Newbie w/ a question

slipe wrote:

Other than economics of scale has anyone heard a reason why full 35mm sized sensors are so expensive? The cost of the materials can't be that large a factor and it seems they would be easier to manufacture with less miniaturization. The way to go up in pixels would be to go to the 35mm sized sensor. Most lenses seem to be made for them anyway.

10Mp isn't really enough to adequately support a relatively inexpensive consumer wide format photo printer. I've been considering a larger pro unit, but it would be a waste of money without more pixels. A Canon EOS-1Ds Mark II would be about right but 7 grand is a little over my budget. Which brings me back to my question – why is the larger sensor so expensive?

Especially since one repair shop charges just $75.00 to replace a 6.1MP Sony sensor with a brand new one. I don't think the high sensor cost is what we've been led to believe. It would make more sense that the higher cost is related to the electronics, driver development and internal mechanics for support of the larger sensor.

[Andrew Waters]

To Stowaway and Peripatetic, thank you very much for the needed information. I had been looking for those answers a couple of weeks ago on this forumbut ran up against a brick wallfrom other sources here,which in turn degenerated into a less than intelligentdiscussion from all of us.:-)

As for the type of picture taking I do it surely isn't a passion for me other than special events in the family, that's all. (This may change now as I'm retired and may look into some of the terrific scenery in the national parks here in Akron, Ohio and other places in the state.And frankly speaking I don't even know if I'll ever blow up a picture larger than 8x10.)

At the same time I'm fully aware of the ''overkill'' aspect of my photography in terms of my pending purchase of the L1, D200, 30D or the Sony Alpha. Suffice it to say, from my position, I can't accept anything less thana DSLR as an SLR is all I've ever owned.

Which brings me to this. I still have a wonderfully performing Canon F1N from the very early eighties. With batteries in the motor drive and an F2.8/80-200 fixed apertureTokina manual focus I'm sure the entire package weighs what seems to be three to five pounds. This iswhy I chuckle to myself when I hear posters and others say how heavy or how noisy a particular, contemporary digitalcamera is. Obviously this shows just how long I've been ''away'' from cameras because I neverthought about a DSLR until a couple of years ago. Now I'm hooked onto the prospect of owning a camera that will let me do so many thingsa film based one can't match.

[peripatetic]

Large sensors are VERY VERY expensive to make.

The smaller the chip the greater the yield.

Consider for example the Canon 700 IS - uses a 1/2.5" chip with dimensions 5.75x4.3mm = 24.5mm^2

The sensor for the Canon 5D is 24x36mm = 624mm^2 - approximately 25 times bigger.

You mention that a small sensor can often be replaced for $75. Well

$75 * 25 = $1875 which probably isn't too far off what replacing the sensor on a 5D might cost. I would expect the cost to be somewhat higher however.

Think about this - on a single wafer - say you get 10 imperfections (which cause you to reject the chip) randomly distributed across the surface. If your wafer has 1000 chips on it - how many do you reject? Max of 10 right? You're throwing away 1% of your chips here.

If your wafer has (1000/25=) 40 chips how many do you reject? Max of 10, though perhaps you get lucky and get 2 or more imperfections on a single chip. You're throwing away perhaps 25% ofyour chips in this case.

There is also the fact that generally the larger the chip the more complicated the design (for many technical reasons) which may contribute to a higher number of imperfections.

Assume then that your rejection cost are 20 times higher because the chip is 25 times bigger and you might get $1875*1.2= $2250. With a profit margin built in.

Anyone fancy giving Canon a call to find out?

A Canon 5D body is available for $2900.

[peripatetic]

At the same time I'm fully aware of the ''overkill'' aspect of my photography in terms of my pending purchase of the L1, D200, 30D or the Sony Alpha. Suffice it to say, from my position, I can't accept anything less thana DSLR as an SLR is all I've ever owned.

I can't see you being disappointed with any of those cameras frankly.

I'm not a big fan of the entry-level models, they're all a bit small and plasticky for my taste. If I were starting fresh I'd be looking at exactly the models you are.

For reference my 20D can print really excellent A3 prints (16"x20") with a bit of interpolation in Photoshop, I was really amazed at how good they are. For 8x10 6Mp is plenty as long as you're not cropping too much.

For my sins I'm saving up for the Canon 5D as that's the one that will satisfy my longer-term ambitions and is a very nice choice for thelandscape and portrait work that I mainly aspire to, with quality very muchin the ballpark of amedium format film rig.

My (slightly out-of-date but I'll be posting to catch up this weekend after getting back from my holidays) photo blog:

http://peri365.blogspot.com



For anyone moving from film to digital (but who still loves film too) I can't recommend http://www.luminous-landscape.com highly enough. There are a vast number of articles there and you will want to read most of them to make the most of the transition from filmSLR to digital SLR. You really are in for the most wonderful time.

[JimC]

As already pointed out by peripatetic, wafer yield is a big factor.

It's my understanding that each time the size doubles, a sensor costs approximately 4 times as much to manufacturer.

That doesn't even take R&D costs into consideration (or the cost of setting up production, where you'll have a smaller number of sales for the more expensive sensor, making it more difficult to "break even" before starting to make a profit).

[tclune]

The sensor isn't the biggest part of the cost difference, if I'm not mistaken. You can see that in the cost of some of the body-only dSLRs. The big increase in cost is in the lens -- you have to scale up the lens size to utilize the larger sensor, and that's not cheap. If you consider the cost of a lens comparable to, say, my FZ20's on any dSLR, you will have to paymore for the lens alone than the FZ20 costs complete.

[DBB]

Speaking of sensor size and cost - I am veeeeeery veeeery sceptical of the claims of cost. But I don't really know. And I suspect those who say it's an arm and a leg don't really know. What I DO know is very few full size sensors are manufactured. And this would make them more expensive.

If anyone really KNOWS, I would appreciate it and some sort of source.

Finally, there is a technological limit that we are close to reaching on sensor size and noise. To change the equations here would mean a change in the existing technology. On other words psysics itself is rearing its ugly head.

For those who think an APS sized sensor is to small and they disdain 10 megs as not enough - I reply, "Uh?"

:-)

Dave

[peripatetic]

The sensor isn't the biggest part of the cost difference, if I'm not mistaken.

When comparing what to what?

For example the Canon 5D and the Canon 30D are very similar cameras except for the sensor.

Yet the 5D is about £1600 and the 30D about £800. I'm willing to bet that at least half of that difference is in the cost of the bigger sensor.

When comparing a P&S to a SLR it's certainly the case that practically every component is different so it's much harder to say.

Speaking of sensor size and cost - I am veeeeeery veeeery sceptical of the claims of cost. But I don't really know. And I suspect those who say it's an arm and a leg don't really know. What I DO know is very few full size sensors are manufactured. And this would make them more expensive.

Eh? I think you've pretty much got that reversed. Once the die has been created the costs per wafer are almost constant. The same equipment that can create the wafers for the small chips can create the wafers for the big chips. So there are no economies of scale to be gained in the production process. Of course the more chips you sell, the thinner you can spread the R&D and die setup costs.

As a visual aid here's a page with a picture oa a wafer that can only hold 8 24x36mm chips, in that same area you could get at least 200 1/2.5" chips. I'm not sure what you could find hard to believe about that statement.

http://www.luminous-landscape.com/essays/dalsa.shtml

In the microprocessor world (and I assume sensors too as the production methods are very similar) the costs of production are essentially constant per wafer. Once the die is set each wafer costs the same. On the other hand the amount of revenue that can be generated is that number of chips that are defect free per wafer multiplied by the price at which the chip can be sold.

You can spend hours researching wafer yields on Google if you remain unconvinced. Let us know what you find out. :blah:

[Andrew Waters]

I had browsed a few other sites before I came across this one and I have to tell you there are some interesting topics to be read.

As for my first purchase of a DSLR I'm waiting for a thorough review on the L1 because, as mentioned previously, I really like the looks of it (even without the live view sensor). Speaking of the L1, I was in Borders Books and Music todayleafing through a digital camera magazinecomparingDSLRs and I noticed the price of the L1 was $1,599. For a minute there I thought I confused it with the previous model but it surely said DMC L1.I'm sure it said with lens. Is this a misprint?

[DBB]

Speaking of sensor size and cost - I am veeeeeery veeeery sceptical of the claims of cost. But I don't really know. And I suspect those who say it's an arm and a leg don't really know. What I DO know is very few full size sensors are manufactured. And this would make them more expensive.

Eh? I think you've pretty much got that reversed. Once the die has been created the costs per wafer are almost constant. The same equipment that can create the wafers for the small chips can create the wafers for the big chips. So there are no economies of scale to be gained in the production process. Of course the more chips you sell, the thinner you can spread the R&D and die setup costs.

I make no claims to knowledge on this. I am saying, that I think the cost of the sensor is not a signifigant part of the camera cost. Do I KNOW what I'm talking about? Nope.

But my question here is do YOU know what your talking about? Obviously the camera manufacturer's try to imply that this is the Big Factor, which is odd, considering the list prices of various Canon full frame cameras.

And a glance at APS sized sensors and we see that camera's containing them have dropped from 5K to $600 bucks. Did sensor cost suddenly go down? (And yes, I'm aware that there are differences in these cameras other then sensor size).

So I'm sceptical - But I admit that I don't know.

Do you?

Dave