Flir E4 Thermal imaging camera teardown

[tequipment]

Hey just wanted to drop you a note.  We use Ingram as a 3rd party warehouse in PA.  They have a shipping company.  All the stock we ship from there is OURS and is not bought from them. 

These units were only ours.

Thanks
Evan Cirelli

Vice President
TEquipment.NET

[pomonabill221]

I really doubt the E4 is 'measuring' the available current capability of the charging source. That would be a neat trick...

Typically (and very common with smart phones, e.g. Samsung S4) is that there are 'tricks' that are used to 'guess' the current ability of the source.

One is obviously with a real USB device, the target will request that the host provides a certain current (up to 500mA for USB2) and the host then informs the target if that is possible. Part of the USB enumeration process.

The common Apple ipod technique to determine the type of charger is to look for certain voltages (pullup/pull down resistors in the charger) on the D+/D-. The Samsung S4 also looks for a USB cable that has the shield connected end to end for the ground path connection. My S4 can charge up to 1900mA with a) the correct charger and b) the correct cable. If I switch to a normal USB cable it drops to around 1100mA. Change the charger (or use a PC) and it drops to just under 500mA. There's a free app that provides this info for the S4 in realtime.

cheers,
george.

That makes more sense since the Flir charger is tethered to the supply, where the usb cable I tried is a micro to regular usb cable for both the computer and the second power supply that I tried.
Other than the E4 loading the supply to the point of overload and droop (not very good), the only other way would be to "strap" some pins on the tethered cable... will have to look into that!

DANG!  I was giving the E4 alot of credit for being smart... oh well... we can always dream and "brag"!

[tequipment]

Im reading this but a little late.  Did it work out for you?
Thanks
Evan Cirelli

TEquipment.NET

[amyk]

I did make a statement about not exceeding 400mA during charge, and the battery doesn't get warm, WRONG!, but IF the temp is monitored VERY carefully (and they do have a thermistor in the battery), then fast charging CAN be done as there is a little heat when using the wall wart at over 1 amp.

If it's a 2800mAh cell 1A is only 0.36C, manufacturers usually recommend ~0.5C (1.4A in your case) as a "standard charge" and most can take a 1C "fast charge". A standard 18650 charger should be able to charge it easily.

[IRspector]

Hi All.
Thanks to Mike, Taucher and Uup !!! I upgrade without problems my E4 to E8+ Special Edition.

Model E4 1.0 S/n 6390688X Made in Estonia
Part number 63901-0101
Software 1.19.8
Calibrate 13.11.2013

 Is it possible to make  60 fps instead 9

[nacke]

Is it possible to make  60 fps instead 9

Use search about that. It is a no no thing...

[ixfd64]

It looks like TEquipment.NET is almost out of stock on E4's again.

[pomonabill221]

I did make a statement about not exceeding 400mA during charge, and the battery doesn't get warm, WRONG!, but IF the temp is monitored VERY carefully (and they do have a thermistor in the battery), then fast charging CAN be done as there is a little heat when using the wall wart at over 1 amp.

If it's a 2800mAh cell 1A is only 0.36C, manufacturers usually recommend ~0.5C (1.4A in your case) as a "standard charge" and most can take a 1C "fast charge". A standard 18650 charger should be able to charge it easily.

Yes, that is true... I don't know where I got my 400mA charge current numbers (must be age... MY age!).
And as long as the temperature is monitored and the high charge rate is terminated when the voltage reaches about 4.1 volts (topping charge start), 1 or in some cases, 2C can be used although the battery life might be shortened if high currents are used.... Li-Ion's are BOMBS if charged too fast/overvoltage, but our Flir's don't have that problem.

[tomas123]

here I found a picture from the flir blackbody refrigerator

http://www.flir.com/cs/emea/en/view/?id=41491

Code: [Select]

FLIR ATS recently purchased a unique blackbody that can calibrate thermal cameras at temperatures as low as -80° C

I don't think that Flir calibrated the E4 at cold temperature

create a panorama image with exiftool, imagemagick, MS ICE  and php

Has anyone tried making a panorama?

[stefbeer]

Some people may already know this but it was new for me and it surprised me a little bit, so I wanted to post it here.
Today I took a look at the datasheet for the E4 and it says as follows.

Measurement
Object temperature range	–20°C to +250°C (–4°F to +482°F)
Accuracy	±2°C (±3.6°F) or ±2% of reading, for ambient temperature 10°C to 35°C (+50°F to 95°F) and object temperature above +0°C (+32°F)

The temperature range is nothing new, but at Accuracy the part after "±2%" was completely new for me.
On the first page of the datasheet on the left side above the address, there's a timestamp: December 12, 2013, 10:43 AM
Is this a short term change of the datasheet or has this part been there all this time and I just never noticed it?

[jlr134]

I dont see that on their product description on Flirs website but on the disk that comes with the camera, in the technical data folder the specs for the E4 do state this.
The document is dated April 13 2013 and that spec applies to all the Ex cameras ( and yes it was new to me and surprised me as well but after looking at the specs of fluke cameras...
it looks like they have similar accuracy tempature limitations)

[trap]

E4 1.0    SN 6390xxxx    SW 1.19.8

Successfully now an E8+ with beta2b menu additions.

But I had a bit of troubles getting it to talk to my PC. Doing the menu additions was a real pain, and I did a few things to allow filezilla to talk to it. I disabled IPv6 with a FixIt from Microsoft (google helped). Then still no communication, although the driver popup seemed to indicate that it was on 192.168.0.2, there was nothing at that IP. Finally ran a network scan and found the cam at 192.168.1.2 !!  My router is 192.168.0.x/24, so why it wound up there makes no sense to me. Once I pointed things there, it talked fine.

There is a webserver on the camera. Wondering what trouble I can get into poking around there.

Many thanks to all who made me buy this. I needed to spend the money, since it was burning a hole in my pocket. Seriously, having a TIC was on my bucket list.

[nacke]

Some people may already know this but it was new for me and it surprised me a little bit, so I wanted to post it here.
Today I took a look at the datasheet for the E4 and it says as follows.

http://support.flir.com/DsDownload/Assets/63903-0303_en_50.pdf Same for E8 too

http://support.flir.com/DsDownload/Assets/49001-2101_en_51.pdf and E50

http://www.trinstruments.cz/data/files/flir-t420-wifi-1113.pdf There is no object temperature limit for T420, otherwise same

I'm surprised too

[Fraser]

@stefbeer,

Good spot. I totally missed that important specification detail ! Or more accurately I misread it as a statement that the +-2.0C applied at above 0C ! To be truthful I wasn't buying for amazing accuracy so it wasn't of that much interest.

It makes it very clear that the camera has a specified accuracy only above 0C. FLIR do not calibrate the E4 below 0C, as others have suggested. 

My industrial Raytek IR thermometer actually refuses to measure any temperature below 0C as it would not be accurate at such temperatures.

From the lack of any accuracy detail below 0C in the specs of the E4, it would appear that any temperature shown is basically a 'best guess' and not to be trusted.

I have also been looking into the behaviour of PVC electrical tape and that is also a source of error in my temperature test. PVC tape changes in emissivity as its temperature changes ! Life is never simple. A true clack body is needed to carry out serious testing of the E4. Sadly I no longer have access to one at work. I may do some more testing on accuracy when time permits.

My FLIR PM695 manual suggests measuring a target surface with a thermocouple and and then adjusting the PM695 emissivity setting whilst looking at the same spot with the camera. this accurately sets the emissivity on that specific surface.

The E4 is still very useful to me as I tend to work with targets at temperatures above freezing  .
I will use my accurate, if somewhat cumbersome, PM695's for any work requiring good accuracy between -40C and +500C.

[Fraser]

I just took a look at my PM695 manual and brochures. The manual states an accuracy of +-2% at all temperatures between -40C and +500C (Optional to 2000C) No limitations on target temperature below 0C are shown.

The brochure states the same, though interestingly, the +-2.0C is added to the accuracy specification.

http://www-old.itl.waw.pl/oferta/wypozyczanie/pdf/g38/Flir-ThermacamPM695LWB.pdf

I looked at the Flir E4 manual and it refers the user to the CD for specs  A bit odd I thought. The E4 brochure is even more evasive on the topic. It states an accuracy of  +-2% and a temp range of -20C to +250C  with no caveats. Naughty.

The marketing boys have had a lot of input to the documentation me thinks !

Well the Ex series is a budget range and, as I have stated before, the industrial preventative maintenance cameras cost an awful lot more money. We should not be surprised to find that the Ex series can't truly compete with its more expensive brothers in some areas.

The low temperature measurement capability, or rather the lack there-of, would not have affected my buying decision and we have to ask, how many of us would actually need to measure sub 0C targets ? I tend to work in the world above freezing 

[nacke]

Well, it's annoying to image buildings in winter conditions when those totally false temperatures run on the screen and pictures.

Development suggestion for Mr. Taucher: option to replace absolute temperatures with temperature difference 

[Fraser]

@nacke,

A very fair point. For winter conditions such a limitation is annoying. It would certainly be an issue if someone wanted to sell their services doing heat and insulation audits involving temperatures below 0C. For my personal heat loss audits in the relatively warm UK winter its not such a problem and I tend to use the image to track down the trouble spots and then investigate them more thoroughly with other tools. it will certainly be a lot more pleasant wielding the relatively light weight and ergonomic E4 when compared to the almost 3kg PM series cameras.

I may be wrong but I believe I read of at least one person 're-calibrating' their E4 to correct the errors that he found in the measurements. Maybe this can be done below 0C to at least improve on the current inaccuracy, if not totally eradicate it. A good source of low temperatures would be needed along with a decent emissivity material and a thermocouple thermometer....then the hard work of plotting the differential as the source warms up from say -20C. I have no idea how you 'tweak' the calibration tables but someone on here may be able to shed some light on the matter. if my old PM695 can be calibrated down to -40C I would hope that something can be done by the user on the nice new Ex series.

[Stuey]

Does anyone else have a dead pixel in their LCD screen? *Frown*

I ran through the modification and had a nagging doubt that it wouldn't work. It worked beautifully.

[Fraser]

Massaging it can sometimes remove a dead pixel depending on the cause. I know it sounds nuts and that is what I thought but it has worked on at least 2 LCD PC monitors that I have worked on. You have to be able to press on the actual LCD panel though and I am not sure that is possible without disassembly of the camera 

If you want to try....just google 'fixing lcd dead pixels'. You have little to lose by trying.

http://www.techradar.com/news/computing-components/peripherals/how-to-fix-dead-pixels-on-your-lcd-monitor-598484

Sadly a single dead pixel is not normally considered grounds for rejecting a product but some suppliers have a no quibble returns policy. You can easily revert your camera to standard E4 and return it ?

[Stuey]

The LCD is protected behind a glass or high-strength composite shield.

I'm not going to go through the hassle of returning the entire unit unless the issue gets worse, nor will I open it up to massage the panel. I did consider that, but the protective cover got in the way. Complaining about it makes me feel a little better though. =)

Mike provided a "remove hacks" firmware option in the download, so it should be possible to restore it factory-fresh.

[john19]

I have no idea how you 'tweak' the calibration tables but someone on here may be able to shed some light on the matter.

I used an IR thermometer to plot the difference between it an the E4 in an excel sheet. Then I used this forumula:
http://u88.n24.queensu.ca/exiftool/forum/index.php/topic,4898.msg23972.html#msg23972
to calculate raw pixel values from E4 temperature values. From those pixel values I then made a new plot of what the difference between the IR thermometer and E4 would have been with new calibration parameters. I used trial and error and watched the curve change. When it was small enough I took the new calibration parameters and entered them in the calib.rsc file, calculated a checksum and uploaded it with ftp.

I tried first to make a new calibration file with the name zcalib.rsc and hoped that it would override the original so that I wouldn't have to overwrite it, but it didn't work. Maybe it only works for cfg files, not for rsc files.

[mrflibble]

It works for .rsc files just fine. Traditional mistakes are either CRC wrong, or wrong file location. You probably know this, but just in case ... for .rsc files you have to use crc32 (not crc01 as for .cfg files).

Also for lazy people like me ... you can use the rdump command to save parts of the resource tree. See post about that not too long ago.

[john19]

It works for .rsc files just fine. Traditional mistakes are either CRC wrong, or wrong file location. You probably know this, but just in case ... for .rsc files you have to use crc32 (not crc01 as for .cfg files).

Also for lazy people like me ... you can use the rdump command to save parts of the resource tree. See post about that not too long ago.

I couldn't get it to work for calib.rsc. The crc32 was correct, because when I removed the old calib.rsc and renamed my new zcalib.rsc to calib.rsc it worked. Both were placed in FlashFS/system.

[john19]

I used trial and error and watched the curve change.

A better way would be to use Least Squares method or something to do this completely automatically. I'm sure excel could do it, but my excel programming skills are unfortunately not the best.

[tomas123]

I used an IR thermometer to plot the difference between it an the E4 in an excel sheet. Then I used this forumula:
http://u88.n24.queensu.ca/exiftool/forum/index.php/topic,4898.msg23972.html#msg23972
to calculate raw pixel values from E4 temperature values. From those pixel values ...

for information
you can simply grep RAW and convert values to temperature with IM convert

convert image to excel readable txt file

Code: [Select]

$ convert.exe 1.tiff -compress none 1.pgm

//open 1.pgm as text file
P2
160 120
65535
12894 12933 12933 12932 12983 12955 12919 12951 13023 13072 13130 13115 13143  || 13 values
13190 13372 13460 13484 13417 13374 13295 13211 13181 13101 13064 13030 13057
12996 12957 12953 12970 12970 12943 12959 12949 12956 12937 12950 12910 12916
12972 12996 13013 13084 13193 13251 13226 13109 13223 13280 13205 13132 13082

grep a single pixel 1x1 (see file above)

Code: [Select]

$ convert.exe 1.tiff -crop 1x1+39+0 -colorspace gray -format "%[mean]" info:
12972

$ convert.exe 1.tiff -crop 1x1+39+0 -colorspace gray -format "%[fx:mean*65535]" info:
12972

convert.exe 1.tiff -crop 1x1+0+0 -format "%[fx:round(65535*u.r)]" info:
12894

convert.exe 1.tiff -crop 1x1+13+0 -format "%[fx:round(65535*u.r)]" info:
13190

convert.exe 1.tiff -crop 1x1+26+0 -format "%[fx:round(65535*u.r)]" info:
12996

convert.exe 1.tiff -crop 1x1+39+0 -format "%[fx:round(65535*u.r)]" info:
12972

instead "1x1" use an area and replace "fx:mean*65535" with the planck formula from here
http://u88.n24.queensu.ca/exiftool/forum/index.php/topic,4898.msg23972.html#msg23972
and you get the real temperature from a fixed spot (need a tripod)