Silvaire wrote:
For a joint using non-flush rivets and plain holes
Unfortunately the majority of the rivets on the aircraft (and almost all of the ones where an edge distance issue arises) are flush rivets in dimples, where the required edge distances are larger.
Yes, the edge distances are larger for flush rivets because the dimpling of the skin disrupts the ability of the skin to resist tear-out: 2-1/2D is typical versus 2D.
2-1/2D for a 3/32 rivet equals 0.234 inches and assuming Vans followed that practice it leaves a distance between the hole and edge of 0.188 inches. When opened up to 1/8, that becomes 0.016 less, or 0.172 inches. This is by coincidence the same as the minimum distance between hole and edge for a 3/32 countersunk rivet, with the 0.219 edge distance that’s called out as the minimum acceptable in the table: 0.219 – (0.094/2) = 0.172 .
As in the non-flush rivet case, I believe it would at least be worth considering whether you need the specified (0.281) edge distance for a 1/8 countersunk rivet when in this particular joint design case the line load on the joint, and on the individual rivets, is limited to that which a 3/32 rivet can carry. The table does make reference to repair practice, i.e. drilled out and opened up rivet holes, but I think in this case specifics of the joint design may matter equally. Also the strength of the alloy used for the skin is not addressed by the table. And don’t forget that with an Oops rivet the dimpling of the skin remains that of the 3/32 rivet. Specifics not considered by general practice sometimes (although not always) allow engineering to save a part.
I don’t know the details of the rivets used by Vans but these are typical ‘Oops’ flush rivets.
This data might be even more interesting because the NAS repair rivets shown are only 1/64 oversize meaning the distance from hole to edge will be reduced by only 0.008 in. and the head is the same size. It appears you’re not the first one to be in a similar situation. If these can be used my thought is that edge distance after the hole surface finish is cleaned up would no longer be the primary issue and that the heat affected zone from laser cut holes may be the more important thing to consider in relation to potential skin cracking as rivets are installed.
Well, sure no molecular expert here, but the intergranular structure of the aluminum must be affected by the heat of the laser cutting process, and lead to embrittlement.
I could accept parts where the edges have been laser cut, but not structural holes, due to reduced integrity of the area around them.
Drilling or reaming to the next size might probably not be enough to eliminate the affected area around each hole…
Oops rivets as such are used to fix isolated mistakes, not entire assemblies. They are more difficult to shop on thin sheetmetal, and are slightly (X1000s of rivets) heavier than their original counterparts. The edge distance problem has already been mentioned above, and I have experienced a lot of very tight edge distances on RV builds.
On top of that, affected pieces or assemblies, such as flight controls, tanks, etc, would first have to be disassembled, i.e. having to drill out the original rivets, then if possible drill or ream to the next size, deburr both sides, before setting the oops rivets.
No cristal ball in front of me, but my guess is that either Vans will try to talk their way out of it by offering replacement of parts on which cracks have occurred during the dimpling process, or then swallow the whole and have all laser cut parts produced and delivered replaced.
Makes me also wonder if any of the delivered quick built kits are affected…
Dan wrote:
I could accept parts where the edges have been laser cut, but not structural holes, due to reduced integrity of the area around them.
Drilling or reaming to the next size might probably not be enough to eliminate the affected area around each hole…
My understanding of the ‘structural integrity’ issue is not enough to make any judgement in that regard. Its not clear whether the issue is contained to a problem with cracking as rivets are installed (which by design involves extreme plastic deformation), or whether in-service fatigue life is also affected. Its also not clear if the cracking phenomenon is caused primarily by rough finish, hardened material or a combination of both.
Dan wrote:
my guess is that either Vans will try to talk their way out of it by offering replacement of parts on which cracks have occurred during the dimpling process, or then swallow the whole and have all laser cut parts produced and delivered replaced
In my experience, that may depend on sample testing to better understand the ‘structural integrity’ issue and how it can or cannot be resolved in conjunction with rework.
@Dan unfortunately just about all production is affected.
There appears to be no real parts traceability, all we’re hearing is “you may have received” -type statements based on our shipping dates. So slow build customers are reliant on memory or having made some notes (fortunately I did for my emp kit, and my wing kit remains unassembled), and quickbuilds are a total unknown.
Any quickbuild kit between the relevant dates is liable to be affected and there is absolutely no way to know short of exhaustively disassembling them. Customers will almost certainly, en masse, reject entire quickbuild assemblies at great cost to Van’s.
They have already agreed to replace any laser cut part on request.
I agree with your statement about rivet work. There are many solutions for upsizing rivets if required for one-off fixes but doing a whole assembly assembly that way is not a solution to a shoddy parts problem. It may not be clear to Peter and Silvaire, but these are not isolated bad holes – generally the entire rib/spar is like that.
Graham wrote:
It may not be clear to Peter and Silvaire, but these are not isolated bad holes – generally the entire rib/spar is like that.
It’s entirely clear to me, and what I wrote above was written with potential rework of every hole in mind. Drilling out holes e.g. 1/64 in. oversize and using the NAS repair rivets in every one would not to my mind be a big issue, if it works to satisfy the design requirements. It is exactly what might be done in a factory environment if Engineering signed off via an approved process.
Once or twice I’ve been personally in the line of fire to come up with rework solutions for manufacturing-induced structural problems, to contain the associated cost and schedule implications on (literally) billion dollar programs. Perhaps as a result my tendency is to think about what can be done within technical limitations, versus automatically scraping away parts. Vans may come up with something, or not, but regardless it has to be a rational process.
Thanks for your clarifications @Graham.
Though not directly involved, but for my duty within our homebuilders association, I’m following both Vans statements on their homepage as well as on VAF.
Graham wrote:
Customers will almost certainly, en masse, reject entire quickbuild assemblies at great cost to Van’s.
The cost will be directly for Vans for now, but rest assured that it will be fully reflected in the next price increase for present and future customers…
Sure, that is naturally an engineer’s approach to it. I’ve been in similar positions trying to find pragmatic solutions to avoid total re-work, although not in an engineering capacity.
Here my perspective is one of a customer. The part is not of appropriate quality, therefore I reject it. It is not my role to find workarounds for their manufacturing mistakes.
It may not be clear to Peter
On the contrary – I think this is crap. But I have not been saying much because I don’t fly a homebuilt (have looked at them many times – in particular Lancairs for performance reasons – but right now Annex 1 would not meet my requirements for flying around Europe) and some think that righteously enables them to attack me, as if building a plane gives you some right of passage into some aviation masonic lodge populated by morally and intellectually superior aviators.
The issue is made worse by the widely known fact that most homebuilders build just the one plane, and a large % of them don’t even finish that one (although few of those go public with that, or are even willing to discuss it privately; I spoke to yet another one recently).
Vans have got a bit lucky here, for a time, because anybody in a manufacturing business will tell you that you can get away with far more “disasters” if you sell to customers who buy just one, versus customers who buy a load. The reasons are obvious; the latter customer has much more visibility of a QA issue, while the former customer is easily bluffed / fobbed-off and is he gets a replacement promptly he will be happy. Well, that’s how it always was before social media 
Some serial RV builder, or really any mechanical engineer, would have spotted holes like that right away. You can’t rivet parts like that (regardless of any undesired “heat treatment”) because a rivet relies partly on being a good fit; it doesn’t just clamp the two parts together along its own axis. If it did just that, it would soon start to loosen.
I am surprised laser cutting is worth doing. CNC punching is a highly developed process and produces perfectly round holes. Especially as, I am told, deburring is left to the builder.
Graham wrote:
I told the laser cutting story to my father who has a long professional background in manufacturing QA. He mused that the person at Van’s feeling most uncomfortable right now would be the person who qualified the vendor.
That’s for sure. For Van’s sake, I hope that they used a huge company with deep pockets. Like many have said, it’s just amazing to me that this happened. Even if the holes were perfectly round, and each cut only took a microsecond, the holes just look ratty.
Having worked with some very small “mom and pop” shops many years ago, I’m guessing that the owner made all the early parts, and they were good, and passed Van’s QA checks. Van’s assumed that they would continue with this same level of quality and didn’t do enough QA on subsequent parts.
