Flir E4 Thermal imaging camera teardown

[Echo88]

Hi,

ordered my E4 Wednesday, eagerly awaiting its arrival. Im already thinking about adding a passive autofocus. Anyone knows if theres software capable of analysing videostreams (per contrastscanning for example) und giving out a signal to control a motorslide which moves a telephoto-lens? Software like http://www.focusmax.org/ (didnt read the whole description, but i think its not fast enough, since its for telescope users). Another question: Is there any information which lens to use as a telephoto-lens for the E4 for ~100m targetdistance? Ive got knowledge about laser-optics, but not about image-optics.

Many thanks to the users on here constantly working on hacking the E4-Firmware 


Greetings,

Steffen

[mikeselectricstuff]

Is there any information which lens to use as a telephoto-lens for the E4 for ~100m targetdistance? Ive got knowledge about laser-optics, but not about image-optics.

You'll need a germanium telephoto lens. Glass is not thermally transparent, so it's unusable. The germanium lenses are quite expensive, though.

For example :
http://www.ebay.co.uk/itm/200mm-F1-3-Germanium-Thermal-Imager-Lens-PPE-Fast-Objective-/281312701181?pt=UK_Lenses_Filters_Lenses&hash=item417f8b2afd

[ixfd64]

I've noticed that pictures taken with thermal scopes typically don't show temperatures. Do people usually not care about temperature readings, or do thermal scopes not have that feature at all?

[Fraser]

The E4 is capable of detecting a person easily at 300m. Much depends upon what you need in terms of target detail. I do wildlife work with a thermal camera and operate a 24 Degree Lens fed from a FLIR X2 telescope. I bought the compact telescope for around $400 on the secondary market where prices can be more reasonable.

The E4 was designed with a 45 Degree wide angle lens (in TIC terms), and this suits its target audience who often wish for >24 Degree lenses when working in cramped conditions. For distance work it is better to find a camera with the narrower field of view though. Adding Germanium lenses in front of the primary lens also adds some loss (approx. 15% on my X2 lens), weight, and costs a lot of money due to the amount of expensive Germanium lens material needed.

Be very careful buying lenses on the secondary market..... they are often still an expensive purchase and there are many traps to fall into, including a totally unsuitable SW lens that massively attenuates the image when used on a LW camera such as the E4 ! That is due to the AR coating differences.

Somewhere I have a table of FLIR detection distances with different lenses. I will see if I can find it.

Do not forget that the E4 can apply a digital zoom of X2 to X8. As you likely know, this is at the cost of resolution, but it provides a larger image on the screen. X2 is useful...... much more than that and pixelation becomes an issue.

Weapons sights do not normally have temperature measurement as they are weapon centric, providing an uncluttered view of the target with the best possible contrast. Such a TIC is not a good choice where temperature measurement is a requirement !

[ixfd64]

I see. But I'm curious: is it possible to hack a thermal scope to display temperatures?

[mikeselectricstuff]

Scopes will often have automatic  gain/zero adjust to maximise the amount of detail in an image. In some cases, they will not, or only rarely rarely, do a shutter calibration, as uninterrupted view is more important than temperature accuracy.
For any scope that is handheld, the movement will mean that for nonuniformity correction, it can be assumed that the long-term avarage scene content will be uniform - I think the FlirOne uses this method, and only needs the manual shutter when you want to measure absolute  temp.
I suspect if you mount a FlirOne on a tripod looking at a static image, the image will fade out, though they may be using the accelerometer as an input to estimate whether the image should be moving or stationery.

Another aspect is that a scope type device would not need to worry about things like long-term drift with temperature and aging, as absolute values aren't important.

[Fraser]

Yes that's the one. You need to translate the lens mm specifications into 'Field of View' though as that tends to be the spec supplied with cameras. The E4 has a very simple and small lens but is still good for 300m spotting human targets.

Aurora

[jjmmss00]

-> In short: Flir is just setting up a challenge instead of ensuring that their customers get maximum satisfaction.

Who's saying that such challanges are not satisfying to some customers

Greetings,

Chris

I have been reading the discussion of the hacking efforts for the new firmware with much interest, as I am wondering what will happen if I someday need to return my new hacked camera for repair, and I receive something much less valuable in return.

However, tiring to look at the big picture, I think that if Flair can make a high quality camera for a low price, other manufactures can also. In the end the camera prices will be subject to competition. Perhaps we will all be able to replace our camera's in the future with something better at a lower price.

[bookaboo]

I just got my hands on a 2.1 unit with permission from the owner to tear down for pictures. I know everyone wants to see what sensor has been used so I will get pics of that, let me know if there's anything else anyone wants to know while I'm in there.

[mikeselectricstuff]

I just got my hands on a 2.1 unit with permission from the owner to tear down for pictures. I know everyone wants to see what sensor has been used so I will get pics of that, let me know if there's anything else anyone wants to know while I'm in there.

Doubt sensor will be different but would be interesting to see if there are any PCB changes, e.g. if the console port FFC connector is still fitted

[bookaboo]

V2.1 Internal photos:
I'm a bit busy today but I hope to have time to put a scope on tomorrow.




 

 

[bookaboo]

I just got my hands on a 2.1 unit with permission from the owner to tear down for pictures. I know everyone wants to see what sensor has been used so I will get pics of that, let me know if there's anything else anyone wants to know while I'm in there.

Doubt sensor will be different but would be interesting to see if there are any PCB changes, e.g. if the console port FFC connector is still fitted

Ugh, I forgot to photo that side of the board and I have to head out now, but ff that's the one accessible by removing the label at the front then yes it's still there 

[DJ]

One wonders if any other products have been studied as intently by such a geographically diverse group.

I've had some designs copied by the chinese, but only saw the results,  not the process.

The good folks at Flir, especially the designers, must be somewhat bemused, as each revision gets trotted out and torn apart.

Sort off odd to have ones work under such a public microscope.  Outside the hacks, the overall execution seems to be holding up rather well.

Flir, if you are reading this, well done. And how's about spreading any bonuses to the design people?

[carlpj]

The E4 has a very simple and small lens but is still good for 300m spotting human targets.

Aurora

Correct me if I’m wrong. Each E4 sensor pixel subtends an angle of 45°/320, or 0.14°. At 300 metres this corresponds to an object 0.74 metres wide (and high; the pixels are approximately square). So a non-obese human will on average partly light up two pixels. Under optimum conditions (think bikini in snow) this will be barely detectable. In hot sun, or among trees, it will be even less obvious.

[mikeselectricstuff]

The E4 has a very simple and small lens but is still good for 300m spotting human targets.

Aurora

Correct me if I’m wrong. Each E4 sensor pixel subtends an angle of 45°/320, or 0.14°. At 300 metres this corresponds to an object 0.74 metres wide (and high; the pixels are approximately square). So a non-obese human will on average partly light up two pixels. Under optimum conditions (think bikini in snow) this will be barely detectable. In hot sun, or among trees, it will be even less obvious.

But they will usually be moving, which will make them easily visible even if they are only a couple of pixels wide

[Zakraf]

I also got a new Flir E4

Model: E4 1.2L
Firmware: 2.1.0
Serial No.: 6392 91xx
Calibration Date: May 23

You can still get into the USB mode menu, but you can’t change to any other mode then “UVC and MSD”

Regarding the USB problem, I also first couldn’t install it on any PC, but FLIR posted a fix
http://flir.custhelp.com/app/answers/detail/a_id/987

[Zakraf]

I did activate the RNDIS_UVC  mode with the “Set_RNDIS_Mode.fif” file from freak_ge

But now I can’t install any drivers for the RNDIS mode and because of that I can’t set back the E4 to “UVC and MSD”

[Zakraf]

Thank you for the PM, it got fixed with manually choose the driver „FLIR Camera Network Device” using the “List of device drivers on the computer” option.
TIConfig is working now, just need someone now adapt the conf.cfc

[JFA]

Has anyone thought of building a focal lenght augmenter on the Ex serie?
My optics classes date from 1984, forgotten everything.
But I was tempted to dust off my old school notes and do some back-of-the-envelope doodling to see if it's possible...

[JFA]

Just found that...
http://www.opticalsoftware.net/index.php/how_to/lens_design_software/winlens3dbasic

[Fraser]

For anyone interested, I have just torn down a Fluke VT02 and posted plenty of pictures here:

https://www.eevblog.com/forum/testgear/fluke-vt02-thermal-camera-visual-thermometer-teardown/msg467641/#msg467641

[ixfd64]

They work for me.

[bookaboo]

Link doesn't work? Not sure what you mean as I can see the photos in your quote. Anyone else having trouble?

[rbg]

I just noticed today that FLIR is using FreeDOS's command interpreter.  I telnet'ed to a 1.22.0 FLIR E4 and was playing with the various commands.  The output of "ver /c" says their command interpreter is from FreeDOS and implies they have modified it (FLIR copyright).  "ver /r" explicitly says it is GPL.  Has anyone asked FLIR for the source of their GPL code?

[Echo88]

Got my new E4 about a week ago and i am very pleased with it. The menu/resolution-hack was done with the TIConfig-Tool without problems, thx marphy!  . At the moment im trying to merge the flir-stream with a C920-webcam-stream and implementing a autofocus-algorithm in OpenCV, just gonna need to find suitable Ge/ZnSe-Lenses first.