Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations SSS148 on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Looking to replace inductive prox switch in demanding application

PaulKraemer

Electrical
Jan 13, 2012
155
Hi,

I am working on some registration controls on a printing press where we use an inductive proximity sensor to detect metal targets on a rotary screen coating station. The targets are mounted on an end-ring that holds the screen and has a gear that allows us to drive it. In the following photo, top/left shows an end-ring and a screen prior to assembly. Top right shows an assembled screen with end-ring. Along the bottom are three different end-rings, with five targets, one with three targets, and one with two targets.

1744739216173.png
When the screen is installed and rotating, these targets pass in front of the proximity sensor, which is mounted in a fixed position. The following illustration shows how the proximity switch is aimed at the targets ..

1744739101999.jpg
The proximity sensor has been installed on this machine for a very long time. We do not know the manufacturer or model number. From its appearance, I believe it is similar to the following Omron part…

https://industrial.omron.eu/en/products/E2E-X1R5C18-2M

… the sensing range of this Omron part is 1.5 mm (just a little less the 1/16 inch). The operator is able to adjust the sensor position closer to or farther away from the target(s). He generally tries to set the distance just under 1/16 of an inch. If he makes it too close, there is a risk of a target hitting the sensor the sensor. Too far away, and targets may not be detected.

The targets are mounted in precision machined slots on the end-ring that are supposed to be exactly 180 degrees apart (for the two target screen). The targets were also precision machined at some point in time, but I feel like the detection surfaces are not in perfect condition. Maybe over the years, they bumped against the sensor at times and/or they may have been slightly disfigured by careless handling. We have an encoder on the screen motor and controls that allow me to count encoder pulses between detection of the targets. For optimal performance in our application, the number of encoder pulses between each target should be the same. I have seen differences in these counts that make me question the proximity sensor.
Right now, we are running a job with the two target screen. With the proximity sensor in its initial position, we were only detecting one of the two targets. By moving it a little closer, we got it to detect both targets. This makes me believe that there is a very slight difference between the sensor-to-target distance for one target relative to the other. If this distance is different, I wonder if one target causes the sensor to activate a tiny bit earlier than the other.

Not knowing the manufacturer / model number of the proximity sensor we have, I have no clue what its specifications and performance characteristics are. This makes me want to consider changing it.

I am wondering if there is a proximity sensor or other type of sensor anyone here would recommend for this application that would give us the most repeatable turn-on, even if there was a slight difference in distance to one target versus another. For mechanical / space reasons, it would be ideal if we can find something with the same threaded barrel shape with the same diameter as what we have now (which I believe to be 8 mm, based on the M8 designation on the Omron part).

Any suggestions will be greatly appreciated.

Thanks in advance,

Paul
 
Replies continue below

Recommended for you

  • Hall effect
  • capacitive
  • electrooptical
are all possibilities, depending on your particular requirements
 
Thanks IRStuff.

I am just hoping to find a sensor that will give me the most repeatable turn-on when the target (which I believe is 1018 steel) comes into it "field of view". Field of view is most likely not the best term, but I am hoping to find a sensor that would be as accurate as possible, and as forgiving of cases where one target might be a tiny bit closer to the sensor than the other target. I need a single NPN open-collector output. Ideally, I'd like something in a 8mm threaded barrel shape like I have now.

Based on this, would you lean towards Hall effect, capacitive, electrooptical, or something different?

I appreciate your help.

Best regards,
Paul
 
The targets were also precision machined at some point in time, but I feel like the detection surfaces are not in perfect condition
And there you have your root cause.
Proximity switches have a detecting surface that looks like a dome if you see them from the side and for them to detect a surface and produce a "1" a certain percentage of the area of the diameter of the sensor and dome seen from above must be covered by the metal target "flag".
So the target or "flagg" must have the same area against the sensor for it to produce a signal at the same distance.
But since you need to move it closer to get a signal from both sides this is not the case.
I would make new targets or "flags".

I don't think a new sensor will solve the problem, even if Omron makes good sensors the problem is the target flag, it can even been seen at the photo at the bottom right, that the one to the left isn't 90 degrees and flat.

Might add we have proximity switches that are more then 50 years old in similar function working fine.
 
Not sure it makes sense i hope the picture helps. :unsure:
The 60% is just made up.
But the closer the sensor is to the flag the earlier it will go "1".

1744744284694.jpeg
 
Yeah, those tabs look pretty ratty. The solution depends on many other, currently undefined, requirements
  • required accuracy
  • temperature
  • contamination
  • lighting
  • cost sensitivity
  • engineering, or lack thereof
  • etc.
Until you can come up with essentially a requirements document, you're just shooting in the dark. Also, while there may be issues with the existing "design", it seems very qualitative. Your missing of one of the two targets suggests that the targets themselves are different, either in positioning or magnetization, etc. Additionally, how the data is being used and processed can matter greatly in performance
 
Well I do not fully agree with you here IRstuff.

Unless the the owner have required anything better then it was when it was new, just looking at the construction of it I can say right away that it wont be less accurate if you make two new targets based on the old ones the construction is quit "none" exact as it is constructed to begin with.

The accuracy is in the placement of the targets 180 degrees apart and the possibility to adjust the gap between the sensor and the target and that the targets are identical in area and length.

Temperature will not affect the accuracy more before then after.
Contamination will not do it unless there is a lot metal dust.
And lighting definitely will not.
 
Your targets need to be identical, dimensions, shape, material, and anything else possible.
A bit of slight difference can cause a lot of issues with triggering.
We had a customer that made systems for paper machines and they used similar flags for triggering.
They sent us some steel bar to anneal. (as I recall it was 0.02% C)
And then after they made flags from them they sent them back to reanneal.
When we asked they said that they estimated that it was 60 year supply.
 
You should be able to set the gap to 0.010 inches and never have a collision.

Has anyone put a dial indicator in place to measure the runout of the tabs?

Such sensors are stupid simple and reliable; you should probably just have a new set of targets made.

The detection volume of an inductive sensor is a tulip shape that tapers as the target is farther away

In the specification for that sensor series as given starting on page 42, the distance from the face of the sensor is given as "X" and the distance from the axis of the sensor is given as "Y"

The "X" value is mainly sensitive to the size of the target, but doesn't change much being off-center. This makes is good for the main use - seeing if something is heading directly at the sensor face.

Your case is an approach at constant distance X (for each tab) and varying distance Y from the axis. If you look at the graphs you can see that even small changes in X will make large changes in the Y value needed for detection. For example, a 0.5 mm change in X may make a 1 mm change in the Y detection.

You could do this experiment on your own equipment by manually turning the end ring while making stepped adjustments to the sensor X gap for just one sensor. Turn it and measure the angle to plot vs. radial gap.

Without knowing more about how much variation you have, how much variation you can accept, what the any other variations there are, or what goals you have expressed in a quantitative way ... kind of stuck.
 
Use the encoder determine the angle between points and a single sensor to determine a zero-reference position, if needed.
 

Part and Inventory Search

Sponsor