iSixSigma

What do other Japanese companies call TPS

Six Sigma – iSixSigma Forums Old Forums General What do other Japanese companies call TPS

Viewing 21 posts - 1 through 21 (of 21 total)
  • Author
    Posts
  • #44971

    Does anyone know what other Japanes companies call TPS – for example: Honda, Mazda, Nissan, Nikon …
    Fujifilm call it Total Quality Control …
    Thanks …

    0
    #145376

    Orang_Utan
    Participant

    Honda calls their CI as TQC, same as Fujifilm.

    0
    #145377

    Thanks …
    I’d also like to ask your opinion about the methods used by other semiconductor fabs prior to Mikel Harry – what is your experience – I can remember MOS 3, MOS 8, MOS 11, and some Phoenix facilities using single-flow processes, one-by-one confirmation of wafers and batches, and JIT, which is not suprising given Moto’s long association with Hitachi and then with Toshiba.
    Looks like Moto should have kept doing what they knew best …
    Cheers,
    Andy
     

    0
    #145379

    Orang_Utan
    Participant

    I was bit late compared with you. I remember almost no wafer fabs exist in Asia outside Japan in 70s.

    0
    #145380

    Marlon Brando
    Participant

    TQC

    0
    #145382

    Eric Maass
    Participant

    Actually, Mikel Harry came to GEG AFTER we had started using statistical methods and focused on variance reduction in our wafer fabs in SPS. J. Ronald Lawson went from the corporate IC pilot line (MICARL) to GEG and worked with Mikel Harry there, so there was some cross-fertilization there.
    Some of our early variance reduction efforts involved comparisons of variance when we implanted into silicon with and without a thin oxide – highly significant difference in variances, minor differences in means. We also reduced the variability of bipolar current gain using MSA (to eliminate emitter home-in) and Breakdown voltages (much lower variance with antimony burield layer than arsenic due to outdiffusion), and threshold voltages of MOS devices (variance was mostly due to oxide/interface charges). 
    The early variance reduction efforts, around 1982-1985, had some spectacular effects on yields that got a lot of attention from management…which led to hiring experts like Mario Perez-Wilson with his structured SPC approach, and perhaps even influenced the hiring of Mikel Harry over in GEG down the road.
    We were using SPC, discovered the need for Variance Components Analysis because of the nested variance issue (die within wafers within batches/lots), and developed control charting approaches based on VCA combined with Multi-Vari charts and I moving R charts. We used DOE to improve yields starting in about 1980, and also had remarkable success using DOE in conjunction with the SUPREM semiconductor process simulator.
    JIT was a mixed bag, but we brought in some consultants that helped us work on controlled WIP, similar to the CONWIP approach proposed in Factory Physics more recently.

    0
    #145389

    Velasquez
    Participant

    Interesting how Japanese companies have turned up their noses at Six Sigma … obviously they are a lot smarter than Americans

    0
    #145400

    Hi Eric,
    I can believe it …
     I think you also had an SPC guy who actually understood rational grouping. Unfortunately, he was killed in an aircraft accident out of Chicago.
    Some of the other innovations out of Phoenix was the use of bar codes to collect wafer fab data. Are we talking about the same fab?
    I think my good friend Simon Thomas later became the ops. manager. (He was the guy who found out how to ‘disclose’ junctions on an SEM.)
    You guys were the original heroes of Moto’s improvements, as described in Keki Bhote’s book, but MOS 8 were the first to achieve Cpk =2 on Leff. (Not such a big problem with  biploar, as you’re more concerneed about paramters than dimensions.)
    By the way, some people still don’t realize that TPS/TQC also uses ‘process excellence,’ as well as ‘one-by-one’ confirmation.
    Cheers,
    Andy

    0
    #145401

    Martin,
    If you strip away all the scientific management stuff and the errors introducted by Dr. Harry, much of Six Sigma reduces to ‘process excellence,’ where the variation in the process is so low (Cpk = 2) and you don’t have to use ‘inspection.’
    TPS also uses ‘process excellence’ as well as Jidoka (100% inspection) which allows them to detect ‘multivariate defects.’ This is called ‘process self-assurance.’
    Cheers,
    Andy

    0
    #145407

    Eric Maass
    Participant

    Simon Thomas is a terrific guy –  I worked some with him and later with his son, who is also very sharp.   Simon pretty much stayed in R & D, ran one of the major labs there, but then left for some other company.
    So, do you work with Simon Thomas now (or did you work with him recently)?  If so – say “hi” for me!
    I guess I’ll have to look up Keki Bhote’s book…

    0
    #145412

    Eric,
    I worked in Austin from 1984 until 1990. Simon helped us solve some of our device structure problems – for example, gate oxide necking near the ‘bird’s beak’ :-)
    Cheers,
    Andy

    0
    #145416

    Praveen Gupta
    Participant

    Hi Andy:
    It has been a while we have exchanged notes here. If I remember, in 1981-84, Motorola Fab was using simple process control techniques, and managers were using normal distribution to predict yields, etc. Instead of DPU, they used to have DPW (Die Per Wafer) type measures.
    In 1984, they launched a Pre-Control initiative in all Semiconductor Sector’s plants, Austin, Phoenix, Japan, Malaysia, etc. Then, they started training in  Advanced Statistics, and Planned Experimentation. I may be missing these names totally. However, the intent was using Dorian Shanin’s variation reduction methods, and design of experiments. These two modules later were split in SPC I, SPC II, and SPC III.
    Then Six Sigma was developed. I was involved in it in 1985 in Austin before I moved to Schaumber and worked with Bill Smith. There it was first a simple six steps methodology supported by all the tools known that time or today. Only new thing that has been added since then is SIPOC, and DMAIC nomenclature. Even the DMAIC, in my observation, evolved from our SPC 373 (or SPC 374) course that has a five phase process improvement methodology.
    Rest is still the same, and packaged differently.
    But now, Six Sigma is a comprehensive system, however, its success or TPS/Lean’s success depends upon management’s intent in using it.
    What do you think?
    praveen
     

    0
    #145429

    Hi Praveen,
    Yes, I remember the metal pre-control flag used in Phoenix. We used Mike Wolfe’s Multi-vari software in Ed Bluestein – except for APRDL of course because they had an ignorant statistician.
    I thuoght the training materials were called Advanced Diagnostic Techniques and Planned Experimentation in Austin.
    If you recall we used ‘Spatial Yield Analysis’ which was first published in Semicodnctor International by a company in Utah, because our greatest yield loss was due to correlated defects, such as dislocations on the edge of the wafer. This is why we put so much emphasis on failure analysis.
    My own opinion is many critical processes are not capable for the simple reason that competitiveness forces many companies to have very tight tolerance limits. If this is true then it is necessary to use the ‘one-by-one’ confirmation method of TPS.
    Cheers,
    Andy
     

    0
    #145430

    Marlon Brando
    Participant

    Andy
    In  the  english  version  of  the  Kaizen  Book (by Masaaki  Imai),in  page  xxi Kanban is  pronounced as  “Kamban”?Which  one  is  correct?
     

    0
    #145437

    Eric Maass
    Participant

    Andy, 
    I remember the “bird’s beak” – an interesting cross section with SEM.  There was also some related issue with a silicon nitride “string” or something along that little edge…We had it with our MOSAIC I process for ECL.
    So, where are you now? It doesn’t sound like you’re still in Austin…but perhaps in a land far, far away…Let’s see, you were in Austin 1984-1990 — I was Device Section Manager and Device Engineering Manager in Bipolar I and II in Mesa, and also led the FACT TOPS Team (first TCS Competition), and was loaned out to the RF Division at 52nd Street to help out the RF1 wafer fab at the request of Tommy George.  I didn’t know that many people in Austin at that time, but can perhaps give you updates on some people if you mention some names to me – I’m at [email protected] (yep, one of the very few who has stayed with Motorola for the longer term). 

    0
    #145438

    Velasquez
    Participant

    While it’s good to see that you see six sigma’s faults, six sigma focuses on counting defects  … THIS IS WRONG
    If six sigma was called “CP2” there would be no problem.

    0
    #145446

    My understanding is Kan = Card and Ban = Signal
    Cheers,
    Andy

    0
    #145448

    Martin,
    The Six Sigma Dr. Harry sold to G.E. is not the ‘Six Sigma’ everyone used. Remember, in 1988 everyone were told to put 6 sigma on their powerpoint slides – perhaps these presentations even included some TPS concepts :-)
    Cheers,
    Andy
     

    0
    #145450

    Q E D
    Participant

    Yes,
    SixSigma – (Harry’s many errors) = TQM

    0
    #145462

    Eric,
    You must be old enough to remember the Dobson Ranch!
    I toured Bipolar 1 and 2, but I can’t remember who showed me around. The SPC guy I mentioned previously was also really helpful.
    I’ll drop you a line.
    Cheers,
    Andy

    0
    #145472

    Marlon Brando
    Participant

    Thank You

    0
Viewing 21 posts - 1 through 21 (of 21 total)

The forum ‘General’ is closed to new topics and replies.