Broken leg sensor holder ===> Broken PLC

During a project for a roller table I am using some proximity switches between the different rolls in order to start the rolls on time and monitoring. Roller table is used to deliver the manufactured 6-12 meter pipes to their testing station. Mechanics advised me to use the below shown method to fix my sensors among the stands and protect them from being harmed by the crossing pipe (sorry I am very week in drawing )

At first I recognized that I do not have a permanent signal from my sensors, pipe jumps up the rollers some time I accepted that and made the program not mentioning the signal cut for short times . but the problem gets worse now . Some of the wrong pipes break the whole steel made holder off (thanks god sensors are happy in their cage).

We must change them it was a fact. Photo electrics are not a good substitute they are sensitive to any moving thing even the laborers moving among the stands they will ruin the sequence. Bigger proximity sensors (e.g. 12mm) are rather big to be placed and of course not economical at all . I am not sure if a 40 mm sensor can stay far enough from the dangerous pipe crossing. We are not very keen on limit switches as we are not prepared enough to build new holders and rough placing plots for them.

We are clueless about the problem. We thought about a flexible or spring leg for our holder in order to avoid break we are working on that but I thought one of the professionals might have a relative experience and solution.

P.S : Sorry for my English

You should look at the newer style photoeye types sensors that have a "teach" ability...ie they can be taught how far to look down to the millimeter range. You can "teach" them to view in a range that begins at 200mm and ends at 300mm as an example. They are capable of being used from 200mm to 50m. We have some on a line that puts paper rolls on a cleated conveyor, these eyes look down at the conveyor belt and detects when there is any object on the belt, these things are very accurate.

This is a link to the brand we use:
http://www.ifmefector.com/ifmus/web/photoelectric.htm

There main page has links to sites for different countries so you may be able to find it in a language you are more comfortable with.
http://www.ifm-electronic.com/ifmint/web/home.htm

As Ron stated, photoelectric sensing should NOT be ruled out. It's probably the most suitable for this application. Have you considered looking across the rollers rather than trying to peer up between them?

You need to research what is available to you. There IS a perfect sensor for this application... You just need to find it!...

You also need to determine WHAT you're trying to sense. You say that the pipes sometimes jump around. How much? If you increase the sensing distance from 2mm to 4mm, will this make a difference? Or, will you STILL miss some pipes?

Once you have this information, you can start researching sensors. Here's links to just a few sensor manufacturers...

Banner, Keyence, Sick, Omron

There are LOTS more!

Happy clicking!

beerchug

-Eric

I will add to that list.

Carlo Gavazzi
Globetron Controls
Allen Bradley
Honeywell
Microswitch
Turck KEP

Oh yeah?...



HERE'S a list of over one-thousand of 'em!... moon2

beerchug

-Eric

P.S. Only kidding, Ron...

This LIST has more detail

Touche Ron!...

But not really.... The link I gave is part of Thomas Register. Your link lists 1019 mfgs. and so does mine!...

beerchug

-Eric

I knew that

<<He made me do it

Sorry but I did not get it right

"they can be taught how far to look down to the millimeter range. You can "teach" them to view in a range that begins at 200mm and ends at 300mm as an example. They are capable of being used from 200mm to 50m"

I visited the Omron site which is widely available here, They are the conjunction of an led and a reciever in the shape based viewed in the picture below they can detect a moving object on a conveyor.



but how can they avoid the crossing workers ?
and how is their price in comparison with the normal photo electric

Using your picture as an example the photoeye should only detect items in the range you set it for. Lets say you mount it above the conveyor (as is done in the picture) 150mm then set it to "see" items on the conveyor. At this point the photoeye should not sense anything before or after the set point, people walking by or near the conveyor should not affect it. Only if they are placing their hand or body between the eye and conveyor will it "see" it.

I have worked in a bottling plant where the photoeyes were used to detect if bottles were filled, were filled to preised point, and if the bottles had caps. They are that sensitive.

I think I mentioned the paper rolls we use them with now where the sensor is mounted maybe 150mm to 200mm above the conveyor input to detect the rolls going onto the conveyor. The operators work around this section all the time with no affect on the sensor.

When you adjust these type photoeyes they are "taught" to look at the area just above the conveyor. Say you mount it 150mm above the conveyor then set it to detect at 100 to 145mm (this range depends on object size etc). I have seen these things detect items as thin as a sheet of paper. The background will not normally have an effect.

Normal (as you call them) photoeyes are just as reliable in most cases, they can be set (usually with a sensitivity dial) to just detect items (which can be metallic or non metallic) which pass in range. We use many of these, our boxmakers use them to detect the cardboard when inserted, for jam detect on conveyor, for diverting, etc etc. In the old days where proximity sensors were used many of those are being replaced by photosensors. An example is a stretch wrapper we have, it uses a photosensor to read the teeth on the sprocket of the drive motor, this is used to count the turns so it will make 2, 3, or 4 wraps at top and bottom of the boxes on a pallet then stop at the same place it started.

I agree with the others in that your problem lies in the sensing method. Writing PLC software to compensate for poor sensing could eventually cause more problems than it solves.

Another thought on sensing---You might consider a "thru-beam" or "retro-reflective" photo-sensor.

The thru beam consists of a transmitter and reciever looking at one another. If something passses in between the two, it senses it.

The retorreflective work kinda the same way but the transmitter and reciever are in the same unit. It transmits to a reflector and recieves the reflected signal. If anything passes between the sensor and reflector, it sees it. One thing to consider with the reflective type--make sure your sensed object is not highly reflective, as the sensor could mistake it for the reflector.

We actually had this reflective problem sensing stretch wrapped pallets on a conveyor. The sensor thought the stretch wrap was the reflector and would trigger 2 or 3 times as the pallet went by. We went to a Diffuse type sensor that sensed about 2 feet, half way across the conveyor, but not people walking on the other side of the conveyor. As eric pointed out--know the entire range of things you need to sense, then look for a sensor that suits your needs.

I think this thru method would work, unless the workers in the area are actually upon the conveyor with the pipes and could get between the eyes/reflector.

Both thru beam and retroreflective are much more robust methods than diffuse reflective sensing, yet a little more trouble to implement because you have to mount and align two things. Although diffuse sensing is more than adequate in many applications.

Good luck.

KISS, use a through beam photo eye.

Now you can detect light to dark and dark to light transitions at a known line-of-sight.

The range sensing detectors seem to be giving you problems because they will trigger on anything that is within range. From your drawing it does appear that your detector is too close the pipes to survive for long.

When using a through beam photo cell, when the signal goes from light to dark you must start a timer. If the photo cells turn from light to dark before the timer times out the light to dark transition was 'noise'. The timer must be set long so only real pipes will get recognized. ( usually I used an encoder to measure length instead of a timer ). The photo eyes source and detectors can be recessed so that the pipe can't possibly hurt them.

In the past I have done saw mill and veneer mill applications that use photo eyes to detect position, lengths and diameter of logs. I have never had any problems except for the initial alignment of the source and detector. In this case I think simpler is better.

Prince said:

Omron....which is widely available here

Click to expand...

Then have an Omron rep come in and show you his "wares"...

Since the machine is "in use", it should be very easy for him to specify the correct sensor for your application. He'll be able to test a variety of sensors for you with no cash outlay on your part!... :nodi:

Consider yourself lucky, you already know what you're looking for.
Try picking a sensor when there are a lot of "unknowns" in the process... Not fun!... banghead

beerchug

-Eric

thnx for all the soloutions

let me repeat what I learnt today , I can not use the metal sensing proximity sensors as they must be placed near the moving pipe. I can use normal or teaching optical sensors in order to fix the problem but I will still have the problem of mis-sensing the unwanted hands and bodies in the area . As I do not have real conveyor I mean a reference plate to adopt my sensors to the normal opto would be enough or they can be used by reflectors.

considering the fact that I want to place the sensors each 6 m in the line which might have continue for more than 200m I can not think about the expensive color sensing or pattern sensing sensors so I must choose one of these.( It is a long roller table and we are just testing a small part )

1 - Normal photo sensors will see the pipe accurate and safe but they will also sense the curious workers.I can write a system like Peter advised to block the noise.( all noises must not be delayed if so they would be considered as the pipe)

2 - metal sensing Proximity sensors will not have the noise problem but then I would have a problem with disconnection ( which easily can be solved by software )and a hardware problem breaking sensors. I have a suggestion from our mech guy to attach each of my sensors - which are now guarded by a steel cage - to a spring like material that can easily avoid the break.

does anybody have another suggesstion or did I get something wrong ?

thnx for the help.
I would inform you about our latest task.

If you could take a picture of the conveyor and parts being detected it may offer a foregone conclusion.

I personally can not determine where the people/workers are a problem with photoeyes. I do not understand the "I dont have a real conveyor" part either, you dont need one using either the "teach" type or rectroflective type, just an area to mount them. Anything that breaks the optics will be detected. I would think that there has to be an area where the objects are loaded, unloaded or moving that is uninfluenced by workers, maybe you need to relocate the sensor?