Wednesday, February 17, 2021

ON Semi Publishes a Full Characterization Report of its 8.1MP Automotive Sensor

Update: The links and the data have been removed at ON Semi request. It is not supposed to be publicly available.

The original post follows:

ON Semi publishes a full characterization report of its relatively new 8.1MP AR0820AT automotive sensor. This is a first time I see such a detailed characterization data in open access with no NDA.

The 1/2-inch AR0820 features:
  • 2.1 ┬Ám Automotive Grade BSI Pixel with DR−Pix Technology
  • On−Sensor HDR Reconstruct with Flexible Exposure Ratio Control
  • Full Resolution Video Capture of 3840 x 2160 at up to 40 fps in 3−exposure HDR and 30 fps in 4−exposure HDR
  • Line interleaved T1/T2/T3/T4 output
  • Sensor Fault Detection for ASIL−B Compliance Support
  • 2 x 2 In−pixel Binning Mode and Color Binning Mode
  • Data Interface: 1.8 Gbps/Lane, 4−lane MIPI CSI−2
  • Selectable Automatic or User Controlled Black Level Control
  • Frame to Frame Switching Among up to 4 Contexts to Enable Multi−function Systems
  • Multi−Camera Synchronization Support
  • Multiple CFA Options including RGB, and RCCC, RCCB


  1. This document should be a template for the whole industry.

  2. Replies
    1. Hi Albert,
      just fyi - it seems more vendors start adding EMVA1288 datasheets. For example I just recently learned that e2v Emerald 67M comes with a EMVA1288 standard datasheet

  3. I'm very happy to see pixel MTF data, but there's a trade-off between peak QE and pixel MTF. I'd like to see a study of object detectability/recognition versus QE/MTF trade-offs.

    Traditionally, sensor companies have fixated on QE specs while paying less attention to MTF. This doesn't guarantee that they at the optimum spot for object detection.

    1. For object detectability versus QE/MTF study, please check the paper "Comparison of Detectability Index and Contrast Detection Probability", equation #6 and #20-32 express the relations between detectability vs. QE&MTF.

    2. Ah, yes! I knew I'd seen a paper related to this topic somewhere. I really should have remembered it was by Robin.

  4. Indeed, props for publishing! May they publish more technical content in the future.


All comments are moderated to avoid spam and personal attacks.