Flir E40 Upgrade & Configuration

[kaz911]

Hi all hardworking people,

got my hands on a E40 mk2 with MSX - sn 645 HW v. 1.2 FW 4.11.0

Managed to get it updated and working in 320 via Parallels Win10 on MBP. Took a while to realise that Flir Tools no longer installs the RNDIS drivers....

Questions: I have the 10mm (45deg) lens for my old E50bx. But when I use it on the E40mkII - MSX can not be enabled? (So add lens and REMEMBER to change lens settings to 45DEG)

2. Does that require additional calibration to get MSX or will it not work at all? (Got 3 options on MK-II in Lens " None, FOV 45d and FOV15d" - but on the MK I -  I have the 45DEG Lens as a 4th option with S/N listed. So "FOV 45d SN 234234234" - I have tried copying the Maps over from the MK I - but no change. So where does FLIR store that info?... 

In general I find the old MK I - more geeky ... but about the same image quality. I have not found an easy way to add more than 1 measurement box onto the MK II?

I do find focus easier with MSX - image kind of "jumps" in place when you have focus.

45 Deg lens seems to decrease temperature readings with about 2 deg C. So ALWAYS remember to set the lens settings correct else your measurements are out of whack!

3. Does anyone have release notes about what has happened since firmware 4.11.0? And is there a >4.11.0 with service menu available somewhere?

The manual now says it has a "Spotting" function which is meant to spot people with high temperatures - but I don't know which release that was in. The Menu for enabling it is not in the 4.11.0 (it disables camera sleep mode as well)

/k

[Fraser]

About supplemental lenses and Exx series......

The E40 is designed to take wide or tele supplemental lenses as you know but, as thermal energy supplemental lenses, they have no effect on the built in visible light cameras field of view. MSX uses the visible light camera for scene edge detection so basically as soon as you change the field of view of the thermal lens, the MSX cannot match the FOV and becomes useless. It is therefore disabled.

The supplemental lenses are normally included in the cameras factory supplied calibration tables but for the most accurate measurements, FLIR say you need to send the camera and it’s supplemental lenses to them for full calibration with the specific lenses fitted. Whether this is truly justified is open to discussion. What FLIR do is install the supplemental lens and carry out calibration to create a specific calibration table for that specific serial number lens. Hence why your MK1 contains the serial number of the 45 degree lens as an additional 4th lens entry. FLIR state that you should keep the lens with the camera to which it was matched at the time of calibration.

The MK1 Exx camera contains the service menu and its calibration routines so a suitably equipped owner can calibrate the camera and any supplemental lenses themselves.You need accurate Blackbody sources for such calibration work. The MK2 Exx series had the service menu removed so sadly you are not normally able to calibrate your own camera or the supplemental lenses. I do not think the calibration tables from your MK1 camera can be used with any success in your MK2 camera.

If you really do need accurate temperature measurements whilst using the supplemental lens, you will have little choice but to send the camera and lens to FLIR for a supplemental lens calibration.

Fraser

[wirbelwind]

I found some interesting pictures in the flir_t6xx_pn55903_04_v5.38.8_update_pack.

[Finderbinder]

Hello.
Just upgraded my E50 1.2, P/N: 64501-0201, SW: 4.11.0 a little bit (resolution and noise).
Would want to upgrade it to a full E60 or so. Thought there are many pages written regarding upgrades I didn't find a complete list (or files) corresponding to my camera's version (hw/sw). Am a little scared to brick or mess my cam acting without a propper guide.

[Finderbinder]

I think my E50 was altered by previous owner. All *.cfg files I got with my cam are attached.

\FlashFS\system\appcore.d\config.d\conf.cfg
\FlashFS\system\appcore.d\config.d\ui_conf.cfg
\FlashFS\system\ui.d\config.d\conf.cfg
\FlashFS\system\services.d\config.d\conf.cfg
\FlashFS\system\service\appcore.d\config.d\conf.cfg

Especially service_conf.txt looks strange for me - a common ending for *.cfg files is missing there.
And the string in it I doubt corresponds to E50:
.caps.config.image.settings.resIR int32 320

May someone with more experience check these? Are these files correct or messed up somehow?
It would be perfect if someone could share original E50 or E60 or Exx config files of the same HW 1.2, P/N: 64501-0201, SW: 4.11.0 version.
Or/and how properly modify these to a full capabilities?

[Userli]

Hey,
I just got hold of a E40 0.10, firmware version 1.24.16 .
I applied the config changes and got the higher resolution and all the other goodies working.
Now I would like to upgrade to the firmware 2.23.14, which unfortunately I can’t find at any of the places mentioned in this thread.
Could somebody please send me or point me to the corresponding file(s)?

[Fraser]

PM with link to the file on my Google drive sent 

[Userli]

Thank you very much Fraser!!!

After a bit of playing around with version 2.23 I noticed that the toolbar is shown vertically on the left, whereas on most pictures here and in the manual I see it horizontal at the bottom of the screen. Looking through the menu and the config files I couldn't see a setting to change this. Is there a possibility to do so?

[Userli]

I just compared the content of the firmware updates 2.23.14 and 2.27.16 and that’s what I found:

6 new files
142 files identical
125 files different
Fpga*.bin files identical
OS version 15.4.12->15.4.14
new strings related to screening
new camtype Z3 in many .asp files
new directory system\services.d\factory.d containing
dcf.rsc
files_count_limit.rsc
new files:
system\web\services\Focus\LensDist.asp
system\sensor.d\ML2.dll
system\web\images\Fenix.jpg
system\web\images\Z3_Astra.jpg

To me this looks mainly like an OS update and the introduction of the new camtype. Is there a specific file, which could hint to more security in the new version?

[Userli]

I checked all the DLLs and didn't find any alarming changes. So I gave it a try and updated to 2.27.16, which worked without any problems. The resolution is still 320x240 and the config files unchanged.

[Userli]

As a proof an interesting result of emissivity: fabric with black printing.

I also copied the medical.pal from the model 2 firmware and added the corresponding translation into the English strings file (strings.en.xml). It now appears as an option. (No real application but I think it was mentioned here somewhere, that it's not possible to add another pallet)

There are two hidden items in the service menu, which you can enable in usermenu.inc: WebCamera and Real Time Video

[jjdemu]

Hi all

I recently got a flir e40

serial: 490xxxxx
part number: 49001-2001
firmware: 2.19.10
model: E40
without MSX

I have tried changing the resolution to 320 and setting noise to 0 in the conf file, but when I reset the camera I lose the cross-hairs. I used crc01 and added that to my conf file. I replace the file back to the original and cross hairs are back. So I am a little lost if someone could chime in on what I must have gone wrong. Still on the fence if I want to upgrade to 2.27.16.

Update 1
Got the flir 2.23.14 update from Fraser, so updating and will be doing a new backup of files and will report back.

update 2
still have no cross hairs. Nm think I see what I did wrong and read more.

update 3
all working great


Thank you all for a great forum and the wealth of information I have found here.

[Userli]

I developed a web page to be added to the system web pages, to browse and edit the resource tree. The resource tree is a data container/hub centralizing all system parameters. The tree is huge and much of the information is quite cryptic. It still provides a good insight into the camera internals. The edit function will most likely not allow for magic changes but the standard modifications like switching to RNDIS mode or even removing the flir logo on the screen are possible.
The installation is quite simple. The attached zip contains the file resources.asp, which needs to be copied to the folder FlashFS\system\web\service\caminfo.
Now the the line

Code: [Select]

menuRow("Resource Tree", "/service/CamInfo/resources.asp", "submenu1", selectedMenu == "mnuCamInfoResources");

needs to be added to the file file FlashFS\system\web\service\inc\servicemenu.inc, like in the example file in the attached zip.
Finally the function

Code: [Select]

ResTree.prototype.getResourceType= function( name )
{
return this.res.GetResourceType(this.expandName(name));
}

Needs to be added to the file FlashFS\system\web\inc\restree.inc, again like in the example file in the attached zip.

I assume by changing the wrong parameters you can brick you camera so use the editing with care and at your own risk.

I attach two images showing the browser in action.

Happy browsing!

[Userli]

I noticed a faint column like structure on the display, which moves with the camera and thus can’t be noise. Looking at some of the pictures posted here, I think to see a similar effect on other cameras.
In order to investigate this phenomenon, I took a sequence of 90 images with the dust cap on. I stacked those to remove noise and used autolevel of imagemagick to make it more visible. The resulting structure resembles the effect I noticed before.

Quantitatively the effect seems quite small though. I see a change of ~30 which, assuming the full range of two bytes is used, is very few but big enough to be visible.
When browsing the resource tree I found background maps referenced (ds120C_we_ap_fi_le_x.bgm) but I don’t find these files on the camera. My guess was that updating those could remove the structure, which most likely developed with age.
Did anybody ever look into this?

[Fraser]

This is column noise and is present on most microbolometer based systems. The columns are read out and small differences in the integrator capacitors produce the differentials that you are seeing.

Fraser

[Bill W]

The term 'noise' may be unhelpful, as 'noise' often gets interpreted as a variable you can average out. 
It is column fixed pattern, and the source is as you say the readout through the slightly differently performing column amplifiers.

The vast majority of the variation is calibrated out, but it is a case of (big number) - (similar calibration big number) and (big number) is highly temperature variable.

The remainder is (mostly) sorted at the shutter calibration but between shutters there is no fix, likewise when viewing conditions (ie amplifier signal levels and temperature) are different to the shutter in place condition.

One reason most cameras will not let span go under 2°C, unless they are so noisy they can hide the fixed pattern anyway

Bill

[Userli]

I started to analyze this problem more in detail. Noise is indeed a random perturbation of the signal and as such can be averaged out. This is not to be confused with systematic effects, which can't.
I started again with a sequence of 90 images with dust cap on. You see the grey scaled version of the image with autoleveling to emphasize the effect at the top. Below you find the distribution of the amplitude of the pixel value variation per pixel. This is the noise level which shows a nice peek (the noise generator is switched off). The system noise is roughly 40 pixel units, which is significant as we will see later.
The plot below shows the distribution of the averaged pixel values. Here you would expect a sharp line at the dust cup temperature.
We can see a random variation around the peek, which makes me wonder if the initial calibration has been done with active noise generator (maybe due to restrictions for civil use?).
The tail to the right must be a drift of some kind most likely due to aging or burn in. This becomes handy as we see later.
The mean pixel values per image column on the top right show the extend of the column effect and that the drift is regional. 
In the next post I'll look into the gains.

[Userli]

I now looked at the content of the file ds120C_we_ap_fi_le_static.gan, which I assume to be the gains of the amplifiers per pixel.

The grey scale image on the top is not very revealing, most probably due to outliers for bad pixels. I should exclude those.

The distribution of gain values in the middle is difficult to predict. This mostly reflects hardware production variations, which might be random or systematic.

The scatter plot on the bottom shows mean pixel values against gains. You would expect a horizontal band with the width of the generated noise (see last post) around a line like the one in green. We can see clearly an additional band in the red square. The latter represents most likely a hardware drift over time. As mentioned already his could be burn in or aging. The interesting part is that, assuming this is a gradual process, this band reflects the function we need to use to compensate for the drift. In visual terms, we need to move those points back to the baseline along the diagonal band.

I fitted a line through this band (visually it's most like a higher order function but let's stay simple for the time being) and found a slope of 0.0119 . This is the ratio delta pixel values over delta gains. Using this ration I now (visually speaking) moved all points to the green line and saved the corrected gains to a new gains file. It seems that this file is only being read at camera startup and it's not sufficient to just restart the application with the corresponding batch file. If somebody knows an easier way to do this, I'd be happy to learn about it.

I'll show the results in the next post.

[Userli]

We see that the column structure is still visible but the regional drift is mostly gone. The "U" shape might be due to the linear correction, whilst at least a second order correction would most likely be required.

However, the regional effects are now in size close to the column effects. The plot of pixel values against gains shows a single band without the drift.

Quantitatively we are in the range of the noise and in principle it doesn't make sense to continue. It would be nice, though, to do a beautification and level out the column noise, to remove the visible effects, which now are already much less visible than before.

To be continued...
 
P.S.: I also looked at the coarse map (3rd image), which I assume to be the pedestals. They don't show any remarkable structure, so I assume it's save to ignore them.

[_Wim_]

We see that the column structure is still visible but the regional drift is mostly gone. The "U" shape might be due to the linear correction, whilst at least a second order correction would most likely be required.

However, the regional effects are now in size close to the column effects. The plot of pixel values against gains shows a single band without the drift.

Quantitatively we are in the range of the noise and in principle it doesn't make sense to continue. It would be nice, though, to do a beautification and level out the column noise, to remove the visible effects, which now are already much less visible than before.

To be continued...
 
P.S.: I also looked at the coarse map (3rd image), which I assume to be the pedestals. They don't show any remarkable structure, so I assume it's save to ignore them.

Very interesting. For further testing you should also make one of these: https://www.eevblog.com/forum/thermal-imaging/anyone-have-an-estimated-emissivity-for-the-ivac-9000-radiators/msg3510108/#msg3510108

Cheap black body to play with...

[Userli]

Indeed I thought about buying or building a bb reference. It's very interesting what Fraser and you have done.
To me, however, it's important to have a geometrical homogeneous measurement, more than a precise one. I will rather look for temperature differences than for absolute temperatures.

Most of the bb sources are too small to cover the whole field of view of the camera and larger ones are quite costly. Having said that, I would definitely be interested to play with such an IVAC unit but whilst they're sold cheap, I'm afraid of the costs of importing an IVAC to Europe.

[_Wim_]

Indeed I thought about buying or building a bb reference. It's very interesting what Fraser and you have done.
To me, however, it's important to have a geometrical homogeneous measurement, more than a precise one. I will rather look for temperature differences than for absolute temperatures.

Most of the bb sources are too small to cover the whole field of view of the camera and larger ones are quite costly. Having said that, I would definitely be interested to play with such an IVAC unit but whilst they're sold cheap, I'm afraid of the costs of importing an IVAC to Europe.

Total cost with import and shipping to Europe will probably be around 70€ (an offer of around 30$ for per unit will probably be accepted):
https://www.ebay.de/itm/203833704424?epid=6034823438&hash=item2f756f7fe8:g:GuIAAOSwbVViDErS

They do appear to be geometrical quite homogeneous if you look at these test shots I made:
https://www.eevblog.com/forum/thermal-imaging/anyone-have-an-estimated-emissivity-for-the-ivac-9000-radiators/msg3504490/#msg3504490

Large dedicated BB options are probably slightly better, but price is at least an order of magnitude higher.

[Userli]

Total cost with import and shipping to Europe will probably be around 70€ (an offer of around 30$ for per unit will probably be accepted)

Thanks for the info! I gave it a try and accepted the counter offer at 40$. The total costs are roughly 80€ delivered. I'll read again your threads in more detail to see what I'll need to do to the device...

On another note:
I just saw that Flir tools indicate a frame rate of 30fps in the live view, whereas the Exx models should have 60fps. Do I miss a setting somewhere?

[_Wim_]

Total cost with import and shipping to Europe will probably be around 70€ (an offer of around 30$ for per unit will probably be accepted)

Thanks for the info! I gave it a try and accepted the counter offer at 40$. The total costs are roughly 80€ delivered. I'll read again your threads in more detail to see what I'll need to do to the device...

Great! If any questions, please feel free to ask (maybe best in the thread about the BB, so the info remains in a single thread).

Total cost with import and shipping to Europe will probably be around 70€ (an offer of around 30$ for per unit will probably be accepted)

On another note:
I just saw that Flir tools indicate a frame rate of 30fps in the live view, whereas the Exx models should have 60fps. Do I miss a setting somewhere?

Good question, never noticed that before. I presume this is related to the usb speed to transfer the images. Maybe it records only in 60fps on the camera itself. I do not know of any settings to change it in flir tools, but as I did not realize before, have not looked hard also 

[cuda12]

My e30 /e60 is 30 fps on flir tools and even on a vid recorded just on the camera (no usb cable) and then transferred and watched on a pc it is still always 30 fps. In flir tools you can choose a lower fps but I have never found a higher choice. Just lived with it as it is actually pretty good.