[Digitalis]

The point is that most of us do not switch systems or brands upon the release of every new model that is better than the offerings of another brand

however there are some people who do, I feel nothing but pity for people like that.

On the subject of Silicon wafers- there have been considerable developments in the size of the wafers - back in 2002-2005 8" were the average, being replaced by larger 12" wafers which are commonly used these days. Samsung and Intel are looking at the viability of producing wafers that are 17.7" in size - you can fit a considerable amount of 24X36mm sensors on a wafer that size.

---

Last edited by Digitalis; 02-21-2013 at 07:06 PM.

[Winder]

I've searched for 15 minutes and can not find anything to support your claim.
So if you've something then please link me to it.

http://www.robgalbraith.com/public_files/Canon_Full-Frame_CMOS_White_Paper.pdf

Maybe I don't understand what you are questioning, but this took about 30 seconds to Google. Is this what you are questioning?
" If the sensors are APS-C size, there are about 200 of them on the wafer, depending on layout and the design of the periphery of each sensor. For APS-H, there are about 46 or so. Full-frame sensors? Just 20. Consider, too, that an 8" silicon wafer usually yields 1000 to 2000 LSI (Large-Scale Integrated) circuits. If, say, 20 areas have defects, such as dust or scratches, up to 1980 usable chips remain. With 20 large sensors on a wafer, each sensor is an easy “target.” Damage anywhere ruins the whole sensor. 20 randomly distributed dust and scratch marks could ruin the whole batch. This means that the handling of full-frame sensors during manufacture needs to be obsessively precise, and therefore they are more expensive."

Once you have the wafer you can cut what ever you want out of it. All you need is a mask for the photo-lithography for the sensor you plan on making. In the past FF sensors were produced with 3 separate masking procedures and if the one mask was misaligned then the sensor was trash. Both Sony and Canon have advanced to the point where they can produce a FF sensor with a single mask which greatly improves yields.

The reason APS-H ever came into existence is because it was the largest size that (at the time) could be produced with a single mask.

According to Canon they could yield 200 APS-C sensors, or 46 APS-H sensor, or just 20 FF senors from a standard 8" wafer.

[Ayoh]

Don't you yourself own essentially all of the newest cameras around though? Seems a bit unfair to criticise.

however there are some people who do, I feel nothing but pity for people like that.

On the subject of Silicon wafers- there have been considerable developments in the size of the wafers - back in 2002-2005 8" were the average, being replaced by larger 12" wafers which are commonly used these days. Samsung and Intel are looking at the viability of producing wafers that are 17.7" in size - you can fit a considerable amount of 24X36mm sensors on a wafer that size.

[mikemike]

however there are some people who do, I feel nothing but pity for people like that.

On the subject of Silicon wafers- there have been considerable developments in the size of the wafers - back in 2002-2005 8" were the average, being replaced by larger 12" wafers which are commonly used these days. Samsung and Intel are looking at the viability of producing wafers that are 17.7" in size - you can fit a considerable amount of 24X36mm sensors on a wafer that size.

Intel just had the first prototype of a 450 mm roll out this year. The camera sensor industry is no intel. A quick search at what kind of fab tech they use and I saw that Canon just started using a 180 nm process in 2012, for reference Intel was doing using that process in 1999. If camera makers used Intel's fabs contemporary processes, we could probably have a gigapixel APS-C or maybe a terapixel 645 camera.

[Anvh]

Maybe I don't understand what you are questioning, but this took about 30 seconds to Google. Is this what you are questioning?
" If the sensors are APS-C size, there are about 200 of them on the wafer, depending on layout and the design of the periphery of each sensor. For APS-H, there are about 46 or so. Full-frame sensors? Just 20. Consider, too, that an 8" silicon wafer usually yields 1000 to 2000 LSI (Large-Scale Integrated) circuits. If, say, 20 areas have defects, such as dust or scratches, up to 1980 usable chips remain. With 20 large sensors on a wafer, each sensor is an easy “target.” Damage anywhere ruins the whole sensor. 20 randomly distributed dust and scratch marks could ruin the whole batch. This means that the handling of full-frame sensors during manufacture needs to be obsessively precise, and therefore they are more expensive."

Once you have the wafer you can cut what ever you want out of it.

It can be me but i'm not reading you can simply cut anything out of it that you want.
If you want to cut 20 large sensor like in your example, then why not cut smaller sensors around the damaged bits?

[Winder]

It can be me but i'm not reading you can simply cut anything out of it that you want.
If you want to cut 20 large sensor like in your example, then why not cut smaller sensors around the damaged bits?

Did you read the Canon white paper I linked. I think page 11 is what you are looking for.

[Class A]

Did you read the Canon white paper I linked.

Note that this white paper in general is a big set of lies intended to rationalise the premium Canon has been charging for FF cameras for years.

Falconeye tore it apart quite a while ago.

---

Last edited by Class A; 02-21-2013 at 11:15 PM.

[Winder]

Note that this white paper in general is a big set of lies intended to rationalise the premium Canon has been charging for FF cameras for years.

Falconeye tore it apart quite a while ago.

Yes. I have read the threads. It still discusses the process of turning a wafer into a batch of sensors. Regardless of cost or yield.

[Kunzite]

It can be me but i'm not reading you can simply cut anything out of it that you want.
If you want to cut 20 large sensor like in your example, then why not cut smaller sensors around the damaged bits?

ICs are not film sheets, obviously you cannot cut how you want.
You can decide to get whatever sensor sizes you want from a wafer, within the equipment's constraints, even mixing them. I guess that's what Winder was talking about? But once the process starts, there is no turning back: cutting the wafer differently than the pre-chosen layout would only result in useless wafer pieces.

[Digitalis]

cutting the wafer differently than the pre-chosen layout would only result in useless wafer pieces

Exactly, and there are always a few dud sensors on the wafer because of impurities in the silicon - and with bigger wafers this is going the be more problematic, not to mention the costs of the machines designed to handle and process the bigger wafers.

[Kunzite]

Bigger wafers are more economical and don't generate higher defects rates AFAIK, but yes, the fab must be able to handle them.

[Winder]

The cost of the really pure polysilicon needed to produce wafers has fallen drastically and the quality of the material has improved. This leads to fewer defects in the wafer and has improved yields.

They make solar panels out of the same material. Has anyone noticed all of the solar panel plants that have gone out of business over the past few year? Some were pretty high profile here in the USA. These plants were developing solar panels from alternative materials. EPA regulations have forced production of polysilicon and other rare earth material overseas. Companies had been working on making solar panels out of other materials. 2-3 years ago a couple of Chinese plants came online and have increased production of really pure polysilicon and the prices have come crashing back down. Plants working on alternative materials can't compete with flood of low cost polysilicon and many have gone out of business. This is also why our government imposed a 30% tax consumers who buy solar products.

This same flood of polysilicon is driving down the cost of full frame sensors and other goods that require high quality polysilicon. I think we are going to see more companies enter into development and production of larger sensors. Toshiba is now providing sensor for Nikon. We should also see an improvement in the quality of sensors.

[falconeye]

According to Canon they could yield 200 APS-C sensors, or 46 APS-H sensor, or just 20 FF senors from a standard 8" wafer.

Note that this white paper in general is a big set of lies intended to rationalise the premium Canon has been charging for FF cameras for years.

Falconeye tore it apart quite a while ago.

Class A, thank's for the quote

I don't remember the post, but somewhere here in the forum I provuded the correct number of FF chips from an 8" wafer, including the wafer placement constraints. It was 66, I believe. And it was a bit less than 200 for APS-C, I believe.

At the time, I made it to predict the end of the APS-C dSLRs in the semi-pro segment. As that has happened now (the D7100 being no exception, positioned by Nikon as consumer camera), no need to argue anymore.

[falconeye]

Wrt K-3 speculation, I wonder if the K-3 would compete with D7100 or rather the D5200.

Obviously, the D7100 is prepared for the battle with its removal of Bayer-AA filter. But how about Pentax to fight with that rather capable 51 point Multi-CAM 3500DX AF module?

Will be interesting times

[Kunzite]

At the time, I made it to predict the end of the APS-C dSLRs in the semi-pro segment. As that has happened now (the D7100 being no exception, positioned by Nikon as consumer camera), no need to argue anymore.

Well, there are the rumored D400, 7DMkII and K-3, so... maybe not yet?