5Dii sensor recycled into new FF entry body?

[psycho5]

I was wondering if Canon would ever do such a thing or would the 1Dx sensor make more sense? Economies of scale already tells Canon to make the same APS-C sensor for every crop body, so why should the FF camp be the exception to the rule?

Logically, Canon would only have to produce 3 new DSLR sensors:

18mp FF (1Dx)
36mp FF (5D3)
XXmp APS-C (7D, 70D, t4i)

If canon will eventually bring us an enthusiast FF, the 1Dx sensor or the old 5Dii 21mp sensor has gotta be the most cost effective choice, because why spend more money on development?

This is all hypothetical and fun to do, but what does your crystal ball predict???

 

[dr croubie]

I was wondering if Canon would ever do such a thing or would the 1Dx sensor make more sense? Economies of scale already tells Canon to make the same APS-C sensor for every crop body, so why should the FF camp be the exception to the rule?

Logically, Canon would only have to produce 3 new DSLR sensors:

18mp FF (1Dx)
36mp FF (5D3)
XXmp APS-C (7D, 70D, t4i)

If canon will eventually bring us an enthusiast FF, the 1Dx sensor or the old 5Dii 21mp sensor has gotta be the most cost effective choice, because why spend more money on development?

This is all hypothetical and fun to do, but what does your crystal ball predict???

Once the 1DX is out, and the future (2D/3D/4D/5Dmk3/6D) (24/28/32/36)MP camera is out at some point mid-next year, they can recycle the current 5D2/1Ds3 sensor into a lower-end, cheaper body with less sealing, cheaper AF, low fps, and they can call it...
...
...a 2nd-hand 5Dmk2.
I'd buy one.

 

[Mt Spokane Photography]

I hope not, its a very old sensor design by now and they can do better without increasing the cost to manufacture it. The design is probably at least five years old. The sensor first appeared in the 1Ds MK III over four years ago in August 2007.

 

[gmrza]

I hope not, its a very old sensor design by now and they can do better without increasing the cost to manufacture it. The design is probably at least five years old. The sensor first appeared in the 1Ds MK III over four years ago in August 2007.

It is not really relevant to the discussion, but my understanding is that the 1DsIII and 5DII use slightly different sensors. Your point about the old sensor design holds for both however. By now, a newer design can probably be manufactured more efficiently/cheaply, and give better performance.
Given where the pricing of the 5DII is currently headed, it gives an indication that Canon may have space for an 18-20MP mid range full frame body, and a 30+ MP body as well. I would see it as more likely that Canon would use a "detuned" version of the 1Dx sensor. That would almost be like the way car manufacturers use one engine block and then add various combinations of turbochargers, superchargers, intercoolers etc. to create engines with different outputs - just look at the VW Golf range as an example.

 

[psycho5]

Given where the pricing of the 5DII is currently headed, it gives an indication that Canon may have space for an 18-20MP mid range full frame body, and a 30+ MP body as well. I would see it as more likely that Canon would use a "detuned" version of the 1Dx sensor. That would almost be like the way car manufacturers use one engine block and then add various combinations of turbochargers, superchargers, intercoolers etc. to create engines with different outputs - just look at the VW Golf range as an example.

I hope not... sure Canon can strip the expanded iso from the 1Dx sensor, but other than that I dont see anything else they can strip from it. The sensor between the 7D and t2i are no different. The turbochargers you refer to are things like full body weather sealing, 12-14 fps, ethernet, etc... maybe even axe one of the three digic processors, but the sensor is the sensor.

Even if Canon did limit the iso to a range less than the 1Dx, it comes back with a "magic lantern" update.

 

[Rocky]

Given where the pricing of the 5DII is currently headed, it gives an indication that Canon may have space for an 18-20MP mid range full frame body, and a 30+ MP body as well. I would see it as more likely that Canon would use a "detuned" version of the 1Dx sensor. That would almost be like the way car manufacturers use one engine block and then add various combinations of turbochargers, superchargers, intercoolers etc. to create engines with different outputs - just look at the VW Golf range as an example.

From a manufacturing stand point, It is cheaper to make the same sensor for 1DX and another FF body. Canon can make a cheaper body with the same sensor by puting less features and lower level of AF, give it ONE Digic 5, platic body, no built-in vertical grip etc.

 

[traveller]

I was wondering if Canon would ever do such a thing or would the 1Dx sensor make more sense? Economies of scale already tells Canon to make the same APS-C sensor for every crop body, so why should the FF camp be the exception to the rule?
,....

The FF camp is the exception to the rule because the number of sensors they get from each wafer is substantially smaller than the number of APS-C sensors. And that's just the raw numbers. Then you've got to allow for fabrication errors, meaning that you will never have a 100% yield.

So whilst the R&D for the sensor may evaporate over time, the actual manufacturing cost does not. Originally the machines they used to "print" the sensors on wafers (steppers) needed to do two passes for FF sensors but only one for APS-C (another source of increased manufacturing cost relative to APS-C.) I'm not sure if this has yet changed.

From Northlight Images (hope you don't mind Keith!) [http://www.northlight-images.co.uk/canon_medium_format_2ff.html]:

"2007 August 18th Canon can now make 35mm full frame chips in one step - and can make larger ones too - more info on the 50D page from a lengthy email we were sent.

"Sensor technology - Canon have mentioned that full frame sensor chips need two stepper passes. Sony now has the technology to do this in one pass and are preparing their own sensors and selling to them Nikon. However, Canon have a new 'one pass' technology too. What's more, it can do even larger than 35mm (36x24mm) in a single pass.
Canon is very excited about a next generation CMOS sensors they are working on. Two full frame versions have 40M and 50M pixels at the -same- noise level as the current 1D series. Low power supply voltages give cooler chips and lower noise even with smaller pixels. Different circuit fabrication techniques also promise to make close to 100% of the sensor area active pixels. This technology won't make any of the upcoming models, but is expected in 2009/10." "

Can't comment on whether this is true...

 

[Haydn1971]

Mindful also of the age of the 21mpx FF sensor, I'm of the view that the 18mpx FF sensor will be pushed into a cheaper body, but also that some photographers are also very keen to see a higher pixel count. My current view is as follows...

1Dx - FF 18mpx Sports and photojournalist
5DIII - FF 35-40mpx studio / landscape camera - 2013 launch perhaps
7DII - APS-C 28-32mpx top end crop - 2013 launch perhaps
9D - FF entry level 18mpx - helps keep cost of 1Dx sensor down - early 2012 launch
70D - APS-C 28-32mpx prosmer small body - 2 year cycle 2012 launch
650D - APS-C stick with current 18mpx crop sensor, pushing up a few mpx with the 700D and so on
1100D as is now, 1200D drops on whichever sensor canon deems cheap to manufacture, 15mpx, perhaps even 18mpx
G1x APS-C mirror less solution using EF-S lenses - 2012 perhaps
G100x compact mirror less solution using bespoke EF-G lenses, 1/1.6" perhaps ? - 2013 perhaps
G1000x compact mirror less solution using fixed zoom lens
Partner with a phone manufacturer for Canon lens/sensor on a phone - 2012/2013 ?

 

[Rocky]

But being able to "print" FF sensors in a single pass rather than two doesn't increase the yield from a single wafer.

Single pass will finitely reduce the cost. just imagine now canon cando 10 stepping instead of 20. You will get the picture. Also there will be a yield inrease also due to cutting the alignment error into half , and less handling of the wafers.

 

[Mt Spokane Photography]

I hope not, its a very old sensor design by now and they can do better without increasing the cost to manufacture it. The design is probably at least five years old. The sensor first appeared in the 1Ds MK III over four years ago in August 2007.

It is not really relevant to the discussion, but my understanding is that the 1DsIII and 5DII use slightly different sensors. Your point about the old sensor design holds for both however. By now, a newer design can probably be manufactured more efficiently/cheaply, and give better performance.
Given where the pricing of the 5DII is currently headed, it gives an indication that Canon may have space for an 18-20MP mid range full frame body, and a 30+ MP body as well. I would see it as more likely that Canon would use a "detuned" version of the 1Dx sensor. That would almost be like the way car manufacturers use one engine block and then add various combinations of turbochargers, superchargers, intercoolers etc. to create engines with different outputs - just look at the VW Golf range as an example.

The silicone is the same, the AA filter and Bayer filters that are placed over the silicone are slightly different, so , in that sense, the finished sensor is slightly different.

I believe that Canon went from 8 in wafers to 12 in wafers which increased the FF production efficiency and speed as well. I've seen mention of using 12 in wafersa in Canon press releases, but no details or date of implementstion, but it was likely needed to mass produce FF sensors for the 5D MK II.

 

[traveller]

Mindful also of the age of the 21mpx FF sensor, I'm of the view that the 18mpx FF sensor will be pushed into a cheaper body, but also that some photographers are also very keen to see a higher pixel count. My current view is as follows...

1Dx - FF 18mpx Sports and photojournalist
5DIII - FF 35-40mpx studio / landscape camera - 2013 launch perhaps
7DII - APS-C 28-32mpx top end crop - 2013 launch perhaps
9D - FF entry level 18mpx - helps keep cost of 1Dx sensor down - early 2012 launch
70D - APS-C 28-32mpx prosmer small body - 2 year cycle 2012 launch
650D - APS-C stick with current 18mpx crop sensor, pushing up a few mpx with the 700D and so on
1100D as is now, 1200D drops on whichever sensor canon deems cheap to manufacture, 15mpx, perhaps even 18mpx
G1x APS-C mirror less solution using EF-S lenses - 2012 perhaps
G100x compact mirror less solution using bespoke EF-G lenses, 1/1.6" perhaps ? - 2013 perhaps
G1000x compact mirror less solution using fixed zoom lens
Partner with a phone manufacturer for Canon lens/sensor on a phone - 2012/2013 ?

I'm not sure if Canon could push the APS-C sensor to 28-32MP for two reasons:

1) The Sony APS-C sensor is already struggling with high ISO noise at 24MP and it's slightly larger
2) In marketing terms what would a 28-32MP sensor say about the 18MP sensor in the 1D X?

I think that sticking at 18MP but improving the noise performance and dynamic range might be a better option (and would fit better with Canon's current 'quality not quantity' megapixel marketing message). Whether this will happen or not is another question; it might depend on how much traction the A77 gets and what Canon think that Nikon will do with the D400.

I like your ideas with regard to the compact camera lineup, but I think that the 'G1X' name is a bit too close to Panasonic's GX1! As for your idea about Canon partnering with a 'phone manufacturer -spot on... In fact, I'd go further: they should buy a 'phone manufacturer.

 

[Rocky]

The real key would be if there was a significant drop in the amount of silicon wasted in the 12" vs 8" wafers as it also stands to reason that the 12" wafers are going to be more expensive than the 8" wafers (a 12" wafer is 225% of the area of an 8" wafer.)

There are 22 FF potential site on an 8 in wafer and 56 potential site on the 12 in wafer. The major (90% +/-) cost of the finished wafer is in the processing, not the starting wafer. So there is 245 % gain in potential site between the 12 in and 8 in wafer. Let us assume that the cost of 12 in finish wafer is 150% of the 8 in wafer (I am being generous here, It can be as low as 125%, depends on the wafer fab). Then the FF sensor cost from the 12 in wafer will be 60% of the 8 in wafer.