Cpk for the Entire Process

Six Sigma – iSixSigma Forums Old Forums General Cpk for the Entire Process

Viewing 9 posts - 1 through 9 (of 9 total)
  • Author
  • #34458


    How would you calculate Cpk for a batch manufacturing process where there are multiple assembly and inspection stations? What would be the overall Cpk?



    Please refer to the posted discussion on calculating the overall Cpk.
    Bob J & Reigle Stewart  gave me good answers on having a Cpk for the entire process.



    explain about CpK and how it is being implemented in a mass production of PCB assemblies amd computers manufacturing?     
     pl. explain


    Jonathon L. Andell

    I have used the Rolled Throughput Yield calculation in the past, with success. If you don’t know how that works, send me your email and I’ll explain it offline.


    Six Sigma well wisher

    In a PCB manufacturing process there are two things you can do to arrive at the Cpk of the process.
    Since it is a series of machines starting from Solderpasting, Chip shooter, IC placement machine, Handplacements if any, IR reflow/wave soldering, Inspection, depanelization, automatic PCB testing.
    The Back end is putting together these PCBs into an assembly to make the end product.
    Each of these machines are processes by itself . You have to do Cpk on each of these critical processes like Solderpaste, Placement, reflow .
    Solder paste – CTQ is the volume and accuracy of paste
    Chip Shooter CTQ – Accuracy of placement, missed placements and scrap generated
    Placement – Accuract of placement
    Reflow – Soldering defects (dry solder, no solder, missing /falled components etc..
    Like this define the CTQ for each of your critical processes and calculate the existing ppm for each of the processes. Make sure you do a Gage r&r before you validate the readings.
    As much as possbile try to work on continuous Ys like volume/height of solder paste, displacement distance from the centre in the placement. Only the reflow ppm will be discrete.
    You can then calculate the Cpk for each of the processes. Work on improvement of the process based on the data that comes.
    Once you have the Cpk for each of the processes , you can use it to calculate the theoretical througput of the processes as well as calculate the throughput (yield) of the enitre process.
    Hope this gives some clarity



    Hi 2 sigma:
    We can category CPK as “process CPK” and “product CPK”, for the multi-assemble process, we only need to know individual process CPK ( generally, not include inspection stations) as to improve,  a overall process Cpk seem no meaning.
    If we want to know overall qulity of final product which passed all assemble process, we also can use term “product Cpk”
    It is easy to get product CPK:
    1. To take a adequate samples from final product
    2. To measure and gain the specific value of one parameter for the product
    3. To calculate specific CPK referring to common CPK formular
    Here, it is need to remind the product CPK indicate the parameter’s CPK , a product maybe have many parameters, at this situation, we can not get a intergrated CPK, if want , we have to use attributed method , e.g. overall Yield, Over sixsigma level(DPPM)



    I agree with Jonothan and the reason is that we cannot assume that each step of a PCB assembly process is random and independent. Neither can we assume that defects are spatially random and independent.
    Accordingly, I would recommend measuring first time rolled yield. If you have a minimum yield target and you want to report a Cpk then convert the former using the usual method, but remember that yields are not nornally distributed – especially at ‘low yield’ or near ‘high yield.’
    Andy U



    Furthermore previous suggestiom, take a look to IPC-9261 (In process DPMO and Estimated Yield). Could be useful for your scope.
    Rgs, Peppe



    One of the ways of estimating the Cpk of the entire process would be to consider the system in its totality. This means that the entire system is considered to be a process that has inputs at one end and outputs at the other.
    One of the underlying facts would be that all the different parameters that could be considered for such an evaluation would have to be treated as attribute type. Thus the final product at the end of the line is either defective (may fail on one or more parameters) or not and basis this data, calculate the total good and the bad parts. Convert this into the dpm and calculate the process capability.
    Hope this helps.

Viewing 9 posts - 1 through 9 (of 9 total)

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