Why are even Chinese thermal cams not a lot cheaper than FLIRs?

[Ben321]

I already have a Seek Compact Pro fast frame that is 320x240 at approximately 15fps. 15fps is definitely better than the 9Hz of the FLIR One, and 320x240 is an absolutely huge improvement. So I'm now looking for my next thermal cam, at 640x480 resolution, but not from FLIR. Their absolute dirt cheapest 640x480 class of thermal cam is about $3000, way above what I'm going to pay. For that I've turned to try to find some 640x480 thermal cams from Chinese companies. But there's 2 problems. First problem is, most of these Chinese companies who's websites I can find do not show prices, nor even name distributors. So to find prices I find myself Googling the exact make and model of camera, along with the word "price" in the search string, for each model of camera from that manufacturer. Eventually I get to some online stores on websites who's names I've never heard of, selling the cameras, and show the price of the devices. Unfortunately though, this is the second problem. The price for 640x480 thermal cams, even from Chinese companies, are bottoming out at just under $3000.

What makes these 640x480 thermal cams so expensive, regardless of what company is making them?
Is it lack of competition? Do they just not feel the need to make the price any less than slightly below the next cheapest competitor?
Is it the price of manufacture? Like to pay the electric bills for running the factories due to the special equipment used to make these cameras (as well as employees who operate the factories), do the factories cost more to run than the typical visible light camera factory?
Is it the price of the raw materials? Are the special materials used in thermal imaging cameras like vanadium oxide and germanium, simply that much more expensive for the manufacturers to acquire, than it would be for them to acquire the raw materials needed to make a visible light camera?

So are they making actually a very narrow profit margin as it is already, and $3000 per camera is simply the actual monetary worth/value of each camera? Or are they making obscenely huge profit margins, and due to lack of competition they simply see no reason to lower the price more, so they keep it right where it is (and thus continue to make obscenely huge profit margins)?

[Ben321]

You're probably right with respect to the factors.

* Limited competition for market share

* Limited competition based on price

* Limited market demand; currently sales are probably 99% to business / government relatively few to consumer uses.  Generally they're used in industry for some test / analysis / diagnostic etc. purpose and if a solution is needed to solve that problem, it doesn't really matter so much if the price was $100 or $100,000; either way if there's a business need to do it it'll make its way into some annual budget and get amortized / whatever over N years if it costs more than a few thousand.  So even if it was only $500 most of the time they'd still have to get purchase approvals and go through the same process to justify / finance it, so $5000 not much difference.
Also most business / industrial applications will probably buy / lease something and expect to keep it in service
at least 5-10 years so there's steady demand but not a huge growth in demand of volume for upgrades.

* Limited technology overlap -- the sensors are custom made down to the process technology, packaging, testing, etc. as far as I know.  It isn't just like some CMOS image sensor with the RGB microfilters replaced by an LWIR transparent layer, it's a totally different sensor technology.  Basically some "fab" factory that makes this kind of stuff and not general purpose ICs so limited "free" Moore's law scaling and ability to share large high volume wafer fabs and such.  Now if someone DID come up with a way to make these work on a high volume commodity CMOS process you'd see them in $10 webcam toys in walmart in a year, but for maybe hundreds to thousands of dollars or more in sensor / packaging / testing cost for the imager core alone there's not a lot of volume demand to make more than thousands per year of high end ones and maybe 10x-100x of that for mid-range & low end.

* Look at CCDs vs. CMOS image sensors -- back in the 1980-2000 time frame most digital video cameras or still cameras well industrial or commercial, usually several hundreds of dollars for even low mid-end VGA/SVGA/PAL/NTSC kinds of units, and several thousand dollar units most common for any moderately serious industrial / videography / scientific sort of camera.  Only a hand full of companies making decent CCD sensors themselves and the sensors and chipset alone often cost hundreds to thousands of dollars.  Most consumer market stuff was using film or vidicon or low end CCD.   Then they started to come out with tolerably good VGA / 2MP / 4MP etc. CMOS image sensors good enough for cheap consumer / small business CCTV cameras, toys, entertainment cameras, and over a few years the CCD market for anything much lower than the higher end industrial sensors vanished and now there are sub $1 VGA CMOS image sensors from the likes of OmniVision, usually two 2MP or more sensors in every one of the several millions of mobile phones / tablets / laptops made every year, and even $5-$10 webcam toys from Walmart etc.   And still today if you want a half decent CCD you'll probably still spend thousands of dollars for your industrial / scientific imager.

* Look at the market share / use of cell phone cameras for mass consumer use vs. people that own and regularly use a semi decent consumer level dedicated video camera or DSLR with proper lenses, much better / larger sensors, etc.  A $50 disposable cell phone gets you a camera good enough for a well lit selfie, and a halfway decent consumer mid-entry level SLR is still more like $1000, more like $2000-$3000 for a much better prosumer or low end commercial camera.  Very different costs of sensors & peripherals (lenses,...)  Very different quality level.  Very different market e.g. 1 billion units / year made vs 100k units / year made or something like that. $10 sensor on commodity wafer process vs. $400-$1000 sensor made on with a custom / niche production design but STILL using (mostly) the commodity CMOS manufacturing fabs / technologies.  Give up using the commodity production lines and the prices go up way more still.

As an amateur astronomer, photographer, scientist, engineer, etc. I'd love to see a THz to UV/X-ray sensitive multi-spectral imager array technology family that could be obtained inexpensively and used for commodity sensing / imaging just like CMOS image sensors have done for RGB / near IR.  Maybe with more novel sensor / fabrication technologies we could be on the cusp of seeing imager arrays with novel spectral capabilities, HDR, TOF, whatever and maybe we'll have $10 toy LWIR webcams in Walmart in 10 years.  But that'll probably be incidental to the technology getting so cheap that even people / applications that don't "need" it gets the capability for so close to free that it is just integrated into lots of things as a matter of course if the sensors are $10 or whatever then most every $50 surveillance camera gets one etc.

People using thermal cams for mission critical applications aren't going to buy a celphone USB dongle for it. Those cellphone thermal cams aren't for industrial/commercial/military type stuff. Those are for hobbyists. There's an extensive hobby market for these, but for some reason, the makers of these (even the Chinese companies) don't seem to realize that 160x120 or even 320x240 resolution is NOT what hobbyists want in their cellphone thermal cams. What we hobbyists WANT is 640x480 or better. There is actually a sizeable hobbyist market for these, but all the thermal cam companies don't seem to have realized it yet. They are still selling these low-res (320x240 or less) cellphone thermal cams for HUNDREDS of dollars each, when my 1080p FULL HD webcam costs less than $100.

The people running these companies are putting high prices on their cellphone thermal cams, that leads me to believe that they are under the mistaken impression that these cellphone thermal cams are mostly being used by professionals. But they aren't. They are mostly used by hobbyists. And the manufacturers seem to really not be properly trying to capture this hobbyist market, by lowering the prices to under $100 like they should be . This leaves me wondering if the manufacturers themselves are hitting a price barrier. Are the companies themselves maybe actually barely making a profit on these cellphone dongles? Does it actually cost the companies more than I'm thinking to make them? Like for a $500 thermal camera for my cellphone, does it actually cost the company almost $500 per unit to make them?

By my estimate, the price to buy for a thermal camera of a given resolution is about 100x what a visible light CMOS camera of the same resolution costs. Is the cost of the manufacturing process for a thermal cam actually 100x what it costs to make a visible light CMOS camera of the same resolution? Or are the companies selling thermal cams just making more obscene profits than the companies selling visible light CMOS cameras?

[james_s]

People using thermal cams for mission critical applications aren't going to buy a celphone USB dongle for it. Those cellphone thermal cams aren't for industrial/commercial/military type stuff. Those are for hobbyists. There's an extensive hobby market for these, but for some reason, the makers of these (even the Chinese companies) don't seem to realize that 160x120 or even 320x240 resolution is NOT what hobbyists want in their cellphone thermal cams. What we hobbyists WANT is 640x480 or better. There is actually a sizeable hobbyist market for these, but all the thermal cam companies don't seem to have realized it yet. They are still selling these low-res (320x240 or less) cellphone thermal cams for HUNDREDS of dollars each, when my 1080p FULL HD webcam costs less than $100.

I think you're confusing what hobbyists want with what *you* want. Higher resolution is always better, but I'd be pretty happy with 320x240, I can't really think of an application I have that really needs more than that. I suspect the vast majority of hobbyists would be happy with a modest resolution, price is by far the most important aspect to hobbyists. You can't compare a mobile phone camera to a thermal camera. A microbolometer is a completely different tech than a CMOS image sensor.

[bap2703]

The real question is: why reverse cost analysis of microbolometers are so expensive ?! Hobbyist want it for free ! 

[Microdoser]

Aside from the drastic difference in economics of scale between a billion people buying a visible light sensor and the number of people wanting an IR sensor, one reason might be the sheer size of the sensor needed to get a high resolution image at IR wavelengths.

I am unsure of the relationship between wavelength and minimum pixel sensor size, but the wavelength at -20c is 0.0014cm so you are looking at a 640 pixel sensor that has an absolute minimum size of ~1cm before photons become smeared over multiple pixels.

Adding in the other required tech needed to focus IR light onto a sensor of that size, and the theoretical minimum size for any device of that resolution is much larger than would easily just plug into a phone. Far larger than a standard camera. Development costs become exponentially greater as you approach theoretical limits...

Modern phones get around this size limit with expensively developed software and other tricks. With the vastly lower number of customers for an IR device, these costs have to be spread around far fewer people, so the cost to the individual user will be proportionately greater.

Just because you are China does not mean you can ignore physics or basic economics.

[Bill W]

By my estimate, the price to buy for a thermal camera of a given resolution is about 100x what a visible light CMOS camera of the same resolution costs. Is the cost of the manufacturing process for a thermal cam actually 100x what it costs to make a visible light CMOS camera of the same resolution? Or are the companies selling thermal cams just making more obscene profits than the companies selling visible light CMOS cameras?

Same answer as every time you ask this question...... yes that is where the cost is at.

Typical phone level sensor is 1.5 um pixels
Thermal camera 12um pixels

80x less sensors per per wafer for the same pixel size
Then add in the complications of fab for bolometers (3D structure)
Then add in the poor yield for bolometers (because of the 3D structure)
Then add in the vacuum packaging needed for TIC
Then add in the lens and flag costs
Then add in the calibrations needed for TIC
Then add in the higher support electronics needed vs a one chip CMOS
Finally you have the lower production to recover design costs in hardware and software.

[Ben321]

Aside from the drastic difference in economics of scale between a billion people buying a visible light sensor and the number of people wanting an IR sensor, one reason might be the sheer size of the sensor needed to get a high resolution image at IR wavelengths.

I am unsure of the relationship between wavelength and minimum pixel sensor size, but the wavelength at -20c is 0.0014cm so you are looking at a 640 pixel sensor that has an absolute minimum size of ~1cm before photons become smeared over multiple pixels.

Adding in the other required tech needed to focus IR light onto a sensor of that size, and the theoretical minimum size for any device of that resolution is much larger than would easily just plug into a phone. Far larger than a standard camera. Development costs become exponentially greater as you approach theoretical limits...

Modern phones get around this size limit with expensively developed software and other tricks. With the vastly lower number of customers for an IR device, these costs have to be spread around far fewer people, so the cost to the individual user will be proportionately greater.

Just because you are China does not mean you can ignore physics or basic economics.

You seem to think that I am talking about integrating it into a phone. I am talking about using it as a USB dongle on the phone (like the FLIR One, and Seek Compact Pro). Yes there are already dongles that are 640x480, but they cost about $3000 (just as if they were a standalone thermal cam), which as far as I can tell is not really what it costs to make. I'm guessing they are just making obscene profits from selling these dongles.

[Ben321]

Aside from the drastic difference in economics of scale between a billion people buying a visible light sensor and the number of people wanting an IR sensor, one reason might be the sheer size of the sensor needed to get a high resolution image at IR wavelengths.

I am unsure of the relationship between wavelength and minimum pixel sensor size, but the wavelength at -20c is 0.0014cm so you are looking at a 640 pixel sensor that has an absolute minimum size of ~1cm before photons become smeared over multiple pixels.

Adding in the other required tech needed to focus IR light onto a sensor of that size, and the theoretical minimum size for any device of that resolution is much larger than would easily just plug into a phone. Far larger than a standard camera. Development costs become exponentially greater as you approach theoretical limits...

Modern phones get around this size limit with expensively developed software and other tricks. With the vastly lower number of customers for an IR device, these costs have to be spread around far fewer people, so the cost to the individual user will be proportionately greater.

Just because you are China does not mean you can ignore physics or basic economics.

Also I think this is a matter of marketing. If they really wanted to sell these to hobbyists. They could advertise on TV ads. You say not as many people want these. I think many people would want them, but just they don't know they exist. I bet people who saw the TV ads could be convinced to want it if the ad showed them what it's capable of. Maybe billions buy cellphones or webcams, but still at least tens of millions will want thermal cam attachments for their phones, if they knew they existed. I have not seen even one ad on national TV during a commercial break, for things like the FLIR One and the Seek Compoct Pro. I wouldn't expect the Chinese companies to advertise directly on US TV, but if they have US distributors I would expect those distributors to advertise on US TV. But they never do. It's like they WANT almost nobody to know about these, so they can remain forever exotic (and thus very expensive) pieces of equipment.

[Ben321]

By my estimate, the price to buy for a thermal camera of a given resolution is about 100x what a visible light CMOS camera of the same resolution costs. Is the cost of the manufacturing process for a thermal cam actually 100x what it costs to make a visible light CMOS camera of the same resolution? Or are the companies selling thermal cams just making more obscene profits than the companies selling visible light CMOS cameras?

Same answer as every time you ask this question...... yes that is where the cost is at.

Typical phone level sensor is 1.5 um pixels
Thermal camera 12um pixels

80x less sensors per per wafer for the same pixel size
Then add in the complications of fab for bolometers (3D structure)
Then add in the poor yield for bolometers (because of the 3D structure)
Then add in the vacuum packaging needed for TIC
Then add in the lens and flag costs
Then add in the calibrations needed for TIC
Then add in the higher support electronics needed vs a one chip CMOS
Finally you have the lower production to recover design costs in hardware and software.

The design costs are eventually paid off, but companies almost NEVER lower their prices. The Chinese company has managed to make LWIR microbolometer arrays at much smaller sizes than 12um. They have gotten them down to 8um, according to this big splash screen on their website (which also says that they are the first company in the world to accomplish this).
https://www.infiray.com/products/
I've linked to this screenshot in case they change this banner at some point.

Hopefully shrinking pixel physical size will shrink the price by the same amount. And yes they claim this new sensor using this new technology also has the massive (for thermal cams at least) resolution of 1920x1080 (which unfortunately raises the cost again).

[Vipitis]

yet there is no product using the 8μm pixel pitch sensor. Or even sample pictures.


You still overestimate how the market is distributed. I linked a report a while ago that answers all your questions and has numbers go back it up. It's not free to read tho, but consumer market for uncooled microbolometers is just a tiny fraction. Hunting and UAV outsells DIY markets already, and they do have a lot of VGA sized products.

The 1920x1200 sensor doesn't exist because hobbyists want larger sensors (which I got told isn't meant for hobbyists), but the British military got an upgrade for their tanks - which is why that product even exists in the first place.

[james_s]

Also I think this is a matter of marketing. If they really wanted to sell these to hobbyists. They could advertise on TV ads. You say not as many people want these. I think many people would want them, but just they don't know they exist. I bet people who saw the TV ads could be convinced to want it if the ad showed them what it's capable of. Maybe billions buy cellphones or webcams, but still at least tens of millions will want thermal cam attachments for their phones, if they knew they existed. I have not seen even one ad on national TV during a commercial break, for things like the FLIR One and the Seek Compoct Pro. I wouldn't expect the Chinese companies to advertise directly on US TV, but if they have US distributors I would expect those distributors to advertise on US TV. But they never do. It's like they WANT almost nobody to know about these, so they can remain forever exotic (and thus very expensive) pieces of equipment.

I think you are borderline obsessed with thermal cameras and *vastly* overestimate the interest to hobbyists. I only know one hobbyist that owns a thermal camera, the other people I know that have them are in the home inspection and energy auditing businesses. I briefly considered buying one but then realized I can just borrow the Flir camera my friend has when/if I have a need for one. It's a very specialized tool that most people simply do not need, even more technical people. They're neat, yes, but for most people the novelty wears off rather quickly.

[Bill W]

The design costs are eventually paid off, but companies almost NEVER lower their prices. T

Only when the production run is beyond the initial estimate and has paid back the costs.

Typical corporate politics usually results in over enthusiastic sales forecasts (as you are indeed making) so the amortisation number is rarely met.  Sales also take the lowest price the accountants will allow to make it easier to sell.

If the amortisation is ever met, either the product is obsolete so you make the next one with more development costs, or prices do indeed drop to increase market share.

FLIR internal economics are somewhat weird though.  As we know E4-E8 for example have the same part cost, but E8 will be making a far greater share of the R&D payback and marketing.

[svgurus]

In my experience, there were 0 photographers or hobbyists who were ready to buy thermal)) I'm making photos with thermal camera almost 2 years already, folks that i shoot ask what camera do i use regularly and i thought to see some Russian copycats or at least someone who would dig for deeper technical details or share something interesting with me, for now i was wrong) one funny case was at studio visit party, i talked with girl artist about my camera and she told her father had similar model and used it for hunting. She is working with tech installations rather than pure visuals but still i was amazed to know that she saw some cool imaging device but didn't research its capabilities.
About phone dongles - this segment of market is really declining, rumors about iray t6s turned out to lead nowhere, opgal stopped manufacturing any therm-apps, thermal expert didn't upgrade their V1 model. some hopes for flir, they recently upgraded tau2 and boson, maybe they will roll out some decent 320 dongle but most likely no 640 ones
What really amuses(sometimes scares/disgusts) me are 80-160pix sensors in some "serious usage" devices, including not only tech inspection toys made by flir/fluke/testo/chinese (wow! this c3 lepton based camera has so robust and ergonomic case and such great reporting capabilities! ) but even hunting gun sights by AGM or some chinese.  but then again, i see some reviews of 384-640 thermals where hunters compare different models by setting some strange digital zoom multiplier(1,3x, 2,4x ) so, either people are dumb en masse or we shouldn't judge them by our standards and experiences
and yeah, to put all woes in 1 post, my favorite notebook review website STILL uses some flir one plus msx to show heating map of notebooks, its 2022 omg these photos looked high-tech in 2014, they have rigorous yet aesthetic approach to making reviews, but this tiny detail freaks me out when i scroll to this part))

[cdev]

Its entirely possible there is a cartel style agreement when one observes what appears to be price fixing. Reading up on the issue, it appears these cartels are very common in various industries and indeed, are fairly normal when big businesses get together. The classic example is the cartel that was formed to fix the prices and lifetimes of tungsten filament light bulbs in the past. When they are set up, generally, they always do the same few things.

In the case of the light bulbs they optimiized the setup for more frequent bulb sales by requiring they burned out in a short period of time. Bulbs lasting years, as one particular bulb has, was internally forbidden. Manufacturers put their heads together and figured out how to do this. And how to keep the existence of this cartel quiet, too.

In the Silicon Valley case, the big firms agreed to forbid their "poaching" one another's workers.  In part to keep wages low. When they got caught they switched strategies to the international staffing market. In essence they moved their activities above the jurisdiction of countries to stop them. By using international staffing deals.

Now we cant stop them.

Years ago when I was involved with consumer electronics, a certain manufacturer was obsessed by keeping their prices high. They did this by only selling to vendors who agreed to do that.

Who knows!? I know that I would buy one of those thermal imagers, if they were cheaper. I could definitely use it for a lot of things. Like improving my energy efficiency!

Peace!

[janoc]

Don't forget that thermal imaging is a dual use technology and tightly regulated. Those imagers are used for everything from night vision devices, targeting pods to missile seekers.

So the better stuff isn't going to get on the mass market where e.g. Russians or Iran could buy it and use it (e.g. the Russian problems with obtaining thermal sights for their military are well known). This applies to the Chinese manufacturers too - they will keep the better sensors for themselves.

[cdev]

My requirements are extremely modest and totally peaceful. I am just looking for air leaks in my house, before the weather gets too warm to see them. Frame rate is utterly unimportant. It could be "one every minute" for all that I care, for example.

[james_s]

My requirements are extremely modest and totally peaceful. I am just looking for air leaks in my house, before the weather gets too warm to see them. Frame rate is utterly unimportant. It could be "one every minute" for all that I care, for example.

Call a home inspector or energy auditor, they'll come and do exactly that for you. Or just buy one of the less expensive cameras and sell it once you've finished checking your house.

[cdev]

That would cost more than a camera would. And be a one time only event.

[Ben321]

In my experience, there were 0 photographers or hobbyists who were ready to buy thermal))

That's because of their price. If the companies wanted to expand their business, they would lower prices, and advertise on national TV, and I'm sure millions of people would want to buy.

About phone dongles - this segment of market is really declining

Again, that's because they didn't MASSIVELY drop the prices for the dongles to help them be affordable to the average person, nor did they advertise them properly so the average person would know they existed.

The failure of consumer level thermal imagers is due to botched business practices of the manufacturers of those devices. The companies in question aren't familliar with how to properly target their products to average people, because average people aren't the people who buy the majority of their products. They usually sell to other companies and the military who use the equipment for mission critical applications. They forget that average person doesn't want to spend $500 on a 160x120 camera at 9fps. They forget that unlike entities buying stuff for mission critical applications, the average person doesn't have unlimited money to spend. So then when they DO make a products for the average person, they still charge WAY TOO MUCH, because they forget that the average person isn't about to spend that much on a camera with a tiny resolution. So they basically have TORPEDOED their own consumer level equipment, with bad business practices, because they don't know how to handle selling it to the average consumer.

[Ben321]

I think you are borderline obsessed with thermal cameras

I just want a thermal cam with DECENT resolution. 320x240 isn't decent. It's basically a joke for any modern camera. Around the year 2000, I got a Logitech webcam for $50, and its resolution was 320x240. Now a webcam with the same resolution wouldn't even bother to be manufactured by any company, due to it being TOO LOW of a resolution. And if it was manufactured now it would cost about $5 to buy one. 640x480 is what I would consider to be decent resolution.

[Ben321]

Its entirely possible there is a cartel style agreement when one observes what appears to be price fixing.

Bingo. We have a winner. Yes, you can't even get a much better deal from Chinese companies. It seems there's an INTERNATIONAL cartel here. Chinese and American companies have colluded to fix the minimum allowed price for a 640x480 thermal imager, to be about $3000. If there was no cartel, while FLIR might sell such cams at a minimum of $3000, the Chinese companies would do as they always do in other industries, and MASSIVELY undersell their American competitors. I would expect a Chinese 640x480 thermal imager to be about $750 to $1500, but instead it was only a couple hundred dollars below $3000. This right here is like absolute smoking gun proof of price fixing.

[Ben321]

By my estimate, the price to buy for a thermal camera of a given resolution is about 100x what a visible light CMOS camera of the same resolution costs. Is the cost of the manufacturing process for a thermal cam actually 100x what it costs to make a visible light CMOS camera of the same resolution? Or are the companies selling thermal cams just making more obscene profits than the companies selling visible light CMOS cameras?

Same answer as every time you ask this question...... yes that is where the cost is at.

Typical phone level sensor is 1.5 um pixels
Thermal camera 12um pixels

80x less sensors per per wafer for the same pixel size
Then add in the complications of fab for bolometers (3D structure)
Then add in the poor yield for bolometers (because of the 3D structure)
Then add in the vacuum packaging needed for TIC
Then add in the lens and flag costs
Then add in the calibrations needed for TIC
Then add in the higher support electronics needed vs a one chip CMOS
Finally you have the lower production to recover design costs in hardware and software.

It may be that the thermal image sensor is significantly more expensive to produce than a visible light sensor, but I can assure you that thermal image sensors are not 100x as expensive to manufacture as visible light sensors. Same thing with the lenses.

My suspicion is that while the they are more expensive to produce (and this should naturally increase the cost somewhat), the reason that they are being sold as expensively as they are right now also is due to price gouging (charging so much more than it costs to manufacture, to the point that the company is basically robbing its customers).

I'm guessing that the actual manufacturing costs for a 640x480 thermal imager core, are probably actually somewhere between $500 and $1000. Possibly even less than $500.

[Bill W]

By my estimate, the price to buy for a thermal camera of a given resolution is about 100x what a visible light CMOS camera of the same resolution costs. Is the cost of the manufacturing process for a thermal cam actually 100x what it costs to make a visible light CMOS camera of the same resolution? Or are the companies selling thermal cams just making more obscene profits than the companies selling visible light CMOS cameras?

Same answer as every time you ask this question...... yes that is where the cost is at.

Typical phone level sensor is 1.5 um pixels
Thermal camera 12um pixels

80x less sensors per per wafer for the same pixel size
Then add in the complications of fab for bolometers (3D structure)
Then add in the poor yield for bolometers (because of the 3D structure)
Then add in the vacuum packaging needed for TIC
Then add in the lens and flag costs
Then add in the calibrations needed for TIC
Then add in the higher support electronics needed vs a one chip CMOS
Finally you have the lower production to recover design costs in hardware and software.

It may be that the thermal image sensor is significantly more expensive to produce than a visible light sensor, but I can assure you that thermal image sensors are not 100x as expensive to manufacture as visible light sensors. Same thing with the lenses.

My suspicion is that while the they are more expensive to produce (and this should naturally increase the cost somewhat), the reason that they are being sold as expensively as they are right now also is due to price gouging (charging so much more than it costs to manufacture, to the point that the company is basically robbing its customers).

I'm guessing that the actual manufacturing costs for a 640x480 thermal imager core, are probably actually somewhere between $500 and $1000. Possibly even less than $500.

Which core / sensor manufacturing facilities have you visited and performed open book costing on ?

As for manufacturing costs, $1500 is probably close as that implies a typical profit margin of 50% on $3000 one-off sale price.  In 1000+ off you might get down to $2000 a core.

That added $1500 (or $500) has to cover sales, marketing, shipping, returns, support, equipment amortisation, R&D amortisation, corporate admin and finally hope to leave a little bit of profit (whether for shareholders or for the Chinese Communist Party)

[Vipitis]

OP acts increasingly irrational

Yes, these cameras are expensive. Take a course on economy or read up on some ideas. imaging sensors are everywhere, that's why they are so cheaply available. The market is larger and two orders of magnitude is easily explained if you look at product volume. Don't act like you know better than the industry veterans that have decades of experience. It's simply disrespectful.
Your viewpoint is strongly biased by the only data point you have, your own opinion. Which is a bad idea for any kind of venture that relies on other people.

See some current overviews listed here http://image-sensors-world.blogspot.com/2022/03/state-of-image-sensor-market.html?m=1
And then look up the latest Yole report on uncooled microbolometers to get an idea of what the market actually looks like and where it's heading. Perhaps that will even show you when 640x512 sensors will be available in consumer products and for prices you can afford.


If you really want to have a high resolution thermal imager, you would manage to do so - by either securing the funding or finding an ebay bargain. I know of several people on this forum that managed to get their hands on VGA sized sensors for well below the 3000$ price tag you experience. Myself included.

[james_s]

I just want a thermal cam with DECENT resolution. 320x240 isn't decent. It's basically a joke for any modern camera. Around the year 2000, I got a Logitech webcam for $50, and its resolution was 320x240. Now a webcam with the same resolution wouldn't even bother to be manufactured by any company, due to it being TOO LOW of a resolution. And if it was manufactured now it would cost about $5 to buy one. 640x480 is what I would consider to be decent resolution.

Many would disagree with you, including me. 320x240 is quite good resolution for a *thermal* camera. You are comparing apples to horses, could that 640x480 webcam see heat? No it cannot, entirely different device with an entirely different purpose. What do you need such high resolution in a thermal image for? High resolution is useful for visible light cameras, but a thermal camera is an entirely different beast.

[Bassman59]

It may be that the thermal image sensor is significantly more expensive to produce than a visible light sensor, but I can assure you that thermal image sensors are not 100x as expensive to manufacture as visible light sensors. Same thing with the lenses.

If you can assure us, then please, by all means, provide evidence to back up your assurance.

And you do know that a camera is more than just the image sensor, right?

[Ben321]

By my estimate, the price to buy for a thermal camera of a given resolution is about 100x what a visible light CMOS camera of the same resolution costs. Is the cost of the manufacturing process for a thermal cam actually 100x what it costs to make a visible light CMOS camera of the same resolution? Or are the companies selling thermal cams just making more obscene profits than the companies selling visible light CMOS cameras?

Same answer as every time you ask this question...... yes that is where the cost is at.

Typical phone level sensor is 1.5 um pixels
Thermal camera 12um pixels

80x less sensors per per wafer for the same pixel size
Then add in the complications of fab for bolometers (3D structure)
Then add in the poor yield for bolometers (because of the 3D structure)
Then add in the vacuum packaging needed for TIC
Then add in the lens and flag costs
Then add in the calibrations needed for TIC
Then add in the higher support electronics needed vs a one chip CMOS
Finally you have the lower production to recover design costs in hardware and software.

It may be that the thermal image sensor is significantly more expensive to produce than a visible light sensor, but I can assure you that thermal image sensors are not 100x as expensive to manufacture as visible light sensors. Same thing with the lenses.

My suspicion is that while the they are more expensive to produce (and this should naturally increase the cost somewhat), the reason that they are being sold as expensively as they are right now also is due to price gouging (charging so much more than it costs to manufacture, to the point that the company is basically robbing its customers).

I'm guessing that the actual manufacturing costs for a 640x480 thermal imager core, are probably actually somewhere between $500 and $1000. Possibly even less than $500.

Which core / sensor manufacturing facilities have you visited and performed open book costing on ?

As for manufacturing costs, $1500 is probably close as that implies a typical profit margin of 50% on $3000 one-off sale price.  In 1000+ off you might get down to $2000 a core.

That added $1500 (or $500) has to cover sales, marketing, shipping, returns, support, equipment amortisation, R&D amortisation, corporate admin and finally hope to leave a little bit of profit (whether for shareholders or for the Chinese Communist Party)

Marketing is very expensive. Yes. But I'm guessing when you take into account the number of cams sold, and divide the price for overall marketing by the number of cameras, the "per camera marketing" isn't anywhere near as expensive as the camera itself. As for shipping, that's typically listed separately as a "shipping fee", at least on every online store I've seen. It's not built into the price of the product itself. Product support and all other company expenses are also probably expensive, but due to the shear number of devices sold, dividing that overall price for running the company by the number of devices sold, is going add only a small amount to the price of each unit of the device in question. I think the vast majority (probably at least 75% to 90%) of the money made per sale of thermal cameras, is profit, and goes straight into the corporate bank accounts of the companies making the cameras. At least that's my hunch based on my own estimates of the various costs involved.

I see this as another example of the problem with corporate greed. And it makes me wish that there was some company out there who would make the mission to provide cheap alternatives to expensive products like this, so that the average person could get their hands on some really cool tech, that would otherwise be unavoidable to the average person if they were to buy it from any other company.

[Ben321]

What do you need such high resolution in a thermal image for? High resolution is useful for visible light cameras, but a thermal camera is an entirely different beast.

To get cool pics in a size that the average person could appreciate, after I post said images online (think posting cool pics to social media). Most people aren't going to want to need to get their face close to the screen to see detail in an image. They want to sit at comfortable distance to view the pics. I'd say that 640x480 is about the smallest size of picture that would be considered reasonable size to the average person. If I got some pics I considered awesome (awesome because they were thermal pics, and not normal pics like most people have) around my neighborhood, and wanted to share them on the socials, I think that most people would laugh if I posted something with 320x240 resolution, because that's so low res. Not exactly a social media worthy resolution at 320x240. Now at 640x480 that's different. I would probably get quite a few more views of my pics if I posted them in that resolution. Yes I can upscale them to 640x480, but that's fake and doesn't actually bring out any extra detail in the image, and in fact too much upscaling can make the picture look blurry.

[Ben321]

It may be that the thermal image sensor is significantly more expensive to produce than a visible light sensor, but I can assure you that thermal image sensors are not 100x as expensive to manufacture as visible light sensors. Same thing with the lenses.

If you can assure us, then please, by all means, provide evidence to back up your assurance.

And you do know that a camera is more than just the image sensor, right?

Yep. Surrounding electronics too. But those probably aren't too different from the average visible light camera.
And of course a lens. Again, my hunch is that germanium isn't too much more difficult to get or work with than glass, but because it happens to be being used in an exotic application, the company accountant is like "Oh this is science equipment, so we can way overcharge for it, because science equipment is supposed to be expensive", and that gives them an excuse to overcharge for it to the point of making an obscene profit, instead of charging an amount that would generate merely a reasonable amount of profit.

[Ben321]

The real question is why are even Chinese companies, which are KNOWN to sell stuff MUCH cheaper than US counterparts, actually not selling 640x480 thermal cams cheaper than FLIR does? Is there some kind of illegal price fixing going on in the thermal camera industry worldwide (or as another poster here called it, a "cartel")? Like is there some kind of international agreement between all thermal camera companies in the world (even in China), to not charge less than about $3000 for a 640x480 thermal camera?

[Bill W]

I see this as another example of the problem with corporate greed. And it makes me wish that there was some company out there who would make the mission to provide cheap alternatives to expensive products like this, so that the average person could get their hands on some really cool tech, that would otherwise be unavoidable to the average person if they were to buy it from any other company.

There is, it is called SEEK.
For historical reasons they would be delighted to wipe out FLIR and the Chinese of they could.  The owners do not need excess profits.  No way would they be in collusion.
That they cannot is the biggest indicator that your presumptions are all

[Bill W]

Marketing is very expensive. Yes. But I'm guessing when you take into account the number of cams sold, and divide the price for overall marketing by the number of cameras, the "per camera marketing" isn't anywhere near as expensive as the camera itself.

Not that simple. 
Advertising to a small market is cheap.  Not that many customers to go at, good coverage of relevant people by specialist media.
Wide scale 'blunderbuss' marketing (eg TV, general media) is expensive and inefficient.
As such you cannot simply lump it into 'marketing per camera'

Equally sales support, running a week long training camp for someone who will buy 10,000 sensors is worth the time and facility invested. 
The same for 'Ben321 Industries' with maximum sales of 1 is not worth it, or adds 10,000 to the price.

[Vipitis]

Again, you shown how little you understand about economics or product development in general, the electronics involved in the integrated readout circuit (yes they are much more complex than CIS) or chemistry. Germanium is nothing like glass at all. You can't mould Germanium. It's a single grown crystal that's cut and machined down to optical shapes. It's more similar to silicon and goes through some of the same processing steps for growing the ingot.

It might also be of interest that SEEK is not a Chinese offering at all, there are threads about the companies origin and history but as far as I know, the sensors are made by Raytheon - in the United States.

There is no international cartel but for DoD contracts there is a bunch of lobbying, do prices will be inflated and probably fund some executes of senators which is just part of the US military complex. However consumer goods prices are very much dictated by supply and demand. The US being hindered by regulations does add to this as well.

Finally, the idea that 640x480 looks a lot better than 384x288 is also misguided. If you are using a lens that can resolve such tiny pixels and get the image processing to get useable results in the end. you can become a published and exhibiting artist with less than VGA resolution. If you were motivated enough you would also find methods with stacking, stiching and processing to end up with useable results. Or make the limited pixel resolution part of your look. Every visible light image you see is 'upscaled' as well in the debayering process, which also means a QVGA thermal camera does look as good as a VGA sized Bayer visible light camera.


All your claims are based on 'I believe' or 'I think' - read the reports I liked from Yole and see how they predict consumer devices to drop in price and rise in resolution.

[ArsenioDev]

 

Like is there some kind of international agreement between all thermal camera companies in the world (even in China), to not charge less than about $3000 for a 640x480 thermal camera?

That's the thing, I got quotes for the latest gen VGA cores from several chinese manufacturers and they're about 1.7K with a pretty good lens on em, if you know where to look on the used market you can score VGA for WELL under a grand these days, I myself have a very nice VGA core I paid only about $650 for and have full documentation on it (That's the gold there).
Lenses for higher resolution imaging get expensive due to the more complex optical paths and more elements required so of course prices will go up, I think OP doesn't quite grok the thermal market and inherently exponentially more expensive tech than standard CMOS vis band imaging.

[cdev]

The arms industry is huge and unless they are boycotted (which maybe against WTO rules, as long as their money is good, any state, including the Russians can quite possibly buy whatever they can pay for. I dopnt know where it stands now. Does anybody? I dont think they are deemed a terrorist by the WTO. Which means I suspect they may be able to replace all the thermal imagers they can pay for. As long as they have money to pay whats demanded, whatever it is. Of course many countries would like to hide the human rights agnostic aspect of world trade. Permanent Normal Trade Relations, and WTO membership, boycotts of any WTO members products, generally, for any reason, are more or less prohibited. But this may be changing, very gradually. It kind of has to, minimally, in response to the public, who learn things like this very very very gradually.

Its a contentious issue for sure. But WTO makes sure that commerce always wins. Sometimes that is preferable. It certainly is for electronics manufacturers, I suspect.

Thats the whole point of the WTO. Stopping boycotts like the one that ended apartheid from inhibiting business.. including the lucrative arms business.. One of the most profitable of businesses.

This is a chapter from an entire free book, written by Sarah Joseph, which you can find onine, its worth reading..

I see this as another example of the problem with corporate greed. And it makes me wish that there was some company out there who would make the mission to provide cheap alternatives to expensive products like this, so that the average person could get their hands on some really cool tech, that would otherwise be unavoidable to the average person if they were to buy it from any other company.

There is, it is called SEEK.
For historical reasons they would be delighted to wipe out FLIR and the Chinese of they could.  The owners do not need excess profits.  No way would they be in collusion.
That they cannot is the biggest indicator that your presumptions are all

[cdev]

What is the situation with thermal imaging equipment and ukraine?

I am generally against wars but realize that much of what rankles me may not technically be illegal or benefit from regulatory attention. Suppose I did want to report "illegal" activity if I saw it, what would/should I report and to whom?

I am not talking about "cheap" thermal imaging which I generally would think positively about as long as it wasn't going to feed the Russian war machine in Ukraine. (or similar things)

What I would envision reporting would be hardware that might be used by terrorists or countries committing crimes against innocent people.

[james_s]

To get cool pics in a size that the average person could appreciate, after I post said images online (think posting cool pics to social media). Most people aren't going to want to need to get their face close to the screen to see detail in an image. They want to sit at comfortable distance to view the pics. I'd say that 640x480 is about the smallest size of picture that would be considered reasonable size to the average person. If I got some pics I considered awesome (awesome because they were thermal pics, and not normal pics like most people have) around my neighborhood, and wanted to share them on the socials, I think that most people would laugh if I posted something with 320x240 resolution, because that's so low res. Not exactly a social media worthy resolution at 320x240. Now at 640x480 that's different. I would probably get quite a few more views of my pics if I posted them in that resolution. Yes I can upscale them to 640x480, but that's fake and doesn't actually bring out any extra detail in the image, and in fact too much upscaling can make the picture look blurry.

I challenge you to find me 10 people who want to buy a high resolution thermal camera to post "cool" pics to post on social media, much less the tens of thousands of people that are required to start to drive down the cost. Once again you are taking your own personal obsession and projecting it on the general population. I assure you, the VAST majority of ordinary people out there don't give a crap about thermal images, and 99% of them are not going to notice the difference between a 320x240 image that has been upscaled and a 640x480 native image. You seem to want a thermal camera as a toy, they are not toys and it is completely absurd to expect a manufacture to cater to this extremely niche use. It's like demanding a 7 digit multimeter to take "cool" measurements and post them online, a handful of voltnuts might find that fascinating but most people are not going to care, and the vast majority of hobbyists see no compelling need for such a precise instrument.

You keep bitching and moaning about "greed" while making baseless assertions of some mass market desire for a very niche item and making comparisons to a completely different and extremely mainstream item that almost everyone owns. Having unusual interests is fine, I've been obsessed with light bulbs and engines my entire life but I am self aware that my obsessive interests are weird and not shared by the vast majority of people in the world, you have apparently not reached this self awareness. Even if thermal cameras cost exactly the same as visible light cameras, I'd wager no more than 0.01% of the population would be interested in buying one. Most people simply do not care, once you've seen a handful of thermal images you've seen them all.

[Bill W]

What is the situation with thermal imaging equipment and ukraine?

I am generally against wars but realize that much of what rankles me may not technically be illegal or benefit from regulatory attention. Suppose I did want to report "illegal" activity if I saw it, what would/should I report and to whom?

You contact the export control department of the state from which you believe that the 'questionable' export is being made.

May be worth adding that while a '<9Hz' imager is not controlled by Wassenaar etc, the state level embargoes such as those generally (US, UK, EU27) in place on Russia and Belarus take precedence.  No <9Hz or even spares to RUS/BLR.

Ukrainian Government controlled Ukraine territory is still (UK) on normal export rules, one suspects that any requested licenses are being readily granted.

[james_s]

But none of those applications really need exceptionally high resolution. 320x240 is high resolution in terms of thermal images.

[Bill W]

But none of those applications really need exceptionally high resolution. 320x240 is high resolution in terms of thermal images.

Indeed there is an argument that thermal, at low resolution, is better for some surveillance tasks as it cannot 'identify' the target, so satisfying data privacy concerns.

One odd aspect to me is that thermal is still coming out in 4:3 ratio.  A 640x240 sensor would be half the price of a 640x480 but nearly as useful for surface based uses.
Even historically few sensors have come in 'wide' formats, yes BST had a 320x120 option for vehicle use, but it was on a full 320x240 sensor.  There was the BAe (UK, ex-Plessey) 256x128 though.

[bap2703]

Did anyone add to the price conspiracy that we might be getting only the low end of binned devices rejected from military use?

[Bill W]

Did anyone add to the price conspiracy that we might be getting only the low end of binned devices rejected from military use?

At least for western suppliers their binned devices with zero defects (A), or zero central defects (B) will either cost more, or a lot more, depending on the whole process yield.

Those suppliers with a relatively small military customer base do supply zero defects at the commercial price, as they have more than they can sell at grade A.  Sometimes you just get lucky.

Why would China be any different ?

Bill

[james_s]

I would think most surveillance applications would benefit from higher resolution than 320x200, actually even higher than VGA would be ideal but of course 320x200 is a valid starting point but if (like NIR/VIS sensors) the incremental cost for VGA, HD, FHD, UHD could end up to be small enough then I'm 100% sure a lot of the mass surveillance market would select the higher resolution chips.

More resolution is always nice, but I'm not convinced that there's a huge benefit even for surveillance. A thermal image is good for identifying that there's something there, but you're not going to be able to see for example a person's face. A thermal sensor paired with a high resolution visible image sensor can provide a nice compromise, overlaying the thermal image over the top of the visible light image similar to the way the chroma bandwidth of a VHS recording is only about half that of the luminance bandwidth. In practice you don't really notice that the resolution of the color is so much lower than that of the base image.

[Microdoser]

You seem to think that I am talking about integrating it into a phone. I am talking about using it as a USB dongle on the phone (like the FLIR One, and Seek Compact Pro).

I'm not though.

You seem to think that an IR sensor should cost exactly the same as the visible light one in your phone or webcam, though, even when the reasons why this cannot be done are repeatedly explained to you.

I think OP doesn't quite grok the thermal market and inherently exponentially more expensive tech than standard CMOS vis band imaging.

This is what I think also.

[bap2703]

Chinese aggressive marketing isn't stereotype though

[polar]

Chinese aggressive marketing isn't stereotype though

 

[cdev]

Military.
UK is heavily invested in the military (Arms) Industry.
especially sales to developing world oligarchs, etc. .
This is absolutely huge in the UK, or so I keep being told.

Its what the money from all that African oil goes to pay for, more guns..helicopters, missiles, rockets

More wars, more moolah...

[All Seeing Eye]

1920x1280 are already products of chinese IRAY,but the price is not affordable to ordinary users.

[Vipitis]

does that 1920x1280 sensor exists in any products? Or is it just a press release.

The 1920x1200 sensor from Fairchild does actually exist and is available as the Vayu HD from Sierra Olympic or for the BAE DVR for tanks... Which isn't that available I guess