There are two more things to keep in mind.
1. Calculation of Cpk is predicated on the assumption that the data are normally distributed. The data under consideration are certainly non-normal, although they may be close enough to be treated as such.
2. Like any measured statistical property, Cpk as calculated from a sample has an associated confidence interval. I have a reference, which I will provide to anyone interested, that shows how to calculate a confidence interval on Cpk.
Needless to say, it is advisable that all parties agree at the start of a project on the calculation and interpretation of Cpk.
Louis Hart
-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Rudolph yu
Sent: Tuesday, December 02, 2008 1:41 AM
To: [log in to unmask]
Subject: Re: [TN] What are the industrial standards for Solder paste printing limits?
Andy:
Check out the website. Actually Cpk 1.66 = 5 sigma, cpk 1.33 is = 4 sigma. One will need a cpk of 2 in order to get to the 6 sigma region.
The averge is about .15mm which is around 6 mil. This is another thing that our process engineer cannot answer me directly why the solder paste dispoisition is 1 mil more than the the stencil thickness on average? Any help?
http://www.isixsigma.com/library/content/sigma_cpk_conversion_table.asp
I dont know what the sample size the software is using. I am still waiting for the answer from the SPI manufactuer.
> Date: Mon, 1 Dec 2008 16:29:30 -0800> From: [log in to unmask]> Subject: Re: [TN] What are the industrial standards for Solder paste printing limits?> To: [log in to unmask]> > Hi Rudolph,> > If I'm not mistaken, a 1.66 Cp means you're satisfying across 10 std > deviations, not 6. Measuring across 6 std deviations would yield a Cp > of 1.0. However even with inclusion of outliers, especially for > printing a high volume batch, reaching a 1.66 using a 25% tol would be > achievable. What is your std deviation and sample size?> > Regards,> Andy> > > Quoting Rudolph yu <[log in to unmask]>:> > > Hi Andy:> >> > In theory, I think the Panasonic SP18 should be able to achieve 1.66 > > cpk. But the data that we collected is not pointing to that..> > One reason I think is due to the outliners that we are having > > difficulties to remove from the calculation. (using the software > > that comes with the SPI machine)> >> > I was just reading couple papers on the internet, and people are > > using +/-50% USL/LSL for their experiments. I just wish I could find > > some literature that says 50% is the industrial standards. If I > > feel more comfortable to set the process control limits at +/- 25%.> >> > Rudolph> >> >> >> >> >> >> >> >> Date: Mon, 1 Dec 2008 15:17:20 -0800> From: [log in to unmask]> > >> Subject: Re: [TN] What are the industrial standards for Solder > >> paste printing limits?> To: [log in to unmask]> > Hi Rudolph,> > > >> What's your mean and std dev? With 5 mils thickness for 50% > >> tolerance, > the USL would be 7.5 mils and LSL is 2.5 mils. This > >> tolerance is a bit > too high. I think 25% is a more reasonable > >> tolerance, unless there are > no-loads that you're taping with tape > >> on the stencil. This will create > a larger gap, and thus much > >> thicker deposits. Even with 25%, you > should be able to get > >> process cap index of 1.66 across 6 std deviations.> > Andy> > > >> > >> > -----Original Message-----> > From: TechNet > >> [mailto:[log in to unmask]] On Behalf Of Rudolph yu> > Sent: Monday, > >> December 01, 2008 3:52 PM> > To: [log in to unmask]> > Subject: [TN] > >> What are the industrial standards for Solder paste printing> > > >> limits?> >> > Hi Technetees:> >> > One of our customers had asked > >> us to show that our solder paste despoisition> > process is capable > >> to acheive Cpk at 1.66 (6 sigma). We just built a> > thousands of > >> pilot boards, and the SPC report I got from the TRI's SPI> > > >> machine indicated that the process is able to reach that > >> goal.......however> > the software sets the upper limit and lower > >> limits for height, area and> > volumn to about +/-50%. Is this > >> reasonable to have such a high variance in> > the printing process? > >> (We use 5 mil stencil with 0201 and fine picthc> > aperature). If > >> I change the limits to +/- 25%, the Cpk will become around> > 1.33 > >> ( 4 sigma)...should I claims that we can acheive 1.66, and point > >> out> > that the wide tolerence was set by the machine expert, or > >> should I limit> > the tolerance spread, and hope that customer > >> will accept lower Cpk.> >> >> > I guess my quesiton is " Is +/- > >> 50% tolerance a resonable limits for paste> > printing process?"> > >> >> >> >> > BTW the AOI and AXI yield is in the high 98 and low 99 > >> range for this> > design.> >> > Thanks> > Rudolph-------------------->
---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 15.0
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to [log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to [log in to unmask]: SET Technet Digest
Search the archives of previous posts at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for additional information, or contact Keach Sasamori at [log in to unmask] or 847-615-7100 ext.2815
-----------------------------------------------------
|
