A true 6x7 CMOS low light sensor camera, can it exist?

[tsjanik]

After seeing the title of this thread I thought you technophiles might enjoy this image.   Here's a shot of a new single crystal x-ray system.  The sensor is a 10x10cm CMOS (large flat square area on the left).  It's only 1 MP, but has very large pixels and well capacity.

Tom

Product link: http://www.bruker.com/products/x-ray-diffraction-and-elemental-analysis/single-crystal-x-ray-diffraction/sc-xrd-components/sc-xrd-components/overview/sc-xrd-components/detectors.html

Edit:  Sigma Merrill added for scale, it's not part of the system.

[BJL]

After seeing the title of this thread I thought you technophiles might enjoy this image.   Here's a shot of a new single crystal x-ray system.  The sensor is a 10x10cm CMOS (large flat square area on the left).  It's only 1 MP, but has very large pixels and well capacity.

I am fascinated by this sort of extreme photographic tech., and having read about many of these ultra-large sensor cameras on manufacturers' web-sites, I note one universal rule: they never mention the price!

[tsjanik]

I'm sure it's quite pricey, this is the first, AFAIK, CMOS for x-ray, previously externally cooled CCD sensors were used.  This sensor is cooled as well, but by an internal system.  The sensor can be purchased as an upgrade for existing systems; if I get a price I'll let you know.

[LKaven]

I've worked in an X-ray crystallography lab (macromolecular protein structure determination), though it was before CCDs were widely adopted.  I believe these devices are way up in the six figure range.  Way up.  They need to withstand the brightest X-ray beams in the world as produced by particle accelerators.

[Chris Valites]

I am fascinated by this sort of extreme photographic tech., and having read about many of these ultra-large sensor cameras on manufacturers' web-sites, I note one universal rule: they never mention the price!

Heh, probably amongst the "if you have to ask..." range of sage advice.

[tsjanik]

I've worked in an X-ray crystallography lab (macromolecular protein structure determination), though it was before CCDs were widely adopted. ..........

Luke:

You would be astounded at the capability of newer systems.  What took four days now takes two hours.  The instrument shown has a shutterless design. i.e., it, in essence, records a video as angles are changing.

Tom

[LKaven]

You would be astounded at the capability of newer systems.  What took four days now takes two hours.  The instrument shown has a shutterless design. i.e., it, in essence, records a video as angles are changing.

Tom, I've heard word of such -- amazing!  That would reduce so much of the hard work in the number of crystals that have to be made, and saves on scarce time on the beam line.  Now if only funding for basic biomedical research could return to mid 1990s levels, we could make such progress.

[mmbma]

of course it's possible. If technology can make a 645 digital sensor, it wouldn't' cost that much more to make it 6x7.

For all the "reasons no" here, more the reason to make one. Instead of 6x7, it should be 6x6 so existing medium format cameras can take advantage of it (hasseblad, HY6, etc.).

[BJL]

of course it's possible. If technology can make a 645 digital sensor, it wouldn't' cost that much more to make it 6x7.

Yes of course it is technologically possible; that has been said many times.  The question is cost, which involves not only the unit cost but how much demand there would be, since very low demand greatly increases the mark-up over unit manufacturing cost needed in order to defray the overhead of creating such sensor and a back using it.

For all the "reasons no" here, more the reason to make one.

I just do not understand that at all.

Instead of 6x7, it should be 6x6 so existing medium format cameras can take advantage of it (hasseblad, HY6, etc.).

That is sadly ironical, coming just a few days after the announcement of the liquidation auction of the factory and tools used to make Rolleiflex cameras like the Hy6, so the last 6x6 format camera is out of production.
http://www.japancamerahunter.com/2015/03/end-rolleiflex/
http://www.proventura.de/insolvenzversteigerung-des-kameraherstellers-dhw-fototechnik-gmbh-ehem.-franke--heidecke-gmbh-salzdahlumer-str.-196-38126-braunschweig/auction/2669/

(The 6x6 format Hasselblad V series was discontinued some time ago.)

So the reality is that all medium format camera still in production are in 645 format (or smaller), and that is the target to which any new MF sensor development will be aimed.

Well, there is one MF camera in a format large than 645 that is not yet officially discontinued: the Mamiya RZ67 Pro IID. That is still available, but described as a "legacy product": http://www.mamiyaleaf.com/legacy_RZ80.asp and I fairly sure that this is just a matter of slowly selling remaining stock of a camera that is no longer in production.

[ErikKaffehr]

Hi,

It seems that there are quite a few photographers interesting spending 100 k$US or so on a camera wit 6x6 sensor. So it should be possible to start a Kickstarter project to finance such a development.

Personally, I am neither interested in 6x6 sensor nor in contributing to such a project.

Best regards
Erik

Yes of course it is technologically possible; that has been said many times.  The question is cost, which involves not only the unit cost but how much demand there would be, since very low demand greatly increases the mark-up over unit manufacturing cost needed in order to defray the overhead of creating such sensor and a back using it.

I just do not understand that at all.

That is sadly ironical, coming just a few days after the announcement of the liquidation auction of the factory and tools used to make Rolleiflex cameras like the Hy6, so the last 6x6 format camera is out of production.
http://www.japancamerahunter.com/2015/03/end-rolleiflex/
http://www.proventura.de/insolvenzversteigerung-des-kameraherstellers-dhw-fototechnik-gmbh-ehem.-franke--heidecke-gmbh-salzdahlumer-str.-196-38126-braunschweig/auction/2669/

(The 6x6 format Hasselblad V series was discontinued some time ago.)

So the reality is that all medium format camera still in production are in 645 format (or smaller), and that is the target to which any new MF sensor development will be aimed.

Well, there is one MF camera in a format large than 645 that is not yet officially discontinued: the Mamiya RZ67 Pro IID. That is still available, but described as a "legacy product": http://www.mamiyaleaf.com/legacy_RZ80.asp and I fairly sure that this is just a matter of slowly selling remaining stock of a camera that is no longer in production.

[BJL]

It seems that there are quite a few photographers interesting spending 100 k$US or so on a camera wit 6x6 sensor.

There are quite a few photographers stating online their enthusiasm for a camera with a 6x6 sensor, but
- a lot of them combine this with expectation of a price far less than US$100,000, and
- even if any accept that price estimate (I do not recall any of these "6x6D" enthusiasts mentioning a price that high), statements of enthusiasm are one thing; realistic willingness and ability to pay such a high price are something quite different.

[ErikKaffehr]

Hi,

Realistically, I would think it is very difficult to make a large sensor camera. To begin with you need a sensor of the right size. Designing a large sensor for a small market is prohibitively expensive, I would think. Than there is manufacturing with all necessary masks. More realistic to find a commercial sensor, but larger sensor are probably made for astronomic observation and aerospace.

Cost is essentially related to sensor surface and growing over proportionally with increasing size due to increased rejection rate. So large sensor -> expensive sensor. And, BTW, small series sensor -> expensive sensor.

Designing the supporting electronics is not child's play. Electronic design for a CCD based camera with good image quality is hard. With modern CMOS the sensor itself is doing all the hard work. Even simple CCD based camera probably needs a display and histograms, so some significant computer programming is needed.

The mechanical manufacturing part may not be very hard. You need a good basic design and it may be relatively easy to build some series using CNC.

After that comes precision assembly.

Now that we have some digital backs, we need to integrate them with camera electronics if such electronics are at hand.

Once we have the cameras we need to get them to customers, so we need distribution channels.

To me all this seems to be a bit unrealistic. High cost, high risk and probable failure.

Best regards
Erik

There are quite a few photographers stating online their enthusiasm for a camera with a 6x6 sensor, but
- a lot of them combine this with expectation of a price far less than US$100,000, and
- even if any accept that price estimate (I do not recall any of these "6x6D" enthusiasts mentioning a price that high), statements of enthusiasm are one thing; realistic willingness and ability to pay such a high price are something quite different.

[EricWHiss]

Erik,
Where do you get your $100k figure for a 6x6cm back?   That seems a tad too high to me, and trust me I've been really looking into this lately.

[hjulenissen]

...
Designing the supporting electronics is not child's play. Electronic design for a CCD based camera with good image quality is hard. With modern CMOS the sensor itself is doing all the hard work. Even simple CCD based camera probably needs a display and histograms, so some significant computer programming is needed.
...

I assume that Canon and Nikon have a large amount of software developers doing all kinds of stuff. They need a smooth, user-friendly, (semi-) cross-platform software base that they can sell in the millions. This means support for quirky languages, country-specific tax laws, being able to run on the very cheapest processor etc.

The camera dreamed of here would be a very different kind of beast, I think. More like a industrial camera featuring some generic embedded x86/arm platform, a large battery/fan(?), running linux and relying upon general open-source components to do as much as possible. If you aim to sell <1000 cameras (at a hefty price and, hopefully, large profit), you make sure that fixed spending is kept low, even if that means per-unit spending increase.

I am not so sure that there is so much "significant programming" before one could have a product that some might consider purchasing. But aiming for a more polished, complex, user-friendly product means spending ever more resources (development, study-groups etc) on software and mechanical design for ever smaller returns. Canikon probably have refined that game to a form of art, and you would need some serious funding to beat them at their own game (especially now that the camera market seems to struggle).


I assume that the camera itself would have only basic exposure controls and readout of histogram, perhaps a crude LCD preview. The main output would be raw files that are developed by e.g. open-source raw development software that is not maintained by the camera developers. This raw development software might be offline, realtime-connected to the camera (on your computer/tablet) or even run locally in the camera.

Or an off-the-shelf tablet could be an integrated part of the product, ensuring that the camera would only have to have basic sensor supporting electronics and the capability to transmit digitized image data to the tablet. Subsequently, user-interaction and image processing could be implemented using high-level languages, libraries and UI familiarity offered on tablets, where there is a large pool of talented developers familiar with the tools.

Examples:
A medical ultrasound probe that connects to your smartphone:
http://www.mobisante.com/products/product-overview/


A digital stomp-box developed (according to the myth) by one guy in a couple of years, beating competitors on a highly specific audio dsp task:

http://www.neo-instruments.de/en/ventilator/ventilator-overview

Now, none of these address the specific challenge of integrating sensors and camera mechanics in a good way. That would be problems to be solved by competent people in such a project. But I guess the general task of integrating such an effort in a box with some user interaction can be done with moderate time spent.
 
-h

[ErikKaffehr]

Simple guess,

based on prices of existing backs. If you have better figure please share it.Existing full frame 645 backs are in the price range 30-45 k$, I think. I guess sensor costs would be significantly higher, especially if development was involved. Also, I would assume low sales volume.

Another point that existing MFD manufacturers seem to have been established at 645 format, so I wouldn't expect them going into larger formats that don't fit their cameras.

A new vendor may sell to a significant 6x6 market, and perhaps have lower margins than Phase or Hasselblad.

It is interesting to note that the Hasselblad FV50c is much cheaper than Hasselblads H-series back using the same technology. And the Hasselblad H-series is cheaper than IQ-250.

Best regards
Erik

Erik,
Where do you get your $100k figure for a 6x6cm back?   That seems a tad too high to me, and trust me I've been really looking into this lately.

[BJL]

Erik,
Where do you get your $100k figure for a 6x6cm back?   That seems a tad too high to me, and trust me I've been really looking into this lately.

What does your research show about likely retail pricing?

What I see is that the lowest price for a back with a relatively puny 54x40mm sensor is around US$27,000 for a Leaf Credo 60, and that back has the relatively large target of use with Phase One/Mamiya bodies plus all the legacy MF systems that it can be adapted to.  The 6x7 format is almost twice the area (56x70mm) so optimistically, unit manufacturing costs would double, and probably far more due to the way sensor yields decline with increasing size (both fatal defects int silicon and the increased risk of failure in the increased number of stitches that must be done on the wafer).  For a hint of the scaling of price with size, observe the pricing trends from 24x16mm ("APC-C") to 36x24mm to 44x33mm to 54x40mm.

Apart from unit costs, the retail mark-up must amplified due to the predictably far lower sales volume of a sensor and back that can only be used with mostly discontinued and entirely manual focus bodies (RX67 etc.) plus view cameras, so US$100,000 seems like a reasonable rough estimate to me.

On the other hand, if you are willing to accept very low resolution (plus stunning dynamic range!), like 50nm pixels at best, then "wafer-scale" sensors in both CCD and CMOS are now available from various sources as special orders, and work continues on them for the sake of medical imaging and such:

http://www.teledynedalsa.com/imaging/products/custom/sensors/
http://iopscience.iop.org/1748-0221/6/12/C12064

[EricWHiss]

BJL,
I'd like to be able to offer a high quality digital TLR in the $13k range.   Still looking for investors….
Eric

[ErikKaffehr]

Hi,

That would be a very nice price.

Best regards
Erik

BJL,
I'd like to be able to offer a high quality digital TLR in the $13k range.   Still looking for investors….
Eric

[S Meyers]

I recently got a quote for a custom large format sensor Dalsa 250 MP CCD camera, for about $160,000.

17,216 x 15,556 pixels, 90mm x 84mm 5.6 micron pixels size.  Didn't follow up to see if it is even possible for field use.  

http://www.ifp.uni-stuttgart.de/publications/phowo11/100Neumann.pdf

http://www.ziimaging.com/en/zi-dmc-iie-250_32.htm

[BJL]

I recently got a quote for a custom large format sensor Dalsa 250 MP CCD camera, for about $160,000.

17,216 x 15,556 pixels, 90mm x 84mm 5.6 micron pixels size.  Didn't follow up to see if it is even possible for field use.

Price seems quite reasonable for the sensor size (and with four other smaller sensors as a bonus), but:
Weight: 65Kg
and that does not include the power supply needed for:
Power consumption: 280W
so, not so good for field use.