Closed Loop Color Control - Heatset Web

[JudP]

Hi All,

Our Goss Sunday 2000 has GMI's closed loop color control system and we recently (about two weeks ago) upgraded this to their latest system and we are having some "interesting" color control issues.

Is anyone else out there running GMI's new Clarios system on their presses?

Jud

[Erik Nikkanen]

Hi All,

Our Goss Sunday 2000 has GMI's closed loop color control system and we recently (about two weeks ago) upgraded this to their latest system and we are having some "interesting" color control issues.

Is anyone else out there running GMI's new Clarios system on their presses?

Jud

Just curious.

Could you comment on what the "interesting " colour control issues are?

[JudP]

Hi Erik, Yes - the press, since installing this upgrade seems less stable. Color fluctuating and density targets randomly changing themselves throughout the press runs.

Jud

[Erik Nikkanen]

Hi Erik, Yes - the press, since installing this upgrade seems less stable. Color fluctuating and density targets randomly changing themselves throughout the press runs.

Jud

In general, close loop systems work by looking for an error. The error is the difference between the target value and the actual value. The magnitude of the error or delta e is used to drive the system for correction. There are usually PID settings which represent Proportional, Integrative and Differential parameters. Each of these parameters has a gain which one sets. It may be that these are not set correctly.

Closed loop systems can be unstable even though the process is not. If the gains for these parameters are set too large, then the system will overshoot and fight itself.

If you turn your closed loop system off while you are running, does the process seem to settle down. If it does, it kind of implies that you are overcorrecting. Reduce the gains down to low levels.

Also in printing on press, the coverage is a factor. High coverage would respond better with larger gains but low coverage will need lower gains in order to not over correct.

If your press is still unstable when you turn off the closed loop control system, then you have a press issue.

Offset is normally a stable process. It may not be consistent but it is stable. This means that if you don't touch it, it should stay where it is most of the time.

I would think there are some manuals and other info that came with the system. You need to read it. I don't know how advanced your system is but one should not think that the engineers who made the system understood everything. You need to know what their system does.

It can't hurt to start with low gains and carefully work up keeping the coverage issue in mind.

Good luck.

[Erik Nikkanen]

To add some info about PID control with closed loop systems. Here is a Wiki page:

PID controller - Wikipedia, the free encyclopedia

It should give you and idea on tuning.

[JudP]

Hey Erik,

Thanks very much and you are definitely correct. We have tested the system as compared with the press to determine if our issue lies from mechanical settings with the press itself or with the CLC system. In this case the problem is a bit of a combination of both.

Last night I had my crews check the strips on the ink and water forms and we did find that they were set too tight however the engineer that it is on site from GMI also found some bugs in the software which need to be fixed which are causing the random ink key settings and ink fluctuations.

This morning things are definitely running smoother. Still have some variation but not nearly as drastic as it was.

Good Wiki page btw... Thanks again.

[Alois Senefelder]

Hello JudP,


I suggest you read the PDF - I've posted

Thread: Lack of Dampening Predictability

Date: 07/20/2010


Regards, Alois

[BeauchampT]

Our plant uses a Grafikontrol system. I'm not sure the differences, but....

We notice that since Grafikontrol uses a camera that simulates a densitometer, it is very unstable. As stated earlier in this thread, high coverage really throws it, any scumming at a startup will cause it to freak out, even a change in the amount of silicone (affecting the gloss of the print) will affect it. We have to keep a close eye on it all the time.

Closed loop control is great - unfortunately, it is still just a computer that cannot think as clearly as a human can.

[JudP]

Our plant uses a Grafikontrol system. I'm not sure the differences, but....

We notice that since Grafikontrol uses a camera that simulates a densitometer, it is very unstable. As stated earlier in this thread, high coverage really throws it, any scumming at a startup will cause it to freak out, even a change in the amount of silicone (affecting the gloss of the print) will affect it. We have to keep a close eye on it all the time.

Closed loop control is great - unfortunately, it is still just a computer that cannot think as clearly as a human can.

I am not familiar with Grafikontrol's system but I am pretty sure that all on press CLC systems do use a camera however, at least in the case of GMI's and Quadtech's system they are actual spectrophometers. These spectros do interpolate to get density readings if that's was you are referring to.

Also, seems odd to me that you would have your camera's installed after the silcone rollers as silcone will definitely have an effect on how the camera sees the color. The only systems I have seen have the cameras in the chills just after the silicone rollers.

Just as an update on our problem here - I've had the GMI engineers on site for the past two weeks and it looks like the problem is solved although unfortunately I don't know what solved it as two things were fixed simultaneously - 1) they did install a software fix but at the same time found a loose connection on the system. They said it was the lose connection causing the fluctuations but I am skeptical. All I know is that the problem is fixed and our new system is working very well.

Jud