Saturday, August 22, 2009

Hynix Image Sensors Progress

It came to my attention that Hynix Newsletter regularly updates its readers about the company CIS development progress.

The January 2009 issue reports that 1.4um pixel development schedule has been pulled-in to meet the market demand. Hynix plans to launch 1.4um pixel 2M/3M/5M products in the second half of 2009. The newsletter also presents the updated Hynix CIS roadmap"

The February 2009 issue reports about the development of 1/5-inch 2MP sensor based on 1.75um pixels with improved low-light performance.

April 2009 issue of the newsletter announces new 1/10-inch VGA sensor YACBAC1S optimized for low light and demonstrating better sensitivity than the old one named YACBAA0S.

The May issue reports that 1/10-inch VGA sensor with 2.25um pixels "successfully entered the Taiwan notebook market, targeting NetBook applications". There is also 1.3MP sensor with 2.2um pixels "under scheduling".

The June newsletter announces 3MP 1.75um pixel YACE4A1SBDBS imager with integrated ISP samples available with mass production start in September 2009. Hynix also plans to pass Mediatek validation by Spetember.

The July issue reports of progress with major notebook makers where VGA "products performance and image quality are tested through a strict evaluation process". The development of 1.0MP and 1.3MP products for notebooks is under way.

Hynix also announces "No-Focus Adjustment Module" using "NeoPAC" plate between a sensor and a lens. I'd guess it works by improving the module assembly precision.


  1. I see Hynix is planning EDoF for the middle of 2010 on 3 and 5 MP parts. Does anone care about EDoF anymore? Omnivision hasn't been able to sell it.

  2. EDOF is coming, albeit slowly. I think by this time next year, mid range cell phones running 3MP cameras will be EDOF enabled. The road has been long because it has taken a huge effort to allow everyone in the supply chain to catch up. There is now some consensus on where to place the EDOF algorithms, ie. on the sensor, on the module as a seperate chip, on the baseband silicon or on the application processor as pure software. The lens makers and module makers have also had some time to play with the technology and create the physical parts needed to realize the feature while giving good performance.

    As for Hynix's No-Focus module, it is not a new idea. The Hynix sensor package is outsourced to Optopac, a Korean packaging company that has a flipchip based packaging structure called "NeoPAC". NeoPAC can obtain very uniform thicknesses from the surface of the cover glass to the sensor image plane due to the electrical connections using uniform gold bumps between the bottom of the etched cover glass and the sensor's electrical pads. After the sensor is mounted by SMT to the PCB, the lens is then mounted to the cover glass directly to obtain predictable optical focus characteristics. This is not possible on traditional module structures because the interface between the lens holder and the module is the PCB, which means one has to contend with solder ball tolerances, flatness of the SMT, PCB warping. Using Hynix's module structure, the only tolerances that remain are the gold bump thickness spread, and the lens manufacturing spread. This makes the no-focus possible because these tolerances are usually small and glass doesn't flex or warp a great deal. It is possible to apply this same technique to CSP sensors or any other packaging architecture with a predictable distance between cover class and sensor image place.

  3. I hear about how 1.4 for a long time, but the QE is a problem. OV has sample 1.4 but now moved to 1.75 BSI. Kodak did 1.4 with clear pixel but loses money and looks for partners. When will 1.4 come to stay?

  4. QE of 1.4um pixels has greatly improved over time. 1.4um sensors are starting to appear in more and more mobile phones these days.

  5. Back to the EDOF comments - based on Dblur basically going out of business, DxO being up for sale, and only Nokia rolling out any products, it would seem that EDOF's attempt at getting in the mobile phone camera business is about over....

  6. Let me disagree. Dblur team and patents are acquired by Tessera and being integrated into its offers. DxO Labs sale does not necessary mean its EDoF death - it may be just the opposite, depending on the buyer.

    Talking about Nokia, while it is driving EDoF in principle, it is not willing to pay much premium for EDoF products. This makes it harder for image sensors and EDoF solutions companies to justify the efforts needed to develop EDoF solutions and products.

    Still, the EDoF idea is far from being dead. Just the EDoF business is more difficult than it was initially thought.

  7. I'm new to this blog (been lurking for a few weeks, I'm impressed) and I'm sorry to comment on an old post, but isn't the Palm Pre using EDoF? I know that isn't driving much volume, but it's a start and shows it isn't just about Nokia.

    If AF doesn't come down in costs and EDoF becomes mature enough, it seems to me that there's still a large volume opportunity for EDoF. Whether the various participants will execute properly to fulfill that promise is another question entirely...


All comments are moderated to avoid spam and personal attacks.