Flir E4 Thermal imaging camera teardown

[Bud]

I recall there is a http web page on E4 where you can read temperature from a few temp sensors in the camera. Should also be readable via telnet session. The stock e4 does not have the web folder , it can be just copied to it, the source was provided somewhere in this thread.

[Fraser]

Most microbolometer based thermal cameras contain the following temperature sensors.....

1. Chassis temperature
2. Microbolometer temperature (at least 2 sensors normally used)
3. FFC shutter flag temperature (or the air around said flag)
4. If a Germanium lens is present ... The lens temperature
5. If a Peltier temperature stabiliser is present ... Hot and cold sides of the Peltier element.

The cameras radio metric temperature measurement routines need much of this sensor data in order to provide accurate and repeatable temperature measurements.

For those wondering why the temperature of a Germanium lens is needed....Germanium changes its transmission spec as its temperature rises. The greater the temperature the lower the transmission percentage until at 100 degrees C it is almost opaque to thermal IR wavelengths. The effect only really starts to become an issue when the lens temperature rises above 50 Degrees C though.

In situations where the Germanium lens is exposed to very high temperatures, such as a furnace, steps are taken to cool the lens assembly. Cooled Air or fluid cooling lens jackets are commonly used.
Aurora

[Bud]

Aurora, you should write a book, man. I would certainly buy one.

[bookaboo]

Most microbolometer based thermal cameras contain the following temperature sensors.....

1. Chassis temperature
2. Microbolometer temperature (at least 2 sensors normally used)
3. FFC shutter flag temperature (or the air around said flag)
4. If a Germanium lens is present ... The lens temperature
5. If a Peltier temperature stabiliser is present ... Hot and cold sides of the Peltier element.

The cameras radio metric temperature measurement routines need much of this sensor data in order to provide accurate and repeatable temperature measurements.

For those wondering why the temperature of a Germanium lens is needed....Germanium changes its transmission spec as its temperature rises. The greater the temperature the lower the transmission percentage until at 100 degrees C it is almost opaque to thermal IR wavelengths. The effect only really starts to become an issue when the lens temperature rises above 50 Degrees C though.

In situations where the Germanium lens is exposed to very high temperatures, such as a furnace, steps are taken to cool the lens assembly. Cooled Air or fluid cooling lens jackets are commonly used.
Aurora

Interesting, I recall seeing a very expensive TI system on a drying furnace which was monitoring for ignited material on the output of the dryer. It had an air blower creating a narrow curtain of air just in front of the camera (though not pointing at the camera). I assume this was to prevent heating of the lens by convection?

(Also +1 to the book suggestion)

[encryptededdy]

For those wondering why the temperature of a Germanium lens is needed....Germanium changes its transmission spec as its temperature rises. The greater the temperature the lower the transmission percentage until at 100 degrees C it is almost opaque to thermal IR wavelengths. The effect only really starts to become an issue when the lens temperature rises above 50 Degrees C though.

In situations where the Germanium lens is exposed to very high temperatures, such as a furnace, steps are taken to cool the lens assembly. Cooled Air or fluid cooling lens jackets are commonly used.
Aurora

That sounds like a pretty big disadvantage... are there any advantages to using Germanium compared to, for example, chalcogenide glass as found in the E4? Does it offer higher optical quality?

At least Umicore's advertising (http://eom.umicore.com/en/infrared-optics/blanks/#GASIR) certainly makes their chalcogenide glass material sound far superior to Germanium.

[Chanc3]

Strange temperature if I compare with my new E30bx and E4 non hacked from my friend, the temperature sometimes is 4 or 5 degrees more. I thing is problem with calibration. I had this problem when I bought it more then one year ago and Flir in that time recalibrated E4 for me but maybe not too god.

Not being funny, but have you setup the camera with the same parameters? Emissivity, Tref etc?

[Uho]

Did the test lens X3 on Flir e5. Works. Not available E4 with friends.

[Fraser]

Uho,

Very nice work 

It is such a pity that FLIR does not give user access to the image reversal function in order to correct the lens inversion. Most thermal cameras have the feature but few provide it as a menu option.

Aurora

[Uho]

I was able to make the lens, without flip the image only X2. If more than the increase - or flips, or loss of sensitivity.
Yes. The menu is do not have enough to flip image.
I tested it on Flir E5. Without camera. Its dimensions are different from the E4. In the company I have not found such a thermal imager. Who knows about the different versions of the E5? Do they really are different?

[encryptededdy]

I was able to make the lens, without flip the image only X2. If more than the increase - or flips, or loss of sensitivity.
Yes. The menu is do not have enough to flip image.
I tested it on Flir E5. Without camera. Its dimensions are different from the E4. In the company I have not found such a thermal imager. Who knows about the different versions of the E5? Do they really are different?

Do you mean i5?

Yes, it can be hacked to the i7 (either 120x120 or 140x140 depending on generation). The actual microbolometer inside is like 280x280

[Uho]

Yes. Thank you. Indeed I5. I was told E5. I ignored the letter. I had a little time .

[MikaJ]

Thanks for all that have worked hacking E4! Just bought from Farnell this week one and in 15minutes it was upgraded with Fubar instructions without any problem.

[fubar.gr]

I've taken some photos with the MSX distance set incorrectly and the optical and thermal images do not line up.

Is there any way to fix it in FLIR Tools?

[Bud]

You can try BFIC

https://www.eevblog.com/forum/testgear/flir-e4-thermal-imaging-camera-teardown/msg610799/#msg610799

I do not think FLIR Tools has adjustment for MSX offset.

[fubar.gr]

Thanks!

[mrweek]

I upgrade my cam successfully with rom E4 1.2L 2.3.0, Thanks for the excellent information.
My Question is the watermark “flir” in the lower left corner is missing after the upgrade,
but some people posted the photos in forum, the watermark is still there although the cam is upgraded.
I wonder to know can I control the watermark display on the screen and how ?
Thanks a lot ^i^

[tomas123]

I've taken some photos with the MSX distance set incorrectly and the optical and thermal images do not line up.

Is there any way to fix it in FLIR Tools?

I would look inside the jpg-binaries and patch the (constant) hex values for distance in the exif header

with exiftool you can watch two flir distance parameters
 

Code: [Select]

$ exiftool -focus*  FLIR0080.jpg
Focus Step Count                : 70
Focus Distance                  : 0.0 mI think, that "Focus Distance" is zero by the fixed focus camera Ex

For Flir Exx I wrote here something to this values:
http://u88.n24.queensu.ca/exiftool/forum/index.php/topic,4898.msg24224.html#msg24224


the simplest way, to found the correct byte sequence is exiftool -v5

Code: [Select]

$ exiftool -v5  FLIR0080.jpg | grep -B300  -i focus
  | | FilterSerialNumber =
  | | - Tag 0x021c (32 bytes, string[32]):
  | |     041c: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 [................]
  | |     042c: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 [................]
  | | PlanckO = -7652
  | | - Tag 0x0308 (2 bytes, int16s[1]):
  | |     0508: 1c e2                                           [..]
  | | PlanckR2 = 0.0276008248329163
  | | - Tag 0x030c (4 bytes, float[1]):
  | |     050c: 20 1b e2 3c                                     [ ..<]
  | | RawValueMedian = 12444
  | | - Tag 0x0338 (2 bytes, int16u[1]):
  | |     0538: 9c 30                                           [.0]
  | | RawValueRange = 624
  | | - Tag 0x033c (2 bytes, int16u[1]):
  | |     053c: 70 02                                           [p.]
  | | DateTimeOriginal = ...R..
  | | - Tag 0x0384 (10 bytes, undef[10]):
  | |     0584: e5 c2 9c 52 2e 02 00 00 00 00                   [...R......]
  | | FocusStepCount = 70
  | | - Tag 0x0390 (2 bytes, int16u[1]):
  | |     0590: 46 00                                           [F.]
  | | FocusDistance = 0


You have luck. FocusDistance has a fixed byte offset to the constants PlanckO / PlanckR2.
It' simple to patch this values in a batch script.

more informations about the tag positions in the exif header you get from exiftool source code
http://cpansearch.perl.org/src/EXIFTOOL/Image-ExifTool-10.00/lib/Image/ExifTool/FLIR.pm
scroll down to 0x390 ff

Code: [Select]

# FLIR camera record (ref PH)
%Image::ExifTool::FLIR::CameraInfo = (
    GROUPS => { 0 => 'APP1', 2 => 'Camera' },
    PROCESS_PROC => \&Image::ExifTool::ProcessBinaryData,
    FIRST_ENTRY => 0,
    NOTES => q{
        FLIR camera information.  The Planck tags are variables used in the
        temperature calculation.  See
        L<http://u88.n24.queensu.ca/exiftool/forum/index.php?topic=4898.msg23972#msg23972>
        for details.
    },
    0x00 => {
        # use this tag only to determine the byte order
        # (the value should be 0x0002 if the byte order is correct)
        Name => 'CameraInfoByteOrder',
        Format => 'int16u',
        Hidden => 1,
        RawConv => 'ToggleByteOrder() if $val >= 0x0100; undef',
    },
    # 0x02 - int16u: image width
    # 0x04 - int16u: image height
    # 0x0c - int32u: image width - 1
    # 0x10 - int32u: image height - 1
    0x20 => { Name => 'Emissivity',                   %float2f },
    0x24 => { Name => 'ObjectDistance', Format => 'float', PrintConv => 'sprintf("%.2f m",$val)' },
    0x28 => { Name => 'ReflectedApparentTemperature', %floatKelvin },
    0x2c => { Name => 'AtmosphericTemperature',       %floatKelvin },
    0x30 => { Name => 'IRWindowTemperature',          %floatKelvin },
    0x34 => { Name => 'IRWindowTransmission',         %float2f },
    # 0x38: 0
    0x3c => {
        Name => 'RelativeHumidity',
        Format => 'float',
        ValueConv => '$val > 2 ? $val / 100 : $val', # have seen value expressed as percent in FFF file
        PrintConv => 'sprintf("%.1f %%",$val*100)',
    },
    # 0x40 - float: 0,6
    # 0x44,0x48,0x4c: 0
    # 0x50 - int32u: 1
    # 0x54: 0
    0x58 => { Name => 'PlanckR1', %float8g }, #1
    0x5c => { Name => 'PlanckB',  %float8g }, #1
    0x60 => { Name => 'PlanckF',  %float8g }, #1
    # 0x64,0x68,0x6c: 0
    0x070 => { Name => 'AtmosphericTransAlpha1', %float6f }, #1 (value: 0.006569)
    0x074 => { Name => 'AtmosphericTransAlpha2', %float6f }, #1 (value: 0.012620)
    0x078 => { Name => 'AtmosphericTransBeta1',  %float6f }, #1 (value: -0.002276)
    0x07C => { Name => 'AtmosphericTransBeta2',  %float6f }, #1 (value: -0.006670)
    0x080 => { Name => 'AtmosphericTransX',      %float6f }, #1 (value: 1.900000)
    # 0x84,0x88: 0
    # 0x8c - float: 0,4,6
    0x90 => { Name => 'CameraTemperatureRangeMax', %floatKelvin },
    0x94 => { Name => 'CameraTemperatureRangeMin', %floatKelvin },
    0x98 => { Name => 'UnknownTemperature1', %floatKelvin, Unknown => 1 }, # 50 degrees over camera max
    0x9c => { Name => 'UnknownTemperature2', %floatKelvin, Unknown => 1 }, # usually 10 or 20 degrees below camera min
    0xa0 => { Name => 'UnknownTemperature3', %floatKelvin, Unknown => 1 }, # same as camera max
    0xa4 => { Name => 'UnknownTemperature4', %floatKelvin, Unknown => 1 }, # same as camera min
    0xa8 => { Name => 'UnknownTemperature5', %floatKelvin, Unknown => 1 }, # usually 50 or 88 degrees over camera max
    0xac => { Name => 'UnknownTemperature6', %floatKelvin, Unknown => 1 }, # usually 10, 20 or 40 degrees below camera min
    0xd4 => { Name => 'CameraModel',        Format => 'string[32]' },
    0xf4 => { Name => 'CameraPartNumber',   Format => 'string[16]' }, #1
    0x104 => { Name => 'CameraSerialNumber',Format => 'string[16]' }, #1
    0x114 => { Name => 'CameraSoftware',    Format => 'string[16]' }, #1/PH (NC)
    0x170 => { Name => 'LensModel',         Format => 'string[32]' },
    # note: it seems that FLIR updated their lenses at some point, so lenses with the same
    # name may have different part numbers (eg. the FOL38 is either 1196456 or T197089)
    0x190 => { Name => 'LensPartNumber',    Format => 'string[16]' },
    0x1a0 => { Name => 'LensSerialNumber',  Format => 'string[16]' },
    0x1b4 => { Name => 'FieldOfView',       Format => 'float', PrintConv => 'sprintf("%.1f deg", $val)' }, #1
    # 0x1d0 - int16u: 0,12,24,25,46
    # 0x1d2 - int16u: 170,180,190,380,760,52320
    0x1ec => { Name => 'FilterModel',       Format => 'string[16]' },
    0x1fc => { Name => 'FilterPartNumber',  Format => 'string[32]' },
    0x21c => { Name => 'FilterSerialNumber',Format => 'string[32]' },
    0x308 => { Name => 'PlanckO',           Format => 'int32s' }, #1
    0x30c => { Name => 'PlanckR2',          %float8g }, #1
    0x338 => { Name => 'RawValueMedian',    Format => 'int16u', Groups => { 2 => 'Image' } },
    0x33c => { Name => 'RawValueRange',     Format => 'int16u', Groups => { 2 => 'Image' } },
    0x384 => {
        Name => 'DateTimeOriginal',
        Description => 'Date/Time Original',
        Format => 'undef[10]',
        Groups => { 2 => 'Time' },
        RawConv => q{
            my $tm = Get32u(\$val, 0);
            my $ss = Get32u(\$val, 4) & 0xffff;
            my $tz = Get16s(\$val, 8);
            ConvertUnixTime($tm - $tz * 60) . sprintf('.%.3d', $ss) . TimeZoneString(-$tz);
        },
        PrintConv => '$self->ConvertDateTime($val)',
    },
    0x390 => { Name => 'FocusStepCount', Format => 'int16u' },
    0x45c => { Name => 'FocusDistance',  Format => 'float', PrintConv => 'sprintf("%.1f m",$val)' },

[tomas123]

two years ago we discussed in the imagemagick forum about a best fit subimage search:
overlay a thermal with a visible image 
http://studio.imagemagick.org/discourse-server/viewtopic.php?f=1&t=23318#p97884

[Bud]

My Question is the watermark “flir” in the lower left corner is missing after the upgrade,
but some people posted the photos in forum, the watermark is still there although the cam is upgraded.
I wonder to know can I control the watermark display on the screen and how ?
Thanks a lot ^i^

Why do you want the logo? It just occupies valuable space.

[mrweek]

Why do you want the logo? It just occupies valuable space.

Because other guy's posts have the watermark, I am questioning is there something go wrong or my cam is angry about the upgrade  

[Bud]

There may be different versions of the files in the update packages. Some may have the logo patched or disabled, others may not.

[marcins]

Hi all
Got today my E4 from PASS, firmware 2.3.0 +5% discount with voucher
and not only
what was in the box :
E4 with battery, charger, and usb cable,call certyfikate,
+ at no extra cost!!!
Amazon Gift Voucher £50
E series Pouch £21.60
E series Battery(extra one) £66
E series Car charger £43.20

is this normal and everyone get extrass??
very impressive first order there:)

now is charging and will do hack on it:)

[Fraser]

FLIR operate specials that include the items you listed.

The USA normally get the best deals first but the E4 has been out For some time so we are now seeing the specials here in the UK. IMHO PASS are a very good company with which to do business. They treated me well in the past.

Fraser

[nightsquint]

Greetings,

Another happy owner of a brand new (now hacked) E4 (bought two weeks ago), thanks to Mike, Fubar, and countless others who made all the hacks work   

Model: E4 1.2L
Serial: 6396xxxx
Part: 63901-0101
Software: 2.3.0
Last cal: July 2015

.info.temporalMk double 13.300059 (quite good)

See attached for dead pixels. E4 screen photographed using a mobile phone camera.
The dead pixels (dark spots) flicker and some supposedly "dead" pixels come to "live" able to register at higher thermal inputs.
But then, other "dead" pixels are really really dead (no amount of resuscitation & slapping helps).

Cheers 

[Fraser]

Those dead pixels are normal and your camera isn't bad at all. The dead pixel map and correction sort all this out for you. The normal spec of a microbolometer is 99.8% active pixels with 99.9% active on the more expensive arrays.

DO NOT SLAP YOUR E4 ! This isn't a kids toy !

Doing so can actually create new dead pixels, requiring the dead pixel map to be re written.

Fraser