iSixSigma

SPC in printing

Six Sigma – iSixSigma Forums Old Forums General SPC in printing

Viewing 21 posts - 1 through 21 (of 21 total)
  • Author
    Posts
  • #36538

    Bob Noga
    Participant

    My company has recently purchased a printing firm as past of a vertical integration.  We have no experience in printing and the firm we purchased has no experience with SPC or Six Sigma.  In order to get the process rolling, I would like to solicit some advice on how to determine capability on a printing press.  Once the copy has been brought into registration, any stoppage will risk the need to re-register.  That being the case, how is subgroup sampling done?  Is it done off-line, after the run?  If so, how can you establish whether the process is in control during a run?
    Any printers out there who would be willing to share some pointers with a printing novice?

    0
    #105660

    Mikel
    Member

    Iam confused. Is your question about Six Sigma or SPC? The two are not necessarily related.

    0
    #105664

    Darth
    Participant

    Hey John, jump in here and help this guy out will you?  Here are a few of my non-expert thoughts.  Apparently the question is more related to SPC issues than SS.  First of all, what metrics about registration interest you?  Are you interested in setup time and how long it takes to hit it?  Are you interested in how long it stays before needing to be adjusted?  Are you interested in measuring drift from some nominal?  Are you interested in how many times you go down and have to re-register?  Some data collection will be done off-line, some during the run and possibly some done on the final printed product.  Is registration your only concern?  How about other issues of print quality?  Are you printing corrugated board or high gloss magazines?  Your question has too many paths to go down for a general post on a Forum but I am sure this John Smith guy can help.

    0
    #105676

    Bob Noga
    Participant

    At this point, the interest is more SPC than SS.  The application is flexographic printing and the specifications are dimensional and “cosmetic” which includes registration.  As a first step, I’d like to establish that the process is under control, normally distributed, etc. and introduce control charting.  With “widgets” all of the above is a piece of cake, but I’d appreciate some advice on how others have approached this with a huge roll of a process that can’t be stopped.

    0
    #105678

    Darth
    Participant

    Thanks for clarifying and hopefully some of the printing spc experts will appear.  

    0
    #105725

    Bob,
    I have assumed that you use gravure printing? Unfortunately, I have no experience of this type of printing – only offset lithography. In my experience (only two years) SPC is not used for ‘offest lithography’ of the fours ‘colours’ because  each image has to be adjusted to target, either automatically or manually due to the presence of ‘run out’ and other sources of variation; in other words there is no ‘natural unadjusted variation’ in the process, which is one of the assumptions of X-bar and R charts.
    I can certainly give you some insight into the causes of ‘overlay’ registration problems and related ‘defects’ such as skew, over-feeding, under-feeding, jamming, etc.
    Cheers,
    Andy

    0
    #105748

    Marz
    Participant

    Bob,
      There are several things you can look at to use SPC. You may want to look at such things as color density, pH of the inks and coatings, dryer temperature, or even COF. You can look at print register (print to print or print to cut.) First you have to establish a timeline on how often you want to check these things and use some measurements to establish targets and control limits for these. We recently started control charting all of these. We were taking the measurements anyway and thought it would be beneficial to track such things as runs and trends on the shop floor by using control charting. If you decide to go this route get ready. This type of thing takes a complete culture change from the top down. If you would like to contact me, post your email adress, and I will help you anyway I can, as time allows.
     
                          Regards,
                              Kevin

    0
    #105793

    Bob N.
    Participant

    Hi Andy
     
    The shop we purchased uses flexographic printing.  The kinds of defects that we see are registration and cosmetic–either voids or spots usually due to transient dust on a plate.  We have a customer that has asked for Cpk on dimensions, though that doesn’t appear to vary:  depending on the die it’s either all in or all out.  Though we have seen variable defects on the depth of the die strike, causing adhesive ooze on some substrates.
    To restate my basic problems 1) I’m not a printer, and the folks in the print shop have no experience whatsoever with SPC.  So, we are often like two people speaking different languages. 2) I can’t imagine how to sample a batch-in-progress, given that stopping the process to sample will induce defects. 
    If you would be willing to give me a few pointers, my e-mail is [email protected]

    0
    #105795

    Bob N.
    Participant

    Hi Kevin
     
    Thanks for your response and offer of help.  I definitely hear you about the culture change.  That will be a major challenge, from a training standpoint.  On the plus side, the folks there seem to have a genuine willingness to learn new things and to grow. My biggest hurdle at this point is that I can’t imagine how to sample a batch-in-progress, given that stopping the process to sample will induce defects. 
    If you would be willing to give me a few pointers, my e-mail is [email protected]
    Thanks again
    Bob

    0
    #105797

    Chris Seider
    Participant

    Bob,
    Welcome to the world of printing which insists most of what they do (if you ask the operators and most of their management) is an art.  We have done SPC on ink density on our presses with mixed success.  We also do SPC on some more critical outputs like repeat length (assuming you have a repeating pattern you are printing).
    Doing SPC on process inputs is very difficult, especially on our presses, since few process parameters are displayed (except temperatures).
    Good luck but just remember to find out what is important to downstream customers (of final end use customer) as a starting point for any SPC.
     

    0
    #105806

    Bob N.
    Participant

    Hi C
    “…which insists most of what they do (if you ask the operators and most of their management) is an art…”
    Boy, I’ve sure heard that more than once!  Of course, I’ve heard it in my current high-tech milieu also.
     
    Cheers
    Bob

    0
    #105811

    Mike Carnell
    Participant

    Bob N,
    There are two types of SPC. Statistical Process Control – which is not what you are asking about. You are asking about Statistical Product Control.
    As long as you are controling your process by stoping the process and checking the PRODUCT you will have this problem. If you understand the way the variables that feed the process affect the quality of the product you do not need to check the product. You control the input variables. That is why we have SPC in the Control Phase. After you have been through DMAI you know what variables are the leverage variables and how to control them and the effect they have on the product.
    You need to control what? Color and registration? Color is not typically a function of the press. You control that at setup. So what variables affect registration? How do you control those? (this is that Star Wars stuff that Yoda was taking about – you don’t need to look at the product).
    This is no different than the people who begin implementing Kanban’s before they have done anything to make the process flow in a predictable manner. The Kanbans don’t work. Neither will your control charts. You will end up checking product deceiding it is bad and then having to investigate it, determine what trashed the product and fix the other variable. Figure that out in the SS project in the first place and control it.
    Just my opinion.
    Good luck.

    0
    #105813

    Bob N.
    Participant

    Hi Mike
    Interesting post.  I would have thought the just the opposite.  That is, if I have no ability to check the process while it is being performed, all I can do is check the product at the end.  I would have called the Statistical Product Control.
     
    Cheers
    Bob

    0
    #105815

    Mike Carnell
    Participant

    Bob,
    Checking the product is checking the product – it doesn’t matter where it is done. If roller speed and tention is critical to registration then I control those two variable but I don’t control them by looking at the product. I control them by understanding if I run them in this window I get good product. Then I never have to concern myself with figuring out how to look at what is being printed because I know that it is good based on my knowledge of the process.
    Just my opinion

    0
    #105827

    Bob,
    I’ve sent you an email with futher infomation …
    Good luck,
    Andy

    0
    #105829

    Scott
    Member

    for printing, it is really difficult to use SPC in process, there are two many element to affect the product quality, such as the operator’s skills, the difficulty of parts and condition of ur facility (Cleaning), capability of ur equipment and the performance of ur printed material.
    I have ever worked for a printing company, and try to use SPC in process, but failed , if you have more action ,pls share with me.

    0
    #105863

    Mike Carnell
    Participant

    Rick,
    You have no idea how many different people on how many different processes have said that same thing. We brainstormed around a tennis ball and had 107 ( “… our best people have worked on this for 9 years” – 6 weeks later!). The thing is there are a very few that have a large effect – you may recognize the term “leverage variable.” The process of going through the Y = f(x) is about going from a point of no substantive knowlwedge (their are so many viables that …..) to understanding which ones affect which parameters.
    Check out Demings “obstacles to improvement”  about the the people who know everything about the process except how to improve it.
    People who make comments like “this is an art” are typically pertetuating a myth for personal security. Here is what I will share with you – learn the SS process – go through the steps and figure out which one are important (quantify not “there are two many element to affect the product quality…..) – get a regression model with a decent r-square value (you put in the work so you understand your process) and then I will be more than happy to sit down with you and work on the control strategy which by the way – using Control Charts is the least desirable.
    Good luck

    0
    #105895

    Scott
    Member

    Mike,
      thanks for ur reply. I agree with u partly. In our company, the printing operator have more than ten years’ experiences in printing aera. and would you like to give me some detailed case for study? you can email to me by [email protected]
     
    Thanks

    0
    #105899

    Mike Carnell
    Participant

    Rick,
    I will get you tomorrow. I am sorry, but this day has gone on forever.
     

    0
    #105901

    Chris Butterworth
    Participant

    Bob,
    Mike Carnell’s information is valuable to you. There are several variables that affect print quality but most of these do very little. There will be one or two that have a significant impact and these are the variables that you monitor using an SPC program.
    You can perform a simple multi vari analysis to assess which variables are worth controlling and then you’re on your way.
    Also, the idea that a culture change is required – well, I disagree with that. A culture change is required if you need the organization to change. But you’re only looking to change the way a few people control their process. Get one or two people on your side and pilot the SPC program.
    Good Luck
    Chris
     

    0
    #105939

    Bob N.
    Participant

    Thanks to all who weighed in on this thread.  Every comment has given me something to think about. 
    As to the latest comments, even as a newbie to printing, I can see that there are just a few key variables that need to be controlled to virtually “ensure” print quality.  I also would agree that a control chart is of low desirability.  One of the first things I did early on was to institute fairly traditional control plans for some of the key jobs, because in addition to the print quality, there are often dozens of other requirements (number of splices, type of splice tape, quantity per roll, core labels, etc.) and little clear idea of how they were controlled or who was responsible.
    But now, we have a large opportunity from a customer that requires X-Bar /R charts and control charts.  That’s what spurred me to Search this forum to see how others in the printing industry may have approached control charting.  I saw enough in the Search to make me think that it was being done, but not enough where I could avoid starting a thread of my own.
    But, the advice has been great, and at this point I will try to identify one or two measureabile process characteristics that drive the primary customer requirements: dimension and registration.
    One other thing, as to the “culture change” comments, I suppose it is easier to characterize it as just convincing a couple of the senior craftspeople.  But the ability to pull even one or two people off the line for some training will involve a change of the management mindset as well.  But, as always, money talks. 

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

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