Flir E4 Thermal imaging camera review

[azinfrared]

New E series as a whole are not great units for most applications. 
They are super low res, no focus ring, no range and span control.   Focus and range/span control are basic functions that are needed for most applications. 

But they affordable to a lot of users who couldn't justify the cost of a high-end TIC, and are MUCH more useful than a spot IR thermometer.
Maybe less so for evaluating heat loss in buildings, but things like electrical faultfinding and condition checking can easily do without high resolution
 the E4 does have range lock, so you can compare images with the same scaling, but you can't manually set the range.
 

Low end cameras are just that, low end.  You will not produce imagery like the ones shown here with these systems.  They really should call these systems, limited IR.  They are IR but the details missed would render them useless to anyone doing IR professionally. 

The FLIR E40 is really the current entry level cam, IMO.

That is a ludicrous statement. Tool snobbery even. TICs have many different applications, and for some users, the E4 is more than adequate for a wide range of tasks.

There is a right tool for the job and a not right tool for the job, not sure how that is "tool snobbery".  I don't think you would be doing PQA with a Fluke 116.......does that make the person that owns a Fluke 435 a snob?  Sounds more like hating to me. 

You are correct with the range/span lock.  It does it.  It isn't just the one thing that makes it a bad cam. 

Look at the end of the day there are two things to consider here.  First, these cams exist because of politics, not necessarily because of the interest of the end user.  I invite you to look at the legal history of Fluke and FLIR.  These cams are a direct attack on Fluke from FLIR.  Second factor.  I am telling you to go buy a $4k camera.  I am just telling you within the FLIR line, you should either not buy a camera or buy an E40.

I am not sure what everyone's time is worth, but as a professional business the time you are wasting screwing around with your lock feature for range and span, someone with a cam that does have it will outperform you over time.  If you bill at $80 an hour it only takes 35 hours of your lock/hold range/span savings to be eroded.  Once that savings is gone, you are left with a lower resolution, lower temp, no span/range, no focus ring, etc camera, while someone that bought the E40 will have all kinds of time savings features and reporting ability. 

I am talking strictly to professionals here.  If you are buying a cam to look at something homemade or your wife naked, then the E4 will do what you want.....unless it goes over 250C.  However, in a professional atmosphere it is a waist of wages and time, in the long run.

[Fraser]

Cleaning your TIC lens....it can get ugly if you do it wrong !

http://www.thesnellgroup.com/Content/infrared-tip-ir-camera-lens-cleaning.aspx?utm_source=Knowledge+Briefs+from+The+Snell+Group&utm_campaign=55959c1605-IR_for_CM_July_2013&utm_medium=email&utm_term=0_a53635059c-55959c1605-&utm_source=Knowledge+Briefs+from+The+Snell+Group&utm_campaign=55959c1605-IR_for_CM_July_2013&utm_medium=email&utm_term=0_a53635059c-55959c1605-99218454

[mikeselectricstuff]

There is a right tool for the job and a not right tool for the job, not sure how that is "tool snobbery".

You seem to have a rather narrow view of what "the job" is.

I am just telling you within the FLIR line, you should either not buy a camera or buy an E40.

I am not sure what everyone's time is worth, but as a professional business the time you are wasting screwing around with your lock feature for range and span, someone with a cam that does have it will outperform you over time.

Again you seem to have a rather narrow view of possible applications.
If your job comprises finding shorts on PCBs or wiring, for example, you really don't care much about absolute readings, just quickly locating hotspots. An E4 is more than adequate for this job, and a more expensive cam would not save a significant amount of time in this application, taking into account the proportion of time for whit it gets used. 

If you bill at $80 an hour it only takes 35 hours of your lock/hold range/span savings to be eroded.

Assuming you care about absolute temperature

I am talking strictly to professionals here. 

There are plenty of professionals in many fields (outside of mainstream/traditional thermography) for whom a $4K cam would not be justfiable, but an $800 one would easily pay for itself.
The E4 is aimed as people for whom the choice is between no camera or a low-end one, not thise whose work would be speeded up signifcantly by having a higher end device.

[Corporate666]

There is a right tool for the job and a not right tool for the job, not sure how that is "tool snobbery".  I don't think you would be doing PQA with a Fluke 116.......does that make the person that owns a Fluke 435 a snob?  Sounds more like hating to me. 

You are right - there is a right tool for the job.  It displays an incredible ignorance of the field of thermal imaging to think that nothing less than an E40 is the right tool. 

Look at the end of the day there are two things to consider here.  First, these cams exist because of politics, not necessarily because of the interest of the end user.  I invite you to look at the legal history of Fluke and FLIR.  These cams are a direct attack on Fluke from FLIR.  Second factor.  I am telling you to go buy a $4k camera.  I am just telling you within the FLIR line, you should either not buy a camera or buy an E40.

A ludicrous statement.  Why should someone spend 400% more on a camera if an E4 will do what they need? 

I am not sure what everyone's time is worth, but as a professional business the time you are wasting screwing around with your lock feature for range and span, someone with a cam that does have it will outperform you over time.  If you bill at $80 an hour it only takes 35 hours of your lock/hold range/span savings to be eroded.  Once that savings is gone, you are left with a lower resolution, lower temp, no span/range, no focus ring, etc camera, while someone that bought the E40 will have all kinds of time savings features and reporting ability. 

I think your understanding of the world of thermal imaging is limited to people who buy these cameras to bill out per hour.  There are many many other uses for them besides that use case. 

I am talking strictly to professionals here.  If you are buying a cam to look at something homemade or your wife naked, then the E4 will do what you want.....unless it goes over 250C.  However, in a professional atmosphere it is a waist of wages and time, in the long run.

Nonsense.  Any businessman knows that you buy what you need, not what you would get "if money was no object".  Buying what you need allows you to buy more of what you need rather than having a smaller number of tools which have capabilities beyond what you need.  I am sure there are video forums where some folks think anything less than a $20k professional video camera are worthless.  As Mike said - that is called tool snobbery.

[WarSim]

New E series as a whole are not great units for most applications. 
They are super low res, no focus ring, no range and span control.   Focus and range/span control are basic functions that are needed for most applications. 

But they affordable to a lot of users who couldn't justify the cost of a high-end TIC, and are MUCH more useful than a spot IR thermometer.
Maybe less so for evaluating heat loss in buildings, but things like electrical faultfinding and condition checking can easily do without high resolution
 the E4 does have range lock, so you can compare images with the same scaling, but you can't manually set the range.
 

Low end cameras are just that, low end.  You will not produce imagery like the ones shown here with these systems.  They really should call these systems, limited IR.  They are IR but the details missed would render them useless to anyone doing IR professionally. 

The FLIR E40 is really the current entry level cam, IMO.

That is a ludicrous statement. Tool snobbery even. TICs have many different applications, and for some users, the E4 is more than adequate for a wide range of tasks.

There is a right tool for the job and a not right tool for the job, not sure how that is "tool snobbery".  I don't think you would be doing PQA with a Fluke 116.......does that make the person that owns a Fluke 435 a snob?  Sounds more like hating to me. 

You are correct with the range/span lock.  It does it.  It isn't just the one thing that makes it a bad cam. 

Look at the end of the day there are two things to consider here.  First, these cams exist because of politics, not necessarily because of the interest of the end user.  I invite you to look at the legal history of Fluke and FLIR.  These cams are a direct attack on Fluke from FLIR.  Second factor.  I am telling you to go buy a $4k camera.  I am just telling you within the FLIR line, you should either not buy a camera or buy an E40.

I am not sure what everyone's time is worth, but as a professional business the time you are wasting screwing around with your lock feature for range and span, someone with a cam that does have it will outperform you over time.  If you bill at $80 an hour it only takes 35 hours of your lock/hold range/span savings to be eroded.  Once that savings is gone, you are left with a lower resolution, lower temp, no span/range, no focus ring, etc camera, while someone that bought the E40 will have all kinds of time savings features and reporting ability. 

I am talking strictly to professionals here.  If you are buying a cam to look at something homemade or your wife naked, then the E4 will do what you want.....unless it goes over 250C.  However, in a professional atmosphere it is a waist of wages and time, in the long run.

Ok I have to reply now. 

Yes FLIR the original maker of thermal cameras has now decided to expand their market share into the consumer market.  I say why not be closer to the optics sources.

Yes the E40 is a great camera but not the only choice.  Initially I also thought the A series was my only option.  After allot of research and drawing on others personal experience I found out my requirements have been served by an E30.  When in manual I can differentiate temperatures of 12mil pitch (4mil trace 8mil isolation).  Some users don't need more.  The one thing I find I have compromised and the E40 users do too is a mounting method to hold they camera still enough to get any finer detail. 

I also paid near 4k for my E30 and I find it to be great.  I am personally following this tread to find out how well this less expensive option performs.  Frankly the sensor in the E30 I have is more than I need for DIY.  I use company cameras at work also, which are less capable than mine unfortunately. 

I do agree with you last statement to a degree.  The attitude of pay up or go home is not fair to "professionals" starting to use IR cameras.  This is a budget issue not a worthiness issue.  This thread has the potential to answer the capabilities questions so a valid cost effectiveness decision can be made.

The one thing that makes this camera a possibility is the short focal length.  Most others I have researched are more for building inspectors rather than electronics.  FLIR has identified near inspection market and I am pleased to see they have included this low cost option in this category. 

Anyone that wants an instant on IR Camera, you will pay allot more with drastically lower quality.  IR cameras use warm or cold detectors to measure IR differentials, then translates them to reference to ambient.  Therefore the sensor needs to change its temp to its relative temp with minimal thermal shock.  The faster turn on low differential sensors simply can not achieve the accuracy that the optics in these FLIR optics can.  With my E30 I can estimate current flow in a trace when set up correctly (no solder resist or conformal, haven't needed to try with coatings yet). 

[Psi]

i was trying to save up money but this E4 TIC is tempting me something chronic.
If mikes poking and prodding of the unit yields info that it really has full res/frame rate data coming out of the sensor then i really cant see how i can resist buying one.

[mzzj]

mzzj,

Yes, such has been a regular occurrence in my use of TIC's. You are correct that a surface coated mirror would be most efficient but remember that glass will reflect my body heat quite effectively as well.

I attach a picture that I just took looking at my mirror....just for you 

Yeah, You are right, the glass itself works as a mirror, backside coating has little to do with it I guess.
"Quite effective" in this case is reflectance of 15% or so. Not a very good mirror but plentiful  enough to see the difference on thermal camera.
85% emissivity gives you 2,4 degree error if your face is +35 degrees Cel and mirror (and the rest of the room) is at 20 degrees. So your mirror-image face should show a temp of 22,4 degrees... 

[mikeselectricstuff]

This is worrying :
Right at the start of theWinCE bootup messages :

dry ice initial status 0x103
dry ice status 0x3

From the Freescale mcimx257 datasheet :

DRYICE DryIce module Security DryIce provides volatile key storage for Point-of-Sale (POS) terminals, and a
trusted time source for Digital Rights Management (DRM) schemes. Several
tamper-detect circuits are also provided to support key erasure and time
invalidation in the event of tampering. Alarms and/or interrupts can also
assert if tampering is detected. DryIce also includes a Real Time clock (RTC)
that can be used in secure and non-secure applications.

However... the brief datasheet I have states that the 257 version of the processor used in the E4 does NOT include the "Security" module. You also have to wonder why there would be an external RTC...

Proceeding with caution.....

[Psi]

Is there complex math involved in processing the image thermal camera sensors?
Or is it relatively simple

If hacking it proves impossible, might be simple enough to solder a new MCU in and write some new code from scratch.

Assuming this protection isn't part of the sensor itself.

[mikeselectricstuff]

Is there complex math involved in processing the image thermal camera sensors?
Or is it relatively simple

If hacking it proves impossible, might be simple enough to solder a new MCU in and write some new code from scratch.

Assuming this protection isn't part of the sensor itself.

I suspect there are a number of proprietory processes between raw sensor data and nice images. At the very least range scaling,  reference level calibration, noise reduction, maybe some edge enhancement, emissivity correction etc. Maybe some geometry tweaking for a cheap lens.
I suspect the raw data would be useable visually, but converting to meaningful temperature could be somewhat involved. Definitely FPGA territory and not MCU.  However as the data timing is already a very close fit for VGA or an LCD panel, it won't be much work to just throw it on a panel to see what it looks like - scaling the 14 bit value to a colour range is easy enough.
As I already have some boards with  an FPGA and  LCD inerface on them, and there are only 3 pins to connect to the E4, I suspect this will happen at some point.....
There is enough space inside that you could even do an internal PCB that could switch the LCD between their  output and "yours", which could make for a rather interesting user-installable "upgrade kit", though you wouldn't easily be able to save high-res images without a lot more complexity.
 Maybe you could even extrapolate some cal data in real time from their image
Funnily enough the abovementioned FPGA boards do actually have two LCD connectors on them. Hmmmmm

[Fraser]

Oh Heck, the spectre of tamper protection and IDS in a product 

As most will know, Credit Card machines have incorporated such defences for many years. I work in a world where such measures are common but thankfully they had not entered the consumer product markets.

For the Hackers and curious amongst us..... imagine consumer grade equipment of the future incorporating tamper traps and IDS in the chip sets. Where removing a screw, opening part of the case, separating PCB's or trying to access a JTAG connector functionality causes a self destruct or disablement routine in the chipset 

Such defences already exist in specialist equipment that the manufacturer wishes to defend against tampering, modification or industrial espionage. It just costs a little more to incorporate into a design and is often made available via chipset features in the more advanced devices. It is hard to interrogate or hack a device if it goes 'poof!' as soon as you open its case or poke it with some exploratory signals/data.

FLIR's most significant sales are in the Military arena. It is not surprising that some of their chips or software may show signs of tamper detection and IDS. Hopefully not enabled on the cheaper consumer grade products.

It will be a sad day when manufacturers use advanced techniques to prevent you even opening an equipment's case. Imagine the iPad....you would not be able to even replace the battery without destroying the OS or chipset ! The manufacturer may say that it is to protect the consumer or their IPR but I am not sure of the legality in causing someone's product to self destruct if the case is opened.

[mikeselectricstuff]

Oh Heck, the spectre of tamper protection and IDS in a product 

If manufacturers really wanted to lock down equipment, there are numerous ways they could do it cheaply and effectively.
In practice most don't give it much consideration. Video game consoles and credit card terminal are the only mass-market items I can think of that have any serious attempt at security ( and even those have been mostly broken, though the skills of some of the breakers are breathtaking!)
   
In the particular case of tamper detection and reaction, the risks of accidental deployment must be carefully considered - an unintended shutdown can cause way more damage to business and confidence than losses due to hacking.
There are also significant issues throughout the whole production and support logistics process- debug and test functionality often provides a way in, but omitting it can lead to more trouble.
Security needs to be designed in from the ground up, and for applications whose core function is not security-focusses, this can easily be neglected at the specification stage, and it is hard to add later, especially as release deadlines approach. Adding a 3rd party OS to the mix only makes things worse.
 
Considering the ITAR issues, I'm surprised that a company selling a product containing a custom made 60FPS 320x240 TIC imager chip with onboard ADC didn't appear to think add even a token amount of encryption or obfucation to a simple binary datastream. Oops - did I give something away there   

[PA0PBZ]

Considering the ITAR issues, I'm surprised that a company selling a product containing a custom made 60FPS 320x240 TIC imager chip with onboard ADC didn't appear to think add even a token amount of encryption or obfucation to a simple binary datastream.

I like the book and can't wait for the movie 

[Fraser]

Oh Mike, you big tease 

I am seriously wondering about this whole ITAR issue. It may be a bit like the reverse SMA connectors fitted to Wi-Fi, a deterrent to misuse but by no means an effective measure against such by skilled persons.

[SeanB]

Itar is not really effective any more, as the technologies are no longer made exclusively in the USA, and the other players are getting better every year.

[mikeselectricstuff]

Oh Mike, you big tease 

I am seriously wondering about this whole ITAR issue. It may be a bit like the reverse SMA connectors fitted to Wi-Fi, a deterrent to misuse but by no means an effective measure against such by skilled persons.

I suspect there is quite a lot of interpretation involved, especially when it comes to component parts and what might reasonably be expected to be modified. Sort of like when does a lump of metal become a gun....

[mikeselectricstuff]

Itar is not really effective any more, as the technologies are no longer made exclusively in the USA, and the other players are getting better every year.

But the people in charge of this stuff move at glacial pace. I think there are still issues with anything that contains encryption, like Bluetooth modules.

[mikeselectricstuff]

Considering the ITAR issues, I'm surprised that a company selling a product containing a custom made 60FPS 320x240 TIC imager chip with onboard ADC didn't appear to think add even a token amount of encryption or obfucation to a simple binary datastream.

I like the book and can't wait for the movie 

I'm taking things slowly and carefully.... will be in the next couple of days. 

[mrflibble]

i was trying to save up money but this E4 TIC is tempting me something chronic.
If mikes poking and prodding of the unit yields info that it really has full res/frame rate data coming out of the sensor then i really cant see how i can resist buying one.

Indeed. Having a low end thermal imager in the toolbox has been on the wish list for a long time now. The < $1000 price point already was pretty tempting. And now Mike teases us with 320x240 @60fps unencrypted bit streams. How the hell am I supposed to hang on to that tax refund now?!?    

@Mike:
Have you checked what happens to the bitstream if you place a horizontal/vertical grating of hot/cold/hot/cold in front of it? Any totally obvious patterns to be seen in the bitstream?

[mikeselectricstuff]

i was trying to save up money but this E4 TIC is tempting me something chronic.
If mikes poking and prodding of the unit yields info that it really has full res/frame rate data coming out of the sensor then i really cant see how i can resist buying one.

Indeed. Having a low end thermal imager in the toolbox has been on the wish list for a long time now. The < $1000 price point already was pretty tempting. And now Mike teases us with 320x240 @60fps unencrypted bit streams. How the hell am I supposed to hang on to that tax refund now?!?    

@Mike:
Have you checked what happens to the bitstream if you place a horizontal/vertical grating of hot/cold/hot/cold in front of it? Any totally obvious patterns to be seen in the bitstream?

It's way, way simpler than that. All you need is a slowly moving hot soldering iron, a can of feezer and a scope.  Imagine the simplest way you could think of to encode a stream of 16 bit words on a clock+data pair. The only complication is there are 2 pins each outputting 160 words per line - I'm currently hooking up a more mechanically stable probing system to see both streams at once to see how they are are interleaved.
Bear in mind that this is raw data - at the very least you'd need to store the reference image during shutter calibration, subtract it out of the live frames and apply a scaling factor.

[Hypernova]

We can just about read the sensor like some cheap ass 3 wire serial ADCs?! 

I received my pay check just this Thursday and I'm starting to feel the burn already.

[mrflibble]

Imagine the simplest way you could think of to encode a stream of 16 bit words on a clock+data pair. The only complication is there are 2 pins each outputting 160 words per line - I'm currently hooking up a more mechanically stable probing system to see both streams at once to see how they are are interleaved.

Damn, that sounds criminally simple. So basically 2 internal ADC's that spit out 16-bit words in SPI fashion? Even taking into account the non-trivial work you'd have to do to get an imagine from the raw data, this all sounds very interesting.

I received my pay check just this Thursday and I'm starting to feel the burn already.

Better get that E4 to properly image the burn.

[mikeselectricstuff]

We can just about read the sensor like some cheap ass 3 wire serial ADCs?! 

Not quite cheap-ass - it does come out at 75mbits/sec

[PA0PBZ]

Not quite cheap-ass - it does come out at 75mbits/sec

Makes sense: 60*320*240*16= 73.7 Mbps

[mrflibble]

Not quite cheap-ass - it does come out at 75mbits/sec

Interleaved over 2 channels at half that rate, right? Incidentally, what IO standard is it?