First screw up - lots of oversize holes

[JonJay]

Be careful. AD4's are much tougher to set than the 3's. It will be a challenge to get them set without damaging the .020 skins. You will be backriveting these, so get a helper to hold you skins tight down to a very solid backing plate and make sure you get the back rivet set flush to the stiffner and skin and the skin is flat against the backing plate. This will be tough to do on the trailing edge rivets but probably doable. It is worth a try, and if it does not work, you are back into new parts anyway, so why not.
Good luck and let us know how it turns out.

[DaX]

Originally Posted by JonJay

Be careful. AD4's are much tougher to set than the 3's. It will be a challenge to get them set without damaging the .020 skins. You will be backriveting these, so get a helper to hold you skins tight down to a very solid backing plate and make sure you get the back rivet set flush to the stiffner and skin and the skin is flat against the backing plate. This will be tough to do on the trailing edge rivets but probably doable. It is worth a try, and if it does not work, you are back into new parts anyway, so why not.
Good luck and let us know how it turns out.

Thanks and I will let you know how it turns out. Just to clarify though, I did not screw up ALL the holes in the rudder, just the stiffener-to-skin holes - the trailing edge and the rest of the rudder will still be AD3's (I hope).

[L.Adamson]

Rick... I think your comment on the rivet strength in tension may be incorrect...

Yes the 1097-4 rivet head is slightly larger than a 426-3 rivet head - by about 0.012 - but the amount of the rivet head that is resisting a tensile pull is way less for the 1097 than the 426, since we are dealing with the amount of material left in the head after the much larger body of a 1/8 1097 rivet vs. a 3/32 426 rivet.

Do we know if the failure mode in tension is in the body of the rivet, or in the head pulling through the material? - since you use head diameter in your comment, I assume you think it is in the head pulling trough failure mode...

The 1/8 rivet will have a much larger shop head, but when you pull the joint in tension, it seems reasonable to me that it will be the much smaller amount of material at the manufactured head side that will fail.

Actually, the head area minus the shank of a 1097 versus the same on a 426-3 is very little difference.

.0140 sq. inches for the 1097 against .0144 for the 426-3. The 1097 also has a dome head for additional material. I also gave the 426-3 a slight edge in my random sampling.

1097's averaged a head diameter of .183 to .185

426-3's averaged .163 to .167 which is a difference of .018 to .020 in head diameter for the larger 1097 rivets.

A few were a bit larger or smaller, but same with the 1097's.

Using the smaller .183 and a middle ground of .165 for the 426-3's, I came up with the above figures.

L.Adamson

PS-- the F1 aircraft rivet heads didn't look any smaller than the 426-3's as I checked one today.

[Kevin Horton]

We know that the joint is strong enough if AN426 rivets are used. If NAS 1097 rivets are used we don't know. We can guess, we can surmise, we can whistle in the wind. But we don't know. I'll take the known strong enough joint over the unknown one any day.

It would be interesting to conduct some pull tests to compare AN426 riveted joints against joints with NAS 1097. Drill, deburr and dimple 0.020" 2024-T3 for 3/32. Rivet half the specimens with 3/32 AN426 rivets. Drill the other half out to #30, and rivet with NAS 1097 rivets. Compare shear and tension strength of the two joints.

Once someone does the above test on a statistically significant number of samples we will know which joint is stronger. Until then we are just guessing.

[L.Adamson]

Originally Posted by Kevin Horton

It would be interesting to conduct some pull tests to compare AN426 riveted joints against joints with NAS 1097. Drill, deburr and dimple 0.020" 2024-T3 for 3/32. Rivet half the specimens with 3/32 AN426 rivets. Drill the other half out to #30, and rivet with NAS 1097 rivets. Compare shear and tension strength of the two joints.

Once someone does the above test on a statistically significant number of samples we will know which joint is stronger. Until then we are just guessing.

I thought about riviting two sheets an inch wide by approx twelve inches, with one rivet each, and then bending sheet (strap) into a loop. Secure the bottom loops to the floor, and then use a cable attached to the 426-3 & 1097 loops over a pully.

Each rivet test would look like a figure "8" with a rivet attaching the two circles. Slowly lift the cable with an engine hoist, and see which gives first.

Using the pully, the tension would be equal. And of course, this would test the rivet heads in tension; as I'm sure that the 1097 does better in shear.

Do the test several times, including additional rivets. Then pull the attached strips straight for shear.

I've also heard that the 1097 does better with dimpled joints, than machined..........but can't be sure on that.

L.Adamson

[szicree]

Originally Posted by DaX

I've already enlarged to #30.

I have emailed Van's about it, but will probably move forward with NAS1097 rivets unless they scream NO! to me.

Setting those rivets takes significantly more pressure and will probably result in some pretty lumpy skins. I'd just bite the bullet and get a new skin. Trust me, this won't be the last thing you have to make twice

[szicree]

Originally Posted by L.Adamson

Actually, the head area minus the shank of a 1097 versus the same on a 426-3 is very little difference.

.0140 sq. inches for the 1097 against .0144 for the 426-3. The 1097 also has a dome head for additional material. I also gave the 426-3 a slight edge in my random sampling.

1097's averaged a head diameter of .183 to .185

426-3's averaged .163 to .167 which is a difference of .018 to .020 in head diameter for the larger 1097 rivets.

A few were a bit larger or smaller, but same with the 1097's.

Using the smaller .183 and a middle ground of .165 for the 426-3's, I came up with the above figures.

Actually, I'm not sure the above calcs tell the whole story. The area of contact between rivet head and dimple is not annular, but rather a frustum of a cone. The surface areas of these, using the same numbers as above comes out to .019170 sq. inches vs .20585 sq. inches, a difference of over 7.4%. If we use the worst case numbers given above we get a 9.7% difference. I sure don't know if this is significant or not, but it is what it is.

[L.Adamson]

Originally Posted by szicree

Actually, I'm not sure the above calcs tell the whole story. The area of contact between rivet head and dimple is not annular, but rather a frustum of a cone. The surface areas of these, using the same numbers as above comes out to .019170 sq. inches vs .20585 sq. inches, a difference of over 7.4%. If we use the worst case numbers given above we get a 9.7% difference. I sure don't know if this is significant or not, but it is what it is.

Yes, I remember looking at the cone surface area under a magnifying glass, and wondering the difference; because the 1097 has that extra "dome" of material on top to smash down, to farther confuse the matter. Does that add back that 7-9% or not....?

L.Adamson

[az_gila]

Originally Posted by L.Adamson

Yes, I remember looking at the cone surface area under a magnifying glass, and wondering the difference; because the 1097 has that extra "dome" of material on top to smash down, to farther confuse the matter. Does that add back that 7-9% or not....?

L.Adamson

Actually, it's a real interesting problem in tolerances... taking the spec. numbers and doing a worse case analysis, it can be +20% or -40% for the NAS1097 head area.

If you assume that the 1097 rivet head squishes out to the entire diameter of the dimple, it becomes +34% and -24% for the NAS1097 head area.

There are quite a few tolerances in place, and all of the numbers came from the MIL Specs here...

http://www.av8design.com/Hardware/MS20426.pdf
http://www.av8design.com/Hardware/NAS1097.pdf

It's not as obvious as I thought, and a nice exercise in tolerancing...

According to the drawings, your 0.163 head AN426 rivets are at the absolute minimum dimension....

thanks for putting my assumption on areas into actual numbers, and for detailing the head sizes.. they sort of look identical to the naked eye, but the NAS 1097 heads are larger than the heads of the next size down AN426 rivets...

gil in Tucson ... enough mathematics for to-nite...

[Rick6a]

Originally Posted by az_gila.

Yes the 1097-4 rivet head is slightly larger than a 426-3 rivet head - by about 0.012 - but the amount of the rivet head that is resisting a tensile pull is way less for the 1097 than the 426, since we are dealing with the amount of material left in the head after the much larger body of a 1/8 1097 rivet vs. a 3/32 426 rivet.......

Gil,

I totally agree with your statement...........if we were comparing apples to apples. In this situation, we are in a sense comparing apples to oranges.... AD3's to AD4's. To wit: The structure is designed to normally accept 426AD3 rivets. By replacing those fasteners with 1097AD4 rivets, we are exchanging a smaller diameter rivet for a larger diameter rivet yet the manufactured head diameter remains the same (though in reality the 1097 is slightly larger). The calculated design forces anticipated on the (rudder) structure remain unchanged either way. All we are really asking those rivets to do is keep a stiffener firmly attached to the skin. Therefore, I expect that in this application, an 1097AD4 rivet is stronger than a 426AD3 rivet....an apples to oranges comparison.