Control > Measurement

In the closed loop color control world, there is a lot of hay made of one side of the equation, with little to no attention paid to the other. This is really pretty stupid, but like many stupid things it makes sense. It is ironic that the arguably more important part of the control system never really gets explained, or even mentioned.

Everyone is obsessed with the measurement part of things. Are you using a camera, or a sensor? Densitometry, or spectrophotometry? Are you measuring targets (i.e. a color bar) or not (i.e. markless, or “in the work”)? Opinions abound on the veracity of the methods, the claims made by manufacturers, and the ultimate suitability of each solution.

People miss that the best measurements in the world make for great reports, very happy prepress folks and very little else. Certainly nothing that ships out the door as a finished product. It is the control part of things that actually affects all those skids of goodies the customer is paying for, yet no one is talking about how the control part is done. They just assume that if you can measure it you have the numbers and they’re all you need, right?

As I wrote some time ago, the problem is you are controlling ink keys at the start of an ink train. The ink train is like a bureaucracy – you cannot control it, and you can barely predict it. Imagine figuring out exactly how much your blood pressure will rise while renewing your green card a the INS in Chicago and you have some idea of the difficulty involved.Dogpic So knowing with great accuracy what your target blood pressure is doesn’t really help you judge how much medicine to take with you to Chicago, does it? The ability to convert a decent measurement into a good ink key move is just as difficult or more so when you consider what adjacent keys are doing. This is where the fancy measurement tire hits the money-making road, and the actual benefit of color control is created.

The reason why these algorithms are not discussed is simple – they’re all software. There’s no cool lighting, or visible machinery to inspect, just software. It’s easier to show someone a stuffed dog than explain an algorithm.

Another reason why algorithms don’t get much mention is that the press OEM may be doing more of the heavy lifting than you think. It makes sense, as the OEM has far better understanding that most of the controls manufacturers out there, certainly on their own presses. But what this means is that the same color control system is mounted on different presses made by different OEMs, you can end up with different performance.
So, if you are shopping for one of these systems, at some point in the process push the prepress folks out of the room and ask about the control side of the product. It might nor might not change your mind but it will add a lot to your understanding of what your buying, which should always be welcome.

4 thoughts on “Control > Measurement

  1. If one does not understand the problem then it can seem difficult. It is like having a combination lock. If you don’t have the combination, it is difficult to open. If you have the combination then it is easy. It is a matter of having the right knowledge.

    Density control is not a difficult problem. The problem is with the capability of the equipment and software used. Press manufacturers do not understand the problem.

    The density control problem on offset presses is not a chemical problem but mainly an ink transport problem. It is not even a fluid mechanics problem. For average steady state density for a particular ink/paper combination, the only variable is ink feed rate. The equation that governs all roller train designs is “ink out = ink in” at steady state conditions. This is based on the principle of “conservation of mass”. The capability of the modern press to feed ink consistently is still not good and therefore variations occure but it is strickly an ink feed problem.

    Presetting ink keys is another issue. The algorithms used today have errors that can be as high as 40%.

    For the best control it is better to have capable hardware and software.

    Erik Nikkanen
    Fountech Inc.

    Like

  2. Are people obsessed with measurement, or is measurement discussed more frequently
    due to the industry moving from densitometry to spectrophotometry (color measuerment)?
    This is a big change and a lot of people have questions.

    Metrology in the print industry is governed by open standards. E.g. CGATS.5
    People can openly debate how and why we measure. This not true for control systems.

    As you point out, control systems are algorithms AKA intellectual property. The methods
    QuadTech uses are probably different than those used by GMI. These processes are what can give a company a
    competitive advantage.

    Does a Print Buyer really care about how you got the color to match? Probably not.
    They just want the highest quality piece at the lowest cost. If they rely on certificates of analysis for quality control and product specification
    compliance, measurement is what determines the pass/fail not the process (control system) used to keep
    the color in spec.

    Like

  3. I would argue that process capability due to what ever control system approach is used is what provides the high quality that would be measured. Measuring on its own does nothing. It can be used in a closed loop for control or for an inherenly consistent process, it can be used to confirm performance.

    Erik

    Like

  4. Are people obsessed with measurement, or is measurement discussed more frequently
    due to the industry moving from densitometry to spectrophotometry (color measuerment)? This is a big change and a lot of people have questions.
    The reason people have questions is because although folks are more interested than even in spectral measurements, all of the new systems showing up are not using spectrophotometers. They’re using CCD cameras. The problem with CCD cameras is that they are not spectral devices, yet many manufacturers claim in literature and presentations that somehow they’re getting spectral data out of these devices. Since this really can’t be done directly, but only through conversion, naturally folks are suspicious. Metrology in the print industry is governed by open standards. E.g. CGATS.5
    People can openly debate how and why we measure. This not true for control systems.
    As you point out, control systems are algorithms AKA intellectual property. The methods
    QuadTech uses are probably different than those used by GMI. These processes are what can give a company a competitive advantage.
    You’re darn right they’re different – ours are ours. GMI, at least in some installations, rely on the press manufacturer to provide the control. This kind of separation opens the door to all kinds of secondary issues, from system response to press-to-press variation. There are other suppliers who don’t do control at all – they have to rely on someone else. Does a Print Buyer really care about how you got the color to match? Probably not.
    They just want the highest quality piece at the lowest cost. If they rely on certificates of analysis for quality control and product specification compliance, measurement is what determines the pass/fail not the process (control system) used to keep the color in spec.
    Have you seen anyone do a color OK through certificates of analysis? Me neither – they look at the impressions, they look at the proof, and they decide if it’s OK. Then they ask the printer to show that the tolerances were held for the rest of the job. Here measurement is of course important, but not in an absolute sense, just in a repeatable sense.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s