Curved CMOS Sensors

[CraigT82]

This is an interesting development...

https://petapixel.com/2020/12/14/the-first-commercially-ready-curved-cmos-sensor-has-been-developed/

Not sure how useful it would be unless your had a scope with a radius of field curvature equal to the sensor though?

[DaveL59]

wow, you'll likely need to buy the matched lenses to go with it but an interesting idea I guess. Ain't gonna be cheap for sure.

[wimvb]

8 hours ago, CraigT82 said:

This is an interesting development...

https://petapixel.com/2020/12/14/the-first-commercially-ready-curved-cmos-sensor-has-been-developed/

Not sure how useful it would be unless your had a scope with a radius of field curvature equal to the sensor though?

 

5 hours ago, DaveL59 said:

wow, you'll likely need to buy the matched lenses to go with it but an interesting idea I guess. Ain't gonna be cheap for sure.

Very interesting development. The curvature radius doesn't have to be an exact match to the focal plane curvature; it only needs to be significantly better than a plane sensor. But I do wonder about long term characteristics. Bending silicon induces stress in the material, and this creates defects in the crystal. This in turn can lead to lower quantum efficiency, and a shorter lifetime. It's certainly not a trivial feat to make a good and stable curved sensor.

[CraigT82]

16 minutes ago, wimvb said:

 

Very interesting development. The curvature radius doesn't have to be an exact match to the focal plane curvature; it only needs to be significantly better than a plane sensor. But I do wonder about long term characteristics. Bending silicon induces stress in the material, and this creates defects in the crystal. This in turn can lead to lower quantum efficiency, and a shorter lifetime. It's certainly not a trivial feat to make a good and stable curved sensor.

That's a very good point. The sensor in the article has a curvature radius of a 150mm which is significantly more curved than the fields of pretty much any scope. Maybe if they produce a sensor with say 600mm curvature radius that would be a good fit for many scopes and put less stress on the silicon? Although saying that I think I remember reading that the curvature direction of refractors is the opposite of reflectors. Maybe the easiest thing is just a flat sensor and a separate field flattener after all! 

[PEMS]

Do refractors and reflectors have the curvature the same way ? Just thinking that if they do not then a scope/sensor mismatch makes things worse. As in effect the curvature doubles.

[Ags]

But... field flatteners work pretty good don't they? What are we fixing here?

[andrew s]

29 minutes ago, PEMS said:

Do refractors and reflectors have the curvature the same way ? Just thinking that if they do not then a scope/sensor mismatch makes things worse. As in effect the curvature doubles.

Yes normal astro refractors have a concave focal plane looking from the objective.

Regards Andrew 

[wimvb]

1 hour ago, CraigT82 said:

Maybe the easiest thing is just a flat sensor and a separate field flattener after all

Probably. Cheapest for certain.

1 hour ago, Ags said:

What are we fixing here?

Square peg and round hole. We're just not sure yet how that hole curves.

[DaveL59]

surely tho you'd need the sensor to be 2-planar curved? Doesn't mention that in the article and the pic looks like its only curved in one plane.

[wimvb]

17 minutes ago, DaveL59 said:

surely tho you'd need the sensor to be 2-planar curved? Doesn't mention that in the article and the pic looks like its only curved in one plane.

In another image on the company web site, it looks curved right, so probably just an illusion or due to imaging angle.

[gorann]

11 hours ago, CraigT82 said:

This is an interesting development...

https://petapixel.com/2020/12/14/the-first-commercially-ready-curved-cmos-sensor-has-been-developed/

Not sure how useful it would be unless your had a scope with a radius of field curvature equal to the sensor though?

In that article there are two cross sections of camera lenses and counting the the optical elements it does not seem to be much of a gain: 14 elements in the flatened image lens and 12 in the curved one... And buying a camera with a curved sensor means that all the lenses you have bough become useless. Interesting idea but I cannot see it take off except in very special instances.

[JamesF]

2 hours ago, Ags said:

But... field flatteners work pretty good don't they? What are we fixing here?

That was my initial reaction, too.  Flatteners generally seem to do a pretty good job, relatively inexpensively.  I'm wondering if there isn't some specific application or group of applications where that isn't true, or where "pretty good" just isn't good enough.  Science research, perhaps?  Military use?

James

[johninderby]

Interesting bit here. So the reduction in dark current means reduced noise.

https://physicsworld.com/a/curved-camera-chips-may-be-the-next-leap-in-astronomical-imaging/

Edited December 15, 2020 by johninderby

[JamesF]

25 minutes ago, johninderby said:

Interesting bit here. So the reduction in dark current means reduced noise.

https://physicsworld.com/a/curved-camera-chips-may-be-the-next-leap-in-astronomical-imaging/

Though they don't seem too certain about the reason for the reduction in dark current at the moment?

Quote

Dave Walton, Head of Photon Detection Systems at UCL’s Mullard Space Science Laboratory in the UK, agrees: “I’m sure there’ll be other space missions that will want to use curved sensors.

Oh, I imagine.  Like, err, spy satellites, perhaps...

James

[johninderby]

Did a bit more reading. Didn’t realise that Sony patented the first medium format curved camera sensor in 2017 shortly followed by Nikon patenting their curved full frame sensor and Canon and Microsoft are working on tjeir versions.

Apparently the curved sensor will allow for simpler camera lenses as less correction is needed and supposed to produce sharper images.

Edited December 15, 2020 by johninderby

[Ags]

9 hours ago, JamesF said:

That was my initial reaction, too.

I think in the old film days people did use curved photographic plates. The curvature of this chip is quite extreme so perhaps in this specific case a field flattener is impractical.

Edited December 16, 2020 by Ags

[wimvb]

2 hours ago, gorann said:

Interesting idea but I cannot see it take off except in very special instances.

They cite astronomy as one possible instance.

Since most cameras are found in devices that are thinner than a centimeter, and lenses need to become more advanced, we may very well see a market for small curved sensors in mobile devices. It's usually the cheapest solution that wins, and a curved sensor with a simple lens may be cheaper than a flat sensor with a more complex lens. The lenses on that web site are probably not typical for the application of curved sensors once the technology matures.

When I worked in the semiconductor industry, colleagues sneered at me when I told them about these exciting new organic led devices I had heard about at a conference. "never" was the general consensus. That was almost 30 years ago. Nowadays we have oled tv screens. I've learned never to say never.

And who knows, we may need them in f/1 fast astrographs. 😁