FLIR E4 circular gradient issue after warm-up

[Signal32]

Hello,

EDIT: Firmware is 1.22. No firmware upgrades / software, resolution,menu, etc hacks have been attempted on this camera.
I have a very strange issue with a FLIR E4 - about 2 minutes from startup, after the camera heats up, it starts displaying a circular thermal gradient. After 10 minutes the gradient becomes really pronounced (see pic at end of post, taken with the lens cover on, same happens if pointed at a wall)

Eventually it will become really pronounced with about 7-10C gradient when looking at a uniform temperature wall.
Here is my understanding of how the camera should get rid of these types of anomalys:  Every ~30sec the camera auto-calibrates, by putting a shutter in front of the sensor. When the shutter is in front of the sensor, the camera assumes the sensor sees a completely uniform heated image with the same temperature as the frame of the camera ( which is sensed via a sensor inside of the camera ). It then uses this baseline to apply corrections to each pixel after the calibration shutter has been removed. Any gradients present on the sensor / any pixels that are more / less sensitive than others will be adjusted correspondingly using the baseline that was taken during calibration.

Here is what I've observed in the case of the defective camera:
- If the calibration shutter is forced to stay always on, then the gradient goes away (of course now you can't use the camera for anything else other than looking at the inside of the calibration shutter ).
- If during calibration, a uniformly heated object is placed in front of the sensor(the calibration shutter is removed from the sensor), then the camera functions as expected, the gradient goes away, and the camera gives good images of the environment.
- This happens even with no lens present, just sensor -> shutter -> uniform-heated wall / paper

This makes me think that somehow the calibration shutter is not uniformly heated, however looking at it through another 320x240 E4 shows no temperature gradient on it.

Anyone has seen this before / have any idea what the issue can be / how to fix it ?

Thanks!

[tomas123]

looks like this story, after a firmware hack:

Good Evening,

I got a fresh Flir E4 with 2.11 firmware.
I downloaded the Flir Tools XP version directly from their site and installed it on XP 32 bit.
After that I followed MATMOE's instructions / the ones that come on a *.PDF file with the downgrade file.
I downloaded the downgrade package - http://s000.tinyupload.com/?file_id=00983732806372988148

Now when the downgrade process to 2.3 is done, the camera itself is working fine but i got this "ring-of-crap" on the screen which can be seen on the photos below. The ring is like static on every picture, no matter what you are shooting.

I havent done any hacks on it yet and aint gonna before this issue gets solved, spose I need to wait a miracle what happened to dza over a night. Was it xmas over there ? =D

[Signal32]

looks like this story, after a firmware hack:

Yes, it does look similar, but I thought that the "ROC" issue that you speak of is temperature-independent, since no one mentioned that it only happens after a while, so I assumed it happens immediately after power on. Mine happens as the camera warms up, the warmer it gets the more pronounced the issue is.

[Dynamics]

Do yourself a favor and put it back to 2.11.0 as that is the latest released firmware, don't let the version numbers fool you, they are asinine. After 2.11 is back on, use the 2lps hack files to properly patch it. Mine is hacked and it was a factory 2.11.0 E4 purchased this year in February.

[Signal32]

Do yourself a favor and put it back to 2.11.0 as that is the latest released firmware, don't let the version numbers fool you, they are asinine. After 2.11 is back on, use the 2lps hack files to properly patch it. Mine is hacked and it was a factory 2.11.0 E4 purchased this year in February.

The firmware version is 1.22. Hardware version 1.1L. I don't think any firmware software upgrades / hacks have been done on this. This is not related to the E4 hack. I did not upgrade / downgrade the firmware or attempt any hacks. I perhaps should have been more clear about that in the original post.

[OrBy]

I get a similar ring when I turn off auto NUC and leave the camera for a while. Are you sure the shutter is moving when the auto NUC takes place? If this camera is unmodified it likely could be sent in for warranty repair.

[Signal32]

I get a similar ring when I turn off auto NUC and leave the camera for a while. Are you sure the shutter is moving when the auto NUC takes place? If this camera is unmodified it likely could be sent in for warranty repair.

Unfortunately it's warranty has expired, it's a rather old camera. I'm sure that the shutter is triggering, I fully disassembled the camera ( down to the bond wires that I accidentally bent and then fixed, oulps ).Now that I looked into NUC( non-uniformity compensation ), it seems that NUC is done based on the current temperature, which may provide a possible cause for the issue. If the tables for the higher temperature ranges are missing(it works as expected for the first 2 min  when the camera is cold), then the camera wouldn't have a valid NUC table when it heats up, causing the issue. BTW how do you disable auto-NUC ?
I'm thinking of flashing an earlier version of the firmware, does FLIR provide older versions of the firmware like 1.19, etc ?

Thanks!

[OrBy]

FLIR has pulled all the old firmware files for some reason or another but many here have archives. I would suggest going up to 2.3.0. I have found it quite stable and feature filled with the menu hack found in the main E4 thread (you should be able to find the firmware there as well). As for disabling auto NUC - you have to enable RNDIS then telnet into it and issue a command.

[Signal32]

FLIR has pulled all the old firmware files for some reason or another but many here have archives. I would suggest going up to 2.3.0. I have found it quite stable and feature filled with the menu hack found in the main E4 thread (you should be able to find the firmware there as well). As for disabling auto NUC - you have to enable RNDIS then telnet into it and issue a command.

I'm not sure if I can get the menu hack to work if I upgrade from 1.22 to 2.3 then try to apply the 2.3 hack( I do know that the menu hack does work on factory 2.3-based cameras ). The resolution hack will probably work, do you know of any cases of 1.22 -> 2.3 upgrade that have gotten the menu hack to work ?

[Bill W]

One possibility is that the flag temperature is now wrong, as you have been in and disturbed things.  Were you really looking at a scene where the coldest part was 80°F ?

Flag is say 80°F
Camera thinks it is 70°F
Camera now adds '10°F worth of data' to the sensor output
This is then multiplied up by the gain per pixel, which is dominated by the lens fall off centre to edge.

Bill

[Signal32]

One possibility is that the flag temperature is now wrong, as you have been in and disturbed things.  Were you really looking at a scene where the coldest part was 80°F ?

Flag is say 80°F
Camera thinks it is 70°F
Camera now adds '10°F worth of data' to the sensor output
This is then multiplied up by the gain per pixel, which is dominated by the lens fall off centre to edge.

Bill

I'm not sure exactly of what you mean.
The issue happened before I opened up the camera, I opened up the camera to try to fix it.
The image in the first photo was taken with the lens cover on(manual shutter thing).
80F sounds about right for the temperature of the lens cover, considering it's part of the camera which warms up.

[Signal32]

I applied the 1.22 hack which worked but the thermal ring issue still happens.
I then updated to 2.3 firmware and again, the thermal ring issue is still there.
Does anyone know what files on the camera hold the non-uniformity compensation / calibration data ?

[tomas123]

there are only a 
GainDeadMap
CoarseMap

search for ds250C_we_ap_fi_le_static.gan inside the "Flir E4 Thermal imaging camera teardown" thread

I think, the NUC is calculated only with a shutter image...

[Bill W]

I'm not sure exactly of what you mean.
The issue happened before I opened up the camera, I opened up the camera to try to fix it.
The image in the first photo was taken with the lens cover on(manual shutter thing).
80F sounds about right for the temperature of the lens cover, considering it's part of the camera which warms up.

OK, my misunderstanding.  However I still reckon there is a bad temperature reading getting into the NUC maths.

Is it OK if started up warm, ie warm it up when switched off ?
If you do tat, do you then get a hot centre as the camera cools ?

Bill

[ollihd]

Any updates on this? I have the same exact issue with some additional lines. No hacks done yet. 2.8 upgraded to 2.11 and the issue persists. I'm on 1.2L hardware. Some images: https://goo.gl/photos/TL43oRGJaUtYBjx98

[Signal32]

Any updates on this? I have the same exact issue with some additional lines. No hacks done yet. 2.8 upgraded to 2.11 and the issue persists. I'm on 1.2L hardware. Some images: https://goo.gl/photos/TL43oRGJaUtYBjx98

Are you seeing the issue immediately after boot or after the camera warms up ?

[ollihd]

What would you say is the warm up time? The horizontal lines I get immediately after cycling power off, power on for the second time. The first time I power it up this does not happen... I don't take the battery out while cycling. The gradient issue seems to happen when i do a "hard reset -> battery off" and power on the first time. After hard reset when I cycle the power the gradient seems to go away, but the lines appear. This is truly strange..

EDIT: the gradient seems to come very randomly / seldomly

EDIT2: Here is a video on how I can replicate the horizontal lines: https://goo.gl/photos/Fn7DwcaZEBNkuoPR7