Thermal camera recommendation

[blueskull]

Well, I never thought I will ask such questions like "what is the best xxx for $yyy if I do zzz with it", but seems I have to ask one here.
Have been looking for a small, low cost thermal camera for a while, but still have not made the decision yet.
Here are my requirements and using cases:

a. Low cost, <$500,
b. Can focus on small objects since I will use it to find hot spots on my PCB, down to 0402 size,
c. Preferably smart phone or PC connected.

And here are my candidates:

a. Flir One Gen 2,
b. Seek,
c. Seek XR,
d. Seek with macro lens mod,
e. (Over budget, but acceptable) hacked E4,
f. You name it.

There are some tests and reviews out there, but I never found any of them focusing particularly at macro mode.
For me, seeing true hot spot (not the MSX interpolated fake spots) down to at least QFN 2*2 is crucial, down to 0402 is better.
I do not care about accuracy or battery life, since I will only use it occasionally for diagnosing only, not final reporting.
If possible, I want to avoid E4 since it is too expensive for me, and even if I need high quality images, I have access to a T640.

So, any suggestions?

[miguelvp]

Seek software is very, how do I put it, green.

But if you can get the SDK it has the highest range but that means a lot of work. I have the PC and Android SDKs that they released to signed in beta users. But things got in the way so I'm not playing much with it.

Flir has a lot of tools and a better SDK for what I hear since I have no direct experience with the tools,  but a lot of members like thomas and others have been doing amazing things. Less range but for electronics it might not matter. Not sure on how they can focus either for close ups.

I had an original Seek Compact and a lens without adapter but now the XR has a narrow field of view. The output is still noisy and not a lot of progress has been done lately but it does have potential. Flir will be less of a headache I think, but still not sure if it can be focused for electronics use, but I guess you can get a cheap IR lens and mod it.

So if you can do image processing (edit: and you can get a hold of the PC or Android SDK), I'll say the Seek XR, if you can't or even if you do, you might improve on the Flir if they have a mod that allows you to focus on small parts.

So at the end, I'm no help because I don't know the capabilities outside of the Seek original Compact and the new XR and I didn't have time lately to see how far they can be pushed.

[frenky]

I have experience only with the first model of SeekThermal.

With 50mm ZnSe lens and with this PC software: https://www.eevblog.com/forum/chat/seekthermal-how-to-correct-the-image-received-from-the-sensor/msg696895/#msg696895
You could get images like this:
(The height of the board is aprox 3cm)


But if the temp range is smaller the images are not that clear:


But on the other hand when you are searching for fault on electronic board, temp range is usually quite big.

If I was in your shoes...

1. i3 ThemalExpert http://www.i3-thermalexpert.com/

Wait a few days to see the review of i3 ThermalExpert from schlafli:
https://www.eevblog.com/forum/thermal-imaging/another-smartphone-thermal-imager-from-i3/50/
If the verdict is positive then you could get smartphone module with resolution 384 x 288 for 820$.

2. Flir E4
This one has been proved to be really great value for the money (after you hack it).
I think you can even get a big dicount on some online stores.

3. Flir One V2
Tomas123 and cynfab are making a huge progress on the software side. See here:
https://www.eevblog.com/forum/thermal-imaging/question-about-flir-one-for-android/125/

Images that they are getting out of lepton 3 are looking A LOT nicer than images from SeekThermal...

[Fraser]

The solution to industrial close up thermography has commonly been the addition of different auxiliary close up lenses. Inframetrics used to produce several focal distances to provide various levels of detail and magnification. My AGA 680 thermal microscope was amazing. It had a dedicated X15 close focus microscope lens assembly. Amazing quality but horrendously expensive. The cheap ZNSE CO2 laser lenses are a good option for use with most thermal cameras. The proper Germanium Close up lenses are great. I have the 6" and 12" versions that can also be attached to my X3 thermal telescope lens. They cost thousands of $ new and around $150 on the secondary market.

I recommend you find a thermal camera that has excellent image quality with low noise content. Then add an auxiliary close up lens if needed. Sadly the SEEK camera would not be my first choice for such when used with the OEM provided software.

You can work out the minimum resolvable target size from the specs of the microbolometer and lens FOV. 320 X 240 pixel resolution would be my minimum for the task that you are suggesting. That's easy for me to say though as I have suitable cameras already.

Fraser

[Fraser]

The F1G2 is at the minimum resolution that I would ever use for any PCB work. I own one and cannot recommend it for your application. The permanent MSX is annoying too.

It is great for its intended purpose though.

Fraser

[frenky]

Another option is Therm-App from Opgal for 960$. http://therm-app.com/product/therm-app-device-with-19mm-lens/

It has the same resolution as i3.
There are lots of photos here: https://www.flickr.com/groups/therm-app-users/pool/

[Bill W]

There's also second hand cameras where you'd get a good sensor / core.  At that level you would also expect a good and more importantly focus-able lens so avoiding secondary lenses.
A common source might be the firefighter cameras as they come up on eBay a bit.

Attached image from an e2v Argus4

Bill

[Fraser]

Just be aware that firefighter cameras are normally fixed focus just like the FLIR E4 and they are normally fully automatic. That is to say, you turn them on and they set themselves up to find people in smoke or fires behind walls. Manual control is not normally offered as that creates the possibility of incorrect configuration in use.

Fire fighting cameras normally need the addition of a close up lens for PCB work.

I own a selection of firefighter cameras and I am conversant with the E2V offerings, past and present. Please note that many more recent firefighter cameras are also ITAR liable, with all that entails. Do not expect to find any deep technical info or schematics for the core.

Fraser

[mikeselectricstuff]

Just be aware that firefighter cameras are normally fixed focus just like the FLIR E4 and they are normally fully automatic. That is to say, you turn them on and they set themselves up to find people in smoke or fires behind walls. Manual control is not normally offered as that creates the possibility of incorrect configuration in use.

In particular they automatically level-adjust, so you can't typically get absolute temperature info. It may have a spot temp indication but this will be oriented towards fire appliucations, so any temp readings at more normal temps may be significantly inaccurate.
Due to robustness & sealing requirements, it's less likely that you'd be able to do things like tweak the lens adjustment for close-up focussing. 

[Bill W]

Fraser,

They are only fixed focus until you undo the locking screw and engage pliers    The image was taken with 100% standard parts.
As soon as you do that then the spot temperature accuracy will get worse but not as bad as adding lenses.  Typical might be +/-2°C as long as lens and sensor are at equilibrium, the fire camera specs will be padded out for reality over the environmental range.

As far as absolute temperature goes, you won't get pretty colour=temperature grades but some have adjustable thresholds or their still images can be analysed either on their own software or through the likes of ImageJ. 

Probably the most plentiful supply of imagers for enthusiasts though ?

regards
Bill

[miguelvp]

Their app is pretty green, unless you have a high contrast on temperatures it's a bit washed off. (and a bit noisy)

But the sensor is capable, just the software that is not there yet.

And I do like Seek, I'm on my 2nd one, just because of the potential.

[miguelvp]

I know, I did one of those versions

Although I did look at Thomas code and he did pretty much added a lot to it. But it originated with sgstairs to begin with.

[-jeffB]

A little problem compared to a FLIR, that it does not have MSX, so IDing cold parts could be hard.

You might be disappointed with MSX for your application anyhow. At close distances, parallax is a big problem, and if you've got (for example) parallel rows of small components, it can easily make things MORE confusing.

I don't use my SeeK much since I got a FLIR E4, but even with the E4 I usually run in straight thermal mode (MSX turned off). I find it's more useful to point into the scene with something at a different temperature -- a finger, or a pen that's been in my pocket -- and use that to make sure I know what I'm seeing on the screen.

[bookaboo]

MSX is highly over rated once you have decent resolution and focus. I virtually never use MSX for anything and I use my camera a lot, it might be handy if you had some thing behind a  large flat surface like a wall and needed to place markers. PCB work would be about the last thing it could ever be useful for.

Although its over your budget I can tell you that a Flir E4 (modified) with one of the lenses from Frasers "Useful information thread" outperforms an E60 for macro work.