Concerning question 2. Depending on whether you have reason to believe that
your smaller data sets (less than 20 or so as you say) contain any
particularly unusual points (extreme outliers, highly attenuating
circumstances, etc), you may be able to estimate what you want by doing a
Monte Carlo simulation using the existing data set.

Yves Dupuis

> -----Original Message-----
> From: Jack Crawford [SMTP:[log in to unmask]]
> Sent: Tuesday, January 09, 2001 12:33 PM
> To:   [log in to unmask]
> Subject:      [TN] DPMO metrics question for member
>
> The following is posted by IPC staff for a member.  Please reply back to
> technet and to Charlie at [log in to unmask] <mailto:[log in to unmask]>
> An additional question that may need to be considered is whether looking
> at only failed assemblies is really doing DPMO or just a statistical
> assessment of failure modes/types.  Jack
>
> #####
> Q.A. is currently trying to follow the IPC standard in regards to
> calculation of DPMO's (using IPC-7912). We are doing this for electronic
> assemblies that have failed at customer sites and have been returned and
> evaluated.
>
> Question One
> We are currently evaluating a sample of our total field replacements
> representing a population of total shipments and assumed installation.
> This is being done for our key PCA's. Do we need to normalize our defects
> per assembly to our total field replacements and total shipments? We are
> currently dividing our total failed components into our total assemblies
> evaluated, not our total assemblies replaced in the field. What methods
> would you recommend in regards to total assemblies shipped for the
> assembly being evaluated?
>
> Have you done DPMO's on returned assemblies and what is your method of
> calculations?
>
> Question Two
> For assemblies with statistically insignificant data, less than 20 or so
> component repairs for the year 2000, is it acceptable to take an average
> DPMO of the remaining assemblies we have already computed DPMO's for and
> apply to obtain an aggregate calculation? The assemblies with significant
> repair data have a range of total opportunities from 175 to almost 600
> opportunities.
>
> Question Three
> We have also taken our total component failures for an assembly and
> separated failures caused by manufacturing defects from component defects
> caused by other system activity. We then divide the total component
> failures by our total assembly failures to get our Manufacturing and
> overall TDU.
> We then use the standard formula by multiplying our TDU by one million and
> divide this result by the assembly opportunities to get our DPMO.
> We then repeat this method for our defects due to manufacturing only, to
> get a TDU and DPMO resulting resulting from manufacturing defects. Are you
> doing similar types of calculations?
>
>

---------------------------------------------------------------------------------
Technet Mail List provided as a free service by IPC using LISTSERV 1.8d
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 delivery of Technet send the following message: SET Technet NOMAIL
Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail Archives
Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional
information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315
---------------------------------------------------------------------------------