Dangerous novices programming plc's

[Matthias von Zorn]

Thats what I thought. Thanks OZEE.

 

[Struggling_forward]

I try to imagine whats the worst that can happen ...............if the PLC out gets "stuck" on, or the isolation relay fails closed, or the Estop sticks.

Or a novice programmer forces the output on and walks away forgetting about it.

 

[icky812]

If so, then the second question would be about traffic lights...

 

[S7Guy]

I can see how shoddy programming could hurt production or even damage the machine. Heck, even experienced programmers can do that.

But if hundreds can die or thousands can be poisoned or fingers can be chopped off or bodies crushed due to a programming mistake, isn't the inherent design of the machine flawed anyway?

Looking at the various machines I have worked on (which are many), I couldn't write a function to hurt someone even if I tried. I mean, if someone has opened a door to work on a machine, how could faulty code make a shear pop up and cut his hand? Personally, I am much more concerned about the qualifications of the electrical and safety hardware designer than the qualifications of the programmer.

As for licensing, I guess it wouldn't hurt in theory, but reality is a different thing. For one thing, it's impossible to even get hardware suppliers to come up with a programming standard, let alone getting the programmers to agree. And who would evaluate the programmers? Some kind of government worker? If he knew anything about programming, would he be working for the government?

So, I think the most reasonable solution is to regulate safety and hardware (as it already is), and let individual plants decide on a programming standard.

 

[Matthias von Zorn]

Some machines can be dangerous. Take an atmosphere furnace. Its only filled with hydrogen, and in a sealed chamber, at 1750F. What could go wrong? Can you say bomb? Its possible, just through incorrect operation to induce air intake into outer chambers. The PLC has code to try to make sure this doesnt happen. Maybe that tech there was annoyed with that delay that would not allow him to open a door till the air is purged and burnoff is established. He is in a hurry, he will just remove some of those interlocks to get that load in faster. A few forces here or there, and its gulping air. Air + H2 + Heat + compression. So,yeah, some machines are dangerous if misused.

 

[n9xcr]

They were just jumping them together at the switch!!!!!!!!!!!

There are ways to put a stop to that.

 

[PLC_Idiot]

First time poster, enjoying the thread.
Certified does not make one qualified. There are some very bad electricians out there who the state has certified. Just as there are some very good electricians out there who the state has not certified.

You could set up a test, but then someone would start a two day seminar to teach you how to pass the test, not teach you what you need to know.
Is the certification to protect the public, or to protect our job?
By reading post, I've determined some of you folks are pretty sharp. But remember you once were a newbie, and dummer than dirt!
I might be very versed on GE and AB but put me in front of a Siemens, I'm going to be a "newbie" for a few days. I will be asking for help and from looking around here, it appears to be a good place to ask.

On Safety
Two phrases come to mind.
1) I am ultimately responsible for my safety.
2). I am my brothers keeper.
So basically we are all responsible for ourselves and everyone around us.

No you should never depend on a software EStop.
I am in my 50's, been programming awhile, I love Xbox, and have been known to call tech support instead of reading a manual. Sometimes because I'm lazy, sometimes because manual really sucks! (stereotypes don't always apply)

 

[PeterW]

2 things that the resurfacing of this topic brings to mind.

Firstly I remember reading the OP post when there was about 2 replies, I was going to respond but was too busy at the time then it faded.

Secondly, seeing the name dahnuguy, what happened there?? he suddenly appeared and was all over the board like a rash, making long posts which led me to believe he was going to be the new 'Terry Woods', then zap.. vanish..


Regarding Goody's original post and in particular this bit

'I saw an incident the other day that heightened my concern - a man and all his tools was thrown off a conveyor and injured because someone else actuated a sensor'

Its quite common for a sensor to wake a conveyor up from 'energy save', I've seen many versions of it, where a sensor can wake up a single or multiple conveyors that have stopped due to no action.

The reason to stop is to save energy, reduce noise and wear and tear.

There is normally no 'warning' sound on wake up (if there was then the reduce noise reason is gone). Usually the conveyors have signs around stating that the system is automatic and will start without warning. Saying that on initial start from a dead system there would normally be a start warning.

I prefer the individual conveyor type where every conveyor has its own time out and restart condition. I also prefer that the sensors do not wake up a conveyor unless the conveyor on which the sensor is fitted is running.

The company I work for usually uses the group versions and the conveyors do not need to be running, in the week that Goody wrote his bit, a worker where I was placed his tools on a conveyor, jumped on and was also taken for a ride, he was lucky, he managed to get off before he was fed into the X-ray machine.

The thing is, he should have known better than that, its down to the person to isolate anything before he decides to get onits his responsibility. The system is automatic after all and safety starts at home.

 

[Goody]

With me being the original poster, maybe I should reply again.

I have liked the way the topic has veered and swayed between novice programmers and safety because I am an absolute believer that any safety discussion is never wasted time. Anything that tries to stop complacency (which I believe is one of the biggest factors in accidents) is a very good thing

At no point was it my intention to put off anyone trying to learn the art and craft of plc programming.
I read a few of the 'we were all new once and shouldn't be elitist' type comments and chuckled.
I also laughed at some of the top people here saying they were once plc dummies - and some still think they are, because thats the mark of a excellent engineer - one that knows that if they work for a hundred years they will still not know everything.

So any new plc programmers here, please keep posting and we will keep answering. The ones I dont like make themselves obvious in their questions anyway

 

[just_lionel]

I agree with the posts here. I started out fresh coming from the computer programming background. I was lucky in the sense that my Dad is a Controls Engineer and helped me alot in my growth and development. I have came along ways from when I started, but I still have much to learn. This is one reason why I love this site. I might not post much, but I do read other people issues and learn from them as well. I have always taken saftey into consideration and I double check my work everytime because I know one small change can make a huge difference.

I just want to say Thank you to all here, it is a great place for all to gather and learn. Heck, I have even been able to help people as well!

 

[lipalipa]

a "new guy"

First things first. Hello to everybody. Excuse me for my English, I’m not an native English speaker. I’m new here and also in the integration field. I’ve spend a little time to read this post because of its tittle. I feel that is necessary to make some clarification or what I see as clarification. The integration field is far from an one-stand-job. If you know economics you can be an economist, if you know electric you are an electrician and so on. But what you have to know to program PLC’s? Well you have to know a lot. Is not enough to know programming, drag&drop, networking, electrical, mechanics, safety system, electronics, cabling, system optimization, system architecture. My point is what do you test for a young guy? Who make the test? You can not, it’s impossible to test all.
Regarding the example with the person who repairs cars or with the driver’s license. With the direction of automatizations the car, driving looking to be fully automated, the future mechanics will be PLC programmers? The future driving license exam will be analyzing, testing and debugging embedded programs? Maybe is to soon to search here the answer for this problems.
Reentering on topic, I found out that 1 week of hard work on site it’s often more teaching than 3 years in school without seeing an working machine. I thing the problems, both the safety an fresh-out-school programmers, are based in economy. We have economy studies and cut-cost directives and deadlines more and more shorter and practically isn’t possible to follow and retest a work.
Overall i find that for really programming an PLC in a automatism system is mandatory:
- a kind of expertise in one of the adjacent filed (mechanics, programming, electrical, .....)
- a good communication skill
- a first project where the "new-guy" does nothing, he learns, watch, ask,...
- a second project with daily supervision
- a 3rd project with overall supervision

and maybe after that he can do a project on his on....

PS Sorry for the long reply

 

[silva.foxx]

First things first. Hello to everybody. Excuse me for my English, I’m not an native English speaker. I’m new here and also in the integration field.

Welcome Lipalipa, you're written English is fine and I'm a rag-arsed spark.

I once worked for a small company that manufactured paper napkins and cups. I went there having self-taught on Mitsubishi Fx. I took on a project of backing up the plant's plcs, over a hundred programs.

I had to research AB, Siemens, Omron, Keyence and Toshiba... to find the neccessary software, comms cable and battery requirements for purchase. I then had to learn a little of RSLogix/Linx, Step5, Step7, Syswin, CX-P, KVLadderBuilder to enable upload.

I went to night school for a 2 year course on AB-slc500, got a course at Siemens for Step5 and a local course for Step7 as a result.

I was about to suck the program out of the Toshiba EX100(?) when I got a shout from an operator. The machine was a paper laminating machine and the Halogen heaters, for drying the gluing process, were on and burning the paper that was webbed up.

The plc was in 'HALT'.

This was the original program. The heaters were through normally closed contacts on two relays in series and the program energised to turn off! Needless to say this was corrected for a fail-safe. Another fault I found on this machine was the 110vac center tapped tranny was only fused on one leg. One day a 110v sol. coil developed a -55v to earth fault and burnt up the earth wire all the way back to the tranny. Smmmokin!

I recently completed my first real 'machine' program as part of a controls upgrade from relay to programmable logic. We were just going to convert but chose to go from scratch with a Siemens S7-300 and Cat4 safety set-up. I'd say I am between novice and intermediate. The program worked fine straight away, I've just to add an interface signal or two then it's done.

My 'license' has been the opportunity to work with all these softwares, see many programs, the continual use of this website, the priveledge of books, working with others who have plc experience, time to learn and the vast data you can obtain from t'internet. I am so lucky to be a rag-arsed spark!

 

[uptown47]

Reading through this thread I think I'm in agreement with the majority of posters in saying that if 'experience' and 'competency' were a pre-requisite for doing any work then how would newbies develop.

Training only goes so far. To develop properly you need to be in the 'thick' of it actually making changes, modifying code, coming across and then surmounting obstacles. This is how we as programmers, and as humans, develop and improve.

I think anyone who has been in PLC's for a while will have had that experience of being a bit 'out of their depth' but you work through it and learn from it.

The problem with 'licensing' would be that as soon as it was introduced it would become a requirement in order to comply with health and safety laws. This would mean that only the people who had attended the course would be in the programming roles.

I've been around long enough to know that just because you've done the course doesn't mean you can do the job!

I think it should be down to individuals to be able to ask for help if they need it or even get someone to check their code if they are unsure of anything.

And, to reiterate what has been said many times in the post, machine safety should never be reliant on the PLC anyway.

:site:

 

[dahnuguy]

still around , busy, one legged man in asskickin contest

Secondly, seeing the name dahnuguy, what happened there?? he suddenly appeared and was all over the board like a rash, making long posts which led me to believe he was going to be the new 'Terry Woods', then zap.. vanish..

Don't know Terry Woods,

Sorry about the rash if I need to be.

No zap no vanish , no email alert that this thread had been responded to and my time has been consumed with applying what I have leared here and been able to figure out.

Basically, I was here allot due to being unable to proceed with my task. It is a point of sucess in the tutelage I received here that I have not been around much of late as I have been swamped with using this new knowledge.

I also have not received any emails up to this one today, so I was quite surprised to find this had grown to this length.

Lastly, I see everyone has a firm grasp of the questions and issues and we all agree there is no solution or none is needed.

As for the why of testing, there are people in charge who know nothing of programming or safety, these persons have hired someone who likes the look of this job and can spell PLC. The tech is pressured to say he can even if he cant to keep his job/ good standing/ pride/ ego. The boss is pressured to get this dangerous thing done without thinking through the after effects and the tech doesn't have the confidance or knowledge to convince the boss it may not be a good idea. So off we go down the "OOPS" highway.

I have been a witness to this , first hand. I was able to illustrate my point by just asking questions until they "figured it out and explained it to me."

"UM I was just wondering, what happens if that valve opens now? Is that bad?"

yes very bad.

"UM ok how is that different from what we are doing in this edit?"

well its completely ............um oh wait nonono that wont work , let me explain why and make it look like I am telling you.

The question remains , with processes that are unsafe if programed incorrectly who will make the changes and who will check the changes and who will say who should be doing it?

I am not so big in ego to get in over my head. And I am always trying to "break" what I write as a test hardening. Doesn't everyone do this? .........No everyone doesn't. Not everyone can even see what might go wrong.

My version of the machine is like a horror show where everything conspires against you and every component fails at worst moment, etc.

I often hear "What are the odds of that happening?"

The odds have to be very small and the danger must require multiple failures before I will even consider leaving it alone.

Even then it bugs me.

An extra valve costs 500 dollars? so............what is a whole batch of product worth? or someones health?

Well that's all I have to say about that..........................today.

 

[bernie_carlton]

Regarding the remark concerning testing and just teaching to the test. As an Air Force instructor the objective for a test was such that, if they could pass it by memorization or whatever then they knew what was needed for the job. If sucessfully answering the test questions, while knowing nothing else, meant that the test taker was unqualified THEN IT WAS THE MAKER OF THE TEST WHO FAILED.

The question according to my Air Force instructor was 'I'm not a mind reader. What BEHAVIOR do I need to see to convince me that the person is qualified?' Answering that question defines the test. If passing a test based on the answer to that question does not make a person qualified then the question was answered incorrectly.