Canon to start using 0.18um (180nm) process for FF?

[jrista]

Chipworks recently released an article analysing the CMOS Image Sensor (CIS) processes from a variety of manufacturers, including Nikon, Sony, and Canon. Historically, Canon has used a 0.5 micron (500nm) process for all of their FF sensors since the original 1Ds. In the Canon analysis, they noted that Canon has a 0.18 micron (180 nanometer) fabrication process (possibly what they used for the 120mp APS-H?) that they may begin using for future FF sensors:

[quote author=Chipworks]Canon does have a 0.18 µm generation CIS wafer fab process, featuring a specialized Cu back end of line (BEOL) including light pipes (shown below). It is possible to speculate that Canon may be preparing to refresh its FF CIS line to supply devices for a new FF camera system.

A move from their 0.5um process to a 0.18um process for FF CIS manufacture would be a fairly significant move for Canon. The accompanying image figure also seems to indicate a double microlens layer...one above the CFA and one below...which could lead to higher Q.E. The article also mentions the use of "Light Pipes", a term I had not heard before. According to a few papers I've read, lightpipes in CMOS sensor design make use of high refractive index materials and a reflective wall in the optical stack the to improve transmission of light from the color filter/microlens to the photodiode, which exists at the end of a narrow tube where all the readout wiring exists (in a frontside-illuminated design). Seems like a lightpipe is an alternative to using a backside-illuminated design that aims to improve Q.E while avoiding some of the complexities and issues with BSI designs. Additionally, the use of copper interconnects should improve efficiency, allowing lower power usage, and hopefully leading to a lower level of electronic noise (the great bane of Canon these days.)

Seems Canon is most definitely not out of the CMOS Image Sensor design game yet. They seem to have some new tricks up their sleeves, and hopefully they will see the light of day in their next FF camera. Ah, competition is good!

Hopefully. Hopefully they take advantage of it and can fit column ADC and stuff on their sensors now.
It sounds like there is some promise there, hopefully it will be delivered.
[/quote]

Yeah, I really hope they implement some kind of CP-ADC. At the very least, it could help them minimize vertical banding noise, and reduce noise introduced by the ADC itself (since each one per-column could run at a lower frequency.) Every time I read the press release about the 120mp APS-H, I swear it vaguely describes some kind of on-die ADC...column-parallel or not, it would at least be a step forward.

 

[Loswr]

A lot of people here are making assumptions about Canon not releasing new products for a variety of reasons, from pricing policy, to milking older technology, to marketing decisions, even to a reluctance to invest capital expenditure in the new machinery required for even thinner chips and sensors, but few people are considering the simple logistical facts: manufacturing enough quantity (inventory) prior to shipping.

Canon is a victim of it's own success, it has made and sold more than 50 million digital-SLR bodies since 2000 (starting with the D30, then D60 up to the 1DX and 6D bodies). In the first 8 years Canon sold 24 million DSLR's or about 3 million per year. Over the last 4 years they've sold almost as many again, now averaging 6.5 million DSLR bodies each year - that is nearly 542,000 per calendar month and more than 27,000 per working day! Assuming that Canon requires a minimum of a 3-month inventory of a new product prior to launch + another 2-3 months supply in-transit (got to ship those boxes half way around the world, time in bonded warehouses, customs etc.), then they have to have a half years Sales Supply in hand before launch. This is probably why the 6D is not being delivered till early-December (which means they could have been manufacturing it since May).

Wow - that's a lot of cameras! But, the vast majority of them are xxxD/xxxxD bodies (and some of the latter are made in Taiwan). The higher-end bodies likely do not need that level of pre-release stock build-up.

In a way, though, we are victims of Canon's success. They're on top, have been on top for a long time, and thus the pressures to innovate are less strong.

 

[caruser]

In a way, though, we are victims of Canon's success. They're on top, have been on top for a long time, and thus the pressures to innovate are less strong.

So we should all publicly hold off from buying a new Canon (FF) DSLR until they migrate to the new process.

Assuming that it will bring the desired/required/imagined advantages (and it's not too late, like for Neuro' .

 

[jrista]

A lot of people here are making assumptions about Canon not releasing new products for a variety of reasons, from pricing policy, to milking older technology, to marketing decisions, even to a reluctance to invest capital expenditure in the new machinery required for even thinner chips and sensors, but few people are considering the simple logistical facts: manufacturing enough quantity (inventory) prior to shipping.

Canon is a victim of it's own success, it has made and sold more than 50 million digital-SLR bodies since 2000 (starting with the D30, then D60 up to the 1DX and 6D bodies). In the first 8 years Canon sold 24 million DSLR's or about 3 million per year. Over the last 4 years they've sold almost as many again, now averaging 6.5 million DSLR bodies each year - that is nearly 542,000 per calendar month and more than 27,000 per working day! Assuming that Canon requires a minimum of a 3-month inventory of a new product prior to launch + another 2-3 months supply in-transit (got to ship those boxes half way around the world, time in bonded warehouses, customs etc.), then they have to have a half years Sales Supply in hand before launch. This is probably why the 6D is not being delivered till early-December (which means they could have been manufacturing it since May).

These logistical supply constraints add up to a considerable lead time. This delay in manufacturing enough new product to physically deliver to specialist camera stores goes a long way to explaining why many CR bloggers feel that Canon has slipped behind competitors in the recent product launch stakes - perhaps the decisions to go ahead with a specific feature-list were made 18 months prior to actual launch, or some figure like that. Whichever way you look at it, Canon has to produce hundreds of thousands if not millions of sensors and digic processors for their cameras, months even before they arrive in your country.

It is a physical impossibility for Canon to launch e.g. a 7D2 in Spring 2013, if they're not already making the parts for that particular body right now....in large quantities.

I'm not sure all of that really qualifies Canon as a "victim" in any respect. Every company wants to be so popular that they can't keep up with demand...that is kind of the holy grail of the free market and supply & demand.

Also, if Canon makes a move to a 180nm process, it won't be in their highest volume products, which are all the APS-C parts. FF sensors are low-yield, just as the product sales are. If Canon has a fraction of their facilities dedicated to 180nm wafer fabrication, I don't see why that would be a problem for manufacturing a low-volume part, such as a 40-50mp FF sensor using new technology. The camera will either be a 1D or a 2/3/4D (although all of those kind of have "dimensional" names, and half of them carry some kind of mythical superstitious connotation to the Japanese, so...I'm guessing 1D line.)

It'll be a costly camera, although probably not as costly as some people seem to think (Canon won't sell a non-flagship body for $9k...it'll probably wind up somewhere between $5k - $7k.) They aren't going to be selling tens of thousands of those a month, nor will they need to stockpile hundreds of thousands before launch. A lower fabrication capacity should do quite well for a low-volume part like that.

 

[jrista]

Chipworks recently released an article analysing the CMOS Image Sensor (CIS) processes from a variety of manufacturers, including Nikon, Sony, and Canon. Historically, Canon has used a 0.5 micron (500nm) process for all of their FF sensors since the original 1Ds. In the Canon analysis, they noted that Canon has a 0.18 micron (180 nanometer) fabrication process (possibly what they used for the 120mp APS-H?) that they may begin using for future FF sensors:

[quote author=Chipworks]Canon does have a 0.18 µm generation CIS wafer fab process, featuring a specialized Cu back end of line (BEOL) including light pipes (shown below). It is possible to speculate that Canon may be preparing to refresh its FF CIS line to supply devices for a new FF camera system.

A move from their 0.5um process to a 0.18um process for FF CIS manufacture would be a fairly significant move for Canon. The accompanying image figure also seems to indicate a double microlens layer...one above the CFA and one below...which could lead to higher Q.E. The article also mentions the use of "Light Pipes", a term I had not heard before. According to a few papers I've read, lightpipes in CMOS sensor design make use of high refractive index materials and a reflective wall in the optical stack the to improve transmission of light from the color filter/microlens to the photodiode, which exists at the end of a narrow tube where all the readout wiring exists (in a frontside-illuminated design). Seems like a lightpipe is an alternative to using a backside-illuminated design that aims to improve Q.E while avoiding some of the complexities and issues with BSI designs. Additionally, the use of copper interconnects should improve efficiency, allowing lower power usage, and hopefully leading to a lower level of electronic noise (the great bane of Canon these days.)

Seems Canon is most definitely not out of the CMOS Image Sensor design game yet. They seem to have some new tricks up their sleeves, and hopefully they will see the light of day in their next FF camera. Ah, competition is good!

If competition is to working then Canon must:
1. Buying steppers from Nikon so that they are able to go down in size and expose a greater surface area than the compact cameras surface with the same accuracy
2.Buy lenses from Zeiss that can handle the resolution to expose a greater area than Canon can today
3. Dedicate a line for the APS and 24x36
4. This is large and heavy investments and are not shown anywhere in press or business reports
5.Gain experience as Sony and others and has refined its manufacturing , today's column ADC sensors from Sony are from the third generation.
Time will tell "what they have in theirs sleeves" , myself, I think they have slept too long
[/quote]

I'm not sure what you mean by "buy steppers from Nikon" for the purpose of exposing surface area. That comes off as rather inane and nonsensical when you consider Canon MANUFACTURES CMOS Lithography units.

As for lenses, Canon lenses are some of the best in the world, and I believe they offer at least as much, if not more, resolution than a lens from Zeiss. Canon is at the bleeding edge of optical science, and they have pushed the envelope in that arena farther than any other company, including Zeiss. They manufacture their own scanners with their own optics, and as far as I know they are capable of scaling well below 100nm these days.

The most significant reason I've stuck with Canon rather than expanding my kit to include Nikon is the quality and resolution of their lenses. They are lighter, use better materials, and often have MTF's that indicate they are perfect or near-perfect wide open (particularly in the case of the lenses I need most, their telephoto and supertelephoto lenses.) Canon has also stated in the past that their newer lenses are more than capable of handling the resolution for a 45mp FF sensor, which is no surprise, as that's about the same pixel density as their now-ubiquitous 18mp APS-C sensor. If there is any area where competition from Canon is impeccable, it is their optics...they are second to none these days, regardless of the industry.

As for "gaining experience" manufacturing sensors, Canon has DEMONSTRATED experience making sensors. It is no mean feat to push a 500nm process as far as they have...they have apparently extracted every last ounce possible from that process, and still remained competitive (thats "competitive", not "best"...not being best does not mean something is no longer competitive.) That demonstrates a particular knack that Canon has for pushing their technology to the utmost limits, and in my opinion, that only bodes that much better for Canon once they do move to a 180nm process. Canon is efficient, efficient to the extreme. They may not stay at the supreme bleeding edge of sensor technology at all times, but they WILL extract every last ounce of benefit they can from it once they get there, and that is why they are the largest, most profitable camera producer in the world.

 

[win nut]

The company I work for used to do 65nm logic devices on Canon steppers, with no problems. I really can't go into details, since I still work there. The ASML steppers (Zeiss) you talk about run about $60 million ea. I am sure Nikons aren't much different. Those guys are for running 14nm logic stuff, not sensors at 180nm. There is a huge used semiconductor equipment market where the steppers that Canon needs to run 180 nm is much cheaper than that. There logic devices we made were larger than camera sensors, roughly 4 die.

 

[jrista]

The company I work for used to do 65nm logic devices on Canon steppers, with no problems. I really can't go into details, since I still work there. The ASML steppers (Zeiss) you talk about run about $60 million ea. I am sure Nikons aren't much different. Those guys are for running 14nm logic stuff, not sensors at 180nm. There is a huge used semiconductor equipment market where the steppers that Canon needs to run 180 nm is much cheaper than that. There logic devices we made were larger than camera sensors, roughly 4 die.

Canon uses their own scanners for their wafer fabs, I don't think they buy much of it from other companies. Canon is in the business of manufacturing Semiconductor Lithography Systems. Not just parts of systems, but entire lithography systems, including the ones capable of sub-100nm manufacture. Canon is more than capable of manufacturing the necessary lenses for a 180nm process. Their glass is impeccable and on the cutting edge, has been for a long time...I don't doubt they have the capability to manufacture a lens capable of etching tiny transistors into a silicon wafer. As a matter of fact:

If competition is to working then Canon must:
1. Buying steppers from Nikon so that they are able to go down in size and expose a greater surface area than the compact cameras surface with the same accuracy
2.Buy lenses from Zeiss that can handle the resolution to expose a greater area than Canon can today
3. Dedicate a line for the APS and 24x36
4. This is large and heavy investments and are not shown anywhere in press or business reports
5.Gain experience as Sony and others and has refined its manufacturing , today's column ADC sensors from Sony are from the third generation.
Time will tell "what they have in theirs sleeves" , myself, I think they have slept too long

We don't need to wait for time. Here is Canon's latest Deep-UV Sub-wavelength Scanner that can be used for sub-100nm manufacture: FPA-6000ES6a Scanner

[quote author=Canon]
* The NA0.86 projection optics system developed through Canon's advanced lens design and measurement technology enables the ES6a to print down to 90nm.

* The AFIS (Advanced Flexible Illuminator System) improves efficiency of special illumination, and its flexible mode-settings allows users to freely design their own illumination modes.

* Its high throughput and high resolution enables ES6a to replace some of the ArF layers needed at 90nm production and even at 65nm production, reducing the total cost of ownership.
[/quote]

I've added emphasis that highlights the fact that these DUV Lithography units use CANON'S own optics, for 90nm manufacture...there is no need for Canon to run to the nearest competitor to "buy up" all the "necessary technology" to be competitive. THIS IS WHAT CANON DOES, and they are very good at it.

 

[jrista]

No
Canon has no steppers who can do the job with larger sensor areas and the cells we are talking about
And no, Zeiss is the only one who can provide lenses for larger sensors areas with smaller geometry we also are discussing

You don't necessarily have to use standard stepper for manufacture of sub-wavelength features. You can use a scanner, or scanning stepper, as well. In Canon's own words:

[quote author=Canon]
The FPA-6300ES6a KrF single-stage lithography tool is designed for the mass production of Dynamic Random Access Memory (DRAM), flash and other memory devices; logic devices such as microprocessors for personal computers; color filters and other imaging sensors and image-processing devices.
[/quote]

Emphasis added to highlight where they explicitly state their DUV scanners can be used to manufacture bayer (color filter) and other types of imaging sensors. Canon isn't manufacturing these things as a single full-sensor exposure in a standard stepper...they use scanning steppers to expand the exposure field:

http://en.wikipedia.org/wiki/Stepper#Scanners

[quote author=Wikipedia]
Modern scanners are steppers that increase the length of the area exposed in each shot (the exposure field) by moving the reticle stage and wafer stage in opposite directions to each other during the exposure. Instead of exposing the entire field at once, the exposure is made through an "exposure slit" that is as wide as the exposure field, but only a fraction of its length (such as a 9x25 mm slit for a 35x25 mm field). The image from the exposure slit is scanned across the exposure area.

There are several benefits to this technique. The field can be exposed with a lesser reduction of size from the reticle to the wafer (such as 4x reduction on a scanner, compared with 5x reduction on a stepper), while allowing a field size much larger than that which can be exposed with a typical stepper. Also the optical properties of the projection lens can be optimized in the area through which the image of the projection slit passes, while optical aberrations can be ignored outside of this area, because they will not affect the exposed area on the wafer.
[/quote]

The DUV Scanner I linked before is what's described above...a sub-wavelength scanning stepper designed to manufacture large area (200mm and 300mm wafers) with details (90nm, or even as small as 65nm it seems) smaller than the wavelength of light used (which in the Case of the 6000ES6a is 248nm from a Krypton-Fluorine laser.)

 

[TheSuede]

One of the main differences between a logic line and a sensor line is in how deeply you can work into the substrate surface.
I'd suggest you take a look at the specs on the machine you mentioned.

Canon's main (almost only...) customer for their stepper line is Canon Semi. They're on a very steep downwards revenue curve, and have been for the last five years. Canon semi's only large customer is Canon imaging. See a slightly disturbing picture here?

The effect this has on available budgets is quite profound:
Sony invest about 1.0-1.5 billion dollars per year in new and improved lithography and processing lines. Having spent almost 7 billion on the Kumamoto TEC site for the last 5 years, next year the smaller Nagasaki TEC is getting 1.5 billion next year.
This IN ITSELF is more than the total revenue of Canon Semiconductor.

And Sony is only the third largest manufacturer (at 18% total market revenue), Both Samsung and Omnivision are larger than Sony. Then comes STM, Toshiba and Aptina. Those big six corner about 80-85% of the world revenue in CMOS imaging sensors. Canon has about 6.5 to 7% of the world market in sales value - though only about 3% in sold units, since they are heavily specialized in large sensors.

 

[TheSuede]

The main problem as I see it is that Canon don't really have any real development momentum (or budget!) on higher resolution processes. They outsource almost everything except for the larger format sensors.

The biggest difference between the others and Canon is that all the other manufacturers are all dominated by their small-sensor image sensors sales, that already now are manufactured at 90 and 110-130nm metal processes on 300mm wafers. Panasonic and TSMC will start volume shipping of sensors made on 65/45nm rules in Q1 2013. Lower mask resolutions than 130nm are not enough to land you any sales any more. Most cellphone and compact camera sensors are manufactured at those levels now, and have been for the last few years. Also consider the fact that some of the others have very large yearly revenues from logic CMOS processes at 45, 32 and even 22nm levels. All of those markets are areas where Canon totally lack any type of experience. Canon outsource all more advanced fabs on their camera bill-of-materials.

The Digic sensors are made by UMC http://www.umc.com/english/class_300/index.asp, and were designed by Texas Instruments http://www.ti.com/lsds/ti/apps/videovision/end_equipment.page
The memory is most often made by Samsung http://www.samsung.com/global/business/semiconductor/product/consumer-dram/overview
The memory/Digic package-on-package mount has to be outsourced, since Canon cannot do it themselves, and it is believed that UMC does the mounting too.
Other peripheral control and logic chips are TI, Mitsubishi, AD and Fairchild.
.........

Going from a 500nm process to 180nm is like going from a 10MP FF camera to a 70MP camera in one generation. This means some really noticeable strains on the process, especially since the CMOS manufacturing process isn't as forgiving as just "taking pictures". You cant "scale to web size" and sharpen in post when you're making CIS wafers.

What you're basically asking from your equipment is to all of a sudden provide compact camera type linear resolution in a FF lens projection coverage - something almost unthinkable in the normal photographic world. The process has to be precise, to a degree where every single contrast and item on the new 70MP image is equal to or better than the 10MP camera - per pixel. In normal photographic resolution usage, we just want the final output to be good enough, which means that we downsample most images - we seldom deliver full-res images to the customers, and we seldom use full-res images in our own output.
Going from a 350nm mask to 250 and then 180nm and 12" wafers was a BIG step for most CMOS manufacturers, and most other manufacturers are a LOT bigger than Canon in this area.
............

So it's not that Canon COULDN'T do it. Even really small (in the imaging field) firms like STMicro can do it by stitching, and thereby tripling the unit prices. For Leica this isn't really a problem since the total BoM on a M series camera is most certainly lower than 2k USD. This gives a healthy margin up to the projected 7k USD end price point.
This isn't what Canon does. They live on volume, not on extreme margins.

 

[aznable]

The effect this has on available budgets is quite profound:
Sony invest about 1.0-1.5 billion dollars per year in new and improved lithography and processing lines. Having spent almost 7 billion on the Kumamoto TEC site for the last 5 years, next year the smaller Nagasaki TEC is getting 1.5 billion next year.
This IN ITSELF is more than the total revenue of Canon Semiconductor.

i get the point, but sony is in the verge on Bankruptcy...maybe they are investing too much

 

[Canon-F1]

The effect this has on available budgets is quite profound:
Sony invest about 1.0-1.5 billion dollars per year in new and improved lithography and processing lines. Having spent almost 7 billion on the Kumamoto TEC site for the last 5 years, next year the smaller Nagasaki TEC is getting 1.5 billion next year.
This IN ITSELF is more than the total revenue of Canon Semiconductor.

and we all know how good sonys profit is.... :

while sony was deep in the reds the last years, canon made profit.

so maybe canon has enough money to spend to do the jump to a smaller process.... now.
without spending money on smaller steps over the last years... like sony has.

 

[Canon-F1]

Well, think know how and it takes time to develope a new sensor, nothing any one do in few years.

what process was the 120 MP aps-h sensor?

 

[jrista]

The main problem as I see it is that Canon don't really have any real development momentum (or budget!) on higher resolution processes. They outsource almost everything except for the larger format sensors.

The biggest difference between the others and Canon is that all the other manufacturers are all dominated by their small-sensor image sensors sales, that already now are manufactured at 90 and 110-130nm metal processes on 300mm wafers. Panasonic and TSMC will start volume shipping of sensors made on 65/45nm rules in Q1 2013. Lower mask resolutions than 130nm are not enough to land you any sales any more. Most cellphone and compact camera sensors are manufactured at those levels now, and have been for the last few years. Also consider the fact that some of the others have very large yearly revenues from logic CMOS processes at 45, 32 and even 22nm levels. All of those markets are areas where Canon totally lack any type of experience. Canon outsource all more advanced fabs on their camera bill-of-materials.

The Digic sensors are made by UMC http://www.umc.com/english/class_300/index.asp, and were designed by Texas Instruments http://www.ti.com/lsds/ti/apps/videovision/end_equipment.page
The memory is most often made by Samsung http://www.samsung.com/global/business/semiconductor/product/consumer-dram/overview
The memory/Digic package-on-package mount has to be outsourced, since Canon cannot do it themselves, and it is believed that UMC does the mounting too.
Other peripheral control and logic chips are TI, Mitsubishi, AD and Fairchild.
.........

Going from a 500nm process to 180nm is like going from a 10MP FF camera to a 70MP camera in one generation. This means some really noticeable strains on the process, especially since the CMOS manufacturing process isn't as forgiving as just "taking pictures". You cant "scale to web size" and sharpen in post when you're making CIS wafers.

What you're basically asking from your equipment is to all of a sudden provide compact camera type linear resolution in a FF lens projection coverage - something almost unthinkable in the normal photographic world. The process has to be precise, to a degree where every single contrast and item on the new 70MP image is equal to or better than the 10MP camera - per pixel. In normal photographic resolution usage, we just want the final output to be good enough, which means that we downsample most images - we seldom deliver full-res images to the customers, and we seldom use full-res images in our own output.
Going from a 350nm mask to 250 and then 180nm and 12" wafers was a BIG step for most CMOS manufacturers, and most other manufacturers are a LOT bigger than Canon in this area.
............

So it's not that Canon COULDN'T do it. Even really small (in the imaging field) firms like STMicro can do it by stitching, and thereby tripling the unit prices. For Leica this isn't really a problem since the total BoM on a M series camera is most certainly lower than 2k USD. This gives a healthy margin up to the projected 7k USD end price point.
This isn't what Canon does. They live on volume, not on extreme margins.

Sure, it's a big step. Which is probably why Canon has held onto their 500nm process for so long, and pushed it to the limits of it's capabilities in recent years. We aren't necessarily talking about Canon making the leap from 500nm to 180nm in one single generation, all or nothing style. As far as I understand Canon already uses 12" (300mm) wafers, so there wouldn't be a need to "switch" to a larger wafer. I gather that they have been producing scanners capable of 90nm manufacture since 2008, so that is not particularly new technology either. Even 180nm processes for CMOS image sensor design are relatively new, and while other manufacturers have already moved much of their CIS design to 180nm, as you stated, it was a VERY COSTLY endeavor that has, in some cases, and very well could, in other cases (Sony?) put them in unstable financial predicaments. I figure Canon will put ONE part on the new process, a low volume part...say, a 46.1mp megapixel monster sensor in a body priced around $5k - $8k a pop? It doesn't even sound like that sucker will be released any time soon, end of next year, possibly later? I wouldn't exactly state that Canon is racing towards a 180nm design and fabrication process for their CMOS sensors at breakneck speed...

All things considered, Canon is not the only IC manufacturer or lithography systems producer or even DSLR company in a bind. Even Sony, the giant it is, is struggling, with numerous bankruptcy rumors floating around. The worlds largest semiconductor manufacturer, Intel, just recently missed estimates and is down in revenue, particularly for larger desktop CPU's as consumers look towards cheaper devices to fulfill their computing needs...Intel may be behind ARM, but that doesn't mean they are inevitably doomed to fail in totality in the near future. Most of all that is thanks to a shitty economy, and not a lack of competitiveness or capability.

Finally, I'd like to know where you get your information. How do you know TI designed the DIGIC sensor, or that UMC manufactured it? According to Canon, Canon themselves designed the DIGIC sensor, and as it stands, I cannot find any explicit information about who actually manufactures them, although Canon generally stands by their "Made In House" mantra for their critical components...sensor and processor. It does not surprise me if the memory chips were made by Samsung, and there is actually some direct evidence for that (link below). Same goes for peripheral control and logic chips, wouldn't be surprising if they were manufactured by other companies, however there is still no definitive information available about who manufactures what that Canon uses in their DSLRs, which makes the explicit nature of your proclamations of "nothing is made by Canon" a bit suspect.

The only piece of evidence I was able to find about DIGIC's design and manufacture was on Wikipedia:

[quote author=Wikipedia]
DIGIC units are made by Canon and used in its own digital imagery products.
[/quote]

I call your claims about the general source of IC's used in Canon cameras into question, particularly regarding DIGIC. Unless you can produce some solid evidence that clearly indicates Canon does not design and manufacture their own DIGIC processors, the facts currently seem to indicate Canon designs and manufactures them. DIGIC itself makes use advanced fabrication processes. This technical analysis of DIGIC 4 clearly demonstrate it used 65nm fabrication technology, and also indicates it does use Samsung memory (not surprising), in a PoP design. At the very least, they certainly seemed to design the DIGIC 4:

Canon Inc unveiled two astonishing features of its in-house designed "Digic 4" image processing LSI for digital cameras to Nikkei Electronics at a new products presentation Sept 17, 2008.

Canon's lithography units are capable of package-on-package and even 3D CMOS manufacture, so I wouldn't be surprised if it was Canon who manufactured the processor and packaged it with Samsung memory.

 

[jrista]

Well, think know how and it takes time to develope a new sensor, nothing any one do in few years.

You are assuming they only just started. Canon demonstrated an extremely high density sensor with 2 micron pixels a couple years ago (the 120mp APS-H) that had on-die image processing (very much like a Sony Exmor). Canon has apparently been developing prototypical 180nm Cu sensors for at last a little while, long enough that they have been thoroughly analyzed by the likes of Chipworks (a painstaking process that takes time and expensive equipment). I suspect Canon started prototyping 180nm sensor design and fabrication a few years ago, demonstrated it with the 120mp APS-H (or a similarly small process...I can't imagine they managed to pack enough circuitry at 0.5 microns such that they could actually produce 2 micron pixels), and have probably continued perfecting the technology since (which seemed to be evidenced by their use of high refractive index light pipes, which is pretty new technology only very recently employed in phone camera sensors.)

 

[Canon-F1]

Now you get your self in a tuff position, is common knowleadge that digit 5 is a texas instrument device
And so also other components in the canon cameras, i think it is time for you to stop thinking so much and controle facts

then it should be no problem for you to present a source?

im more into photography then what company produces what part for my cameras... so no it is not common knowledge i fear. :

 

[Canon-F1]

Well by some of us

so where is the source ?

i looked at the provided links in "thesuede´s" posting but i did not see any infos relating to that.

but i found this about the digic 4:

http://www.canon.com/technology/interview/digic4/digic4_p1.html

Am I right in saying that Canon has always had a policy of developing its image processors in-house, right from when the first digital cameras started being developed? Given that, today, most other camera makers customize processors provided by specialist LSI chip suppliers, what is Canon's stance regarding in-house development now?

Ikeda : I think it's more than that. There's also the pride that we feel at being the world's leading camera manufacturer. Canon has built up a great deal of knowledge with respect to image capture. However, with traditional cameras, camera makers didn't really need to have much to do with camera film or developing. With digital cameras, on the other hand, we are in a position to handle everything ourselves, right up to the "developing" stage. There is no way we are going to just abandon all of the possibilities that this opens up by entrusting it to someone outside the company

so is this a fake development team.... or has canon abandoned it for the digic 5?

if this is common knowledge it must be easy to provide a link to a source for his claim?

The Digic sensors are made by UMC http://www.umc.com/english/class_300/index.asp, and were designed by Texas Instruments

 

[marekjoz]

The main problem as I see it is that Canon don't really have any real development momentum (or budget!) on higher resolution processes. They outsource almost everything except for the larger format sensors.

The biggest difference between the others and Canon is that all the other manufacturers are all dominated by their small-sensor image sensors sales, that already now are manufactured at 90 and 110-130nm metal processes on 300mm wafers. Panasonic and TSMC will start volume shipping of sensors made on 65/45nm rules in Q1 2013. Lower mask resolutions than 130nm are not enough to land you any sales any more. Most cellphone and compact camera sensors are manufactured at those levels now, and have been for the last few years. Also consider the fact that some of the others have very large yearly revenues from logic CMOS processes at 45, 32 and even 22nm levels. All of those markets are areas where Canon totally lack any type of experience. Canon outsource all more advanced fabs on their camera bill-of-materials.

The Digic sensors are made by UMC http://www.umc.com/english/class_300/index.asp, and were designed by Texas Instruments http://www.ti.com/lsds/ti/apps/videovision/end_equipment.page
The memory is most often made by Samsung http://www.samsung.com/global/business/semiconductor/product/consumer-dram/overview
The memory/Digic package-on-package mount has to be outsourced, since Canon cannot do it themselves, and it is believed that UMC does the mounting too.
Other peripheral control and logic chips are TI, Mitsubishi, AD and Fairchild.
.........

Going from a 500nm process to 180nm is like going from a 10MP FF camera to a 70MP camera in one generation. This means some really noticeable strains on the process, especially since the CMOS manufacturing process isn't as forgiving as just "taking pictures". You cant "scale to web size" and sharpen in post when you're making CIS wafers.

What you're basically asking from your equipment is to all of a sudden provide compact camera type linear resolution in a FF lens projection coverage - something almost unthinkable in the normal photographic world. The process has to be precise, to a degree where every single contrast and item on the new 70MP image is equal to or better than the 10MP camera - per pixel. In normal photographic resolution usage, we just want the final output to be good enough, which means that we downsample most images - we seldom deliver full-res images to the customers, and we seldom use full-res images in our own output.
Going from a 350nm mask to 250 and then 180nm and 12" wafers was a BIG step for most CMOS manufacturers, and most other manufacturers are a LOT bigger than Canon in this area.
............

So it's not that Canon COULDN'T do it. Even really small (in the imaging field) firms like STMicro can do it by stitching, and thereby tripling the unit prices. For Leica this isn't really a problem since the total BoM on a M series camera is most certainly lower than 2k USD. This gives a healthy margin up to the projected 7k USD end price point.
This isn't what Canon does. They live on volume, not on extreme margins.

+100 for this info. I was looking for this kind of information and couldn't find anywhere. Appreciate it a lot!

 

[Mick]

All i can say is that when i left making semi conductors in 1998, we were making critical dimentions on 8 inch wafers down to 250. Theory and experiments already had us down to 180. Guess whos steppers we were using? Heres a clue, the companys name begins with a C and they make a lot of cameras. That was 14 years ago. I wont bore you with the technical stuff about how its done etc as thats boring.

 

[Canon-F1]

Heres a clue, the companys name begins with a C and they make a lot of cameras. That was 14 years ago.

casio?