Countersinking versus Dimpling

June 23, 1995

Question
Machine countersinking appears to be another one of those things to be avoided. Reports of cracked skins alone has caused me to change my plans of machine countersinking my inboard wing skins. Yet again, I think about the material and technique here, and this stuff has been done for a long time.

What's going on here?

Answer
RV-listers,

I knew I could get some real facts and data to go along with this discussion.

MIL-HDBK-5F - Metallic Materials and Elements for Aerospace Vehicle Structures

This is a design guidebook of material strengths and fastener strengths that is approved by the Army, Navy, Air Force and the FAA.

Shear strength of 3/32 MS20426AD flush rivets (values are ultimate strength in lbs.)

Sheet thickness	Dimpled	Countersunk
0.032	217	178
0.040	217	193
0.063	217	216

This shows a 18% loss of strength for every countersunk rivet in 0.032, and an 11% loss in 0.040 material. The Handbook also had a warning that in 0.032 material, this was a "knife edge condition" and was undesirable, and _not_ approved.

To answer the question above, basically, we homebuilders are not following approved practices in this area, and we are seeing the effects.

For those of us substituting 1/8 rivets at the floorpan/firewall joint, the equivalent shear strength of a 1/8 dimpled rivet in 0.040 material is 388 lbs., a 79% increase over 3/32 dimpled rivets.

These figures are valid for 2024-T3 material, and probably would not be valid for other grades of Aluminium. Here is some more data of interest.

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. The previous post's figures were for ultimate strength (failure), while the above are those loads that would cause a small permanent set.

YIELD strength of 1/8 MS20426AD flush rivets in 2024-T3 (values in lbs.)

Sheet thickness	Dimpled	Countersunk
0.040	367	231 (*)
0.063	506	321

... Just some data to show that those that say always dimple are right :^)

.. Gil Alexander, RV6A, #20701 ... fitting cockpit stuff

PS The gentleman who requested material, rivet and bolt useage and strength information should get a copy of this document .... it is about 5 inches (double-sided) thick though! It is the first time I have seen it, and it is really a wealth of knowledge on materials. It is copyright free, and I presume can be obtained through the Government Printing Office.

Gil Alexander gil@rassp.hac.com

Go back to Hovan's Home Page