cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Question on cost drivers of Full Frame Cameras

SGFFX
Enthusiast
There is much discussion about how a full frame must be more expensive than the APC-S (a smaller sensor). I don't understand why this should be the case from a manufacturing perspective. Here is my logic and a question for Canon Engineers. The ongoing discussion of chip yields for computer memory and processors has to do with number of devices per die (chip area). More devices equals poorer yield leading to more complicated manufacturing processes to increase yield. In other words more devices per die means more expensive chips. So here is my questions, the Canon 70D is has a 20.2 Mpixel CMOS APS-C size chip and the Canon 6D has a full frame (35mm) chip but still 20.2 Mpixels. And the 70D has the new dual pixel auto-focus which seems like it would add complexity (more devices per chip). Both use the Digic 5 processor chip and the size of internal memory should be same/similar. In other words component costs for processor and memory should be equivalent. If I make the assumption that a pixel takes as many devices no matter the size of the chip then on the basis of 20.2 Mpixels per larger die (35mm chip) the cost of producing the 6d should be less than that of producing the 70D. So my conclusion is that the full frame sensor is not a cost driver. What am I missing? What drives the greater cost of full frame cameras from a component manufacturing perspective? Steve
5 REPLIES 5

ebiggs1
Legend
Legend

Well you certainly did a thoughtful, albeit incorrect, synopsis of the process.

The full frame sensor, itself is physically larger. Thus the pixel size is different. If you compare a 20 meg pixel APS-C sensor to a 20 mega pixel FF, on the FF, pixels have to be larger, right? They are more costly to produce.

Take an example from the iPhone some have huge pixel counts. Cheap! Smiley Sad

Then, there is the economies of scale. There are way more crop size sensors sold than FF. 

A sensor can hold many smaller APS-C sized pixels, and if a few are defective it's not a big deal. The same pixel count FF sensor with the required larger sized pixels because of the lower yield of pixels per sensor real estate makes a defect or two a problem. And contributes to making costs even higher. The few reject pixels are a BIG deal. Thus the higher reject rate on FF sensors.

 

And of course there is the ancillary electronics to run them.

 

EB
EOS 1DX and 1D Mk IV and less lenses then before!

The smaller runs of bigger sensors no doubt cost a bit more to make, but it can't be more than like 50 bucks a unit. If you back the parts cost out of the price of the camera, after a profit (40%?) to Canon, and a profit to the retailer (10%?) , and after the cost of the other 2000 parts (body, shutter, chip, and 1,997 others) of the camera, and the cost of assembly, and the R&D cost, and after the cost of the marketing, there is not much room left for the maximum possible cost of the sensor, period.

A big part of the price markup has to be marketing. Charge serious and wealthy customers more than casual and budget customers. It is not anything sinister; blender companies and coffee maker companies do the same thing. They sell an entry level camera, a mid range camera, a high-end model, and a pro model. The added features of the more advanced models add cost, of course, but the differentiation of the product lines al by itself allows for some of their ability to charge more too.
Scott

Canon 5d mk 4, Canon 6D, EF 70-200mm L f/2.8 IS mk2; EF 16-35 f/2.8 L mk. III; Sigma 35mm f/1.4 "Art" EF 100mm f/2.8L Macro; EF 85mm f/1.8; EF 1.4x extender mk. 3; EF 24-105 f/4 L; EF-S 17-55mm f/2.8 IS; 3x Phottix Mitros+ speedlites

Why do so many people say "FER-tographer"? Do they take "fertographs"?

TCampbell
Elite
Elite

Hi Steve,

 

It's a good question and I realize the analysis sounds reasoned enough... but it's a bit more complex than that.

 

First (and this is really a nit)... when you manufacter a "wafer" full of chips, typically not every chip on that wafer will be useable.  Some will fail.  So the question isn't how many chips fit on a wafer... it's how many "working" chips can the average wafer yield and does the increased size increase the probability that *somewhere* on that chip there will be a problem resulting in that particular chip not being useable.

 

Second... Canon (and Nikon, Sony, Pentax, etc.) are all in business to make money.   As a consumer... we WANT them to make money.  While it's short-sighted to think that they should give the camera to us for free, they could only do that so many times before they'd go out of business and then we'd be stuck with cameras that have no support.

 

Third (and this is the biggie)... there's more to the camera than the sum of it's parts... plus mark-up.  It takes people to develop all of these products... it takes a LOT of people.  Those people get a compensation/benefits package.  There is real-estate, utilities, etc.  The cost of the product has to be based on the cost of it's parts, plus the loaded cost of the army of people space, equipment, all the other resources and consumeables that it took to invent all those parts and processes as well as the army of people needed to manufacturer it, distribute it, and sell it. 

 

When you take a camera, add up the total cost of everyone and everything needed to bring that camera from concept to production plus the ongoing costs of production and then DIVIDE that by the number of units of that camera that they'll sell... you get a different picture.  

 

The costs of everything that isn't technically a "part" of the camera but for which the camera could not exist were it not for those things, creates a "loaded" cost -- and that gets added into the cost of each unit they sell.

 

If you take a cross section of 100 Canon DSLR buyers... then consider how many of them buy a Rebel vs. a mid-level and how many purchase an APS-C camera vs. a full-frame, you'd probably find that if you group together all the "full frame" buyers (the total number of 1D X buyers plus the total number of 5D III buyers plus the total number of 6D buyers in the room) then I'm guessing it's still a relatively tiny number compared to the mob of people who buy a Rebel or mid-level APS-C body.

 

So the loaded cost to produce and sell a 5D III isn't the higher cost of the chip... it's the fact that a much smaller percentage of people are buying that 5D III.  

 

In summary, you are probably right in that the actual "material" that goes into making a full-frame chip is not really that much expensive than the material that goes into a crop-frame chip -- so it seems reasonable that it should only cost a little more.  But when you think through the "loaded" cost of bringing a camera model from concept to production compounded by the fact that the cost is divided amongs fewer units of high-end bodies being sold, it's not surprising that the high end bodies cost significantly more and not just a tiny amount more.

 

Tim Campbell
5D III, 5D IV, 60Da

amfoto1
Authority

Actually the cost of the larger sensor itself is a primary consideration...

 

A typical wafer can accomodate 80 APS-C size sensors.... vs only 20 "full frame" size sensors.

 

There's also greater loss to quality control with FF... If, say, there are two flaws on that  wafer, it makes for a failure rate of 2.5% among the APS-C sensors.... or 10% among the FF sensors made from the same theoretical wafer. 

 

In  other words, that same wafer might produce 78 usable APS-C.... or 18 usable FF image sensors.  That's a big difference and no doubt really effects the final cost of the camera.

 

The above was detailed in a Canon white paper some years ago. I don't know if it can still be found online anywhere. But it basically appears to cost about 4X as much to make a FF sensor, as it does an APS-C.

 

Likely aside from the sensor the other electronics in the camera are similar in cost... though there may be some economies of scale for higher volume manufacturing of less expensive models.

 

But some other parts of the camera also have to be scaled up to accomodate the larger sensor: the pentaprism, mirror, mirror box, shutter, etc. are all larger and cost more to make. The same is true of lenses.... in order to throw a larger image circle, FF compatible lenses need to have larger diameter elements.

  

Meanwhile, a 6D costs about 1.5X as much as, say, a 70D. That seems reasonable to me. (Note: There's no direct one-to-one model comparison possible, since features vary between crop and FF. For example in some respects, the 70D has a superior AF system, in other respects the 6D's is superior. Prices used are from B&H Photo: approx. $1800 for 6D and about $1200 for 70D.)

 

***********
Alan Myers

San Jose, Calif., USA
"Walk softly and carry a big lens."
GEAR: 5DII, 7D(x2), 50D(x3), some other cameras, various lenses & accessories
FLICKR & PRINTROOM 

 





 

""There's no direct one-to-one model comparison possible,..."

 

You are right, of course, but the 7D and 5D Mk III are very close! Smiley Very Happy

EB
EOS 1DX and 1D Mk IV and less lenses then before!
Announcements