FLIR E4 Teardown Q/A and newbie questions

[Fraser]

Anyone who charges for a service that relies upon calibrated equipment would need to think carefully before using any equipment that has been 'hacked' or used outside of the OEM specification for the unit. If the E4 were used to just track down faults, leaks etc. it is being used as a source of data for the operator to interpret and calibration is not such an issue. Quite simply, if a plumber or insulation engineer uses it to find leaks, it will be fine. Conversely, if a an engineer is auditing motor bearing temperatures for a formal report using a hacked E4, he had better be sure that the measured figures can be substantiated using some form of calibrated reference.

In case some readers are not aware, not all thermal measurements require a calibrated camera, some utilise a calibrated thermal black body reference that is set at the acceptable threshold temperature. The camera views the black bod and the target and any differential is immediately obvious. These days it is more common to use the cameras built in temperature measurement capability though. During the Bird Flu outbreak I travelled to Tokyo and the airport had thermal cameras set up with a black body reference in the field of view to provide the operators with a visual comparison of the reference against the crowds of people moving through the check-point. The reference was presumably set at a humans nominal face temperature and anyone glowing hotter than that source got stopped for fever checks.

Legality only comes into a thermal cameras usage if you commit fraud or negligence. If your thermal survey proves to be inaccurate the least you can exec t is a demand for a refund from the customer and no further work with them. The right tool for the job applies in such a case. In commercial work, a thermal camera should pay for itself anyway....it is only hobbyists who have to bear the full outlay with little chance of recovering the money through casual usage.

Aurora

[Iphone_hack]

Thanks
It is much clear
The reason I asked regarding legal issues
because I have home inspectors asking if they can do this in their E4s
Of course I didn't tell them yes or not
But in that industry, you don't need a perfect accurate temp
Just need the general difference in temp so they can see air leakage or moisture
Or even if the floor radiant heat is working or not

By the way
Do you know if the E4 needs calibration once a year ? Once every 2 or 3 years maybe?
Or it is one of those type that don't need it?
Thanks

[Fraser]

FLIR will advise that annual calibration is recommended. Their calibration service also includes performance testing and upgrading to the latest firmware.

[DJ]

Thermal imagery for disease screening is looking for people with fevers. As only a few degrees difference is important, resolution,  as well as absolute accuracy, the known reference is required.

[Baya73]

I own an E6 (E6 1.1 - 1.19.... I've got the menu hack and resolution hack working for 3 months now, but can reliably use 192.168.250.2 only. Also, The camera only works with RNDIS, not RNDIS + UVC. I'm looking for a way to make it work on RNDIS + UVC so I can use Joe-c's excelent remote control app.

As of right now RNDIS + UVC gives a picture signal to the software but disables all commands to the camera. If I use RNDIS only. I get no image but the remote commands work.

hoping someone knows what else I could try.

Thanks !

[Baya73]

I fixed my problem with using RNDIS+UVC. Running TIConfig (even if already upgraded somehow fixed the issue). But now when I try to use the Internet... no Internet. If I disable the camera.. Internet connection backup and working. I tried moving all the adresses to a different segment (192.168.102.x) but that doesn't work.   

[AintBigAintClever]

Try assigning a static address to the USB NIC associated with the camera, but with no gateway. Use IPCONFIG /ALL from a command prompt to get the address you were doled out, you need to use the same IP address and subnet mask. Nothing else.

[Baya73]

I'm doing the tests on two different computers running windows 8.1, so far I get the same results.

My network is running DHCP at 192.168.102.1, all computers and devices are getting adresses from the DHCP (192.168.102.range 150-250)

Ok, so I assigned the FLIR NIC at 192.168.0.1 subnet 255.255.255.0... nothing else. When I plug the camera in, the window at the lower right reports "camera at IP address 192.168.0.2"

If I open FlirInstallNet it reports 192.168.0.2 / Flir USB Video

using ipconfig /all:

Computer IP: 192.168.102.181 subnet 255.255.255.0   (autoconfig: yes DHCP: yes)
Flir IP: 192.168.0.1 subnet 255.255.255.0  (autoconfig: yes  DHCP: no)

Results: using Joc-C's program I get "telnet failed" but UVC (video) works + Internet connection works
deactivating computer NIC.... program works in every mode !! I seems the computer's NIC blocks the RNDIS path somehow.

Thank-you for helping me out, I greatly appreciate it.

I will continue trying different configs. I hate windows networking. 

These computers are laptops using Wifi. I will try a Ethernet wired connection on another computer today...

Thanks again !

[Baya73]

... Tested using a wired Ethernet connection and same results ...

I tried to reinstall the Flir device drivers, but it says drivers already installed.

[joe-c]

I have a E45 used in the past. Because that I have installed a "thermal connect 3" or something that sounds like this.

I could remind on connection problems with the E4, so I uninstalled the program.

Do you have a running connection program from FLIR like this? I ask because the notation you told. My PC says nothing while connecting to the E4... but in the past he told the connected E45 with IP. Both cameras use USB.

When I plug the camera in, the window at the lower right reports "camera at IP address 192.168.0.2"

[Baya73]

No, this is my first product from flir. The only other products I use all connect via a RS232 emulator (USB)

 

[joe-c]

Hmm, ok... but why you have this window that says the IP? Or looks it like a windows balloon notation?

I have a screenshot from my device manager for RNDIS_UVC and my camera works well. How this is shown by yours?

[Baya73]

Yes, it's a small windows that rises up from the bottom of the screen.

[Sofia]

Hy again,

with my E4 is something wrong.
When I lock the temperature the colors start changing even if I hold camera in the same place (more and offen then they should - every second).
The screen is going darknes and whiter like when is kalibration but there is no kalibration at the moment.

https://www.youtube.com/watch?v=MYU86l0FLdI&feature=youtu.be

What happend and how I can fix the problem?

What if I will make factory reset by settings? It will be like before upgrade?

Make the menu hack again (without factory reset!). Verify the files you had copied to the camera. I think any one file was copied incorrectly.
Factory reset does not change the hack. The configuration is reset to default values (language, USB mode, color mode, everything can be changed in the configuration).

I made like you say but nothing happens.

Such a fault has not been reported yet. And all other functions are running as well? I have no idea! 
Battery removed? Which firmware? Resolution hack OK? Do you have a backup of the original files? Which way you got the hack done?
While the error persists, can you change other settings in the menu? (manual temp scale, color mode) What are the actual settings (msx, color, measurement...), can you made and show a photo (photo and lcd screen same or different)?
I want to reproduce the situation.

Frimware 1.21.0
MSX and Resolution Hack - ok
I have only original config file, I don't have menu file before upgrade.
I made hack by Fillezilla and cfc-generator and after there wasn't fail message. While I was making hack the camera was in standby.
While error presist I can work normaly.
I work on MSX, Rainbow, lock temperatures but in other option it is harder to find problem but there is also something wrong.

This is the movie when camera work incorrectly. 

https://www.youtube.com/watch?v=MYU86l0FLdI&feature=youtu.be

Hi, Please advice.
I have the same problem as jumbo. Is there a solution?

[jumbo]

Hy again,

with my E4 is something wrong.
When I lock the temperature the colors start changing even if I hold camera in the same place (more and offen then they should - every second).
The screen is going darknes and whiter like when is kalibration but there is no kalibration at the moment.

https://www.youtube.com/watch?v=MYU86l0FLdI&feature=youtu.be

What happend and how I can fix the problem?

What if I will make factory reset by settings? It will be like before upgrade?

Make the menu hack again (without factory reset!). Verify the files you had copied to the camera. I think any one file was copied incorrectly.
Factory reset does not change the hack. The configuration is reset to default values (language, USB mode, color mode, everything can be changed in the configuration).

I made like you say but nothing happens.

Such a fault has not been reported yet. And all other functions are running as well? I have no idea! 
Battery removed? Which firmware? Resolution hack OK? Do you have a backup of the original files? Which way you got the hack done?
While the error persists, can you change other settings in the menu? (manual temp scale, color mode) What are the actual settings (msx, color, measurement...), can you made and show a photo (photo and lcd screen same or different)?
I want to reproduce the situation.

Frimware 1.21.0
MSX and Resolution Hack - ok
I have only original config file, I don't have menu file before upgrade.
I made hack by Fillezilla and cfc-generator and after there wasn't fail message. While I was making hack the camera was in standby.
While error presist I can work normaly.
I work on MSX, Rainbow, lock temperatures but in other option it is harder to find problem but there is also something wrong.

This is the movie when camera work incorrectly. 

https://www.youtube.com/watch?v=MYU86l0FLdI&feature=youtu.be

Hi, Please advice.
I have the same problem as jumbo. Is there a solution?

Did you check camera before upgrade?
I dont have solution for this jet.

[Fraser]

That is a really strange fault that is being shown in the you tube video. Sorry to say I have no cause to propose as it does not look like a compensation table related issue.

I do have to wonder though if we are seeing the occasional micro-bolometer that has 'issues' but is adequate for use in the low resolution E4 as that operates a 4x4 pixel averaging algorithm to lower the resolution to 80x60. It may well be that some cameras exhibit unfortunate image issues when the full resolution is released and pixel averaging is removed. Regret this is pure conjecture as I have no inside line of why some cameras are not working well after the 'upgrade' There was previous discussion regarding whether FLIR use 'binning' to select which micro-bolometers are fit for use in certain models of Ex series cameras. It was never proven one way or the other and it would certainly be interesting to see the dead pixel map of an E8 camera. Dave has an E8 

On e test would be to revert the cameras that are misbehaving to standard E4 state and seeing whether the issue disappears. If it does, the resolution hack ONLY should be applied. If the issue returns, I believe those cameras have limitations in the hardware that make them poor candidates for the upgrade. Return them for an exchange or refund if possible. Sadly I understand that FLIR are inspecting some cameras before refunds are issued, so some owners could be in for a rough ride 

[Sofia]

That is a really strange fault that is being shown in the you tube video. Sorry to say I have no cause to propose as it does not look like a compensation table related issue.

I do have to wonder though if we are seeing the occasional micro-bolometer that has 'issues' but is adequate for use in the low resolution E4 as that operates a 4x4 pixel averaging algorithm to lower the resolution to 80x60. It may well be that some cameras exhibit unfortunate image issues when the full resolution is released and pixel averaging is removed. Regret this is pure conjecture as I have no inside line of why some cameras are not working well after the 'upgrade' There was previous discussion regarding whether FLIR use 'binning' to select which micro-bolometers are fit for use in certain models of E4. It was never proven one way or the other and it would certainly be interesting to see the dead pixel map of an E8 camera. Dave has an E8 

On e test would be to revert the cameras that are misbehaving to standard E4 state and seeing whether the issue disappears. If it does, the resolution hack ONLY should be applied. If the issue returns, I believe those cameras have limitations in the hardware that make them poor candidates for the upgrade. Return them for an exchange or refund if possible. Sadly I understand that FLIR are inspecting some cameras before refunds are issued, so some owners could be in for a rough ride 

Thank you for the quick reply.
Do I want to ask whether I could to taste factory set to "E8 +" and what the consequences.

Thank you.

[JFA]

For the fluctuation problem...
I suggest you test your cam freshly turned on, vs turned on for quite a while.

Reason is, I observed that there is fast thermal drift if the cam was cold and I use it in my warm hands in a warmer environment, enough to produce significant variation (at least on the color high contrast palette).  Images taken before and after show that the temperature readings are OK, that it is simply the color space that is being re-adjusted very quickly.   This virtually does not happen with the regular palettes, but the high contrast color changes color space so fast with such a small delta-T that it is sensitive to it.  After the cam has been on for a while, it sorts of stabilises.  I've used two E4 up until now and they both do that, but it causes no problem as far as measuring temperatures (with Tools and/or with the in-camera cursors (H, L, Spot) is concerned.

I suspect the effect (perception) is is a conflation of eye sensitivity to colors, camera temperature drift, and possibly some minor software adjustment hunting for a set-point. 

[JFA]

Yup, look at the max temp value on the scale. You'll see that it drifts down while the image is darkening, then it fast climbs back up when the image is made paler (sawtooth pattern). 

Did you store your cam in a significantly colder or warmer place than where you ran the test?

Maybe it's the thermal adjustment loop (software) that is not updated often enough?  Maybe there is a software variable for the update speed...

(I haven't looked at the config files yet)

[Sofia]

Yup, look at the max temp value on the scale. You'll see that it drifts down while the image is darkening, then it fast climbs back up when the image is made paler (sawtooth pattern). 

Did you store your cam in a significantly colder or warmer place than where you ran the test?

Maybe it's the thermal adjustment loop (software) that is not updated often enough?  Maybe there is a software variable for the update speed...

(I haven't looked at the config files yet)

Thanks for the advice.
It is a fact that in yesterday's first test was in the hall very hot 38c. I will try to get her to a cooler environment and test.
I know. :-)

[JFA]

It's not a matter of cooler or warmer, it's a matter of having a camera more or less thermalised (equalised in temperature) with the environment.
Note that your hand influences it too, makes it warmer.  If you stored the cam in your basement at 19C, brought it at +38 and ran some tests, expect major drift.

It seemed like the speed of the drift exceeded the software temperature compensation interval.

[joe-c]

In the Video the thermal Scale was locked so it was easy to see, that the whole image was drifted fast.

I don't think this comes from the detector. Maybe the internal temperature sensor has a defect.
The internal calculation I don't know. But I know, that the viewed image was a result of a detector pixel value and the thermal reference (and the image from closed shutter additional). If this reference not work correctly, the result (whole thermal Image) will also not correctly shown.

Maybe just a setting problem... I will take a look into the resource tree.

Edit:
Interesting... there exist a "antiRamp" node.
I tried some changes but could see a result...
.calib.detector.data.ds250C_we.settings.antiRamp.ODLA                           8
.calib.detector.data.ds250C_we.settings.antiRamp.ODLB                           7
.calib.detector.data.ds250C_we.settings.antiRamp.ODRS                           7
.calib.detector.data.ds250C_we.settings.antiRamp.ODSTART                        0
.calib.detector.data.ds250C_we.settings.antiRamp.RBA                           28
.calib.detector.data.ds250C_we.settings.antiRamp.RPA                           46

You could try to deactivate the TSens usage:
rset .image.flow.TSens.partOfImgFlow false
the TSens seems to be frequently used (.image.flow.TSens.updateInterval  500).
Is see no changes on my Camera... but have you after turning off the same problem?

If I make: rset .image.flow.baffle.partOfImgFlow false
The thermal image looks like a view through a small plastic sheet.
After NUC and turn on again I have the normal Thermal image but the Temperature seem to be shifted by +46°C...additional NUC and it was normal again.

In the Image node also a "antiramp" exist.
Here I could see changes:
.image.flow.detector.antiRamp.ODLA                           8
.image.flow.detector.antiRamp.ODLB                           7
.image.flow.detector.antiRamp.ODRS                           7
.image.flow.detector.antiRamp.ODSTART                        0
.image.flow.detector.antiRamp.RBA                           28
.image.flow.detector.antiRamp.RPA                           46
Maybe the Values above was for Restart or factory reset.
however, what are your values?

[Sofia]

In the Video the thermal Scale was locked so it was easy to see, that the whole image was drifted fast.

I don't think this comes from the detector. Maybe the internal temperature sensor has a defect.
The internal calculation I don't know. But I know, that the viewed image was a result of a detector pixel value and the thermal reference (and the image from closed shutter additional). If this reference not work correctly, the result (whole thermal Image) will also not correctly shown.

Maybe just a setting problem... I will take a look into the resource tree.

Edit:
Interesting... there exist a "antiRamp" node.
I tried some changes but could see a result...
.calib.detector.data.ds250C_we.settings.antiRamp.ODLA                           8
.calib.detector.data.ds250C_we.settings.antiRamp.ODLB                           7
.calib.detector.data.ds250C_we.settings.antiRamp.ODRS                           7
.calib.detector.data.ds250C_we.settings.antiRamp.ODSTART                        0
.calib.detector.data.ds250C_we.settings.antiRamp.RBA                           28
.calib.detector.data.ds250C_we.settings.antiRamp.RPA                           46

You could try to deactivate the TSens usage:
rset .image.flow.TSens.partOfImgFlow false
the TSens seems to be frequently used (.image.flow.TSens.updateInterval  500).
Is see no changes on my Camera... but have you after turning off the same problem?

If I make: rset .image.flow.baffle.partOfImgFlow false
The thermal image looks like a view through a small plastic sheet.
After NUC and turn on again I have the normal Thermal image but the Temperature seem to be shifted by +46°C...additional NUC and it was normal again.

In the Image node also a "antiramp" exist.
Here I could see changes:
.image.flow.detector.antiRamp.ODLA                           8
.image.flow.detector.antiRamp.ODLB                           7
.image.flow.detector.antiRamp.ODRS                           7
.image.flow.detector.antiRamp.ODSTART                        0
.image.flow.detector.antiRamp.RBA                           28
.image.flow.detector.antiRamp.RPA                           46
Maybe the Values above was for Restart or factory reset.
however, what are your values?

Hello
I do not know if I'm in the right way "Google Translate"
The picture's how I came to the data. Data are from backup TICconfig.
Factory settings I've not made. The camera in my work and I'll get to it tomorrow.
Otherwise, the camera is 1.22. # Generated 05/15/2014 at 9:54:00

[Sofia]

In today's test at 25c after 10min still the same :-(

[joe-c]

In today's test at 25c after 10min still the same :-(

have you changed some of the settings (over Telnet with RNDIS mode)?