Wednesday, March 02, 2011

iSuppli: BSI Sensors in 90% of Smartphones by 2014

ForexPros, Balkan.com and many other sites published an article on BSI sensor becoming dominant in the near future. Some interesting quotes:

"Needham & Co analyst Rajvindra Gill expects BSI sensors to be the predominant image sensors for smartphones within 2 years or so.

By 2014, 9 out of 10 smartphones and feature handsets will have BSI image sensors, up from fewer than one fifth today, predicts IHS iSuppli's Pamela Tufegdzic.

Analysts reckon the CMOS image sensor market is worth around $3 billion, and BSI sensors make up about 5 percent of that now."

"OmniVision, which said in November it was already pursuing design wins for its second-generation BSI chip, is about 6-9 months ahead of its nearest competitors."

16 comments:

  1. "Apple has popularized BSI technology and a lot of competitors aren't going to want something that's not BSI because it's not as fancy, nice, efficient or wonderful," said Raymond James analyst Hans Mosesmann.

    Wow... This technology is fancy, nice, efficient or wonderful... Incredible, everyone must follow Steve Job's intuition about BSI.

    IT IS FANCY!!!

    Seriously, it is only a "marketing technology", I'm afraid...

    ReplyDelete
  2. The job of a market analyst is to produce a steady stream of comments like this one so that investors who understand even less about a particular industry will feel comfortable letting the analyst's company (or a company buying the analysis) handle their money.

    The job of an engineer, on the other hand, is to figure out why things work - or don't work - and how to make them better, based on physical principles.

    Anyway, don't let it get you down. Have a laugh, and keep on engineering.

    ReplyDelete
  3. Well I agree with you. The job of an engineer is to make things ("fancy technology") work based on physical principles.
    These things ("fancy technology") are strongly related to enthusisastic opinions of market analysts based on proselytizes (from Steve Jobs and others).

    Ok, ok I stop my sarcasms...
    I laugh and get back to my studies and engineering!

    ReplyDelete
  4. Does anyone really believe Omnivision is 6-9 months ahead of Sony?

    ReplyDelete
  5. I think what they are saying here is OVT/TSMC is ahead of others in terms of the BSI fabrication process and volume.

    ReplyDelete
  6. Dr. Fossum, i am so sick of hearing that ovt is ahead of sony on bsi. sony was first to market w/ bsi made w/ soi for the high end dsc and camcorder market. ovt was 1st to market w/ bsi made from bulk for phones. normally, you start at the high end and work your way down if you have to start at one end. i prefer the sony route.

    so if ovt is 6-9 months ahead of sony in the low end (which is no great accomplishment imho) then sony is at least 2 years ahead of ovt on the high end (which is a great accomplishment imho).

    then there's the whole issue of how did ovt beat samsung and aptina to the low end biz? could other players just use fsi w/ 1.4um and larger pixels along with tech like light pipes(aptina) or SEES(samsung) to solve sensitivity issues? did ovt bring bsi to market before it needed to just so it could claim it was 1st to market w/ a new manufacturing bulk process for the low end,i.e., phones?

    and then there's the legal issue of whether ovt stold ip from ziptronix to enable it to get to the low end quicker. i heard that ovt's process uses van der waals force only to bind the active wafer to the carrier wafer so it doesn't infringe on zip's ip. is that or can that be true? or do the two wafers need to be bonded w/ epoxy, some other chemical, or plasma? would using van der waals force only conceivably cause yield issues?

    please understand that as is obvious, i am not an engineer, and i do not know how to converse with engineers on this subject. i am hoping some of you can critique what i've said in simple lay language.

    thanks in advance,

    EK

    ReplyDelete
  7. Omnivision is kicking everyone's ass because they have a better product.Apple knows a winner when they see it.

    ReplyDelete
  8. To set the record straight, BSI has been around for decades. It is not a secret. Sony's BSI are based on this well-known older technology. The reason few embraced it, is because it was far too expensive. Sony uses SOI which is the more expensive method. OVTI along with TSMC have patented a MANUFACTURING method that is much more cost effective than ANYTHING in the entire history of BSI, and this is what is such an amazing accomlishment. And their BSI2 should increase their lead in this BSI MANUFACTURING technology.

    To recap, it is not just the BSI itself...it is the technology that brings it to market at such low cost and volume.

    ReplyDelete
  9. samsung uses bulk and tsv's to get a better yield. plus, since they use tsv's. they don't have to worry about getting sued by ziptronix for patent infringement like ovt.

    also, sony was so successful in selling bsi into the camcorder, dsc/slr market in 2009 that it allowed sony to leapfrog over ovt for marketshare based on revenues. i'm interested in seeing how it did in 2010. btw: last i saw based on a tsr survey, aptina beat out ovt for 2010.

    http://image-sensors-world.blogspot.com/2010/12/sony-to-buy-back-image-sensor-fab-from.html

    http://1.bp.blogspot.com/_PsITwyTOc4Y/S3sEFCLvWzI/AAAAAAAAAYc/Jricln1z0UY/s1600-h/Yole+CIS+MArket+Shares+2009.jpg

    EK

    ReplyDelete
  10. @ "i am hoping some of you can critique what i've said in simple lay language."

    If you post a specific technical question, Eric or anyone else with enough time and interest may give a best-effort answer here.

    My impression of the questions you've asked above is that some of them are rhetorical and others are really general business questions. Overall, I would say that you do not want real criticism, but are looking for validation of your quick take on some aspects of the image sensor industry.

    If you want Eric's expert opinion in this regard, you should approach him privately and ask if he is available and willing to give it.

    ReplyDelete
  11. cdm, if i wanted your advice on board etiquette, i'd ask for it. the question about whether a carrier wafer can be bonded to an active wafer with only van der waals force is not a rhetorical or general business question but rather a question for an engineer. if you don't know the answer or do but can't share the answer for whatever reason (e.g., conflict of interest) than say so if you wish.

    i believe that van der waals force alone is not enough unless you're willing to deal with low yields. this would explain why samsung used tsv's w/ their bulk product.

    so, i did post a specific question. cdm, you just aren't answering it for whatever reason.

    ek

    ReplyDelete
  12. Regarding the bonding process, I did quickly skim through Ziptronix patent claims. My limited understanding of wafer bonding is that generally, two clean oxidized wafers will stick to each other, esp. if a little extra pressure and heat is given. I think Zip's claim is that you etch back some of the oxide and it works even better with "annealing". I suppose this increases the number of unnaturally terminated bonds at the surface. Over time these bonds will join with "crap" in the atmosphere like water vapor and hydrocarbons and become less inclined to stick to another oxidized wafer. I don't know if it is Van der Waal bonding or other charge-based bonding that makes the two wafers stick together but it has been well know for many many years that recently "cleaned" SiO2 surfaces with the right time/temp/pressure profile will stick together quite well. I don't think any other type of material is needed at all. I also know that this type of bonding recipe is generally a trade secret with some elements publically known and some not disclosed anywhere. (We considered bringing such a process up at Siimpel for the MEMS work).

    I am sure that Omnivision and esp. TMSC bring some new tricks to the table for their process in order to improve yield. As I stated a couple of years ago, kudos were then due to that team for bringing up a high volume BSI manufacuring process. I am not an OVTI fan, really, but you have to respect their accomplishment even if you don't respect their PR campaign.

    ReplyDelete
  13. thank you for your response Dr. Fossum. you answered my question. i do respect the work tsmc and ovt did to bring a relatively inexpensive bulk bsi product to the consumer electronics market. but i am disgusted w/ the pr campaign that belittles sony and does not give kudos to samsung for its bulk product made w/ tsv's.

    thanks again.

    ek

    http://image-sensors-world.blogspot.com/2010/02/omnibsi-package-analysis-by-chipworks.html

    ReplyDelete
  14. @ EK

    I am giving you grief here because I felt you were rude back in January. Then, I expressed an opinion that disagreed with what you wanted to hear, and your immediate response was sarcasm and a suggestion that my opinion was tainted. To me, this was very insulting, so I didn't feel motivated to follow up with details you might have wanted in that thread or in this one.

    If you want, we can bury the hatchet now. If you make an effort at courtesy going forward, I will reciprocate, and it will save us both some misery.

    ReplyDelete
  15. cdm, i am truly sorry that you felt i was being rude back in january. thanks for offering to bury the hatchet. i will watch myself and try to avoid any appearance of rudeness in the future.

    ek

    ReplyDelete

All comments are moderated to avoid spam and personal attacks.