Friday, February 11, 2011

Chipworks: Sony-Ericsson SE006 16MP Sensor is Based on 1.12um Pixel

Chipworks reverse engineered the new Sony-Ericsson SE006 phone and revealed that its 16MP BSI sensor is based on 1.12um pixels. Chipworks writes:

"Sony is again first to market – this time with a next generation 1.1 µm class pixel. Sony announced the IMX081PQ Exmor-R CMOS image sensor and IU081F camera module in October 2010. Featuring 1.12 µm pixels, the camera modules are claimed to be the smallest and thinnest modules of this resolution for mobile phones. These devices mark Sony’s first deployment of Exmor-R (back illuminated) sensors in mobile applications, and are part of Sony’s strategy to aggressively pursue design wins in high-end mobile phones.

The 10.5 mm x 10.5 mm x 7.9 mm module features a chip-on-board (COB) mounted image sensor, an autofocus mechanism, and a printed wiring board (PWB) with embedded components – something which Sony has been a bit of a laggard on until now. The IMX081PQ uses Bayer patterned color filters and microlenses on a 1.1 µm pitch.
"


8 comments:

  1. Beautiful. This is going a lot further in my opinion than orange and blue renderings.

    ReplyDelete
  2. Sony is always Sony...

    ReplyDelete
  3. curious about what is the max number of connections you has seen on the flex connector of a lens module? can i put a couple of MIPI phys on it (10 diff lanes)? what is the biggest die you have ever seen used on COB? what is the max power consumption seen used on COB? just looking for rules of thumb....

    ReplyDelete
  4. What a waste of 16MP sensor... such a cheap lens.

    ReplyDelete
  5. Mark my words. Either OV and Aptina merge or they will die from Sony(high end) and Samsung(low end) attacks.

    ReplyDelete
  6. One neat thing about this chip is that if you work backwards from the published frame rates and resolutions, it appears that when the resolution gets low enough, outputs from multiple rows undergo simultaneous A/D conversion, probably by dint of passing outputs to the per-column converters of unused rows. It looks like this gets around the per-column throughput limitation, albeit at the cost of reduced overall throughput. (I'll concede that some of the aggregate throughput drop may result from the marketing convenience of nice round multiples-of-base-fps figures for popular formats.)

    With some fairly simple modifications to their basic structure, imaging chips in this class could probably get to much higher (> 60 fps) base frame rates, higher ADC precision (> 10 bits), or both.

    Commercial pitch: If you are at Sony, Samsung, etc., and have some pull when it comes to picking up outside ideas, I'd be happy to tell you how.

    ReplyDelete
  7. Sorry, that should read "unused columns", not "unused rows".

    ReplyDelete

All comments are moderated to avoid spam and personal attacks.