non contact wall thickness of fibreglass pultrusion

Pultrusion?

yes...like extrusion only pulled rather than pushed

makes (in our case) structural elements of FRP

I need to measure the wall thickness of a RHS without contact...on all sides

Or at least without gel between the transducer and material

And while I am here;

Pull speed is measured by 2 encoders, 1 on each puller (they look like Dynapar cubes of some sort), through a Red Lion ADC to a Koyo PLC.

Its all over the place and longer averaging time does not help...might it be affected by noise?

As far as I can see the signal lines run with 110V control lines

DAC rather

2 encoders read into PLC using 1 analogue channel

conversion math performed, output to HMI

So you have an analog signal (the wheel position) which you convert to digital (via the encoder) then you convert it to analog (via the Red Lion DAC) then it goes into a PLC where it is converted to digital so the PLC can use it?

God I hate excess conversions.

Are you sure the reason that the reading is all over the place isn't that the actual speed is all over the place? The fact that longer averaging doesn't help indicates that the noise (or actual speed variation) is of lower frequency than the cutoff of the filter -- I can't say what that means in your context but perhaps it suggests something.

Ultrasonics is a good way to go, but expensive to set up the appropriate hardware. For truly non-contact measurements, I'm thinking Laser Generated Ultrasound would be your best bet (I should admit that this is my field of study so i might be a little biased). You could use a Nd:YAG or similar laser to generate the ultrasound and an interferometer to pick it up. Unfortunately, if the thickness is small (roughly less than a quarter of an inch) the types of waves generated will be very complex (they are called Lamb waves). This makes the signal processing very difficult.

Is this meant to be an on-line application or post manufacturing measurement? Doing the meaurement in a manufacturing environment leads to noise issues.

I suggest you do some research on Google or similar to find past methods for measuring wall thickness in fiberglass laminates.

Also, check out

for background info on nondestructive measurement techniques.

Hope this helps,

Matt

The measurement system seems to be ad-hoc idea piled on ad-hoc idea, turtles all the way down. To answer the question put, if the shape comes through a die with a core, the core can be extended and used as one plate of a capacitor. Provided the material's dielectric constant is really constant, a good measurement can be had that way.

Thickness can vary not only because of changing pull speed, but because of temperature variation in the feed stock. Drawing is much more sensitive to that than extrusion. (Extrusion forces the material against the die, but drawing pull the material away.)

Jerry

Yes, it may have been done to make programming a bit easier

And the PLC is full, the designer left no room for expanding / modifying the machine

Yes...it may vary a small amount but not as much as indicated (+- anywhere up to 100mm/min every update) The speed control valving is of good quality and independant to PLC...originally it was controlled by the PLC and the pull was jerky and unusable so was replaced with an independant unit...probably jerky due to unstable speed information!

Hmmm, goes to investigate noise and filters

Thanks Tim

it does

good idea! goes to investigate

they do like to play with resin formulation and curing temperatures though

would capacitance vary with temperature?

thickness varies if the mandrel moves or wet out fails causing dry fibres inside the section...an indicator would be nice rather than relying on the operator to check

On Sun, 24 Jul 2005 13:41:10 GMT, fulliautomatix proclaimed to the world:

A Beta gauge

Be well,

HoP

The preceding message represents personal opinions and/or advice that may prove incorrect or harmful. But then maybe not. Feel free to disregard.

------- Words have no Warranty ------ ------- No View without Merit ------

yes

this is good

thanks

There are certainly ultrasonic thickness testers around, but for an 'outside-the-box' idea, have a look at Ultrasonic Level Transmitters.

If you pick a high-frequency short-range one and install it above the web looking down, some (eg. IIRC

can output the difference between two boundaries in air and may, with a bit of tweaking, be able to read continous thickness of the sheet.

Make sure you get the rep to demonstrate whether or not he can do it before purchase, tho'...

HTH, Cameron:-)

You might try here:

--Gene

thanks Cameron

they are local even!

Thanks Gene

saw that, looks like it's all surface contact type stuff

HaHa!!!! Now we're onto something

thanks Hop

This can be made with ultrasound, but - as far as I know - nobody offers ready made solution.

Normal ultrasonic measurements (they are standard for example for pipes made with extrusion method) are using water as contact medium. Measurement in the air (contactless) can be made too, but requires a relatively expensive development.

Because we have developed some ideas for this kind of measurement (based on our experience with water based measurements), I have even tried to convince some companies to finance such development, but without succes.

It is also possible to make such measurement using x-rays, based on tomografic principle, but this can be even more complicated.

Wieslaw Bicz

---------------========== OPTEL sp. z o.o. ===========---------------

------===== R&D: Ultrasonic Technology/Fingerprint Recognition ====------ ul. Otwarta 10a PL 50-212 Wroclaw Tel.:+48 71 3296854 Fax.:+48 71 3296852

--------==== mailto: snipped-for-privacy@optel.pl -=-

====-------

Can this be done online in a continuous process? By what method?

They'd probably be willing to pay for non contact void detection in large areas of infusion

I will write you directly about the method, that I would propose.,

Void detection can be principally much easier, than thickness measurement - the information must be not so precise collected.