Does anybody know how a bolt-hole eddy current inspection probe works?
If you have a varying magnetic field next to a loop of metal, it induces a current through the metal.
But creating this current requires energy. So the system generating the magnetic field can detect if it in the air or next to the loop of metal.
In practice, you can detect change much more finer than metal/no-metal and you can detect cracks as they basically break the loop or reduce the cross-section of the loop.
Depending the frequency of the varying magnetic field, you can go more in the depth of the material at the cost of resolution. Low frequency= more depth, less resolution. High frequency= surface detection, more resolution.
The active element of a probe can be fairly small around 1-2 mm in diameter.
2 main types of probes:
Surface and rotoprobe
With a surface probe, you would go along the edge of a bolt head or nut/washer or along the edge of a top hat bush installed in a lug.
With a rotoprobe, you remove the bolt, choose the probe of the same diameter and rotate inside the bolt hole. So you’re able to say if the crack is going up/down/fwd/outboard.
Then you have to re-install the bolt and that might mean reaming the bolt hole oversize and using slightly larger bolts.
You need some calibration block representative to the material and the type of cracks you’re trying to (not) find to setup your equipment.
As mention in the NTSB page you can find cracks down to .8-1mm. But a more realistic size if you have a fair number of inspection to do is 2mm.
the crack of 0.04 inch is just over 1mm. My inspection above wouldt not have spotted the issue if under the bolt head.
There is more to the equation:
The structure when subjected to the maximum limit load (max G at max take-off weight at Va) can still be ok despite a crack longer than 1mm. Possibly the structure is OK with a crack grown to the edge on one side (1/2"). That would be our critical crack size.
The crack will take some time to grow between when detected and that critical crack size. That will depends on what kind of flying but in 1st order by the number of take-off/landing or aerobatic use.
Then you take the propagation life between a practical inspection detection length and critical crack, divide by a safety factor and you get your inspection interval.
So the better your inspection, the larger the inspection interval.
Super informative post, Xtophe.
If the FAA try to make this an AD, it will be resisted by US AOPA, due to the huge fleet size and the rarity of it. But who can tell what will happen? Is there a precedent for this sort of statistic?
That is really useful Xtophe
It seems that the critical part one needs to inspect is the bottom outboard bolt holes. Here is a photo of the area I borescoped last week
The closest bolt in the photo is the outboard bolt. I am very interested in the Metallurgy of this. If I did have a crack under the bolt head, are you saying that an inspection every 25 hours would be highly likely to pick up the issue before it became critical? I change the oil every 25 hours and to remove the wing root inspection panel and put in the borescope during the oil change might take 20 minutes per side.
I totally know I am being over the top about this. There are a huge number of Pipers out there and there are far bigger risks but for some reason this issue seems more worrying as it is totally out of one’s control.
This is a very interesting video on Piper wings. I wish I’d seen it a few weeks ago as it would have saved lots of time trying to view/expand some poor quality Piper drawings in the Maintenance Manual. All much more obvious when one sees photos/ videos
Very interesting, and also explains what the likely cause for the somewhat loose wing we saw in the other video was.
He is not going to run out of PA28 wings anytime soon, there must be well over 50 stacked up behind him…
Just had an interesting conversation about this with the repair and service firm where we have our Beech, who are also servicing Pipers.
There are key differences in design and maintenance intervals regarding the wing spars and bolts. Key differences are
a) the wing bolts of the Beech are more than 16mm in diameter instead of roughly 10mm in the Piper;
b) the wings of the Beech are being attached at two main wing spars front and aft, instead of the one main spar and the two very small ones fore and aft in the Piper; and
c) the maintenance manuals of the Beech requiring to check (and if necessary, replace) the wing spars and bolts every 5 years and replace them every 10ys.
This is a very interesting video on Piper wings
Interesting indeed. As pointed out earlier, the 4 inner bolts do nothing here. A very odd design. The normal way to do this is one bolt on the upper flange, and one bolt on the lower flange. Also, the attachment points gradually taper on/off from the main spar and webs. The attachment itself is double, essentially doubling the strength of the bolt.
I am not sure you can say the 4 other inner bolts do nothing when all operating as designed. For example, when connecting steel I beams on building sites they are always connected with plates and lots of bolts each side so I would say the design is not that unusual. A structural building engineer told me that the joins were as strong as the I beam itself.
Having said that, if it fails on the outer hole I agree they do nothing! Perhaps Xtophe could comment on how quickly a crack could propagate.
It strikes me that whilst the crack is under the bolt it might creep slowly as it still has some support. I would hope that it would be progressive until it suddenly reaches the web of the I-beam. At that point I could see how things could go downhill very quickly ………… but I have no idea! It is a shame the NTSB didn’t publish the photos of the other wing root which also had cracks (but was obviously still holding on)
I am not sure you can say the 4 other inner bolts do nothing when all operating as designed
A beam takes bending moment and shear. In an I beam this translates into tension and compression in the flanges, and shear in the web (often called shear web for this very reason). In reality it is a gradual thing, both the web and the flanges takes a bit of both, but the principle is correct. That is why it is shaped the way it is.
In normal construction steel using construction bolts, the bolts themselves works only in tension. What holds the parts together is friction. Therefore you need a whole bunch of bolts and equally large splice plates to get enough bolts and area so the parts won’t slide relative to each other. This must also be done on the web, not just the flanges. A much better method would be to weld them together; no need for splice plates or bolts, and full strength is achieved.
For the bending moment, the Piper solution probably does what it’s supposed to do I guess. But the same bolts must also take all the shear forces. At positive G, the frictional connection on the lower bolts is unloaded, at least the outer ones, and once this is done, the outer-most bolts will take all the tension in the flanges, through the bolts themselves. The flange, the metal between the two outer most bolts will take a much larger part of the load, both shear and bending. I would believe it is designed for it with enough metal and so on. But it is an odd way of doing it, and it obviously doesn’t work all that well for every wing. The error could also be wrong torque on the bolts, or the holes are misplaced a tiny amount and the bolts are forced in.
I have seen similar arrangements as the Piper. But on those each wing spar go all the way to the other side of the box. The spars are cone shaped so they overlap. On the other side they are simply hold in place by a pin. All the pin does is just to prevent the wing from sliding out, but could also take some real forces. The bending moment and shear forces are perfectly taken by the box for all loads in such an arrangement.