Should I use a PID or...?

Hello everyone.

I'm currently working as an intern for a company that has assigned me a project to work on. In my last post I went more in detail about the project, but long story short, my job is to make my company's ped welding stations safe, by being able to detect any obstruction between the weld tips, such as an operator's hand.

The only thing I have to work with is an LDT that tells me position of the weld tip. Other than that I wasn't given a whole lot of instruction. I've come to the conclusion that since I know the distance between the weld tips, and I know the time it takes to extend the weld tip (I can adjust that with a pressure regulator), that I can figure out velocity.

My thoughts are since the LDT has a linear rate of change, any obstruction between the weld tips should cause the rate of change of the LDT to slow down for a split second, making it non linear.

My initial thought was to repeatedly run a short timer as the weld tip extends, say 50ms. At the start of the timer, record the position of the LDT, and at the end of the timer record the position of the LDT. If the LDT didn't move between a certain distance in that time, say 28 - 30 bits of change, then something had obstructed the weld tip, and a fault had occurred.

I'm not sure if my intital idea will work or not, but the other day, one of my co-workers suggested maybe trying to use a PID. Being only an intern, I have never had to use a PID. I understand the general idea behind them on how they work, but when it comes to setting one up and actually using it, I am pretty clueless other than having read the instruction help file and several other articles on the web.

I am stuck using only the LDT and time to figure this out. I am not allowed to add anything else to the ped welder.

Any thoughts or suggestions on how I should approach this would be greatly appreciated. Thanks!

Hodor,

#1 A PID is for an extended period of time (minutes to hours), not seconds.

#2 You should NOT put a regulator on the spot welder for controlling pressure, your welds will be cold and inconsistent based on my experience.

#3 I'm sorry, OSHA rules when it comes to two hand anti tie down requirements. If there was some way to lower a guard when you pressed the foot pedal and GUARANTEE the guard was lowered all the way and that no hands were caught, you would then be able to activate the welder. The rules have changed since I saw this operation and I am no longer certain if it's even allowed.

As I said in your last post, Know the rules (all of them), Anything you come up with MUST be approved by the operators, maintenance, supervisors, safety, and management and everyone must agree !! Accidents happen everyday and as the person in charge of the project, you must defend how the system works. If you can ask WHAT IF, there better be an answer. a 1 in a billion chance of an accident will happen within the first day of production (learned that the hard way).

james

A PID function is the wrong tool to apply to the task you described (calculating speed). Assuming you know the pressure, your method has a chance of being able to detect a deviation from the expected speed.
Having said that, do not for one second consider your method acceptable to protect an operator.If you have to ask why its not acceptable, you haven't done enough research into current safety standards.

first things first ...

I'm currently working as an intern for a company that has assigned me a project to work on.

Click to expand...

is this supposed to be a "real world" project - that will eventually be incorporated into the plant's operation? ...

or ...

a "let's see what the new intern can come up with" type project - which is supposed to be mostly a learning opportunity - and which will never be put to actual use? ...

knowing the answer to that question might make a big difference into how you approach the project - and to your final "grade" on the solution that you eventually come up with ...

Ron,

Here's a link to Hodor's other thread about this project. http://www.plctalk.net/qanda/showthread.php?t=103730
We've had reactions similar to yours.

thanks, Steve ... I hadn't seen that one ... looks like the forum has already "been there and asked that" ...

party on ...

Thanks for the responses everyone.

It sounds like a PID is not the way to go then.

When my boss first told me about the project, he said it was more of a test project and not that pressing, but he did say that is was something that the company was interested in. For those who didn't read my original post, basically my company has several hundred ped welding stations that currently use palm buttons to operate. My company is interested in going back to using foot pedals instead of palm buttons, but to do this, the welders must be safe from crushing an operator's hand between the weld tips. That's what I am tasked with doing.

I do fully believe that this is something that my boss actually wants to achieve. I find it hard to believe he would go through all the effort of having the shop spend a week tearing down an existing ped welder and adding a foot pedal to it, mounting an LDT to it, and adding a whole new panel and PLC if he wasn't serious about this project. There is also a set of custom CAD prints drawn up for the setup as well.

That being the case, this very well may be an example of let's give the intern an complex task and see how he does and what he can come up with. After I go back to college, he may hand over the project to someone more knowledgeable...or for all I know, he may have someone else currently working on the same thing.

I was initially given the CAD prints and the PLC logic for the current weld stations and told to look through them. After that, my boss sent me the following links and told me to replicate with similar results. After that I was pretty much on my own to come up with whatever I could.

(http://www.unitrol-electronics.com/soft-touch.html)
(http://www.entroncontrols.com/en/smart-detection.html)

(The second link contains a short 1:20 video called "SD Obstruction Detection" that shows an operator holding a part being welded. Various objects such as a pencil and a piece of vinyl tubing are then stuck between the weld tips and they are not crushed. This is the exact thing my boss wants me to come up with)

Given very little instruction on where to begin, I started reading up and thinking of of sensors that I could add to the welder to aide in making the welding operation safe. That's when I made my original post asking for help and suggestions, but when I later brought up adding anything to the welder he just looked at me and essentially said no.

Given all of your responses and how quickly my boss shut down the thought of added anything to the welder, I'm beginning to think to think that he already has an idea of what he wants done, and that this whole thing is merely a learning experience / test for me to show what I can do. When I did express concern early on with me working on a safety system for a process I do not fully understand, my boss seemed rather unconcerned and told me not to be shy and have at it. A fellow classmate of mine actually interned for the same company last summer and they offered him a job at the end of his internship, even though he never finished the project they gave him because the machine he was supposed to work on wasn't built until the last week of his internship. He currently works full time sitting a few desks away from me.

But regardless of if this is merely a test for me, the intern, or this is something my boss actually wants me to get done, I'd like to get something working, even if my boss or someone else ends up shooting a hole in my design. I'd like to have something to report by the end of my internship rather than coming up empty handed. So if anyone has any ideas or suggestions on how I should proceed, I would be more than grateful to hear them. Again the only things I have to work with are the position of the weld arm from the LDT, and time.

Thanks again for all the responses and support, and for the responses to come.

Sincerely,
The Intern

What you describe in the video sounds a lot like a mold protection circuit in an injection molding press. In that case, the mold closes rapidly and without protection until it gets close to closing, and enters mold protection. In the mold protection zone, the press closes with only enough pressure to overcome gravity and whatever friction exists. If a metal case insert is placed off center, or a part of the mold or a bolt fell out and is laying across the mold, it will stop the press from closing and after a short time a timer will time out if it had not reached the end of the mold protection zone and the press will fault out. Once the press reaches the mold protection end switch, it closes normally and with full tonnage. This switch needs to be set as close to the physically closed position as possible for maximum machine and tooling protection.

I would in NO way advise doing this as a means of personnel protection, as others have mentioned.



Also, Hold The Door.

Thanks for the idea, It's something I might try next, if my current plan doesn't work. I see how both aren't foolproof, but having something to report to my boss is better than saying "I got nothing"

Hold the door, indeed

If you haven't already done so, I suggest you talk to your facility's safety compliance person. Whatever solution you come up with, this person will have to sign off on it before it can be deployed. Best to get his or her input early in the design stage.
I also suggest you see contact the local representative for the stand-alone products offered by both of the vendors you provided links to (and any others that may offer similar products). See if you can get the representative to come in to your plant for a demonstration. The products they offer have already been vetted. They wouldn't be offering them if their lawyers and insurance carriers hadn't had their say. For you to try to design something similar from a clean sheet of paper would be akin to reinventing the wheel. The only reason to go that route would be if there was something about either product that rendered it unsuitable for your application.