Flir E4 Thermal imaging camera teardown

[difxati]

Does anyone know how to get the manual temp scale on the menus?

Thanks again

I have hacked my E4, it is 1.2L with firmware 2.11, both resolution and menu hack work perfectly, including the manual temp scale and zoom !
Thank you everyone!

For resolution hack:
I follow the guide ( http://fubar.gr/hacking-the-flir-e4/) and use the script pack from 2lps (https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/msg921880/#msg921880).   

For menu hack:
I didn't  use the python script. I modify the config.cfc as 2lps  (https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/msg975993/#msg975993) and replace it manually.

[peperoca]

I have hacked my E4, it is 1.2L with firmware 2.11, both resolution and menu hack work perfectly, including the manual temp scale and zoom !
Thank you everyone!

For resolution hack:
I follow the guide ( http://fubar.gr/hacking-the-flir-e4/) and use the script pack from 2lps (https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/msg921880/#msg921880).   

For menu hack:
I didn't  use the python script. I modify the config.cfc as 2lps  (https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/msg975993/#msg975993) and replace it manually.

That's great difxati! Should I unhack my 2.3 E5 before upgrading or could I just upgrade to 2.11 and then re hack the res or the menu depending on what's missing?

I can't find the cfccfg_v2.py. I only manage to find the cfccfg.py


UPDATE:

After update from 2.3 to 2.11 (previously un hacking menu in 2.3) the reshack survived.

I use the 2lps to hackmenu with the scripts but the reshack dissapears and the camera turns to black after loading bar ends. The thermal image is displayed while the bar loads, but then fades to black. I run the menu restore.fif and the camera works (without image mode) and if I replace the conf.cfc with the previous version I restore the reshack and the image mode.

I'm not being able to perform the menuhack... Any ideas? Should I downgrade to 2.3?

UPDATE 2:

Manage to find the cfccfg_V2.py on the thread. However I cannot make it work.

Here is the screen output:

"C:\Python27>cfccfg_V2.py 0x2A2A5802 0x02D55113 conf.cfg conf.cfc
usage: C:\Python27\cfccfg_V2.py suid fileIn fileOut"

Where am I messing it up?

Best,
Pepe

[difxati]

UPDATE 2:

Manage to find the cfccfg_V2.py on the thread. However I cannot make it work.

Here is the screen output:

"C:\Python27>cfccfg_V2.py 0x2A2A5802 0x02D55113 conf.cfg conf.cfc
usage: C:\Python27\cfccfg_V2.py suid fileIn fileOut"

Where am I messing it up?

Best,
Pepe

you need your "suid" to run the cfccfg_V2.py  script for decrypting the cfc to cfg or encrypting cfg back to cfc

how to find your suid:

Probably for the menu, you can just apply the menu hack.

The ftool is for the 1.x version of conf.cfc. For 2.x you can use cfccfg_V2: https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/msg596959/#msg596959 (there is a link to the original post of how to use it - https://www.eevblog.com/forum/testgear/flir-e4-thermal-imaging-camera-teardown/msg594600/#msg594600)

To get your SUID:
0. Put the camera into RNDIS mode
1. Open the command prompt and telnet to the camera:
telnet 192.168.0.2

You will gain access to the FLIR Command Line Interpreter

2. CD to FlashFS/system/

\>cd FlashFS/system/

3. Execute suid.exe
\FlashFS\system\>suid

The SUID necessary for cfccfg_V2.py will be printed.

I use "Putty"(a software) to send telnet command.

[peperoca]

you need your "suid" to run the cfccfg_V2.py  script for decrypting the cfc to cfg or encrypting cfg back to cfc

Hi difxati, I got my SUID while decrypting the conf.cfc isn't it something like: 0x2A2A5802 0x02D55113?

[sterlinggold]

Amazing, I got an E4 with 2.11 from Amazon Warehouse for 740. Used the FLIR_E4_2.11.0_hack.zip from 2lps. High res is Da-Biznez and the full menu with zoom and rainbow color scheme is outstanding. I used the latest flir tools package and latest 64 bit drivers on win 7 64. Followed the standard steps from fubar and all worked great.
Made my day Ya'll.
Thanks so much.

[difxati]

my ds250C_we_ap_fi_le_static.gan and ds250C_we.crs
 

[tomas123]

very interesting

How good works the bad pixel mapping with of whole line?

[OrBy]

my ds250C_we_ap_fi_le_static.gan and ds250C_we.crs
 

Owch - that's from the factory?

[Chanc3]

That doesn't look good!

[difxati]

There are lots of vertical bands when shooting at area with even temperature. The corrected dead pixel line becomes one of them so you won't differentiate it(pic 114).

pic115 is the best scenario that no vertical bands at all
pic114 is the worst case
pic136-138 are the results after turning off the dead pixel map

There is always a black point on the lower part (dust?)
I am not quite happy with my E4

[difxati]

my ds250C_we_ap_fi_le_static.gan and ds250C_we.crs
 

Owch - that's from the factory?

from the factory

[tomas123]

Thanks for the great images.

After playing around with Flir Tools I say, that you have more trouble with the large dust grain, as with the bad pixel line.
The Flir bad pixel mapping works great.

The vertical bands in image FLIR0114.jpg are typical an all Exx / Ex images with low temperature span.
I'm not sure, if this is a result of the NUC (non uniformity image correction) or of the column noise from microbolometer.

Maybe the lines are a little bit  curved...

[difxati]

Thank you for your explanation. That makes me feel better if the vertical bands are normal.
BTW, can I set when cold boot up it jumps to thermal mode directly instead of MSX mode?
And is it possible to assign short cut to a key such as locking the temperature scale or change to "no measurement"?
Thanks.

[joe-c]

BTW, can I set when cold boot up it jumps to thermal mode directly instead of MSX mode?

maybe, but i am not sure.

And is it possible to assign short cut to a key such as locking the temperature scale or change to "no measurement"?
Thanks.

with the FW 1.22 and below, yes. You have to extract the facet_z3.rcc files and than you can change files like "Button.qml".
for example, a large press on the left button (archive) will perform NUC. a large press on the right button (back) does nothing.
there you can implant a shortcut... but its many work:
-change something
-repack to .rcc
-upload to camera and replace the existing
-restart the camera
-try
... sometimes maybe a black screen if you make a mistake.
but i am not sure... maybe FW 2.xx prevent changes like this, i don't tried 

I think its more easy to implant a small micro controller board with a uart connection to the Camera pcb (debug port). than you can build in your own extra buttons with extra functionality.

[difxati]

After tearing down the e4 into pieces and blow the sensor and filter with direct air, I can still see a dark spot at the lower part.
Can I use 70%IPA to clean the sensor and the filter? Will it damage the coatings?

[joe-c]

Can I use 70%IPA to clean the sensor and the filter? Will it damage the coatings?

i don't know... i did never try it.
But if you can't see dirt on the Sensor window, maybe the calibration is the problem.

If flir has dust or something else on the sensor while calibrating, this will filtered in the calibrated image.
but if the dust fall down, you see the spot again.

download your map from camera an analyze it...

[Bill W]

Is it a hot spot looking at a cold object and cold spot when looking at a hot object ? (this is dust on, so you should see it as Joe says)
or the reverse  (this is dust falling off after calibration)

Otherwise it could just be an area of 'very different' on the sensor where the factory calibration is doing so much correction that residual errors are left visible in the low contrast / high gain scene.  )114 looks more like an uncorrected sensor image than a corrected image.

Bill

www.fire-tics.co.uk

[difxati]

Can I use 70%IPA to clean the sensor and the filter? Will it damage the coatings?

i don't know... i did never try it.
But if you can't see dirt on the Sensor window, maybe the calibration is the problem.

If flir has dust or something else on the sensor while calibrating, this will filtered in the calibrated image.
but if the dust fall down, you see the spot again.

download your map from camera an analyze it...

I have uploaded the map
Can I modify the calibration map to mask that spot?

Is it a hot spot looking at a cold object and cold spot when looking at a hot object ? (this is dust on, so you should see it as Joe says)
or the reverse  (this is dust falling off after calibration)

Otherwise it could just be an area of 'very different' on the sensor where the factory calibration is doing so much correction that residual errors are left visible in the low contrast / high gain scene.  )114 looks more like an uncorrected sensor image than a corrected image.

Bill

www.fire-tics.co.uk

When my E4 just come out from factory, the spot disappears on cold object( internal of a fridge) and hot object (a tea pot) but appears as a fade cold spot @2x°C.

After cleaning (just blow the lens, the silvery window and the sensor with air. By doing so I can confirm the spot comes from the sensor), it appears as a solid cold spot @2x°C, but almost disappear on hot object and appears as a " very solid cold spot" on cold  object( internal of a fridge)
I can see it even turning off the dead pixel map. Is it possible that I make the dust embed "deeper" on the sensor surface by blowing air on it? Or the air pressure intensify the sensor defect on that spot?
Can I use other mechanical method(wipe it with alcohol or using adhesive tape to stick away dust) to clean the sensor surface ?

update: I used alcohol and adhesive tape to clean the sensor with power on. The spot is not related to dust as it cannot be wipe or stick away. I think it is a bunch of defective pixel that measure temperature lower than the normal. So anyways to update the calibration map or data(attached are the 4 map file and the calib.rsc) to at least make the spot less visible? Or it can only be done by flir? Thanks!

[tomas123]

I know, you cleaned your sensor from dust
... but I found an old post as entry point 

dust on image see here

https://www.eevblog.com/forum/testgear/flir-e4-thermal-imaging-camera-teardown/msg329239/#msg329239

https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/msg337392/#msg337392

I took the camera apart as per mikes instructions and used a camera sensor dusting ball to clean it after which the spot was gone.

and
https://www.eevblog.com/forum/testgear/flir-e4-thermal-imaging-camera-teardown/msg345877/#msg345877

from Ex manual:

[difxati]

I know, you cleaned your sensor from dust
... but I found an old post as entry point 

dust on image see here

https://www.eevblog.com/forum/testgear/flir-e4-thermal-imaging-camera-teardown/msg329239/#msg329239

https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/msg337392/#msg337392

I took the camera apart as per mikes instructions and used a camera sensor dusting ball to clean it after which the spot was gone.

and
https://www.eevblog.com/forum/testgear/flir-e4-thermal-imaging-camera-teardown/msg345877/#msg345877

from Ex manual:

how about the sensor? not the lens, not the covering filter.
I used 70% IPA to wipe the sensor and also adhesive tape to stick away any dust on surface when power on. I thought the sensor would die instantly but it survived. Its too bad that the spot remain there. The dust particle may be inside the sensor or it is not dust at all?
And can I tweak the calibration file to make it less visible?

[Fraser]

Please be aware that you are not touching the sensor itself when cleaning a microbolometer. You are touching the window that sits in front of the very delicate VOx microbolometer pixel. This window is made from similar material to the lenses and must be treated with respect. Saying that IPA may be used. I recommend you use pure IPA as the 70% stuff is rubbing alcohol and contains water. Water is not a great contaminant to have in the IPA.

I have never applied any form of tape to a lens of sensor. Depending upon the adhesive, such could be a bad idea leaving residue or, if very sticky, apply unwelcome forces to the microbolometer window. There may be specialist tapes for such use but I would not know where to buy it.

One way to determine whether you have contamination of the microbolometer or defective pixels is capture an image that shows the defect and zoom in to it to see the pixels. Look at whether it is a defined cluster of pixels or more of a blurred edge blob. The latter can indicate dust contamination but sadly the cameras image processing can make this test challenging. I used this technique to prove that my TESTO camera has 3 visible dead pixels. They stand out plain as day when I zoomed into the image to see individual pixels. They were so well defined that they could not be dust. Dust casts a shadow on the pixels below. Dead pixels are very sharp edged.

I have a cluster of 4 dead pixels on one of my E4 cameras and the dead pixel map has captured them. As such they cannot normally be seen in images.

Fraser

[bookaboo]

I've had success cleaning dust of my E4 sensor using a standard digital camera cleaning kit.
http://www.ebay.co.uk/itm/Full-Frame-CCD-CMOS-Digital-Camera-Sensor-Swab-Cleaning-Kit-24mm-Vacumm-DC581-/161959710582?hash=item25b58d0f76:g:aDQAAOSwZetXOVm3
Or similar.

[difxati]

Thank you for all of the above suggestions.
I clean the sensor surface  many times and it looks clean and dust free. The spot is still there. May be it is inside the sensor module.

[Chanc3]

Have you checked the lens?

[difxati]

Of course.
You can see the photo above, without lens and the windows, only the sensor and the spot remains there