Luminous Landscape Forum
Equipment & Techniques => Medium Format / Film / Digital Backs – and Large Sensor Photography => Topic started by: pcunite on August 31, 2010, 08:47:25 am
-
Now that Canon has developed an 8x8 inch CMOS sensor all we need is for this technology to be made available to us in various sizes. Along with LiveView, which is a given in the CMOS world, this would truly revolutionize high end photography. 35mm will always be around for its nimbleness but 4x5 just got put on Canon's road map.
-
Now that Canon has developed an 8x8 inch CMOS sensor all we need is for this technology to be made available to us in various sizes. Along with LiveView, which is a given in the CMOS world, this would truly revolutionize high end photography. 35mm will always be around for its nimbleness but 4x5 just got put on Canon's road map.
Link: http://www.dpreview.com/news/1008/10083101canonlargestsensor.asp
-
S...t and I just sold my Sinar P2 8x10. ;)
Canon rules!
-
Speculation about global shutter:
http://www.eoshd.com/content/326-Have-Canon-made-global-shutter-breakthrough-with-experimental-sensor
-
Finally! A Canon thread which actually belongs in this forum ;D
-
ASTRONOMY! ASTRONOMY! ASTRONOMY!
...was my first heart-pounding reaction on seeing this. That was even before I got to the quote from Canon Japan: "Potential applications for the new high-sensitivity CMOS sensor include the video recording of stars in the night sky and nocturnal animal behavior." Normal video rate astro-imaging is not often useful to the professional astronomer - when we do need high time resolution imaging e.g. for pulsars, we need millisecond or better - but the huge area and sensitivity are. And one benefit of the video rate is that the observing duty cycle, defined as [exposure time / [exposure time + readout time]], will be close to 100% even on brighter objects where it usually suffers in astronomy - as the CCDs are readout slowly (seconds to minutes) to lower the readnoise.
This technology could overtake the giant CCD sensor mosaics being constructed for the likes of the LSST: http://www.lsst.org/lsst/science/concept_camera (http://www.lsst.org/lsst/science/concept_camera). Especially if Canon can deliver their trademark low readnoise (much better than CCDs) on a platform that big.
-
When will they make a budget 5 * 5" 10 micron sensor (on a Sinar board) for use with large format film lenses?
-
<< ASTRONOMY! ASTRONOMY! ASTRONOMY! >>
The price of such a sensor is certainly astronomical.
PdF
-
Very interesting!
And yes the price... would have to be at least 40 times the cost of the 5D2 sensor since that's what they can get from the same wafer.
-
Very interesting!
And yes the price... would have to be at least 40 times the cost of the 5D2 sensor since that's what they can get from the same wafer.
Maybe, but a 4x5 would only be about 20 times, so about $20K... I have no idea, just guessing here of course.
-
Maybe, but a 4x5 would only be about 20 times, so about $20K... I have no idea, just guessing here of course.
All of it's a gross simplification - who knows what the yield will be, but with bigger chips, they will have to toss a higher percentage that's for sure. Plus who knows the market and other costs for this one?
All aside, if they can go big that is super news. I'm not as keen on a 8x10 as I would be a 56mm x 56mm chip for my Rollei.
-
<< ASTRONOMY! ASTRONOMY! ASTRONOMY! >>
The price of such a sensor is certainly astronomical.
The price of a single piece of silicon that large is still ridiculously cheap, as I mentioned before. We are talking about roughly 50 US$ per 200m wafer... you might need 10 of then to get one perfect one, but even that is doubtful as the processes have been perfected to incredible levels.
So a large sensor might be expensive to product but silicon is a small contributor to the final price of the device, although some vendors have obviously gone to great lenght to make us feel otherwise.
Cheers,
Bernard
-
What a paradox!
Silicon is the most common element on earth.
Eduardo
The price of a single piece of silicon that large is still ridiculously cheap, as I mentioned before. We are talking about roughly 50 US$ per 200m wafer... you might need 10 of then to get one perfect one, but even that is doubtful as the processes have been perfected to incredible levels.
So a large sensor might be expensive to product but silicon is a small contributor to the final price of the device, although some vendors have obviously gone to great lenght to make us feel otherwise.
Cheers,
Bernard
-
What a paradox!
Silicon is the most common element on earth.
I would speak of paradox if these wafers were expensive, but in fact they are not. ???
Cheers,
Bernard
-
What about the bandwidth limitation for communication between camera back and monitor?
Do we need to replace firewire 8oo with something faster, or is the limitation in the camera computer?
Perhaps Dr Roadknight can enlighten us.
John
-
What about the bandwidth limitation for communication between camera back and monitor?
Do we need to replace firewire 8oo with something faster, or is the limitation in the camera computer?
Perhaps Dr Roadknight can enlighten us.
John
I imagine it would have to be directly connected by fibre channel, like in supercomputers.
-
What about the bandwidth limitation for communication between camera back and monitor?
Do we need to replace firewire 8oo with something faster, or is the limitation in the camera computer?
Perhaps Mr Roadnight can enlighten us.
John
I do not know.
"Bandwidth" means data transfer rate, and it is the data transfer rate between camera and computer that would be of concern.
If you knew the video frames-per-second and file size per frame, and the actual "real world" data transfer rate of firewire 800, it should be simple arithmetic.
The file format for the transfer from back to camera would probably be a compressed raw format similar to Hasselblad's 3FR,,, and then you would have to do similar calculations with the audio signal.
...and I have no PhD or MD, and there is no "k" in my name.
-
The price of a single piece of silicon that large is still ridiculously cheap, as I mentioned before. We are talking about roughly 50 US$ per 200m wafer ...
The price of the blank wafer is clearly a very small part of the sales price of the sensors made on them. That 200mm wafer would hold about 20 36x24mm sensors, for a silicon cost of about $2.50/sensor, way under 1% of the price of such sensors.
... you might need 10 of then to get one perfect one, but even that is doubtful as the processes have been perfected to incredible levels.
Do you have any basis whatsoever for that one in ten figure? We are talking about needing
1. free of fatal defects over almost the entire wafer, whereas with most chip fab., each defect only causes one of many chips to be rejected.
2. massive stitching; maybe 100 fields or so, and thus greatly increased opportunities for errors there, and complexity (repeated careful alignment after moving the stepper to each new field) in the process.
Given that puny little 44x33mm sensors cost several thousand dollars, and the yield with clearly be lower for this far larger size, we can safely scale by a factor of well over the area ratio (about 30), getting well past $100,000 I would say.
P. S. Dalsa already offers CCD sensors of about this size on a custom basis: the Canon claim is mostly big CMOS. No camera maker has ordered a 5"x4" sensor from Dalsa (or even 6x7 or 6x6 or full 645 size) again hinting that cost is a major barrier.
P. P. S. It is probably 10MP or less, going by the hint that photosite area is 100 times greater than on Canon 22MP, 36x24mm sensor.
-
Do you have any basis whatsoever for that one in ten figure? We are talking about needing
1. free of fatal defects over almost the entire wafer, whereas with most chip fab., each defect only causes one of many chips to be rejected.
2. massive stitching; maybe 100 fields or so, and thus greatly increased opportunities for errors there, and complexity (repeated careful alignment after moving the stepper to each new field) in the process.
Given that puny little 44x33mm sensors cost several thousand dollars, and the yield with clearly be lower for this far larger size, we can safely scale by a factor of well over the area ratio (about 30), getting well past $100,000 I would say.
P. S. Dalsa already offers CCD sensors of about this size on a custom basis: the Canon claim is mostly big CMOS. No camera maker has ordered a 5"x4" sensor from Dalsa (or even 6x7 or 6x6 or full 645 size) again hinting that cost is a major barrier.
Well, your guess is a good as mine regarding how many wafer would be required to produce one of these.
Now that I think aout it, there is almost no cost involved in fact. It would be fairly easy for a company like Canon to identify in advance those perfect wafers and to use these for large chips, while using non perfect wafers for smaller chips. From then on it is just a matter of process control to make sure that at least some wafers are perfect.
This is not a genius idea. Intel does the same with their CPUs. Higher Mhz units are often perfectly identical to lower Mhz ones, they just pass the tests better. :)
This approach would only cost as much as the process to identify the perfect wafers from the others. A trivial cost amidst a large series.
So my prerogative is that getting perfect 200m wafers suitable to large sensors has basically no cost if you produce enough smaller sensors a year.
As far as stitching is concerned, sensors are pretty low tech devices in terms of lithography, and I don't expect stitching to have particularly low yields even if a large amount of passes are required.
So I am not convinced that all the talk about large sensor costs is not mostly driven by a marketing urge to sell the resulting products at undeservely high prices. This, again is the same with CPUs, where the very same design rated at a higher Mhz is selling at 3 times the price mostly because we are being told by Intel/AMD it is a better part.
Cheers,
Bernard
-
From D.Roadnight "I have no PhD or MD, and there is no "k" in my name."
Apologies. I should be more careful first thing in the morning!
John
-
Now that I think about it, there is almost no cost involved in fact. It would be fairly easy for a company like Canon to identify in advance those perfect wafers and to use these for large chips, while using non perfect wafers for smaller chips. From then on it is just a matter of process control to make sure that at least some wafers are perfect.
This is not a genius idea. Intel does the same with their CPUs.
I am tempted to suggest that the testing for defects and such is done electronically on chips after they are fabricated, not by optical examination of blank wafers in advance. But what does either of us really know about the details?
What we do know is that several companies compete in the market for large sensors (most famously Kodak and Dalsa) and so if it really were as easy as you say, and if we avoid internet cynicism about industry leading companies being more ignorant thus us internet pundits about how to reduce costs or increase quality, or conspiracy theories about Kodak and Dalsa fixing prices and in the price strangling the DMF market to their mutual detriment, then one or more of these competitors would have done what you suggest, and reduced its costs substantially (they both make lots of smaller CCDs too, so your idea fits their situation), and then reduced their prices in order to gain market share, and so on.
The actual evidence of MF sensor competitive market pricing instead indicates that with the current state of the art, the costs of MF sensor design and fabrication is such that the price to customers for even a 44x33mm sensor is well over US$1000, and prices will scale at least in proportion to area. More according to every source I have read on chip yield.
And in case anyone suggest it, large CMOS sensors are not cheaper than equally large CCDs to make; that is an internet myth, based on falsely extrapolating the cost savings for tiny "camera on a chip" cell-phone and web-cam modules. Note well that the last remaining DSLR CCD's are clustered in the cheapest models!
Short version: I do not believe any internet forum proposal about dramatically improving price/performance beyond what competitive market forces are currently producing unless it is backed by evidence of a technological break-through that it too new to have been commercialized yet.
-
What we do know is that several companies compete in the market for large sensors (most famously Kodak and Dalsa) and so if it really were as easy as you say, and if we avoid internet cynicism about industry leading companies being more ignorant thus us internet pundits about how to reduce costs or increase quality, or conspiracy theories about Kodak and Dalsa fixing prices and in the price strangling the DMF market to their mutual detriment, then one or more of these competitors would have done what you suggest, and reduced its costs substantially (they both make lots of smaller CCDs too, so your idea fits their situation), and then reduced their prices in order to gain market share, and so on.
The actual evidence of MF sensor competitive market pricing instead indicates that with the current state of the art, the costs of MF sensor design and fabrication is such that the price to customers for even a 44x33mm sensor is well over US$1000, and prices will scale at least in proportion to area. More according to every source I have read on chip yield.
Short version: I do not believe any internet forum proposal about dramatically improving price/performance beyond what competitive market forces are currently producing unless it is backed by evidence of a technological break-through that it too new to have been commercialized yet.
BJL,
Neither Kodak nor Dalsa produce many imaging chips, even without assuming price fixing (which is most probably happening by the way), they wouldn't be able to deploy the strategy I am suggesting. In the world of consumer electronics they are tiny tiny players compared to Canon, Sony and now probably Nikon if rumors are to be believed about Nikon producing the D3100 chip themselves. Any series less than 1.000.000 parts is not significant in the grand scheme of things.
What I am saying is that producing large chips would have no significant cost for a company producing many small chips.
Cheers,
Bernard
-
Neither Kodak nor Dalsa produce many imaging chips, even without assuming price fixing (which is most probably happening by the way), they wouldn't be able to deploy the strategy I am suggesting.
Bernard,
if you scan this product list
http://www.kodak.com:80/global/en/business/ISS/Products/index.jhtml?pq-path=11937/11938
you will see a great variety of smaller CCDs from Kodak, so as I indicated in my previous post, it seems fairly clear that Kodak does make a great many more of those small sensors than it does MF size ones. (There is a lot more to the CCD market than consumer cameras).
But perhaps I should revive my previous comment that, as far as I know, there is no evidence that your idea of testing wafers before deciding which ones to use for big sensors is used, or is usable: all the testing I know of is based on the electronic performance of semi-conductor devices once they have been fabricated. That is true in your example of processors, which are "speed rated" after testing, and Kodak sells evaluation boards for testing some of its sensors, like this one:
http://www.kodak.com:80/global/plugins/acrobat/en/business/ISS/eval/MTD-PS-0881_KAF-39000_User_Manual.pdf
Do you have any evidence that your wafer choice by visual inspection of blank wafers is done, or doable?
-
Bernard,
if you scan this product list
http://www.kodak.com:80/global/en/business/ISS/Products/index.jhtml?pq-path=11937/11938
you will see a great variety of smaller CCDs from Kodak, so as I indicated in my previous post, it seems fairly clear that Kodak does make a great many more of those small sensors than it does MF size ones. (There is a lot more to the CCD market than consumer cameras).
Does Kodak manufacture the sensors of their compacts, or do they use Sony parts? I was under the impression that they were using Sony parts but I could be wrong in which case my nice theory would become somewhat less credible indeed. :)
But perhaps I should revive my previous comment that, as far as I know, there is no evidence that your idea of testing wafers before deciding which ones to use for big sensors is used, or is usable: all the testing I know of is based on the electronic performance of semi-conductor devices once they have been fabricated. That is true in your example of processors, which are "speed rated" after testing, and Kodak sells evaluation boards for testing some of its sensors, like this one:
http://www.kodak.com:80/global/plugins/acrobat/en/business/ISS/eval/MTD-PS-0881_KAF-39000_User_Manual.pdf
Do you have any evidence that your wafer choice by visual inspection of blank wafers is done, or doable?
Nope, but I'll ask some contacts knowledgeable in the field.
If this cannot be done yet it does for sure look like a tremendous business opportinity if you are right about the current extremely high costs of perfect wafers. Jee, do I hate this forum software.
Cheers,
Bernard
-
I've been told the Kodak chip used in the Pentax 645D costs $1000 in quantity.
-
Does Kodak manufacture the sensors of their compacts, or do they use Sony parts?
That question is irrelevant to your idea about using lesser wafers for less critical needs like smaller sensors; all that matters is that Kodak has a use for lots of those lesser wafers, which it does because it make lots of small CCD's, as indicated by its product list.
Kodak gives some details of where its sensors are used at http://www.kodak.com/global/en/business/ISS/Applications/
where the headings are, in order, "industrial, medical, photography, scientific and security"; photography only ranks third place! As I said previously, there is a wide range of markets for CCDs other than consumer cameras.
Let me know what you find out about wafer/chip inspection!
-
That question is irrelevant to your idea about using lesser wafers for less critical needs like smaller sensors; all that matters is that Kodak has a use for lots of those lesser wafers, which it does because it make lots of small CCD's, as indicated by its product list.
Kodak gives some details of where its sensors are used at http://www.kodak.com/global/en/business/ISS/Applications/
where the headings are, in order, "industrial, medical, photography, scientific and security"; photography only ranks third place! As I said previously, there is a wide range of markets for CCDs other than consumer cameras.
There is a difference between selling a wide array of speciality products and producing many parts.
I'll let you know what I find out.
Cheers,
Bernard
-
There is a difference between selling a wide array of speciality products and producing many parts.
What matters for your "wafer pre-selection" idea is that the larger sensors are a small enough proportion of total production, which only requires that Kodak's very numerous smaller CCDs outsell it larger CCDs by a sufficiently large ratio. I do not have the numbers, but that seems a very good bet: when you look through the various applications described and sensors suggested for each, it looks as if most of the work would be handled by the smaller, far more affordable sensor models.
Look at it this way: Kodak cannot be selling more than about 5,000 sensor for MF back each year, given that the whole MF back market is only about that size, and Dalsa has a healthy share of it. It needs to be selling many hundreds of times as many of its smaller sensors for that smaller sensor business to have vaguely competitive economies of scale. (Total annual sensor sales are well over one billion!) And that means many wafers full of small sensors for every one "perfect" wafer needed for your high yield scheme.
-
What matters for your "wafer pre-selection" idea is that the larger sensors are a small enough proportion of total production, which only requires that Kodak's very numerous smaller CCDs outsell it larger CCDs by a sufficiently large ratio. I do not have the numbers, but that seems a very good bet: when you look through the various applications described and sensors suggested for each, it looks as if most of the work would be handled by the smaller, far more affordable sensor models.
Look at it this way: Kodak cannot be selling more than about 5,000 sensor for MF back each year, given that the whole MF back market is only about that size, and Dalsa has a healthy share of it. It needs to be selling many hundreds of times as many of its smaller sensors for that smaller sensor business to have vaguely competitive economies of scale. (Total annual sensor sales are well over one billion!) And that means many wafers full of small sensors for every one "perfect" wafer needed for your high yield scheme.
Yep, that is indeed correct. Now it is probably easier to implement when you produce many sensors of the same type instead of having a large portfolio of small series items that might not always fully use full wafers. But anyway, the idea remains applicable if you move enough wafers around.
Cheers,
Bernard
-
Bernard and JBL,
Just a few comments to the ongoing discussion..
1) Michael Reichmann had an interview with Claus Molgaard, (Phase One VP R&D) and also with Jesper Sandager, VP Manufacturing - Phase One, LLVJ 13 & 14. I got the impression that Phase One had a significant reject rate on sensors (from Kodak). Also I got the impression that many of the sensors within specs still could have dead pixel columns, which were taken care of in postprocessing. This is from memory, I have not re-viewed the videos.
2) I guess that cost is not that much dependent silicon wafer prices. The process is applied to the wafer and the chip is tested. So it's not the wafers that are rejected but the complete chip.
At that time Phase indicated that the sensor cost was about half of the cost of the back. Now, cost and value is not the same. And that was a few years ago.
Best regards
Erik