Tip: Riveting technique

[Rick6a]

It is a given that .032 is the minimum material thickness acceptable for machine countersinking aluminum sheet to accept an AN426AD3 rivet which is set in a standard 100 degree countersink. Still, with .032 it is all too easy to generate the dreaded knife edge...that is...no bearing surface if the countersink is made the slightest bit too deep. A knife edge is a bad thing and does nothing for structural integrity. That is why we routinely dimple thin material. That said, some of you may remember the late Tyler Feldman's Grand Champion RV-6 of a few years ago. He elected to machine countersink the .032 wing surfaces to provide the smoothest distortion free surface possible. He knew a dimpled hole simply cannot match the smooth look of a machine countersink. It is that kind of attention to subtle detail that raises the bar and takes home the top trophy. I did something similiar on the upper tank skin surfaces on my RV-6A. For the most part, I dimpled most of the wing skins per plans but did machine countersink the upper surfaces of the .032 tank skins....with a small difference. To insure that I produced adequate bearing surface for the rivets set into the skins, I intentionally made slightly shallow the countersinks for the AN426AD3 rivets. After setting the rivets, I used a high speed rivet shaver to flush the rivets to the skin. The picture attempts to emphasize the point. The rivets to the right of the tank have been dimpled normally, all the holes on the tank itself have been machine countersunk and most of the rivet heads slightly shaved. This is a common sheet metal technique routinely used in aerospace production. A 2D picture simply cannot convey the obvious differences between the two styles of countersinks which is especially noticeable with changing light and viewing angles, but I hope you get the idea.

Rick Galati RV-6A "Darla"

[mlw450802]

Originally Posted by Rick6a

... For the most part, I dimpled most of the wing skins per plans but did machine countersink the upper surfaces of the .032 tank skins....with a small difference. To insure that I produced adequate bearing surface for the rivets set into the skins, I produced countersinks for the NAS470AD3 rivets made intentionally shallow. After setting the rivets, I borrowed a pricey high quality, high speed rivet shaver to flush the rivets to the skin. The picture attempts to emphasize the point. The rivets to the right of the tank have been dimpled normally, the holes on the tank itself have been machine countersunk and the rivet heads slightly shaved. "

It does indeed make a very nice looking, smooth surface. Nice job!

Although, as we should all be aware, the structural strength of the countersunk joint is significantly less than that in skins where we have nesting dimples.

-Mike

[uk_figs]

Does anyone know what tecnique Van's uses for the QB kits, I have QB tanks and the rivets are perfect without any of the slight skin indentation that I have on the wing skins using the C-frame hammer method (I am using the spring back dimple die set). I also noticed that the rivet line on the main spar looks much smoother than the wing skin to ribs and assume this is due to the increased stiffness of the spar flange when you rivet.

[dav1111]

If you go to the Philippines web site and look at picture #3 and #23 you will see what they are using for dimpling:

http://www.rv8.ch/gallery/view_album...bonanza&page=1

Lots of good looking RV-10 QB Fuselage and QB Wing pictures also.

Russ Daves
N710RV - Fuselage Stage

[az_gila]

If you shave the rivets, be carefull how much you take off. A 3/32 rivet can only protrude 0.006 and still meet specifications after shaving.

This specification is an important one, since it is directly related to the final strength of the rivet, and as previously posted, a countersunk rivet is a weaker joint than a dimpled one.

If you read the MIL Spec. (which all of the strength calculations are based on for structural analysis), you will find that they don't recommend a countersink in 0.032 for 3/32 rivets due to a knive edge condition being created - Interestingly enough, the FAA used to allow it in the old version of AC43.13, but now they just refer to the MIL HDBK 5 that is based on this rivet specification document.

I have a copy of the MIL-Spec on my web site...

http://www.home.earthlink.net/~gilal...ec/rivet_a.htm

The metal repair section of the AC43.13 document can be obtained here..

http://www.faa.gov/certification/air...3/Ch_04-04.doc

For those of you who want actual strength numbers... this was an old Matronics posting of mine from 1995... This came out of the very thick MIL HDBK 5 document...

----------------------------------------------------------

Following is the equivalent data for YIELD strength (in this case, defined
as a permanent set of 0.005 inch)

YIELD strength of 3/32 MS20426AD flush rivets in 2024-T3 (values in lbs.)

Sheet thickness ---- Dimpled ---- Countersunk

0.032 -------------- 209 -------- 132

0.040 -------------- 209 -------- 153

0.063 -------------- 209 -------- 213


This is actually a worse % loss than the ultimate loads, giving a 33% loss
in 0.032, and a 27% loss in 0.040.

--------------------------------------------------------

Keep riveting, but go for dimples in those thin sheets. Is it really worth a 33% loss of strength just to look fractionally smoother? - It's your choice...

gil in Tucson EAA Technical Counselor (and a spec. reader...:^)