Lists

Saturday, March 21, 2009

Weekly News Log

Yahoo: Omnivision continues to exploit its 1.75um pixel design, this time with 9MP OV9810. The imager also provides 1080p HD video at 30fps. At full 9MP resolution its speed drops to 8.9fps. The 1/2.3" OV9810 delivers 960mV/lux-sec sensitivity. The OV9810 is immediately available for customer sampling with volume production slated for the second half of calendar 2009. EETAsia adds that the sensor's pricing is under $10 in volumes of 100K.

Yahoo: Aptina announced MT9V117 VGA SoC in 1/6" optical format, allowing 4mm-thin module designs. Its target applications are notebooks and webcams. The sensor's speed is 60fps in VGA mode and up to 120fps at lower resolutions. The MT9V117 is currently sampling with production expected in late Q2 2009.

Digitimes published a couple of rumors about recent Hynix successes. The first one says that HP asked its OEM Inventec to add Hynix' products to the testing list of sensors for its upcoming 10" netbook expected to begin volume production in May or June. Initially HP planned to use OmniVision's OV7690 in these netbooks. Inventec does not rule out the possibility of switching supply to Hynix from OmniVision, if testing results are good enough.

The second rumor says that Hynix will begin shipments of sensors for Samsung and LG handsets in the third quarter this year at the earliest. Initial orders include VGA and 2MP products, according to Digitimes sources. The company also plans to begin sampling 3-, 5- and 8MP products with customers in Q3 2009 with shipments to follow in 2010. From Digitimes' picture below it looks like Hynix was unable to fit all the decoupling caps inside the module and has added an outside tail for them:



Another Digitimes article tells that TSMC got orders from Himax and Advasense for 2, 3, 5 and 13MP sensors. Meanwhile, SMIC has secured CIS orders from most related IC design houses in China.

2 comments:

  1. Does the OV9810 use subsampling for video? Or does it read and resample the entire sensor?

    ReplyDelete
  2. I'd guess it uses a combination of subsampling and binning.

    ReplyDelete

All comments are moderated to avoid spam and personal attacks.