Reading mode: Light Dark

FF sensors expensive to make? Think twice

Go to last post
afterburn said:
Canon states that zero pixel defects are allowed on their sensors. If
a single pixel is out in your camera, they will replace the sensor
free of charge regardless of warranty status.
Click to expand...
ROFL!!!

You obviously haven't been reading the defective pixel mapping threads.
 
Thanks for that little insight in how it really works, thats very interesting.

I wonder if it is possible to see the stitching process in RAW pictures of older Canon. Commenter cited this on Chipworks:
http://www.brisk.org.uk/photog/d3greensat.html

So with some effort it seems to be possible to reveal stitching in D3 pictures (not relevant in normal photos).
Its RKM said:
afterburn said:
That's pretty cool (it is), but the production cost of a wafer is
about the same whatever you put on it.
Click to expand...
Only if the wafer goes through the same number of process steps, in
the same foundry machines. The whole point about the OPs message is
that to date an FF sensor could only be made using a 3-step reticle
stitch, increasing the number of process steps for the wafer,
reducing yields and significantly impacting costs. Reticle stitching
is not required for APS sensors using stepper kit available in most
fabs. If Canon have, and there is limited evidence of this,
eliminated the stitch process then they will reduce their costs to
levels commensurate with conventional chip yield estimates.

Your earlier point is quite correct, the larger the device the fewer
you can get on each wafer and the higher the probability of a
catastrophic defect being included in any device. This determines
yield and hence, given fixed wafer process costs, the device cost.
However, reticle stitching changes that model significantly - wafer
processing costs increase and yields reduce, both due to the
additional number of processes being implemented for reticle
stitching.

I have been involved in several designs of sensors requiring reticle
stitching and I can assure you it has a major impact on yield, and
hence cost, over and above the area effects that you refer to. On a
single stitched device, depending on the level of circuit complexity
in the stitch region, you can expect yield to fall by as much as 50%
(I have seen worse, but that was atypical), while process costs
increase by around 20%. With a 3 stitch process I would not be
surprised to see yields fall and process costs increase further.
That accounts for a significant cost hike over and above the area
effects that you are considering.

The OP suggests that Canon may have eliminated the stitching process
for FF sensors using their new stepper and hence brought the costs of
FF sensors into line with the predictions that your simplified view
of life suggests.

The linked article states that recent devices on 1Ds cameras show no
visible stitching artefacts (which are usually pretty easy to spot
visually) whilst devices from the D3 still clearly show 3 stitch
interface artefacts (and shows them in photos).

Irrespective of what you would like everyone to believe, stitching
does impact costs - and significantly. Eliminating the requirement
for reticle stitching through the use of a larger field coverage
stepper, would have a major impact on the cost of FF sensors.
Click to expand...
 
blabla0815 said:
afterburn said:
Canon states that zero pixel defects are allowed on their sensors. If
a single pixel is out in your camera, they will replace the sensor
free of charge regardless of warranty status.
Click to expand...
That is very interesting, never heard of that, can you name a source?
Is this only true for the pro cameras?
Click to expand...
It says so in FAQ on the Dutch canon website. I would assume (but you know what they say about assumptions) a similar statement could be found on the Canon USA site.
--
Gijs from The Netherlands

Canon 5D + EF 24-105/f4 L IS
Canon 30D + EF-S 17-55/f2.8 IS

Got the equipment, still looking for talent. Please contact me if you find some cheap.
 
I only repeated what I read on the (Dutch) Canon website. Maybe Canon USA is more forgiving.
--
Gijs from The Netherlands

Canon 5D + EF 24-105/f4 L IS
Canon 30D + EF-S 17-55/f2.8 IS

Got the equipment, still looking for talent. Please contact me if you find some cheap.
 
gml said:
Here is a link I found on the FM forum:
http://www.chipworks.com/blogs.aspx?id=4626&blogid=86

"Canon appears to be the only lithography equipment manufacturer to
offer an i-line stepper configured for 200/300mm wafers with a 50mm x
50mm field. "
In other words, Canon have the technology to make FF sensors in a
single lithography pass. And that's on 300mm (12") wafers.
So, FF sensors are not that difficult and expensive to manufacture -
as some marketing white papers would like us to believe.
Here is a link to the stepper itself:

http://www.usa.canon.com/opd/controller?act=OPDModelDetailAct&fcategoryid=2402&modelid=9165
Click to expand...
and how many hundreds of millions or more did it take canon to come up with that?
 
Where on the Dutch site? URL?

I don't believe that Canon select their perfect sensors for exclusive export to Holland!
 
Its RKM said:
Where on the Dutch site? URL?

I don't believe that Canon select their perfect sensors for exclusive
export to Holland!
Click to expand...
But you know that the Dutch become VERY angry if their well earned money is not spend on perfect products, don't you? :-)

Roberto
 
afterburn said:
Well, see my own guesstimates before I quoted the listed document.
Regardless, it does not change my point one bit.
Click to expand...
You are right that the yield has a significant effect to those huge FF chips prices. I remember hearing figs like c. 80% in APS-C size sensors and something like 20% for FF chips. 1:4 ratio and counting the 1:2.5 size this means from a single wafer you get about 1 10th of FF sensor compared to APS-C sensor and thus the 10 times price. I have understood the APS-C sensor price is around $50 (plus company profit) so FF sensor would cost around $500 (plus company profit). And when put that into a product with all the sells network costs/profits and taxes, you get around $1000 add-on to product price because of FF sensor, and something more because of the FF size mirror and viewfinder.

What comes to the OP post, the single pass litography machine is very significant when the yield is tried to get lower. No more junk because of micrometer scale offsets in litography passes. Eventually we can expect the cropped to FF sensor cost ration to gome down - perhaps to min 1:5 level - meaning the minimum camera sells price difference would be "only" around perhaps $500. Today lowest cost new APS-C cams sells around $500, which could hint tomorrow lowest cost FF cams around $1000...$1500 (perhaps $1290 just to throw a wild number).
 
I find it interesting to compare the two main boards for the D3 and the 1D III shown in the article. I can't see the D3 PCB close enough to read the part numbers but if those six AD chips there are really AD9974's it is interesting, each of those things is a dual 14 bit AFE (analog front end) consisting of signal conditioning amplifiers and a 14 bit analog to digital converter.

I am surprised that Nikon doesn't have this function integrated somewhere (as Canon obviously has done). Perhaps they have done this for performance reasons, I don't know. It looks to me like Nikon's MB is somewhat more expensive than Canon's. This is a big maybe because integrating that functionality doesn't come for free either, but it is usually less than having separate chips.

Six AD9974's gives them a total of twelve 14 bit AFE's! I wonder what all of those are used for? They may be combining them to pick up a couple of extra bits of noise reduction (like the digital audio guys do).

To me that was the interesting part.
 
gml said:
In other words, Canon have the technology to make FF sensors in a
single lithography pass. And that's on 300mm (12") wafers.
Click to expand...
As others have noted, this isn't the primary determinate of sensor expense. Basically, single step versus multiple step doesn't change yield by enough to change costs significantly; the only thing it impacts is how fast you can get a wafer ready, which in turn impacts how many wafers you can run a day on the fab. So it might allow you to recover the cost of your fab slightly faster, and it might produce slightly lower overhead expense, but the bottom line is still dependent mostly on wafer cost and wafer defects (yield).

Many of us who were trying to estimate FF versus APS sensor costs actually were already factoring out stepping (i.e. assuming that you'd get the FF in a single shot).

--
Thom Hogan
author, Complete Guides to Nikon bodies (18 and counting)
http://www.bythom.com
 
Its RKM said:
The OP suggests that Canon may have eliminated the stitching process
for FF sensors using their new stepper and hence brought the costs of
FF sensors into line with the predictions that your simplified view
of life suggests.
Click to expand...
The stepper in question appears to have become available in 2001. So it is hardly new. It is limited to relatively coarse lithography, but that shouldn't be a big issue for making sensors.

-Z-
 
David Hull said:
...
I am surprised that Nikon doesn't have this function integrated
somewhere (as Canon obviously has done). Perhaps they have done this
for performance reasons, I don't know. It looks to me like Nikon's
MB is somewhat more expensive than Canon's. This is a big maybe
because integrating that functionality doesn't come for free either,
but it is usually less than having separate chips.
Click to expand...
AFAIK Canon has not integrated their analog front end, ad-converters as well. I am pretty sure about that for previous generation cameras. I think Sony mentioned something about sensors with integrated ad-converters, but i am not sure.
David Hull said:
Six AD9974's gives them a total of twelve 14 bit AFE's! I wonder
what all of those are used for? They may be combining them to pick
up a couple of extra bits of noise reduction (like the digital audio
guys do).
Click to expand...
Pretty simple, most DSLRs nowadays have multiple channel ADCs. One reason is the faster read-out and the lower noise. When you read out the same sensor with a higher frequency (and less ADCs) the noise will increase. I think Sony announced (introduced?) a sensor where each column has its own ADC.
 
Cost and selling price are not directly related.

Selling price is determined by the market (demand)
gml said:
...Yet, the 1Ds still has a ridiculous price tag - despite Canon's own
technology advances.
Click to expand...
 
blabla0815 said:
David Hull said:
...
I am surprised that Nikon doesn't have this function integrated
somewhere (as Canon obviously has done). Perhaps they have done this
for performance reasons, I don't know. It looks to me like Nikon's
MB is somewhat more expensive than Canon's. This is a big maybe
because integrating that functionality doesn't come for free either,
but it is usually less than having separate chips.
Click to expand...
AFAIK Canon has not integrated their analog front end, ad-converters
as well. I am pretty sure about that for previous generation cameras.
I think Sony mentioned something about sensors with integrated
ad-converters, but i am not sure.
blabla0815 said:
Six AD9974's gives them a total of twelve 14 bit AFE's! I wonder
what all of those are used for? They may be combining them to pick
up a couple of extra bits of noise reduction (like the digital audio
guys do).
Click to expand...
Pretty simple, most DSLRs nowadays have multiple channel ADCs. One
reason is the faster read-out and the lower noise. When you read out
the same sensor with a higher frequency (and less ADCs) the noise
will increase. I think Sony announced (introduced?) a sensor where
each column has its own ADC.
Click to expand...
That would make sense. As for Canon, I don't see the converter chips on their MB. Maybe they are somewhere else. I suspected them to be inside the DIGIC III part.
 
David Hull said:
Six AD9974's gives them a total of twelve 14 bit AFE's! I wonder
what all of those are used for? They may be combining them to pick
up a couple of extra bits of noise reduction (like the digital audio
guys do).
Click to expand...
The sensor has 12 channel read out as stated in the initial press releases. They simply use one ADC per channel. But you're right, it is an expensive circuit, a total of $150 parts cost.

--
Bob
 
gml said:
Here is a link I found on the FM forum:
http://www.chipworks.com/blogs.aspx?id=4626&blogid=86

"Canon appears to be the only lithography equipment manufacturer to
offer an i-line stepper configured for 200/300mm wafers with a 50mm x
50mm field. "
Click to expand...
Chipworks have added two and two and got five. Imagine how p~ssed you'd be if you'd paid thousands for the actual report. Canon state clearly that they use stitching in their full frame sensors, their two lines are both 8".
gml said:
In other words, Canon have the technology to make FF sensors in a
single lithography pass. And that's on 300mm (12") wafers.
So, FF sensors are not that difficult and expensive to manufacture -
as some marketing white papers would like us to believe.
Here is a link to the stepper itself:
Click to expand...
Not in their production lines, they don't, and one stepper a fab process does not make.

--
Bob
 
blabla0815 said:
AFAIK Canon has not integrated their analog front end, ad-converters
as well. I am pretty sure about that for previous generation cameras.
I think Sony mentioned something about sensors with integrated
ad-converters, but i am not sure.
Click to expand...
The Sony IMX021 used in the D300 and A700 has integrated ADC's in each column.

--
Bob
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top