Production line hole detector.

[Carl Kremer]

Hello.

First I am a noob to electronics but can solder and short a circuit.

Outline of problem.

I work for a company that makes plastic drainage membranes. These have to be guaranteed watertight.
Our production machine is 2.20 meters wide. Producing 2 meter wide rolls of membrane.
These membranes are either flat or studded with a thickness of .6 to 1.5 mm depending on product.
Due to impurities in some plastics, contamination etc there can suddenly appear small holes. 2mm to 1cm are common.

Current solution to problem (Which are terrible)

The membrane is extruded and shaped around rolls. It moves across a table which has a light tight box with webcams looking up at
the membrane. These are connected to a motion detection software which sounds an alarm when 1 or more cameras detect light/motion.
There are numerous problems with this setup:

1. Light leaks. (False alarm)
2. Some membranes are thin and light shines through. (False alarm)
3. Low frame rate and quality of image. (No alarm)
4, Dirt can cover lenses
5. Sometimes the computer doesn't communicate with the alarm controller. (Which is an adam thing)
6. Takes a lot of space.
7. The thing cries wolf and is not taken seriously.
8. Sometimes we run clear membrane. (Can't be used at all)

My idea for a solution:

I want to run 2 long strips of conductive brushes across the width of the membrane. (Anti static / conductive filament)
One mounted on top of the membrane and one on the bottom. The membrane keeping them apart.
These brushes would be charged somehow and if there is a hole the bristles touch giving a arch/short and sound an alarm.
Please see my totally awesome schematic (Done completely in MS Paint) at the bottom

Some requirements:

This needs to be somewhat safe. I can enclose somewhat so no hands go between but I don't want to kill anyone. A little shock is ok.
When a hole is detected the alarm will sound and the detector/power will be turned off (So it doesn't burn up I guess)
It will need a button to turn off the alarm and on the detector to detect eventual next hole (Reset button?)
Maybe a variable control to tune the power so that it will arch even if the bristles don't completely touch (So it can be tuned to thicker materials)

Possible problems that I can see.

Sometimes the hole is not complete. (The membrane might just be very thin)
Make it safe from people hurting themselves and fire.


Looking for some creative ideas for a setup like this. Wan't to make a prototype from off the shelf/cheap parts. I got faith this thing could work and be in one word "AWESOME" Please help me realize it. I will keep everyone posted of the eventual build.


Here are some parts I'm looking at (Don't know if they work or how to implement them)
Brush for prototype. (**broken link removed**)
How to power it? ( **broken link removed**) maybe this
(**broken link removed**)

My schematic :

 

[Reloadron]

Just off the top the conductive brushes you are looking at are very high resistance. Brushes like this are generally nylon or other synthetic that is impregnated with carbon particles. Thus a short is not quite a short but a very, very high resistance path to ground. Not saying it is a bad idea as it shows you are thinking. **broken link removed**

Using light isn't a bad way to go. Maybe with improvement it could work. Possibly a color of light with a sensor that only responds to that color? There are likely scanning solutions out there but that can get real expensive.

Ron

 

[BobW]

I think light is still the correct approach, but not as complicated as cameras and motion detection. Just a bar of LEDs on one side with a reasonable light shield around them, and an array of photo cells on the other side. Modulate the light intensity at some frequency such as 45kHz (which I think is close to the standard for IR remote controls) and then pass the output of the photo cells through a 45kHz bandpass filter so that they are not sensitve to ambient light. BTW, I would avoid using infrared, as it tends to reflect off many surfaces (making it good for remote controls but bad for your application). Blue or UV LEDs and the corresponding sensors would probably be the least susceptible to scattering problems.

 

[rumpfy]

The problem sounds to me like it is a problem related to an extruded material.
You say the holes can be up to 10mm (diameter?) and that to me sounds like the extrusion process is not well controlled and maybe the material you are using is of poor quality or is reclaimed/reprocessed material.
You dont say where you are. This might affect the approach you take.
You say you want a hole detector, But I think you need a process which doesnt put holes in the product in the first place. So if you were able to improve the process, what would be the size of the smallest acceptable hole. If you produced product with a small hole, could you detect it with the methods proposed here?
Good manufacturers employ professional people to solve these problems, and when you say the machine is 2.2 metre wide, then you are talking about a seriously big investment.
Assuming the product is extruded, then I would be looking at some kind of voltage withstand capability test. Such in process testing is done with the manufacture of insulated wire for electric cable, and the equipment is generally readily available. You would need to know the dielectric characteristics of the material before going further.
Hope this helps.

 

[dr pepper]

We have polymer extrusion machines here, fortunately holes are not an issue for us.

What you suggest sounds like a good idea, you just need a simple circuit to control the sounder/relay, as carbon brushes have a resistance in the order of k ohms.

Morgan carbon make a range of antistatic brush bars that might do the job.

There are industrial versions that use infra red, but your talking lots of cash.

 

[Carl Kremer]

Thank you so much all your insight.

First I might have painted a bad picture of our hole detector. It detects holes and serves it purpose. But I want to improve the system to be close to fool proof and without false alarms.

The ideas about using light / improving on the system we have are good but are alittle to much of an investment for me at this time.

Rumpfy: You are completely right. We extrude the material and in some of our products we use recycled material. These can contain particles that will produce these types of holes. They are easily detected.

Reloadron : I have had a look at these **broken link removed** They state 100% conductive (Guess low resistanse)

The brushes I were looking at are just for proof of concept. (Will have to make with my own cash) Will they work for this? With the 30000 vdc (See ebay link above) generator connected to a 12 volt power supply?

I have a buzzer, I have a 12vdc power supply.

$20 2x bruses
$20 HV generator
$20 "a simple circuit to control the sounder/relay"

All i need is some info on setting up a circuit like this. (If this would work of course)

And yes as I can see the brushes have a resistance of 1010 Ohm. How does this affect the arch I want to create between the brushes? I mean a spark plug wire has alot more?

And the more I am thinking about this. The spark may have to jump almost 1 cm in some cases. Would it jump though a hole? I know air insulated pretty well but I bet plastic insulates a little more.

 

[rumpfy]

I used a value of 100 volt per 0.001 inch as a value for the dielectric strength of air, For most of the plastic materials, I used a value of up to 10 times greater for virgin grade polyethlene.
As your process improves, you will be required to improve the hole detection capability.
There are so many unknowns/variables in your question that it is hard to know where to begin.
Materials;
what are they? Polyethylene/propylene, EPR/ CSP/natural rubbers/polyurethanes?????
reclaimed material; what is it?/do you reprocess it yourself/what controls do you have for controlling impurities/is the granule size consistent/ do you treat the recycled material as a masterbatch or do you use it 100% concentration/is there a purchase specification for this stuff???

With your existing hole detector;
Is it your own design or part of the original machine supply?
You say 'fool proof'. To me this says your hole detector is unreliable/ Is it responding to electrical interference/is it triggering on holes that arent there? is it triggering on thin sections of material that arent holes but where the material thickness is below spec?

At this stage to be talking about brushes etc is interesting but ignores the system problem/solution. If you talk about these things I guess you are committing to a dielectric test; and if so, the method of establishing the electric test field is relatively unimportant. A chain mail electrode is good enough if your production arrangement can accomodate this setup. As far as test voltages are concerned, 10 to 15 KV DC would sound OK to me and for this voltage level you can buy off the shelf gear and it is commonly available.
Hope this helps.

 

[crutschow]

A high voltage brush touching the moving plastic could leave a static charge on the plastic surface. Could that be a problem with the subsequent processing of the plastic?