Statistical process control?

[tmiller_iluvprinting]

How, who, and what, would be my questions to anyone willing to share where does a company start with statistical process control in the press room? How do you keep statistics, software, manually, or other? Who is responsible, a statistical process engineer, the press room operators, prepress, the lean team? What does a company keep statistics of, incoming consumables, do you send inks to private labs for analysis, different substrates and how they react when exposed to water and ink, fount pH & conductivity, roller life, blanket life, press settings, do you measure every job printed, or just some of the jobs printed? Seems extremely daunting, but also extremely necessary. Has anybody participated in IDEAlliance's G7 Process Control Training, and was it worthwhile? What about RIT's PSA Certification, worthwhile, or not?
Regards,
Todd

 

[Deleted member 16349]

where does a company start with statistical process control in the press room? How do you keep statistics, software, manually, or other?

"Statistical process control" is not a valid method if one can not control the variables. As a process control method, statistical process control is not so suitable in the press room because the print industry does not believe it needs to control the variables.

Often people plot histograms and think of them as control charts but they are not useful for control. They are basically only a history of what happened.

There are books and I am sure there are on-line tutorials related to statistical control that can explain what is required and how to calculate the statistical data.

Control of processes comes from controlling the important variables with capable process technology and not from measuring results.

 

[tmiller_iluvprinting]

Erik,
How does a company know if the variables are in a controlled state if the variables have not been measured?

Control of processes comes from controlling the important variables with capable process technology and not from measuring results.

Again, measurement must enter into this somewhere, or do you just take a 'leap of faith' that the technology is capable? I understand your frustration with the Print Industry from your past posts. Controlling the variables is the key to success, and as a company, or maybe even an industry we have failed miserably at this. Maybe this failure will be the reason for the downfall of offset printing. Papers have more filler, inks have less pigment, fountain solutions have no alcohol, Komori presses have a reverse nip dampening system, yet, many times we are still searching for solutions that we used from a manufacturing model that was in use 20 years ago. We need to figure out how to print fast and accurate with the tools and supplies that we have today, and forget about the good old days. Capable process technology, and the people to use it? I am not sure it exists in the press room?

 

[Deleted member 16349]

Erik,
How does a company know if the variables are in a controlled state if the variables have not been measured?

You are right. Measurements and statistical methods will tell you that a process is in a controlled state or not. But the measurements will not ensure that the process is in control.

The idea of "Capability" is a statistical one, but a process has to be designed or modified to obtain or improve capability.

If you just want to know "statistically" what the variation is in your results, then yes you can apply the statistical process control math to do that. It will show the variation in statistical format. It is helpful to determine where you started from especially if you want to make improvements . To improve the results, you need to make some changes to the process.

Engineers should be able to help with these improvements but I have not been too happy with the role of engineers have had in the printing industry. They have not really been doing their jobs well.

 

[gordo]

How, who, and what, would be my questions to anyone willing to share where does a company start with statistical process control in the press room? How do you keep statistics, software, manually, or other? Who is responsible, a statistical process engineer, the press room operators, prepress, the lean team? What does a company keep statistics of, incoming consumables, do you send inks to private labs for analysis, different substrates and how they react when exposed to water and ink, fount pH & conductivity, roller life, blanket life, press settings, do you measure every job printed, or just some of the jobs printed? Seems extremely daunting, but also extremely necessary. Has anybody participated in IDEAlliance's G7 Process Control Training, and was it worthwhile? What about RIT's PSA Certification, worthwhile, or not?
Regards,
Todd

You can make things complicated or you can make them practical.

Many of the shops that I've visited and the one that I worked at had a basic approach.
First, you need to establish specified standards with associated tolerances. Then plot and graph your measurements. You can then compare deviation across parts of the process.
For (simple) example.
Your specification for CtP plates is a 50% dot +/- 2% - i.e. you expect to measure a 50% dot on the plate and accept a range between 48% and 52%.
Lets say that you measure and graph one plate a day. Any significant event - e.g. new batch of plates, processor replenishment, etc. is noted on the graph.
Each month you'll have a chart that shows some 30 data points plotted including events that impact the integrity of the plate.
You could see how consistent your plate imaging is. How changes like new plate batch impacts (or doesn't) consistency.
If you were doing the same thing for say, dot gain, on the press, you could compare the two charts to see if there's any correlation between the plate imaging and the presswork. E.g. if a change in plate batch impacts the presswork.
I've seen shops do this for a variety of process variables and post the charts, when appropriate, in places like the shop cafeteria. One example was a shop that plotted make-ready time to press run for day and night shifts. This information can help identify areas that need correcting. It may also help determine why presswork does not meet expectations.
Typically this work is done by a dedicated QC manager, production or plant manager.
Some of this information can be extracted from on press CPC systems.
You may not be able to completely control aspects of the process, but you can monitor and correlate the impacts of the variables and as a result potentially diagnose problems and/or fine tune the process.

best, gordo

 

[michaelejahn]

What Gordo said.

I like Ian Reid's pressSIGN, if you happen to be just shopping for tools. Otherwise, just hire Gordon, as it is not just about tools, but changing the way you manage your press(s)

Hope that helps !

 

[Kaoticor]

How do you keep statistics, software, manually, or other? Who is responsible, a statistical process engineer, the press room operators, prepress, the lean team? What does a company keep statistics of, incoming consumables, do you send inks to private labs for analysis, different substrates and how they react when exposed to water and ink, fount pH & conductivity, roller life, blanket life, press settings, do you measure every job printed, or just some of the jobs printed?...
Todd

Hi,

A few examples of what we personally have found helpful to keep records of are the following:

1 - Machine Performance Efficiency - Data logged automatically from the presses and a brief summary report of the months efficiency is prepared by a designated manager.

2 - Paper is sent out to private labs once or twice a year depending on production to compare against COA's that the paper company gives us. This is done by a designated manager who also might pull a guy from the presses to help him briefly compile the necessary items to send out.

3 - Roller and blanket life are all logged by the press operators, who offer feedback to management if performance declines. Job settings are documented by the operators and kept on press.

4 - Fount pH and conductivity are not logged. Press operators know the proper pH and conductivity to run, so unless there is a problem or we are running tests, we find no need to keep a record of it.

5 - Inks are sent to private labs once every two years by a designated manager. Samples are compiled by a designated pressman, then the manager will follow through. Since we rarely change suppliers, we haven't found a need to do this with more frequency.

6 - Every job's inks and targets are measured and logged, then an automated program compiles a trend of current targets compared to actual press readings.

Hope this helps!

-K

 

[tmiller_iluvprinting]

Thank you all for your excellent responses and advice.
Regards,
Todd

 

[Dan Roll]

You can buy a college level statistics textbook (Amazon has these at very reasonable prices if you are patient) that comes with a free copy of StatLab, an extremely useful program for those who are not statistics professionals but would like to analyze their data. No substitute for an advanced degree in math though......