROR store: in wrong state
       -d   Difference FFF (add -j for JPEG)
general options:
       -a   Averaged image                                             >store -a \storagecard\34.fff  >store -a \storagecard\42.fff
       -p   PNG compressed pixels
       -l   JPEG-LS compressed pixels
       -u   Use last grabbed image
       -i   Add image(s) [and form instance] to new IIG    > store: cannot add IIG to db, error 0x64012013
       -t   Timing mode                                             >store -t \storagecard\35.fff  time: 543 (427) ms
JPEG options:
       -o   No overlay (keine Scala etc)                     >store -j -o \storagecard\36.jpg
       -r   Only overlay                                             
       -g   Use POGO (if available)         weniger Displayelemente als overlay                   >store -j -g \storagecard\37.jpg
       -c   Cut image with overlay to IR image size     >store -j -c \storagecard\38.jpg

The second method is the rset command.
The Flir Tools App use this method.

see this Wireshark TCP recording as snippet of used commands:

Code: [Select]

root.3vlig
image.services.store.overlay
image.services.store.irCut
image.services.store.overwrite
image.services.store.filenameW
image.services.store.stage.start
image.services.store.bgActive
image.services.store.busy
image.state.channel
image.services.store.format
image.services.store.format.JPEG
image.services.store.filename
image.services.store.filename./Temp/irviewerimg.jpg
image.services.store.commit
image.services.store.bgStoreNow