Fixing loose rivets so they stay fixed

Some rivets attaching the underside horizontal stabilizer skin to the HS rear spar on our RV-6 are loosening and smoking. I understand this is a pretty common problem if the skins were machine countersunk too aggressively.



I want to deal with this before it gets much worse, but it's not totally clear to me how best to attack the problem. Anyone with experience on this I'd like to know what, short of reskinning, fixes this so it stays fixed.

Here's one question. If the rivet head is sunk below the skin, does it make sense to ream out to #30 and replace the AD3 rivet with an AD4? I realize you wouldn't deliberately machine countersink for a 4 rivet in .032 sheet, but if it's already almost there, would it make a better fix at this point? Or, is it better to stay with 3's, and just use a slug when squeezing to make sure they are set tight?

Here's another. Would it make sense at this point to bond the skin to the spar? What I have in mind is drilling out all the rivets, cleaning the faying surfaces, laying on Scotch-Weld 2216, and re-riveting. It seems to me that would keep the joint from working loose again. But maybe there's a good reason not to do this?

Thanks in advance for any help!

--Paul

Paul,

There are things that can cause smoking rivets including poorly prepared or oversize holes. I have difficulty understanding how too deep countersinks have much of anything to do with smoking rivets. The first thing I tend to suspect when "smoking" rivets appear is that the rivets were installed too short in the first place, were poorly bucked or underbucked. There are places in the plans where the rivet length callout tends to be on the short side. For that reason and more often than not I usually install rivets a half length longer than called out in the plans. I do that to insure the rivet shop head I just bucked is long enough and substantial enough to really swell and grab ahold of its dimple. Then too, underbucked rivets (even correct length rivets) will not swell enough to sufficiently to fill in and seal its dimpled hole. Over time, such rivets can work loose.

Can you provide us a picture or two peering inside the horizontal stab in an effort to assess the quality of the shop heads of the smoking rivets?

Rick, Thanks for the help, I appreciate your insights.

One thing, if it makes a difference for what you would advise, the skins here were machine countersunk, not dimpled. The builder followed the RV-6 construction manual which says "The recommended procedure for the horizontal stabilizer is to dimple all the rivets except those along the rear spar. It is difficult to get a squeezer alongside the HS-409s to dimple the rear spar." The skins are .032" here, pushing it for machine countersinking AN426-AD3's. Then consider this is Subassembly 1 in the whole process, so the builder was probably still on the steep part of his learning curve. Over-countersinking would produce an oversized hole in the skin, and would make it hard to set the rivet tight, since the factory head could move below the skin.

It's kind of hard to get a good view of the shop heads without removing the elevators, but I tried taking a couple photos. This is the shop head on the rightmost rivet in the photo in post #1 above:


And this is the middle rivet:


Thanks again for your help...

--Paul

Paul, the following is me thinking out loud, not from direct experience with this particular problem, so take it for what it's worth.

I think that no matter what you do with the existing rivets, or with new up-sized rivets in the existing rivet locations, they will likely never form as strong a joint as it was (or should have been) originally. I'd consider installing new rivets half-way between the existing rivets along that row if there is sufficient spacing to do it without violating the hole-to-hole edge distance requirements.

Hey Roee --

Those rivets on the skin to rear HS spar are on 1.25" pitch. So installing another rivet between each pair would make for 5/8" center to center spacing. That is supposed to be enough for AD3's (even AD4's).

I guess I'm going to have to drill out all the existing rivets and disassemble that joint in any case. I'll need to get access to both sides of the skin and the spar web for deburring if I'm going to be drilling new holes...

--Paul

This should work

Carefuly clean out the countersink by hand, open the hole to #30 and install a 4- rivet. Then shave or file flush. This is a common problem on many production aircraft (and is actually called out for in the Lear 24 SRM). Another solution is to dimple a AN 960PD4 washer and shave or file after instalation. My experiance with any type of structural adhesive is not good for this kind of repair.
Good luck
Andrew
-9 preview
AP IA
23 years structures

Rick6a said:

Paul,

There are things that can cause smoking rivets including poorly prepared or oversize holes. I have difficulty understanding how too deep countersinks have much of anything to do with smoking rivets.

Click to expand...

It's actually a not to uncommon of a problem when an .032 skin has been machine countersunk for 3/32 rivets. If it is done perfectly, .032 is marginal at best. If it is even slightly over done for depth the hole will be knife edged and over sized.

I think the plan of adding additional rivets between is a good fix. A trick you can use to install new rivets in the locations of the smoking ones is to punch some small aluminum spacers (.020 - .025 thick) with a whitney punch. Lay it on the rivet head when squeezing the rivet to make sure the rivet gets set tight to the bottom of the countersink. This will help with the problem of the countersink being slightly too deep and help compensate for the thickness of paint adjacent to the holes.

Garage Guy said:

Hey Roee --

Those rivets on the skin to rear HS spar are on 1.25" pitch. So installing another rivet between each pair would make for 5/8" center to center spacing. That is supposed to be enough for AD3's (even AD4's).

Click to expand...

Yep, you're good on spacing.

Garage Guy said:

I guess I'm going to have to drill out all the existing rivets and disassemble that joint in any case. I'll need to get access to both sides of the skin and the spar web for deburring if I'm going to be drilling new holes...

Click to expand...

No, I'd leave the existing rivets in and not disassemble the joint. Leaving it assembled will ensure that you don't introduce any misalignments, twist, etc. when installing the new rivets. And in this case disassembling the joint shouldn't be necessary anyway. Since you'll drill the new rivet holes through the skin and spar in assembly, i.e. as one tight sandwich held together by the existing rivets, there will be no burr on the mating surfaces between the skin and spar. You should just drill the holes (first drill slightly undersized, then ream to full size), countersink the skin, and deburr the back side of the spar where the shop heads will go. Then go ahead and install the rivets.

And if there are any smoking rivets that you still want to remove and replace, then again, I would do so in assembly, *after* the new reinforcing rivets have been installed.

I agree with Roee but dimple instead of countersink the skins.

roee said:

No, I'd leave the existing rivets in and not disassemble the joint. Leaving it assembled will ensure that you don't introduce any misalignments, twist, etc. when installing the new rivets.

Click to expand...

I disagree

If the rivets are smoking, they already aren't doing much.

Drilling out a few rivets is not likely to cause any misalignment. If you really do have a concern about it, drill and cleco the new holes first before removing the bad rivets.

I would not recommend you attempt to dimple countersink for the new rivets unless you use a pneumatic squeezer. You will be dimpling two layers of .032 simultaneously which doesn't work that great with a pneumatic squeezer, but it will look ok. If you try it with a hand squeezer it will look terrible (unless your name is Hulk Hogan)

awm said:

Carefuly clean out the countersink by hand, open the hole to #30 and install a 4- rivet. Then shave or file flush. This is a common problem on many production aircraft (and is actually called out for in the Lear 24 SRM).

Click to expand...

Just to confirm, this is called out even for .032" skins?

roee said:

Since you'll drill the new rivet holes through the skin and spar in assembly, i.e. as one tight sandwich held together by the existing rivets, there will be no burr on the mating surfaces between the skin and spar.

Click to expand...

But the sandwich isn't all that tight, anymore. You can feel a little relative movement between the skin and the spar. You can't see it very well in the photos, but some of the 'smoke' (aka finely divided aluminum oxide) is present in the seam.

--Paul

Garage Guy said:

But the sandwich isn't all that tight, anymore. You can feel a little relative movement between the skin and the spar. You can't see it very well in the photos, but some of the 'smoke' (aka finely divided aluminum oxide) is present in the seam.

Click to expand...

In that case, I would also clamp it for drilling.

Cusom rivet

What you will end up with is a rivet head that is dimensionally some where between an AN426 and an NAS1097. The smoked rivet has removed some of the skin. .032 is not enough to machine countersink for a 4- rivet, hence clean an reshape the countersink by hand, open up to remove the knife edge, and drive a AN426AD4-rivet. Shave what sits proud and its as good as new (kinda). The determining factor on what thickness is allowed is the avoidance of a knife edge condition with rivet diameter. When a machine countersink is to deep for a rivet, the head can't fill the hole the way it is supposed to and it ends up smoking (Cessna firewall ect...)
Its a shame, I can get oversize shank, undersize head, but no oversize flush head without blowing open the hole.
Andrew

awm said:

What you will end up with is a rivet head that is dimensionally some where between an AN426 and an NAS1097.

Click to expand...

OK, that makes sense.

Now reading MIL-R-47196A, it says


Looks like I have some thinking to do about which rivets to fix, since if I follow that, I won't be able to do them all. Wondering, does the Lear SRM say something different? Because Learjet standards would be good enough for me...!

--Paul

No worries

Hey,
I no longer have access to the Lear Manuals. I am not familiar with the purpose of the document you are quoting (type and speed of aircraft). In addition, most repairs that I have done that involve DER's the engineers are more concerned about material type for the fasteners than diameter. The thinking is how is the joint going to fail, and will one or two steel fasteners not give a little before they go and then start a progressive fail situation. The other document to check would be AC43.13-1B para 4-57 (4,b) and figures 4-9 and 4-10. Granted the tables are for protruding head rivets, the things to look at are the engineering notes about %used and will fail sheet or fastener. From the Convair 580 SRM ( I wish I had a copy) Dimpled are the strongest joints, then protruding, then machine coutersunk. My thinking is as long as you have edge distance, go ahead and open them up. If there are one or two existing fasteners in a line that are O.K. but by themselves, blow them open. What I try to imagine is what does it do, and how does it break. Can you imagine something else letting go before this joint fails? What else would have to happen?

Andrew

A link to the MIL-R-47196A document here: http://www.vansaircraft.com/public/Specs.htm

As far as scope of the specification, it just says: "1.1 Scope. This specification covers the preparation for and installation of buck rivets." Full stop. On the other hand, it also says "approved for use by US Army Missile Command", so I guess it's possible some of its stipulations are overkill for RV's. It's referred to a lot around here, though. I was looking at it because it mentions over-countersunk repair specifically, where AC43.13-1B doesn't, as far as I can see.

--Paul

why smoking rivets?

I've just gone through replacing a "bunch" (not telling #) with an A&P of smoking rivets on the horizontal stabilizer, and even need to replace some of the replacements. Had assumed it was from all the acro (+5.5, -2.7 Gs typical flight), but it may be from moving the airplane around by pushing on the horizontal and vertical stabilizers at the ribs and stiffeners. Since the airplane is basically white, the smokers show up easily. I haven't seen any smokers anywhere else on the airplane. I'm planning on using a tail wheel tow bar to see if this is indeed the problem. For other than light colored aircraft tails, the smokers may not be showing up.

RV-4
Slider
485 hours, at least 350 acro hours.

Just a thought....
I would mix up some ProSeal and spread between the skins if I could get a little in there and also use on the rivets as I set them. Much like a fuel tank except you dont need to cover the buck tail. This will keep the skins from moving and stop the smoking rivets. This is done on pressurized aircraft as a matter of course.
Dave (Swift Driver)

I'll bet the proseal is used in pressurized aircraft to "create a seal"; not to adhere anything.

A common misconception amongst the homebuilder community is that proseal is an adhesive. It's not, it's a sealant. Their purposes are completely different.

You could use something like Metalset A4 instead. It's similar to JB weld, but available in larger quantities and is stronger. It can be sanded, drilled and tapped, etc.

I would never trust any sealant to bond anything...

Phil

Wow

Phil said:

I'll bet the proseal is used in pressurized aircraft to "create a seal"; not to adhere anything.

A common misconception amongst the homebuilder community is that proseal is an adhesive. It's not, it's a sealant. Their purposes are completely different.

You could use something like Metalset A4 instead. It's similar to JB weld, but available in larger quantities and is stronger. It can be sanded, drilled and tapped, etc.

I would never trust any sealant to bond anything...

Phil

Click to expand...

Phil,
So you're saying you would never use ProSeal to "bond" anything.
Gosh.

Ok, I hear you and believe you when you say it's a sealant not an adhesive.
But in my experience ProSeal is a better "adhesive" than most adhesives!!

Mark

Bond with

scrim cloth and Hysol 9309 na. Cessna does it for leading edges and "waffel" cowls. For smoking rivets, it is almost allways hole prep or prop wash in my experiance.
Andrew

Phil said:

I'll bet the proseal is used in pressurized aircraft to "create a seal"; not to adhere anything.

A common misconception amongst the homebuilder community is that proseal is an adhesive. It's not, it's a sealant. Their purposes are completely different.

You could use something like Metalset A4 instead. It's similar to JB weld, but available in larger quantities and is stronger. It can be sanded, drilled and tapped, etc.

I would never trust any sealant to bond anything...

Phil

Click to expand...

I agree with Phil in concept, but not entirely in practice.
I know of RV's that have been flying for more than 10 years with the NACA vent housings attached with only proseal. I have also used it as an adhesive for other things.

With that being said, I think builders need to be very selective in what adhesive job they choose to use it for.

There is also 3M 2216 http://www.aircraftspruce.com/catalog/wppages/scotchweld.php. That is some no-foolin' adhesive. A local aerospace guy and RV-7 builder swears by it and suggested it to me for this repair. But it doesn't seem to be generally accepted best practice for this particular problem...

--Paul

You said that better than I did, Scott.

I agree. Just don't use it any place you're looking for a structural bond.

(Or is it really sealing the NACA vents)

Phil said:

I'll bet the proseal is used in pressurized aircraft to "create a seal"; not to adhere anything.

A common misconception amongst the homebuilder community is that proseal is an adhesive. It's not, it's a sealant. Their purposes are completely different........I would never trust any sealant to bond anything...

Phil

Click to expand...

Sorry Phil. I'm just not buying it. "Not adhere to anything?" Painting with so broad a brush is fraught with hazard. In the world of production aircraft manufacturing, proseal is used in many different applications.

A common misconception that circulates among the RV community and frequently perpetuated by more than a few of its leading voices is that proseal is not an adhesive. If you want to spit semantic hairs in a very narrow sense that is strictly true but also happens to be terribly misleading. Granted, the word "proseal" itself is our shorthand for a whole range of polysulfide sealants and many of the mixes including the CR3202 and CR3204 type sealants we use in fuel tank construction do in fact possess superior adhesive qualities and a major reason why it does its job so superbly. So tenacious is its natural adhesion, I had no problem using proseal to "glue" certain parts directly to the airframe with no backup "chicken rivets" whatsoever and that includes the NACA vents, rudder cable exit fairings and even the aluminum fairing underneath the horizontal stab. Just ask anyone who ever removed a fuel tank access cover how easy a task it was to pry loose from a gasket of long cured proseal!

I have been up to this point futilely searching for the official polysulfide manufacturer's document that outlines a compelling case that proseal can also serve as an adhesive. The information contained within that document is completely at odds with those RV builders who blithely maintain "proseal is not an adhesive." I do submit the following manufacturers document for your consideration and contained in several places within that document, highlight the "excellent adhesive" properties of some of its "proseal" mixes. The document may also give the casual reader a general idea of just how versatile "proseal" really is:

http://www.nslaerospace.com/SA_FlameMaster.asp

rvbuilder2002 said:

With that being said, I think builders need to be very selective in what adhesive job they choose to use it for.

Click to expand...

Well stated.

Sure it adheres to stuff, it's sticks to everything. But that doesn't mean it was meant to glue two items together.

All of the references in the document talk about it's ability to stick to various materials, that's fine. The stuff will adhere to both metal surfaces, but what actually occurs between them? There's several molecules of preseal in between and you're counting on the structural strength of proseal bonding to proseal. This is the exact reason we use cotton flox in the composite world.

How much load will that carry and how do you know?

I've yet to find any document from the manufacturer or the FAA that says they'll support it carrying load. The stuff just wasn't designed to operate in shear, tension, or any other loads. It's designed to plug holes and prevent leaks.

I guess I should explain my reason for cringing every time I hear someone say, "I'll just slap some proseal on it and stick it in there."

In multiple hangar visits with [some] RV builders, this comment has come up. The two comments that really got to me going 1) involved a bracket to hold a wiring conduit deep in the tail (close to the rudder cables and elevator horns) where riveting was nearly impossible. 2) Gluing a stiffiner rib into the inside of a control surface to prevent oil canning. There are others, but these two really that stuck with me.

I just think folks should really think about what they're doing before they do it. And just because something is "really sticky" doesn't mean there is any real strength between the two sticky bonds.

Phil

The FAA approved repair manual...

Phil said:

....

I've yet to find any document from the manufacturer or the FAA that says they'll support it carrying load. The stuff just wasn't designed to operate in shear, tension, or any other loads. It's designed to plug holes and prevent leaks.
.....
I just think folks should really think about what they're doing before they do it. And just because something is "really sticky" doesn't mean there is any real strength between the two sticky bonds.

Phil

Click to expand...

...and Service Kits for the Grummans use "proseal" as an adhesive, sometimes in conjunction with 3/32 NAS1097 rivets, sometimes without, as a repair method for the original bonded structure.

A lot of the previous pro and con comments are based on the proseal being a sole structural component.

A smoking rivet is one that is still intact, and holding the parts together, but also allowing a little play between the parts. Redoing the joint with proseal and new rivets could be looked at as just removing the "little play" portion, and not taking 100% of the structural loads. If the parts can't move, then the rivets won't smoke.

It's a combination approach...

custom rivets

OK, here's what I've ended up doing.

Taking everyone's suggestions to heart (including phone conversation with Van's tech support), I liked awm's idea of 'custom rivets'. The idea is to have the effect of a 1/8" shank rivet with a head size between that of a AN426AD-4 and a NAS1097AD-4. But what should the custom head size be, if you're going to machine countersink in .032" skins?

Consider the basic geometry of a 100 degree flush head rivet:


Now define a quantity "B" as the "factory head bearing area", the difference between the area of the factory head and the cross sectional area of the rivet shank: B = (A*A - D*D) * pi / 4. Larger B means increased resistance to pulling through the skin under tension loads. On the other hand, for a fixed shank diameter D, larger B means larger A which means larger H, and H should be less than the skin thickness, to keep from knife-edging the countersink.

For some standard rivets (dimensions in inches, except B in square inches; head diameter A is measured from examples on hand):

426AD-3's are marginal in .032" skins, 426AD-4's are obviously too big, 1097AD-4 "oops rivets" have a very small factory head bearing area. But suppose you shaved a AN426AD-4 rivet to make a custom rivet with head diameter .200, .195, or .190. You would have:
Without doing any actual strength analysis, the custom-.195 looks like a good compromise, with both better head bearing area and better knife edge margin than the AN426AD-3 rivets that they will be replacing.

One approach to accomplish this repair would be to drill out the existing rivets, ream the holes to #30, countersink the .032" skins to depth 0.029", set AN426AD-4 rivets, and shave them flush, creating "custom-.195" rivets in place. However, none of the local A&P's I asked have a rivet shaver ("I wish I had one!" was a common response), and anyway the skins are painted and I didn't want to strip the paint.

So what I did instead was manufactured custom-.195 rivets by chucking AN426AD-4-4 rivets in my mini-lathe, turning the heads to .195" diameter, and facing ("pre-shaving") the head 8 thousandths. The result looks like this (left to right: AN426AD-3-3.5, NAS1097AD-4-3.5, custom-0.195, AN426AD-4-4):


I also made a 0.195" diameter steel slug to use on the head when squeezing these rivets, to account for the fact that they will be recessed below the paint on the skins.

Yesterday I finished replacing the 38 AD-3-3.5's on the right horizontal stabilizer rear spar lower skin with these custom 0.195" head diameter rivets. The result looks quite good, in my opinion. I'll keep an eye on it to see how it holds up, but I'm optimistic that this will be a satisfactory repair, without using any adhesives or drilling holes for additional rivets. Avoiding the problem in the first place would be preferable, of course... I'd recommend dimpling instead of machine countersinking here when building, even if it means having to grind down your dimple die to avoid interference with the spar web.

--Paul

Phil said:

Sure it adheres to stuff, it's sticks to everything. But that doesn't mean it was meant to glue two items together.
.......

I just think folks should really think about what they're doing before they do it. And just because something is "really sticky" doesn't mean there is any real strength between the two sticky bonds.

Phil

Click to expand...

in composites the resin tends to be brittle, and not transfer loads, so flox or some sort of chopped fiber is used to transfer loads through the resin. pro seal is about the opposite of brittle as we go.

Phil do you know of any examples where proseal was used by itself and failed from holding the parts together? I mean failed as in not allowing a leak, but losing adhesion to where the parts fell apart?