The Missing 360 Degrees

The black box

The whole thing had started with a black box called the A34.

The A34 controlled the infeed, and it had one particularly stupid weakness: every revolution of the press operated a mechanical timing relay.

Click. Click. Click.

At 10,000 impressions per hour, that poor relay was making and breaking almost three times every second. The relays were East German, and they were not exactly famous for longevity. Some lasted only a few months. Then the infeed would begin acting up again.

Uwe Peterreit had mentioned the problem not long after I came to Hamburg.

That was how it started.

A better system

Initially, I considered simply rebuilding the A34 with modern electronics. Then I thought: Stuff this. Why reproduce an old black box when I could replace the whole thing with a programmable controller?

IFM had an office nearby. Sommer was the engineer there, and Becker was his technician. I would go over and explain what I needed the press to do. I knew the machine, the timing, what the old electronics were trying to accomplish, and where they were failing. Sommer knew how to make the IFM controller do it.

It became a very good partnership.

The basic retrofit replaced the infeed control and the side-guide checking electronics. Once those old components came out, the original analog speed indicator lost the electronics that drove it, so I replaced that too. Instead of an old 10-milliamp moving-coil meter, the operator now had two seven-segment displays showing press speed digitally.

The system eventually grew into several options: the infeed upgrade, ink ductor timing, and a timed air-blast system for presses running heavier board stock.

The air-blast version was quite sophisticated for its time. The operator entered the sheet length in millimeters on a thumbwheel switch at the infeed console. The controller calculated the required angular position and opened the appropriate air valves during the correct portion of the sheet travel. The blast stopped roughly eight to ten centimeters short of the tail; blow underneath the very end and you could lift the sheet and create exactly the problem you were trying to prevent.

Kösel was running lighter paper, not board, so they did not need the air-blast option. They wanted the infeed upgrade, the digital speed indication that came with it, and the ink ductor timing.

Kempten

The ink ductors had to operate at very specific mechanical positions. The press had plate and blanket cylinders of normal circumference, but the impression cylinders were double circumference: two sheets, two gripper systems, one impression-cylinder revolution for every two revolutions of the plate and blanket cylinders.

I used an incremental encoder mounted to the Einturenwelle in the infeed. The encoder produced one degree per count: zero, one, two, three, all the way through 360 degrees. Then its index pulse appeared and reset the counter to zero.

Perfect.

Or so I thought.

The printing units were mechanically phased 120 degrees apart. Units 1 and 4 shared one angular relationship, 2 and 5 another, and 3 and 6 the third. Those timing windows themselves fit neatly into 360 degrees.

The controller counted properly. The encoder was mounted properly. The degree values were correct.

We started the press.

Everything worked.

Then we shut it down.

The intermittent fault

Later we started it again, and some of the ink ductors did not work. One might stay locked. Another would not lock. Then we would try again and things would work. Then they would not.

Nothing dramatic happened. The ductors were mechanically cam-controlled, so they were not going to suddenly throw themselves into the wrong position and destroy anything. They simply stopped behaving correctly.

To the pressmen, and initially to us, it looked intermittent.

And there are few things more annoying than an intermittent problem.

We checked everything: encoder, counter, position, inputs, outputs, timing, wiring, program. I did not leave a damn thing unchecked.

Saturday disappeared. Then Saturday night disappeared.

Around three o’clock Sunday morning, we finally left the plant. The press had to run Monday. The customer did not care about my controller or my lack of sleep. They wanted production.

Fair enough.

The napkin

We went to a Gasthaus in Kempten, something like a small Hofbräuhaus. Big wooden tables, Bavarian atmosphere, food, beer, the Allgäu around us.

We sat across from one another at one of the tables.

And I started doodling on a napkin.

I drew the cylinders. I drew the encoder relationship. I wrote down the degrees.

The encoder was not wrong. The program was not wrong. The counting was not wrong.

The whole system simply had no idea which half of the impression cylinder it was looking at.

The encoder made one revolution while the plate and blanket cylinders made one revolution. But during that same time, the double-circumference impression cylinder had only made half a revolution.

So the encoder’s index signal arrived every 360 degrees. But the ink ductor function belonged to a 720-degree mechanical cycle.

The first side of the impression cylinder might require a function at 120 degrees. The corresponding position on the other half was not 120. It was 480.

Same encoder position. Completely different position of the impression cylinder.

When the controller powered up, it could find 0 to 360.

Once I saw it, I knew the controller needed a reference pulse only once every 720 degrees.

The encoder’s index pulse had to go.

That’s my pin

We went back to the plant. Now I needed something on the press that physically followed the correct 720-degree relationship.

I did not know exactly what I would use. I figured I might have to fabricate a metal flag on some rotating component and put a proximity switch over it.

So I started looking around the machine.

Then I saw it.

A pin.

Part of the large delivery-chain sprocket assembly. It was already there, and it rotated in exactly the relationship I needed.

I called Manfred Pilz. Manfred was the electronics man with Werner Grossmann’s group down there. Grossmann ran the agency operation. Herzing was the electrician with his own installation company. Hiltenbeutel was the mechanic. Pilz was the electronics genius.

Manfred was one of those people you could call at any hour. I do not think he slept much anyway. The man bought electronic components in such quantities that I once joked he could use integrated circuits for traction on the ice outside his workshop.

Naturally, he had the proximity switch I needed.

Naturally, he had shielded cable.

Probably enough to reach Hamburg.

He came in. I fabricated a bracket. Manfred helped crawl into the machine and mount the sensor. We positioned the proximity switch over the existing pin so the pin passed directly beneath it once during the correct mechanical cycle.

Then we ran the new signal back to the IFM controller. The old encoder index signal was disconnected. The new proximity-switch pulse went into the same input.

Now the controller did not reset every 360 degrees.

It reset every 720.

There was no suspense. I knew it would work. I would have bet my house on it. There was nothing else left.

We started the press.

The ink ductors worked.

We stopped it. Started it again.

They still worked.

The machine finally knew where it was.

Monday morning, Kösel got its press back.

Years later

Years later, I was working on a completely different Planeta press in St. Louis, Missouri.

That machine was hydraulic from one end to the other and leaked oil everywhere. We called presses like that an Ölschwein.

The problem I was there to fix had nothing to do with ink ductors. But while working on it, I noticed something interesting.

There was a large exposed gear with two colored regions marked on it. Before starting the press, the operators had to inch the machine manually until the gear was sitting in the correct colored region. Otherwise the ink ductor relationship was wrong.

I looked at it and smiled.

The difference was that years earlier, in Kempten, with a napkin, a proximity switch, and one conveniently located pin, we had actually fixed it.