(7 Updates, still needing help fpga bin file) Hacking my flir i5

[sam1275]

Hi everyone.

I got a used i5 mk2 from ebay with FW 6.5.3, it's very good condition and works fine, I want to hack it to i7, and I followed this:
https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/msg326229/#msg326229

I finished 1-6 but stuck at step 7, I cannot ping or connect to my camera, I cannot even get an IP address auto-assigned, and the virtual network interface for the RNDIS always have 0 received packets. I tried manual assign an IP address but no effect at all. My main PC is a clean Windows 7 x64.

I then tried it on another virtual machine running POSready(a WinXP x32 edition), I factory reset the camera before, made sure the OS is clean, the firewall is turned off, driver is official and rebooted after install, I tried manual assign 192.168.0.X to the pc side, then the "thermacam" says the camera is on 192.168.0.Y(Y=X+1), but then still no traffic, and themacam always throw a error 67 after a minute or two saying "the network name cannot be found".

Then I uninstalled the standalone driver, and try to install Flir tool instead, now the virtual machine is in a endless reboot loop, every time logged in, it copy a few files and says need restart to install, always the same...

What could be wrong? Please anyone give me a hand! Thanks!!!


Update1: I restore the virtual machine to the clean OS state(I have the original HDD container backed up), then installed all dotnet and VC++ library, then install flir tool, this time takes 2 hours because of the online installer(first time took 10 minutes then boot loop). But I found I still have to install the device driver because the flir tool doesn't contain one.
And after all those, the same problem still happens, I use the flirinstallnet.exe in the flir tool menu to do more test, this program can see my camera and show it in the list, but every instruction ended with
"Connection timeout
Operation aborted
Close TELNET link
NoAdapter"
No matter I manually set the ip address or not.

Update2: I think I can try to modify the file through the fif file update method:
https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/msg321956/#msg321956
I can update it through SDcard.
But that file given is for EX, can anyone give me the original ix's file: "flashfs\system\appcore.d\config.d\conf.cfg" ? Thanks!

Update3: SUCCESS! I setup another windows 7 virtual machine, without useless virtual hardware such as sound card and printer, then I install as much network service component as I can, then set all net work to home zone to enable share, OS firewall leave untouched, then install flir driver, reboot, connect my camera, and success!!!
However I have a few more questions:
1. How can I check if the resolution/FOV is extended? I have some picture before and after my hack, they do seem better, but not that big, how can I make sure the hack is working?
2. Are there any menu hack for ix? I tried a few settings in different config files to enable more features, but unsuccessful, I download the e4 menu hack and found there are many files to be overwritten, including some I don't recognize, so I'm not dare to just use them because they are for ex...

Update 4: I'm answering my first question in update3: I set the emissivity to the lowest 0.10, because I found the lower the emissivity, the higher the ISO will be, so there will be a lot of noise at 0.10 setting, that is very good for me to determine the pixel size, then I found my hack real works!
Then I tried different ".caps.config.image.settings.resIR int32" value, then I find it have limitations, the real resolution will not lower than 60x60 no matter how low I set the value, and will not higher than 120x120 no matter how high I set it, I even set both ".caps.config.image.settings.resIR int32" and ".caps.config.image.settings.fpgaResolution.resIR int32" to 240, but no difference than 120.
About the menu hack, I tweaked many setting that enable the measuring function, but the menu still not showing them, I think the menu item is different to the internal function settings.

Update 5: I found a file called "fpga_isc0601_120x120_c16.bin" under "\FlashFS\system\", I think it's limiting my resolution to 120x120, I think I have a mk2 device, but I cannot get 140x140 because this file? Could anyone have a mk2 ix camera check your similar file, and upload it if it's 140x140? Please also provide "fpga.castor.bin". Thanks!

Update 6: Menu hack success! Just add the gui.cfg.

Update 7: I'm tweaking "\FlashFS\system\calib.rsc" and found many use of it, Now I can:
1. Extend temperature range -273.15 - 4726 Celsius;
2. Modify minimal temperature-span to 1 Celsius, so I can see much better detail in some cases;
3. Disable artificial noise generator, or set to lower value;
4. Set auto-NUC's start-up calibrate count, min/max calibration interval, max allowing temp delta;
5. ...And many more...
I modified calib_org.rsc as well, I don't know if it's necessary.

[Ice-Tea]

Eh, the i5 !=the E4?

[sam1275]

Eh, the i5 !=the E4?

Thank you for replying, I already knew, the thread I linked is how to hack a ix.

[Ice-Tea]

Missed that, sorry...

[sam1275]

Missed that, sorry...

No problem 

[tomas123]

https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/?action=dlattach;attach=85490

[sam1275]

https://www.eevblog.com/forum/thermal-imaging/flir-e4-thermal-imaging-camera-teardown/?action=dlattach;attach=85490

Thank you, but I already saw that manual before, and that cannot help me, I cannot even get a single packet receive on the RNDIS NIC.

[tomas123]

You are not the first user, wo has troubles with RNDIS drivers.

I think, that a USB over a VM is not the best way (two handshakes).
Try to install the Flir Tools on the host machine (make a snapshot, if you hate this).

[sam1275]

You are not the first user, wo has troubles with RNDIS drivers.

I think, that a USB over a VM is not the best way (two handshakes).
Try to install the Flir Tools on the host machine (make a snapshot, if you hate this).

Thank you, I did try it, as I said in my first post, but not work.

[Bud]

You can try couple things:
- unplug the  PC from the network before connecting the camera, to eliminate possible IP address conflict
- install Windows XP, that worked for many including myself.

[sam1275]

You can try couple things:
- unplug the  PC from the network before connecting the camera, to eliminate possible IP address conflict
- install Windows XP, that worked for many including myself.

Thank you, I already success, please see update 3.

[tomas123]

Update3: SUCCESS! I setup another windows 7 virtual machine, without useless virtual hardware such as sound card and printer, then I install as much network service component as I can, then set all net work to home zone to enable share, OS firewall leave untouched, then install flir driver, reboot, connect my camera, and success!!!

As I wrote, most troubles comes from the configuration: VM with virtual hardware  + RNDIS driver.
You wrote, that your windows host machine with Flir Tools + RNDIS drivers also doesn't work. This is unusual...

[sam1275]

Update3: SUCCESS! I setup another windows 7 virtual machine, without useless virtual hardware such as sound card and printer, then I install as much network service component as I can, then set all net work to home zone to enable share, OS firewall leave untouched, then install flir driver, reboot, connect my camera, and success!!!

As I wrote, most troubles comes from the configuration: VM with virtual hardware  + RNDIS driver.
You wrote, that your windows host machine with Flir Tools + RNDIS drivers also doesn't work. This is unusual...

Yes, and it might because I'm keeping my host system as secure and light as possible, I uninstalled unused windows components, close all unused or unsafe services, use Kaspersky and turn all settings as high as possible...

[Platonas]

Update 7: I'm tweaking "\FlashFS\system\calib.rsc" and found many use of it, Now I can:
1. Extend temperature range -273.15 - 4726 Celsius;
2. Modify minimal temperature-span to 1 Celsius, so I can see much better detail in some cases;
3. Disable artificial noise generator, or set to lower value;
4. Set auto-NUC's start-up calibrate count, min/max calibration interval, max allowing temp delta;
5. ...And many more...
I modified calib_org.rsc as well, I don't know if it's necessary.

The file Calib.rsc has a CRC32 and not CRC01 checksum.  Does the crc01.exe works for this in order to creare a CRC32 code?
Do you have the exact lines that can be modified so as to perform the above "improvements"?
In my camera (i5 i think is gen1) ".calib.ccase.0.TSpanMin double 4" is the value for "2" statement of yours.  I will make it 1 but need CRC to be evaluated again I guess.

BTW I did the whole mod/hack and the camera works fine but... The saved images when viewed in Windows Explorer, are reported as 240X240 pixels.  The same applied for all images I had on my card before the hack.  I am confused, ia my camera's resolution 240x240?

[sam1275]

Update 7: I'm tweaking "\FlashFS\system\calib.rsc" and found many use of it, Now I can:
1. Extend temperature range -273.15 - 4726 Celsius;
2. Modify minimal temperature-span to 1 Celsius, so I can see much better detail in some cases;
3. Disable artificial noise generator, or set to lower value;
4. Set auto-NUC's start-up calibrate count, min/max calibration interval, max allowing temp delta;
5. ...And many more...
I modified calib_org.rsc as well, I don't know if it's necessary.

The file Calib.rsc has a CRC32 and not CRC01 checksum.  Does the crc01.exe works for this in order to creare a CRC32 code?
Do you have the exact lines that can be modified so as to perform the above "improvements"?
In my camera (i5 i think is gen1) ".calib.ccase.0.TSpanMin double 4" is the value for "2" statement of yours.  I will make it 1 but need CRC to be evaluated again I guess.

BTW I did the whole mod/hack and the camera works fine but... The saved images when viewed in Windows Explorer, are reported as 240X240 pixels.  The same applied for all images I had on my card before the hack.  I am confused, ia my camera's resolution 240x240?

Yes it need CRC32, so CRC01 will not work, I use this tool to calculate CRC32: https://www.digitalvolcano.co.uk/hash.html .

I'll list a few useful lines here:
".calib.ccase.0.TMaxClip double" --- Max temperature in Kelvin.
".calib.ccase.0.TMinClip double" --- Min temperature in Kelvin.
".calib.ccase.0.TSpanMinAuto double" --- Min temperature span in Kelvin, I also modified ".calib.ccase.0.TSpanMin double", don't know if nessersary.
".calib.nrdp.ds250C_we_ap_fi_le.alt.0.noiseGen.enabled bool" --- Noise generator switch.
".calib.nrdp.ds250C_we_ap_fi_le.alt.0.noiseGen.level int32" --- Noise generator level.
".calib.tune.autoNuc.ds_we_ap_fi_le.maxInterval double" --- Max interval between NUC shutter triggers in seconds.
".calib.tune.autoNuc.ds_we_ap_fi_le.minInterval double" --- Min interval between NUC shutter triggers in seconds.
".calib.tune.autoNuc.ds_we_ap_fi_le.gradientChangeTLimit double" --- Trigger a NUC if sensor's temp change reach this value (I cannot confirm this)
".calib.tune.autoNuc.ds_we_ap_fi_le.caseChangeCount int32" and ".calib.tune.autoNuc.ds_we_ap_fi_le.restartCount int32" --- Force NUC shutter trigger count at power on, I don't know what's the difference between this two values, I change them both to 1 (default 7) but still get 3-4 trigger when poweron, but much lower than default 15 triggers.
".calib.tune.autoNuc.ds_we_ap_fi_le.caseChangeInterval double" and ".calib.tune.autoNuc.ds_we_ap_fi_le.restartInterval double" --- Force NUC shutter trigger interval at power on, I don't know what's the difference between this two values, and they seems to have little effect because when I change the min interval to 10, and this two stay default 15, the start up interval is 10.

Mine is always 240x240 in jpg size, I think the camera just take a part screen shot whenever you take a picture, it have nothing to do with the sensor resolution. To confirm this, don't disable the noise generator, and set the emissivity to lowest 0.10, then set the reflected temp to very high or very low(-100 or+300 or so), it's even much better to modify the temp span to 1 degree, then keep the lens cap on, so there should be tons of noise... Now take a picture and copy it to your PC, zoom it to see the noise pixel, it's much larger than the jpeg's pixel.

Please feel free to ask more questions if you have:)

[tomas123]

Are you sure, that you noticeable extended the temperature range.

My experience is, that Flir doesn't cut the temperature range in which the sensor values are suitable.
I think, that Flir cameras with a second extended temperature range have a hardware modification (ND filter, ADC).
see Flir Exx (-20°C bis +120ºC / 0ºC bis +650°C)

[sam1275]

Are you sure, that you noticeable extended the temperature range.

My experience is, that Flir doesn't cut the temperature range in which the sensor values are suitable.
I think, that Flir cameras with a second extended temperature range have a hardware modification (ND filter, ADC).
see Flir Exx (-20°C bis +120ºC / 0ºC bis +650°C)

Flir ix have a accurate measuring range from -20 to +250 Celsius, and the camera will limit the display number from -40 to +270 Celsius, so you already won't get a accurate measure at all display range. After my modify, the accurate range is still -20 - +250 C, but you will get a wider display range, I think it's better to display a near value, than just stuck at the highest/lowest value.

[sam1275]

I just asked FLIR for the bin file, and as expected, they refused to give me

Hello,
We cannot send these files.  Please contact our Service Center if you are having difficulties with the camera:
ccare@flir.com
XXX-XXX-XXXX (Option X)
Regards,
XXXXX

So... Still need the 140 bin files, anyone have it?

[ignat750]

Имею i7, скину оригинальные файлы.