NDI - Tap Testing 2

[midsidenode]

'currently having non-destructive inspection performed on composite sandwich structures. We are using various ultrasonic techniques. But I was wondering if anyone has a decent tap test procedure that they would be willing to share?

thanks
msn

 

[wes616]

Take a hammer.... wack hard enough to create a delam. Scrap part.

Seriously though, tap testing can only find large voids and delams. When it comes right down to it, these are usually quite obvious. It's essentially like tapping your drywall with a small hammer to find a stud.

Wes C.
------------------------------
No trees were killed in the sending of this message, but a large number of electrons were terribly inconvenienced.

 

[blakmax]

I think that what wes616 was subtly trying to say is that the tap test is extremely limited in value. It can only find the presence of delaminations and disbonds in relatively thin materials and even then it is not sufficiently reliable to determine the extent of the damage. I'd persist with ultrasonics or examine thermography.

 

[blakmax]

I forgot. NDI can not guarantee the integroty of adhesive bonds. It can only tell you if the bond is defective, but can not assure you that the bond is good. I have seen tap tests rated as a pass when they have been stuck together with double sided adhesive tape. Most tests do not interrogate the interface until it has separated, and in many cases that is the critical locus of bond failure.

 

[RPstress]

You need trained personnel with a standard hammer. The training is on panels and parts with known defects in them. If a defect is suspected, then a destructive examination and repair is usually necessary. I'm unaware of any public standards. I believe companies develop their own in-house standards. Delams can often be caught by tap testing. Voidy parts are much harder to catch.

Such 'mandraulic' methods are being overtaken by things such as thermography, as blakmax suggests. Tap testing is usually only used on parts too big or too awkward in shape/surface contour for, e.g., ultrasonic examination. Even so, manual A-scan reflection ultrasonic is more likely to be used than tap testing.

Most sites still use automated through-transmission ultrasonic, albeit with more and more sophisticated software to assess the results.

Kissing bonds referred to by blakmax are a major issue, and usually very thorough process control, particularly of surface prep, is necessary where secondary bonded joints are critical. Periodic tear down of good parts is necessary, initially as much as one in ten or even five, through the rate can drop as confidence in the process is built up.

 

[berkshire]

midsidenode
Tap testing is extremely subjective and depends entirely on the skill of the user, RPstress has pretty well covered it. I know from personal experience on honeycomb panels that a tap test that would indicate a defect of a couple of square inches would often reveal damage of a square foot or more once the skins were stripped back.
Here is an excerpt from one aircraft companies training manual: " TAP TESTING TECHNIQUES
Tap testing is widely used to evaluate the condition of laminated and bonded structures. The method consists of lightly tapping the surface of the part with a coin,light hammer, or other suitable object. The acoustic response is compared with a known good area. A "flat" or "dead" response is considered unacceptable. The acoustic response of a good part can vary dramatically with changes in geometry, in which case a standard of some sort is required. The entire area of interest must be tapped. The method is limited to finding relatively shallow defects.
Where multiple bondlines exist over core, the core bond cannot be evaluated. In a honeycomb structure, the far-side bondline cannot be evaluated. Thus two-sided access is required for a complete inspection of honeycomb structures."
I personally feel you are better off with thermographic and ultrasound methods.
B.E.

 

[PanelGuy]

There is a specification on how to build the hammer for tap testing. I can't find my copy of Bitzer's book right now or I would let you know a spec.

Talk to the folks at Wichitech. They have several devices that are automatic tappers with acoustic sensors. As well, the only place that you can find a "Boeing approved" tap test medallion.

Is it cheesy? yes Is it effective? it can be. I have used coins, hammers, screw drivers and golf balls. Golf balls are the best for a visual. I had one bounce back perfect until it hit a delam on a four layer panel. The ball hit the skin and actually stopped like it was stuck with tape. That one was hard to see as it was beteen the 2nd and 3rd layers.

Depends on your budget and how much time you get to do your testing, but a little tapping never hurt.

http://www.cnde.iastate.edu/cnde_news/e-newsletter0201/story_of_the_tapper.htm

 

[blakmax]

Geez PanelGuy, you are opeing a can of worms. I have seen some stupid specifications for the tap test that insisted on using ONLY a US quarter, which in other countries may be difficult to find. So, with that in mind, what brand/number etc. of golf ball were you using? (Only kidding!)

On a more serious note, I have real concerns about the applications to identify defect sizes on in-service parts, no matter what the test method. In essence, I have reservations about damage tolerance with regard to adhesive bonds per se. Most analyses are conducted either analytically using FEM or by structural tests with artificial defects. Inherent in these methods is the asumption that the surrounding adhesive maintains full strength. In practice, this aprach is valid for production defects, but has significant limitations for service defects.

Adhesive bonding is dependent on chemical bonds formed at the interface at the time of adhesive cure. In some cases where processes are less than optimum (in particular for metals) the bond formed to oxides on the metal can degrade in service due primarily to hydration of the oxides (e.g. Al2O3 forms Al2O3.2H2O) and in the process the adhesive disbonds. The consequence of this is that the interface degrades over time and the bond strength also degrades.

Now if the damage tolerance analysis assumes (or validates by testing) a certain defect size in the bondline and it assumes that the surrounding adhesive maintains strength, then these assumptions are not valid for service disbonds where the interface of the adhesive bond surrounding the identified defect may also be degraded.

So, the question is "What is the value of damage tolerance for adhesive bonds?"

Can of worms numer two is opened.

Regards

blakmax

PS there is a solution.

 

[PanelGuy]

Golly blakmax...I do agree and that is where our product differ to the world...we have a bit more tolerance of damaged bondline in most of our application. I gotta tell ya, if I saw the technician walking down the wing bouncing a golfball, I would probably get off the plane.

The flip side is, my internal view anyway, is that ultrasonic is nothing more thn a more accurate better monitored tap test...using computers to read high energy "tapping". This is ultimately the third can...Even the best we have, though good, cannot detect everything AND certainly cannot tell a tech on the ground that "well, yes it has a spot here see, but it is not that bad and the structure is fine"...Even Rutan uses a safety factor...albeit small.

It ultimately is that hard part of anisotropic materials...I have had the situation where a detected minor defect was OK as long as it was in the right orientation and location. Don't even start with process induced microcracking at the interphase. I had a lay up once where we had the wrong fiber sizing, everything looked right, just didn't work right when loaded.

I guess the end is that each test is good for certain things, but none will cover all and not everything can be detected. The CCM at University of Delaware has done a lot with embedding strain grids (Smart Composites)in civil structures to detect when things are moving and failing. In the long run I think that self monitoring systems will be key. Consider that a series of embedded strain gauges in a wing could be monitored in the beginning of service to create a "normal" and over time a sudden change would indicate pending failure.

Don't know, but I agree that there is much to learn yet.

Take Care

PanelGuy

 

[blakmax]

Hi Panelguy. I think I hear a can of worms opening. Self monitoring structures are fine in theory, but when you find damage and you repair it, there is an extreme probability that you will cut the sensor wires or optic fibres in the process of repairing the damage. You then have a zone outboard where your sensor system is useless. Now you have to revert to the old tap hammer or ultrasonics as well as the self monitoring system. Eventually you end up with the same old system, and further your structure is carrying heaps of useless embedded instrumentation that just adds weight.

Stick to the tap hammer.

Regards

blakmax

 

[rerig]

Tap Hammers can be fabricated per MIL-HDBK-337.