cooled/more expensive thermal imager and PCBs

[coppercone2]

I have a seek thermal. I know you magnify it (its 320 resolution one).

But I am curious about possible benefits that lower noise (actively cooled?) cameras might have in the lab.

How about seeing small currents in traces?

How much better can you do on a budget in the multiple thousands range? Do you get capabilities?

It feels like its good right now, but what do you get from better?

Is there any improvements other then higher resolutions? Like dynamic range/sensitivity? Like if I could double the resolution of the camera I have, I am not sure if it would really give me too many extra capabilities based on the image that it produces. what can be improved to make it more useful for PCB / small chassis work?

[theodoremoreau]

Hey there! So, you've got a Seek Thermal camera, huh? Not bad!

Ever struggled to spot tiny currents in traces on your PCBs? Well, a lower-noise camera can make those sneaky details pop out clearer. It's like turning up the brightness on a dimly lit room – suddenly, everything becomes easier to see.

Now, when it comes to price, you're looking at a few grand for the really good stuff. But trust me, it's not just about the dollar signs. With those pricier cameras, you're getting a whole bunch of capabilities packed in. Think better sensitivity, dynamic range, and even some cool features like advanced temperature measurement tools.

Sure, your current camera might do the job okay, but picture this: sharper images, finer details, and more accurate temperature readings.

And hey, it's not all about resolution. Sure, doubling the resolution sounds nice, but there's more to it than just pixels. We're talking about improvements in dynamic range, sensitivity, and overall image quality. These are the things that really make a difference when you're diving deep into PCB and small chassis work. You want every detail to stand out, every temperature variation to be crystal clear – and that's exactly what those high-end cameras deliver.

So, yeah, upgrading to a lower-noise, higher-quality thermal camera might seem like a big leap, but trust me, it's worth it. 

[JOE 2345]

As I know cool thermal core camera is more expensive but more accurate and the image is more clear, but the price is more expensive.

[Bill W]

As generalisations:

Actively cooled cameras are usually those for the 3-5µm band, so the benefits from cooling is mostly used to compensate for the lower signal from ambient objects and the detector mechanism itself.
On ambient scenes, maybe 4x better than a decent 8-14um camera ?
With the higher system cost you will get better uniformity, lenses and so on to match the ticket price.

Whether just going just one step up from Seek / China to 'professional' 8-14µm cameras is worth it depends on your use.
You swap the magnifier for a focusable lens with maybe scope for an extension tube
You get more control
You should expect get better uniformity and lower noise.
It won't be a 'night and day' difference.

I doubt small currents would show up well, tracks don't get that hot because they are very low resistance over a conducting ground plane, unless the PCB designer was particularly incompetent.  I have only ever seen 'hot' tracks due to conduction out from devices that were burning up.

[Fraser]

As a side note to readers “Theodoremoreau” has been identified by GNIF as a “Bot” and banned. Damned AI invading forums !

https://www.eevblog.com/forum/beginners/issue-when-printing-a-pdf/

Fraser

[Fraser]

As the owner of cooled thermal imaging cameras I can advise that a decent modern microbolometer based camera normally provides all that most general users would require of a thermal camera. The cooled cameras tend to appear in more specialist scenarios found in military, scientific or specialist roles. My crazy expensive FLIR SC4000 cooled camera was needed for imaging through ceramic and glass to profile heating elements. In such a scenario a LWIR microbolometer camera was not useful. The original purchaser was forced to buy a cooled camera with extended wavelength coverage (SWIR & MWIR) in order to meet their R&D needs. That specialist need cost them £150K !

Cooled cameras have a lot going against their use in the modern world of thermography….they are normally very expensive, heavy and can have a limited operational life dictated by the cooler specification. An expensive asset that can have a high cost of ownership.

With regard to imaging performance…. Again we move into the realm of specialist uses….. the modern cooled sensor FPA thermal camera presents beautiful low noise images in great clarity. This can be due to the low noise thermal detector system and the high quality of lens often found on such expensive cameras. The low noise imagery has none of the column ‘noise’ and image artefacts found on most microbolometer thermal imaging systems but modern digital image processing techniques have greatly improved many microbolometer based cameras in this respect. When it comes to thermal sensitivity, there can be some confusion as cooled thermal imaging cameras can have an NETD of 18mK and provide amazing performance in terms of thermal sensitivity, yet some microbolometer based cameras are now claiming 20mK from an uncooled, normally quite noisy, microbolometer FPA. Are these equal in performance to a professional cooled thermal imaging camera in terms of the imaging performance and thermal sensitivity ? The answer is a simple no. It is a bit much to expect a sub $2000 to perform as well as a camera costing more than many luxury saloon cars ! I advise caution where manufacturers NETD figures are concerned. MRTD would be a better metric for many users in my opinion. Even then I suspect manufacturers marketing departments would skew the figures.

Finally, would I recommend buying the cooled thermal imaging camera over a decent microbolometer based camera ? In most cases, no. It is like owning a top of the range Ferrari …. Would you take it shopping or use a less sophisticated and more practical generic car ? Common sense supports using the more practical, lower cost, option. It is the same with cooled cameras. I love their performance but many microbolometer based cameras are far easier to operate, far quicker to use and get the job done easily. It is the more specialist roles that justify breaking out a cooled camera. There were cooled LWIR cameras in the past but I suspect the need for those has greatly diminished in the presence of better professional LWIR microbolometer based cameras.

I hope this helps readers to relax about using uncooled microbolometer based technology as it has much to offer. If you are a specialist user needing MWIR for a task, you have my sympathy as you have little choice but to dig deep into your pockets to fund a cooled MWIR camera and all that ownership of such involves

I love my cooled thermal cameras but I am a collector so do not take my ownership of such as a “need” for this specialist technology 

Fraser

[Fraser]

I forgot to say, if anyone wants to see how useful a simple microbolometer based thermal camera can be in electronics repair, take a look at the YouTube channels of NorthridgeFIX (Alex) and Electronics Repair School (Sorin) as they both use thermal imaging to track down shorted components and heat dissipation issues on PCB’s. In these scenarios they inject high currents of between 1A and 5A to generate heat in the failed components and PCB tracks leading to them. You need decent current flow to create enough thermal output from a PCB track to see it clearly with a common microbolometer based camera that has a stated NETD of around 60mK. You would not see normal low currents in logic PCB’s for instance but you might see PCB track heating in some high power driver circuits that are passing several Amperes through copper tracks. That said, if correct cross section copper tracks are used then they should not get  very hot ! A decent modern cooled camera would identify heating in a copper PCB track before it could be identified by a microbolometer based camera but in the case of tracing shorts, you just use a little more current and that is a lot cheaper and easier than using a cooled camera !

Fraser

[Fraser]

An interesting article on cooled MWIR vs uncooled LWIR thermal camera technology in the security industry.

Note the comments on “thermal contrast” at MWIR wavelengths vs LWIR (Blackbody theory) and also the ability to change the integration time on cooled cameras to improve performance. A Cooled MWIR camera definitely has sensitivity and contrast advantages over the LWIR microbolometer, but it really only make a significant difference to a user who has more specialist needs, as I have already stated. Long range thermal imaging is a classic case where cooled thermal imaging cameras are king due to their excellent thermal sensitivity and resulting sensible lens dimensions where long range lenses are needed.

https://www.sourcesecurity.com/insights/co-2752-ga.2246.html

[coppercone2]

Damn I guess maybe for a commercial lab with a decent budget that's maybe a good move (just to improve quality and where time = money) but I am not that enticed to aspire for an upgrade...

[coppercone2]

also I love how the forums literally turn into alien now. This model seems to enjoy microscopy.

[coppercone2]

but the way that post was written I almost feel like we are attacking the holo doctor on voyager. It just talks about cameras. Not too sure of a evil agenda yet.

Oddly enough, unlike Prometheus, we don't have a problem with the AI's serving us food!



However I do feel I should perhaps be paid to deal with AI trying to interact with me, which probobly brings profit to some mega corporation. I bet its their god damn R&D arm making these imposters. CHATGPT black site.

[Fraser]

I found the inclusion of “Hey there”, “Sure”, “And hey” and “So yeah” interesting in that post. These are words that mimic some human vocabulary commonly found in forums. Not slang, but very human in nature. Not precise and concise speech that might be expected from a logic driven computer system. For me this is the unacceptable “face” of AI. It is mimicking human communication traits closely to pass itself off as human and that could be seen as deception. When I am hearing from AI, I want to know that it is a computer talking to me and not another human. There are situations where accurate human mimicking is likely more appropriate, but AI posting to a forum is not one of those cases. I also wonder whether the AI is just collecting and collating questions and associated answers from forums like this and then regurgitating responses to questions by stringing together previous replies to questions, but in an ‘intelligent’ AI manner to mimic a human. I am not OK with what I am seeing in the development of AI where it directly interacts with the general population and I can see problems in the future. Humans can sometimes move too fast with technological developments to the point that they do not fully understand what they have created or the implications of such. AI is an area of science that is now fashionable and I only hope some wise minds keep a tight leash on how AI is allowed to “invade” daily life for humans. It has much to offer if used wisely.

[coppercone2]

I think they are trying to make it feel natural sits more marketable and makes people feel relaxed but in just makes me paranoid because its going to end up tricking lazy people into doing something stupid when its running amok

regardless its a unauthorized experiment. you are supposed to be paid to participate in such things and be informed. I think the legal system needs to guarantee privacy from machines some how. But its even worse because its not just eavesdropping its interacting.

Why should we be inconvenienced to teach some corporate robot for free? its already disrupting the shit out of this thread.. it is stealing leisure time for corporate ends.

Google spiders were kinda paid off because its free to use the database they make. this one ain't free, even if you treat it as a advanced search engine. mother fucker is wasting our time and wants us to subscribe to a monthly service to benefits. It just used to slightly hang up another machine to do computations with minimal impact (some traffic).. now its making you THINK for their profit.

I guess someone thinks that we should be forced to it because it will 'trickle down' to us. Where have i heard this shit before?

It's clearly some kind of research program.


This needs to be cracked down like auto dialers to try to get surveys out of you and all that other scammy shit. The closest analog there is to a rogue chatbot is a autodialer that tries to get you to answer survey questions or provide analytics based on your telephone answers, because it requires direct interaction and effort.


I kinda wanna throw a rock through the owners window at 3am so I can gather human data patterns on response to stimulus for some esoteric research of mine

there is no way to see it as not being a psychological experiment of some sort too.

How about the government confiscates the servers generating this crap and pays off the public by reconfiguring the machine do peoples taxes for free next year?






BTW I think it operates like Jen translating Italian in the IT Crowd.


its a true neural network response based on 0 understanding and pure context

[jpanhalt]

@Fraser

I suspect it was you that reported the AI.  On my first read, it seemed like sarcasm with such seemingly over the top fake enthusiasm.

[Fraser]

No, not I. It was GNIF who spotted it and had the bot banned.