Equipment Thermal Imaging - The Want is Strong

So, I do thermal imaging work 'cos readers like purty colours and sometimes it actually conveys useful information too.

Tweet

— Twitter API (@user) date

I use a Flir C2 for the thermal imaging, strapped to a modified mobile phone tripod adapter on a copy stand. The thermal data is then processed and overlaid on an edge-enhanced greyscale visible light image to get what you see above.

It works, but the thermal resolution leaves a lot to be desired: it's 80x60, or 4,800 individual readings. Enough to get a rough idea of what's going on, but you're not going to see individual resistors glowing happily in the dark.

Trouble is, there's a big-ol' gap between the entry-level Flir stuff and the pro-level Flir stuff. Most obvious upgrade route would be the Flir One Pro (non-LT), which has a 160x120 thermal sensor for 19,200 individual readings - exactly four times the resolution of what I have now. Trouble is, it is an accessory device for a smartphone - which I've always wanted to avoid, 'cos why complicate things unnecessarily?

Then I found the Flir ETS320.



It's basically what I've built myself, except with a screen that means you don't have to be hovering over the top of the thing. Captures images in the same format as my current camera, too, which means I can continue to use the same post-processing workflow - something I'd have to give up if I went with a cheaper brand like Seek (notorious for noisy sensors.)

320x240 thermal resolution, 76,800 individual readings. 11.5 times higher resolution than my C2. Absolute beast. No visible-light camera, so it doesn't do in-body MSX blending like the C2 or Flir One Pro would, but I don't use that anyway 'cos the quality's pants.

There's only one real downside... price. It's £2,748. There's a dude on FleaBay in Belgium selling a suspicious amount of test gear, has an as-new used-once one for £1,640 delivered - but a whopping two feedback (for a fibre splicer and a Fluke energy monitor, both £2k a throw).

It's all academic at the moment anyway, because I'm completely and utterly brassic - but am I mad to be considering the upgrade? My clients like the thermal stuff I do now, but d'you think I'd be able to charge more (or get more thermal work) to earn back the cost of the ETS320?

Pipedream or plan? You decide!

 

Unlikely with some of the deadlines we've got recently. Getting products with 4 days before the review

 

It's becoming increasingly common since the knockoffs hit the market and the entry cost fell through the floor. They're not usually... fantastic, though. This is one from a certain B2B electronics giant:



Same camera as mine, funnily enough.

This one came from a colleague for a web outlet:



Some flavour of Seek, that one - probably one of the ones that dangles off your smartphone.

Then Pimoroni's started playing, too:

Tweet

— Twitter API (@user) date

That's a Seek again, but apparently you can't lock the thermal scale - which is why it looks like the USB controller has got *hotter* in the second image, even though it's actually considerably *cooler*.

Mine are better, though.

"How long is a piece of string?" Having better-quality images for the mags won't help, 'cos I get page rate regardless. It might mean I can sell more to broader outlets, though - and I could always have a try at offering thermal imaging analysis as an outsourced service to businesses (and other journos, for that matter.)

That'd certainly make more sense than shipping a bunch of motherboards or what have you to me!

 

I'm about to do *ten* thermal images, and this would have been super handy. Sadly, I'm not getting paid anywhere close to enough to justify it. Boo-urns. Back to the document stand, hacked-about smartphone tripod adapter, and the ol' faithful FLIR C2 it is!

 

Eh, not so much. If you're spending £3k on a piece of lab equipment, you're usually A) not spending your own money, and 2) need a VAT invoice, certificate of conformity, and calibration certification. Saving £1k (or whatever) by buying second-hand sounds good, but when it's not your money and you don't get any of the latter - nor any real guarantee it's not been ridden hard and put away wet - it's a tougher sell.

There's literally only one showing in Sold Listings on FleaBay, and that was sold in the US for just shy of £1,500 - which would leave me in the hole for £1,300-ish if I flipped it, and that's even assuming I was lucky enough to find a buyer.

There's a BNIB one from Canada pretty (relatively) cheap buy-it-now, which is tempting, but it wouldn't arrive in time for this particular project - the deadline's the 12th.

 

I mean, it's not like I can't get anything decent out of what I'm using now...





That's the Raspberry Pi 4 at idle and load, using the latest firmware.

 

There's a firmware coming out that makes things even beterrer:




Bear in mind, this is what it looked like with the launch firmware installed:

 

What, the FLIR C2? Nah, hasn't had a firmware update for about four years. Doesn't really need one, though: it spits out low-res JPEGs with embedded thermal data, I run 'em through a script to interpolate that data into a bigger PNG with a scale of my choosing (while also compensating for environmental and reflective temperature, thermal emmisivity, and humidity, 'cos those are the kinds of stupid things you have to do if you want actual accurate readings instead of just pretty colours), then it goes into The GIMP to be overlaid on a visible-light image of the target device.

 

They do - it looks like this:



They're not designed for close-up work: the absolute smallest distance you can align the "multi-spectral imaging (MSX)" at is a metre. Less than that, and you get the above: the IR image being pushed above the visible-light image. Couple it with the fact the visible light image is a 640x480 webcam shot, and... yeah. There's a reason I do it by hand.

The above £3k actually-designed-for-what-I'm-doing camera? No visible light capture at all, so no MSX.

I just skipped through some of that, and I'm already annoyed. You can't do what they're doing: the surfaces they're measuring have completely different thermal emmisivity. The "data" they're getting from that camera - several models above mine - are junk.

Here's what I mean, borrowed from FLIR itself:



See that "hot" spot? That's Kapton tape on a shiny metal can filled with hot water. It's the exact same temperature as the can it's stuck to, but the tape has a high thermal emmisivity and the can a really low thermal emmisivity. The result: it looks like hot tape on a cold can, even though that's absolutely not what it is.

That's what they're showing in the video: there's no way of knowing (aside from common sense, like "yeah, the brakes they were just using to brake with are probably hot") whether you're seeing a real heat difference or a difference in thermal emmisivity.

 

Oh, it's a total pain. If you just want pretty colours, it's point and shoot. If not, you need to compensate for thermal emmisivity, environmental temperature, reflected temperature, and even humidity. Massive ball ache, but it's the only way!

 

Had an email from a PR this morning saying they're trying to get FLIR some column inches, and was I interested in reviewing anything?

"Funny you should ask..."

Not too sure FLIR's going to fancy putting a £3k camera on long-term loan, but you don't ask and all that!

 

Handy for finding grow-ops, too...