OT : Thermocouples

[leitmotif]

Kid

No matter how it gets done it is going to take time to get the TCs connected correctly.

Remembering back a little yes the dies are heavy and awkward. Yes if they roll over a connector that is the end of the connector.

Think on this for a second.
Time 0 start zone #1. Sense current
Time 0 + 30 second look for temp Z 1
+ 60 second measure temp z 1
All you are looking for is a temperature increase even if only one half degree.
Compare T 60 to T 30 AND if 60 more than 30 then turn on zone two repeat for two. IF NOT turn on check TC connection screen.

Dan Bentler

 

[Honga]

RFID - I'd go with readers in a base, beacons on the plug. When you plug it in get your plc to compare the plug. If wrong or missing then give an alarm or something.

Just trying to think of a more elegant solution, so you use less readers.

 

[Honga]

Thinking about this further.. You could even program the plc to change it's config so the plug orientation doesn't even matter! The idiots could plug any of them into any plug.

 

[Lancie1]

At 19 or 20 years old, you have virtually ZERO experience. Be glad you have a job.

I believe the Kid has bigger thing in mind. He will be running a company like this someday....

So, to summarize ideas from many users:

On each line there is a extrusion dies that has about 12-20 zones and a TC for each zone. So you have up to 20 zones and thermocouples on each line. How many lines are there?

On the lines that have plc controlled zones, I limit the setpoint and I limit the output. So not all lines have PLC controlled zones? How many do?

These dies have power and TC plugs on them because they are moved to other lines and changed out a lot.

From the jack back to the plc card is 50-70 feet on most lines. From the jack to the die it is 10-18 feet.

Hardwiring is not a option as TC stays with the die so it has to be unglugged and they move 2-3 times a week on average.

All zones control and go to the plc TC input cards. So all thermocouples go to PLC input cards?

There are already policies in place to check things as they have been done. This is done by the line tech but most of the time done half a%&ed.

I was also thinking of running a loop for a dc digital input through the connector with the TC and use a different Pin on each and not allow heat contactor to engage unless the correct one is plugged in? What do you guys think? Great idea.

We have grounded TC that get that way by being crushed from time to time and cause problems also. My TC input cards detect TC opens but not shorts i don't know what could be done to detect that.

But it does get old idiot proofing everything to the T and believe me i try. Some parts of the process you just can't idiot proof any further just the nature of the process and sometimes the management here that just goes right over their head. Actually I have found that there is always at least one more way to idiot-proof just about anything. The key is to find the best, most practical method.

I like the RFID idea. RFID tags are cheap. Readers are a little on the expensive side. But if the pockets are deep enough and you ever wanted to play with cool new technology here is your chance. I like it too, Alaric.

When it comes to idiot proofing, what I do is look at what the operator error is, and then what I can do to eliminate it. It sounds simple but a lot of times it isn’t. Great technique, Clay.

Note: the color coding made the operators responsible for the hook up. Calling maintenance when you had tried to cross colors guaranteed the operator received disciplinary action. No way you could argue that if you had been paying any attention whatsoever you got the colors mixed up. Coded, keyed, or colored connectors should do the job, but would require the approval of some higher up to fund the cost. I am not sure the Kid has such cooperation on this project.

Starting and testing the zones one at a time just won't work on this process. It would take too long. These are very large extruders and it takes a bit of time to see a good degree of temp change. It normally takes 5-6 hrs from cold for these boys to heat up and soak in and be stable so all zones have to be started at once. Could the tests be done after a zone operation has started? If error found, alarm sounds, zone shut down, connections corrected?

Wireless would have to stay with the die and get torn up the first day here. I am going to have to stay with a plug-in solution. We do have a couple things on banner sure cross though they are really neat. Use the PLC if possible to help solve your problem. You can control the program and the costs.


Think on this for a second.
Time 0: start zone #1. Sense current.
Time 0 + 30: second look for temp Z 1
Time 0 + 60 second: measure temp z 1
All you are looking for is a temperature increase even if only one half degree.

Compare T 60 to T 30, AND if 60 more than 30, then turn on zone two, repeat for two. IF NOT, turn on check TC connection screen.

RFID - I'd go with readers in a base, beacons on the plug. When you plug it in get your plc to compare the plug. If wrong or missing then give an alarm or something.

 

[REHJ]

Just a thought? Would it be possible to add a "dummy" key to the system. More than a color match or a tag match but a physical extra keyed plug that should be plugged up to insure a matching TC to a Plug? It would run with the TC and attached with a short enough lenght so that it could only be plugged into the matching TC-Plug. To make it really effective it would not necessarilly have to be a dummy but a real input to checked against logic so that it has to be correct?
Also the approach of having people held responsible for their actions as a way of controlling a process!!! That system is long gone... Should not be but just is.

 

[OkiePC]

The sad part about making engineering solutions to people problems is that when the trained apes foul something up, it is now the engineers' fault. Not to mention the degree to which productivity suffers due to all the extra steps, extra hardware. It's just one more step on our inevitable evolution toward Idiocracy.

 

[Lancie1]

Another great idea, REHJ. It may be possible. I think that part of our problem at this point is not having a clear definition of what is "possible".

Sometimes better ideas are generated if we stop looking at the BIG PROBLEM, and instead focus on the SOLUTION. This is a brain-storming technique, so things get a little quirky. We don't know the solution, but what do we know about the solution?

Lets try to put the solution in a box. Often when the solution is defined in terms of What is Possible given the present real situation, budget, personnel, equipment, knowledge, free time, and so on, many possibilities are eliminated, leaving only those that can be done.

The Solution to PLC Kid's Thermocouple Problem:

 

+-----------------------------------------------------------+

| Cost: How much available to pay for the solution?         |

|                                                           |

| Time: How much time is there to design, build, install?   |

|                                                           |

| Personnel: Who is available to help with solution?        |

|                                                           |

| Equipment: What stuff might help, and is it available?    |

|                                                           |

| When can the solution be done: anytime, only during       |

| night shift, weekends, holdidays?                         |

|                                                           |

| (Other restrictions and constraints on the solution here) |

|                                                           |

|                                                           |

+-----------------------------------------------------------+

The questions in the box lead to many other questions, and when all the questions are answered, the obvious solution to the problem is usually staring you in the face.

About costs: My experience is that if the payback period (time for the Solution to pay for itself) is more than 3 years, you can forget about that particular solution. Company presidents do not look much beyond 3 years, and if you want to make sure it gets done, the payback must be 1 year or less.

 

[Lancie1]

Okie,

That is true. Once you become the champion of a solution, you had better make sure it is the best, cheapest, and quickest, or your name will be mud and they will never listen to you again.

One way to get ideas adopted is to go to the guys doing the work and find out what would work for them. This concept of going to the final user is also part of developing teamwork. Once people buy off on an idea, they will either help you make it work, or at least not try to prevent it working.

 

[Paul T]

Kid

No matter how it gets done it is going to take time to get the TCs connected correctly.

Remembering back a little yes the dies are heavy and awkward. Yes if they roll over a connector that is the end of the connector.

Think on this for a second.
Time 0 start zone #1. Sense current
Time 0 + 30 second look for temp Z 1
+ 60 second measure temp z 1
All you are looking for is a temperature increase even if only one half degree.
Compare T 60 to T 30 AND if 60 more than 30 then turn on zone two repeat for two. IF NOT turn on check TC connection screen.

Dan Bentler

Yeah... something like this is exactly what I was thinking. The logic would turn on the zone output at 100%, with the PID in manual for the test. Then, you monitor the zone feedback. How long does it take for the zone to change 0.5 or 1 deg? One minute or less would be very reasonable, I hope. That's all I'd be looking for. If you can detect the temperature change in 15 or 20 or 30 seconds, so much the better... you end the test after a 0.5 deg or 0.75 deg or whatever temperature rise, and go on to the next.

Aside from that, you'd probably want to check the zones in some order that would minimize crosstalk. That is, rather than checking 1-2-3, etc, check something like 1-4-7-10-13, 2-5-8-11-14, 3-6-9-12-15. That would help to minimize the influence of the previously heated zone on the one next to it.

I suppose that with a little more work, it would be possible to figure out where the incorrect TC's really were, if some were found. But then you'd have to present that info in such a way that the people who weren't paying attention the first time did better the second time. That... would be hard, too. So I'd still go with a message like "TC/Heater Mismatch, Check all TC's for correct placement" and let it go at that.