Flir E4 Thermal imaging camera teardown

[Clint]

Thanks, guys. My wallet really needed this.

Mr Wallet will love the fact everywhere are sold out

[ovnr]

Thanks, guys. My wallet really needed this.

Mr Wallet will love the fact everywhere are sold out

Still in stock a couple places here in Norway. But then I end up paying $1600 for it with VAT. :/

[edavid]

The General Tools GTi10 has a 160 x 120 IR resolution and 30 Hz refresh rate, and doesn't cost too much. I wonder if it's possible to hack it into the high-end GTi50 with 320 x 240 IR resolution?

Any idea who makes this?

[mrflibble]

If you want to have a quick look at the dead pixels on yooooouuuur E4, do the following:

Code: [Select]

# check for yourself that I'm not making stuff up. :P
rls .image.flow.maps.combGainDeadMap.pixReplace

# show dead pixels
rset .image.flow.maps.combGainDeadMap.pixReplace false

# hide dead pixels again (default)
rset .image.flow.maps.combGainDeadMap.pixReplace true


[ixfd64]

The General Tools GTi10 has a 160 x 120 IR resolution and 30 Hz refresh rate, and doesn't cost too much. I wonder if it's possible to hack it into the high-end GTi50 with 320 x 240 IR resolution?

Any idea who makes this?

Um, General Tools?

[mikeselectricstuff]

If you want to have a quick look at the dead pixels on yooooouuuur E4, do the following:

Code: [Select]

# check for yourself that I'm not making stuff up. :P
rls .image.flow.maps.combGainDeadMap.pixReplace

# show dead pixels
rset .image.flow.maps.combGainDeadMap.pixReplace false

# hide dead pixels again (default)
rset .image.flow.maps.combGainDeadMap.pixReplace true


Quite a lot....

[XOIIO]

Saw this on hackaday and am watching mikes videos, man this is awesome.

How long do you think it will take for them to fix it? I've always wanted a thermal imaging camera but I'm in a bit of a financial squeeze right now (19, unemployed, hard to find work) so I probably couldn't but one to hack for some time.

Also has anyone "bricked" theirs trying this? It seems unlikely but you never know.

[PA0PBZ]

Not too bad actually.

[mrflibble]

That is not a nice image 

If all of those dots are dead pixels, that is a lot worse than Mrfibble's image.

That is indeed a crapload of dead pixels. On my E4 that setting does show precisely the same bad pixels as what I extracted earlier on from the calibration picture. Translation: I'm pretty sure that setting shows the dead pixels as is.

I suddenly feel a lot better about my 99.919% operating pixels. 

[mrflibble]

Have you been exposing your unit to lots of X-Ray or very high intensity heat sources by any chance?

That shouldn't matter, as this setting shows the dead pixels as determined at the time of calibration. AKA: before Mike had a chance to stuff it in the X-ray to find where those pesky screws are.

[SeanB]

So it seems a lot of the units get the less than stellar detector arrays, while the good ones tend to be used for the high end, unless they have a batch that is spectacularly good yield wise and then they just rebin the lower end for the cheap ones.

[okent]

Exactly!  I spent a full day running around looking through mine at 80X40 and at 320X240 it's amazing!  No worries about any dead pixels for me.

As to the "bricking", I took the advice of Mike and others and just did the simple file upload and have stayed away from other advanced options.
It's so much more functional with that hack I just can't see tinkering around with anything else as of right now.
I have not heard of a single person "bricking" theirs to date.

One other thing of note, there has been occasional talk of the possibility of hacking this to 30hz or even 60hz.
With 60hz systems the battery life is significantly diminished.

[G0HZU]

I think I'll give the pixel checker a miss as I can't see it gaining me anything but varying degrees of woe over the quantity of duds...

Note:
I'm not sure if I'm missing something or doing something wrong but my saved images appear to show some lossy compression artefacts which is a bit of a shame. Is there a way to save the image in raw format or to reduce the amount of compression?

[mrflibble]

missing something or doing something wrong but my saved images appear to show some lossy compression artefacts which is a bit of a shame. Is there a way to save the image in raw format or to reduce the amount of compression?

Use exiftools.

http://www.sno.phy.queensu.ca/~phil/exiftool/

There's also a gui for it.

[Clint]

Saw this on hackaday and am watching mikes videos, man this is awesome.

How long do you think it will take for them to fix it? I've always wanted a thermal imaging camera but I'm in a bit of a financial squeeze right now (19, unemployed, hard to find work) so I probably couldn't but one to hack for some time.

Also has anyone "bricked" theirs trying this? It seems unlikely but you never know.

Aside from the novelty there isn't really much use for one unless you're doing serious inspection work or building PCBs...

Not much use for one, are you joking Once you have one you can't understand how you lived without one......

[Clint]

*for an unemployed student   

LOL - Even more time to play....... it would bring a whole new meaning to 'hot or not' !

[G0HZU]

If you open it in FLIR Tools the artifacts go away, but you have to screenshot to save.

Wow thanks!
I've now started playing with the Flir tools and I didn't realise you could examine the saved images in such detail.
The image below was taken using my latest focus tool to let me zoom in really close to a PCB.

[Rasz]

Have you been exposing your unit to lots of X-Ray or very high intensity heat sources by any chance?

That shouldn't matter, as this setting shows the dead pixels as determined at the time of calibration. AKA: before Mike had a chance to stuff it in the X-ray to find where those pesky screws are.

Does the camera generate dead pixels map every time it calibrates? I dont think so, I think its factory calibrated = Mike couldnt make new ones, as they would show up.

Not to mention customs Xrays everything anyway.

[mrflibble]

The dead pixels shown are those found during calibration by the Elbonian tech.

[tanntraad]

Hi!

I found your awesome forum after seeing Mike's post over at Hack a Day and, after a few hours of reading, I ordered the E4 from Tequipment. Impressive work you guys do!

[mikeselectricstuff]

I suppose bad pixels can be ID'd by simply looking at the offsets required for shutter cal - if above a certain threshold, then mask the pixel.
This may miss the odd one that happens to be stuck at the same value as the shutter temp.

[Mavro]

Please take a look at the two pictures of dead pixels that I attach here.....notice anything different ? except the obvious pixels of course !

Take a look at the thermal span on the right hand side. There is something not right in Mikes picture. The span goes way too wide and to -40 degrees Celcius  I wonder if this is the reason for his very ugly image ? No idea how the span is set but -40 Degrees seems unusual to say the least ! Has anyone else checked their E4 ? If so, please advise of span used by the camera. A detector tested over a smaller range of temperatures would presumably have less 'out of tolerance' pixels than one tested over a massive range that includes an out of spec -40 degrees  I smell something fishy here.

Mike, with all the experimentation on your unit (for our benefit) maybe you have corrupted something ? The -40 degree temp could have been caused by such ? She's only spec'd to -20 !

That happened when Mike blasted the temp sensor with "freezer"?

Watch it happen @ 38:01:
 https://www.youtube.com/watch?feature=player_detailpage&v=NtqUE67BUDI#t=2281

(embed code won't respect the youtube seek of "t=2281" so you have to manually seek to 38:01 -- bummer)

[mikeselectricstuff]

The span goes way too wide and to -40 degrees Celcius  I wonder if this is the reason for his very ugly image ? No idea how the span is set but -40 Degrees seems unusual to say the least ! Has anyone else checked their E4 ? If so, please advise of span used by the camera. A detector tested over a smaller range of temperatures would presumably have less 'out of tolerance' pixels than one tested over a massive range that includes an out of spec -40 degrees  I smell something fishy here.

I assume it's simply that the bad pixels are making it think there are some -40 deg spots in the image, so it's setting that as the minimum temp when the bad pixel masking is disabled. 

[equinoxe]

Obviously we can only know whether that is the case when there are more maps known.
Both hypothesis are plausible on first inspection.

If there are other maps that are "as bad" as Mike's and have the same temperature span, then it's not unlikely that the pixels are causing the temperature. But if on the other hand there are other maps out there that show no correlation between a large number of bad pixels and the temperature, then you at least have falsified one hypothesis..

[mikeselectricstuff]

Also notice my pic is in the high contrast palette so slightly bad pixels may be a lot more noticeable than with the standard one.
I can't see how freezing the temp sensor could have anything to do with it, as bad pixel mapping would all be done before any correction for ambient - I didn't freeze the image sensor.