400 Commonwealth Drive, Warrendale, PA 15096-0001

AEROSPACE INFORMATION REPORT

AIR5065
Issued 1999-08

Influence of Grain Flow on Bolt Integrity

INTRODUCTION Many of the military and aerospace company bolt specifications have a requirement concerning grain flow in the head of forged bolts. The requirements are of two types. One uses photomicrographs to illustrate what is considered "acceptable" and what is "unacceptable". An example of this is AS7471, which covers Waspaloy, but the photomicrograph is of a steel bolt. Waspaloy is a vacuum melted material that is solution treated after it is headed. This recrystallization thermal process wipes out grain flow pattern as the deformed grains are eliminated. In addition the relatively small amount of inclusions, compared to an air melt steel, make it difficult to determine where the flow lines were. For these reasons an air melt steel bolt was used for the photomicrograph examples. This difficulty of determining grain flow in a finished bolt exists for all vacuum melted materials such as IN718 and titanium alloys. Other specifications avoid the difficulty of determining the flow pattern of heat treated parts from a photomicrograph by using sketches. The origins of the sketches go back over forty years when they were drawn for commercial cold headed bolts. Prior to the development of cold headers just before WWII, all bolts under about 1/2 inch were made on screw machines. The steels used were "dirty" which allowed very fast cutting. These steels were unacceptable for cold heading because the stringers caused splitting. The cleaner steels required for cold heading began to be preferred by the users due to the superior characteristics of the clean material. Sketches were added to the specifications showing an upset pattern for a headed bolt. These sketches became part of the military and early aircraft specifications and are unchanged today. A close look reveals that there are no "flow" lines near the outer diameter of the heads. The reason is that they never were made to illustrate exactly how the metals flowed but merely that the part was forged. As new alloys began to be used the sketches did not change to reflect any change in forging characteristics. The actual flow pattern of materials depends on the lattice structure such as bcc, fcc or cph and this is further complicated by texturing developed from coldworking of the thermo-mechanically treated materials such as A286, IN718 and MultiPhase alloys.

SAE Technical Standards Board Rules provide that: This report is published by SAE to advance the state of technical and engineering sciences. The use of this report is entirely voluntary, and its applicability and suitability for any particular use, including any patent infringement arising therefrom, is the sole responsibility of the user. SAE reviews each technical report at least every five years at which time it may be reaffirmed, revised, or cancelled. SAE invites your written comments and suggestions. Copyright 1999 Society of Automotive Engineers, Inc. All rights reserved. QUESTIONS REGARDING THIS DOCUMENT: TO PLACE A DOCUMENT ORDER: SAE WEB ADDRESS: (724) 772-8510 (724) 776-4970 http://www.sae.org FAX: (724) 776-0243 FAX: (724) 776-0790 Printed in U.S.A.

SAE

AIR5065

INTRODUCTION (Continued) A part of the flow line requirements is the prohibition of cutting "excessive" flow lines in the underhead area. Some specifications, such as AS7471, show an example of unacceptable grain flow cutting. This allowance of some cutting is due to the need for underhead removal on bolts in order to have a bearing face close to 90 degrees to the ground shank. To define "excessive" is difficult and possibly needless. It is standard practice to start with a diameter 0.010 inch over the nominal for most "short" bolts. Due to a swelling of 0.005 inch in the shank, this results in grain flow cutting when the shank is ground to size by the removal of 0.016 inch from the diameter. For longer bolts, such as those with a length ten times the diameter, the cutting becomes more significant. Greater stock must be provided to assure clean up during shank grinding. The starting stock can be up to 0.030 inch over nominal and an additional 0.005 inch increase in diameter can be experienced during forging. The need for more stock removal is due to some materials memory as a coil and also due to some warpage during heat treating. So the prohibition of "too many" flow lines cut would have to differentiate between long bolts and short bolts. All long bolts produced by all the fastener companies for the last 30 years have always had more "flow" lines cut than the amount cut on short bolts. This prohibition is a costly item for aerospace companies because the bolt suppliers are required to maintain an excess of raw material sizes in order to make the oversize shank parts needed for repair (oversize holes). It is also costly because much time and effort is made trying to bring out flow lines to determine if too many were cut. This determination in a finished IN718 bolt can be a Sisyphean task. Figure 1 illustrates the type of sketch used to show cut flow lines in bolt heads and show the "flow lines" for bolts in this program. The Aerospace Metals Engineering Committee (AMEC) initiated this program to determine what effect, if any, cutting flow lines would have on aerospace bolts. To assure the maximum cutting of flow lines, this program proposed fabricating the test bolts from bar stock without forging. The proposed program was circulated for comments to individuals and groups including SAE Committee E-25.

-2-

SAE

AIR5065

FIGURE 1 - Examples of Flow Line Cutting Due to Grinding of Shank and Underhead Surfaces of Aircraft Quality Bolts

-3-

SAE

AIR5065

ACKNOWLEDGEMENTS The following companies made contributions to this program which AMEC gratefully acknowledges. Federal Mfg. - Chatsworth Fabrication and testing of 3/8 inch 200 ksi A286 Valley Todeco-Sylmar Fabrication and testing of 7/16 inch 220 ksi IN718 and Beta C and the testing of 10-32 A286 Bristol-Brea Fabrication of 10-32 A286 Deutsch-Gardena Fabrication of 1/2 inch 8740 Automotive Racing Products-Santa Paula Testing of 1/2 inch 8740 Russell G. Sherman Coordinator

-4-

SAE

AIR5065

1. SCOPE: Performed tensile and fatigue tests on bolts to determine if cutting of materials flow lines at the bolt underhead would degrade bolt integrity. Five different alloys used in aerospace industry fatigue critical applications were employed to fabricate 5/16 to 1/2 inch diameter bolts. These bolts were machined from bar stock producing fully cut material flow lines under their heads. Tensile testing with and without underhead washers was performed along with industry standard fatigue testing. 2. APPLICABLE DOCUMENTS: The following publications form a part of this document to the extent specified herein. The latest issue of SAE publications shall apply. The applicable issue of other publications shall be the issue in effect on the date of the purchase order. In the event of conflict between the text of this document and references cited herein, the text of this document takes precedence. Nothing in this document, however, supersedes applicable laws and regulations unless a specific exemption has been obtained. 2.1 SAE Publications: Available from SAE, 400 Commonwealth Drive, Warrendale, PA 15096-0001. AS7471 Bolts and Screws, Nickel Alloy, UNS N07001 Tensile Strength 165 ksi, Corrosion and Heat Resistant Procurement Specification

2.2

NAS Standards: Available from Aerospace Industries Association, 1250 Eye Street NW, Washington, DC 20005. NAS1348

2.3

Other Publications: Petersons Stress Concentration Design Factors, John Wiley & Sons, 1966

-5-

SAE

AIR5065

3. PROCEDURE: All the bolts fabricated were machined completely from bar stock to a 12-point configuration. Such a configuration is used by the aerospace industry in fatigue critical applications. The 12 points, of course, were not machined but this area was machined to the minimum of the 12-point diameter. The bolts were then processed in the normal manner, heat treated and then thread rolled and fillet rolled. The bolts were made, in most cases, with a shank length equal to two times the diameter. This is the shortest part which can properly be fatigue tested. This length also imposes the highest bending stress level to the head and the threads from any misalignment during testing. Test bolts of the following materials and sizes were fabricated. TABLE 1 Size 5/16 inch 7/16 inch 7/16 inch 10-32 1/2 inch Material A286 Beta C IN718 A286 8740 Strength 200 ksi 200 ksi 220 ksi 130 ksi 180 ksi Part No. MS9036 12 pt. MS14181 MS9556 MS21250 Company Federal Mfg. Chatsworth Valley Todeco Sylmar Valley Todeco Sylmar Bristol Mfg. Brea Deutsch Mfg. Gardena

The tests performed, except for those on 10-32 bolts, were fatigue and tensile tests. Five parts were tensile tested normally and then additional tests were run with a 5-degree washer under the head. The fatigue tests were standard tests at the normal specification loads. If no fatigue failures occurred, the loads were increased to force failure. In addition, the engine bolt configuration A286 10-32 bolts were stress rupture tested. These tests were not stopped at 23 hours but were continued to failure. Some of the tests were witnessed by AMEC members.

-6-

SAE

AIR5065

4. RESULTS AND DISCUSSION: The most significant result was that no parts failed in the head. In many cases the parts were subjected to loads 10 to 20% higher than normal, especially in the fatigue tests. The results are tabulated in Tables 2 through 6.
TABLE 2 - Test Results on MS9036/BACB30LE 200 ksi, A286 5/16-24 Bolts Tensile Load Pounds 14,350 14,340 14,125 2 deg. underhead washer 14,040 14,000 5 deg. underhead washer 13,400 13,760 13,740 (1) 10% overload (2) 20% overload Tensile UTS, Failure ksi 224 thread 224 thread 221 thread 2 deg. underhead washer 219 thread 219 thread 5 deg. underhead washer 209 thread 215 thread 215 thread Fatigue (R=.1) Cycles at Load 5760 lb 130K nf 130K nf 130K nf 130K nf Fatigue (R=.1) Cycles at Load 6336 lb(1) 130K nf 130K nf 6K thread 44K thread 100K thread 94K thread Fatigue (R=.1) Cycles at Load 6912 lb(2) 170K thread 79K thread

TABLE 3 - Test Results on MS14181/BACB30US, 220 ksi, IN718 7/16-20 Bolts Tensile Load Pounds 2 deg. underhead washer 32,500 32,200 32,500 32,700 5 deg. underhead washer 33,400 (1) 15% overload (2) 30% overload Tensile UTS, Failure ksi 2 deg. underhead washer 252 thread 249 thread 252 thread 253 thread 5 deg. underhead washer 259 thread Fatigue (R=.1) Cycles at Load 14,700 lb(1) 130K nf 130K nf 130K nf 130K nf 130K nf Fatigue (R=.1) Cycles at Load 16,700 lb(2) 51K thread 40K thread 88K thread 107K thread 52K thread

-7-

SAE

AIR5065

TABLE 4 - Test Results on MS9556, 130 ksi A286#10-32 Engine Bolts(1) TABLE 4A Tensile Load Pounds 3650(3) 3675 3705 3700 3710 Tensile UTS, Failure ksi 162(1) 163 164 164 164 Fatigue (R=.1) (2) Cycles at Load 1630 lb 130K nf 130K nf 9K 35K 45K 44K 34K thread thread thread thread thread Fatigue (R=.1)(2) Cycles at Load 2000 lb Fatigue (R=.1)(2) Cycles at Load 2300 lb

(1) This part has a reduced shank diameter (pitch diameter). The specification for this bolt calls out the H-28 diameter as the stress area (0.0200) which is considerably smaller than the 0.0226 area (pitch diameter) invoked in NAS1348 and used for this report. The reported tensile strengths of 162/164 ksi would be 190/192 ksi using the smaller area. The larger area was used for consistency with the other bolts tested for this program. The use of different stress areas in various specifications in this country is the source of many troublesome misunderstandings and should be addressed. (2) This part has no fatigue requirement. 1630 lb represents load for 160 ksi bolt; 2000 lb for 200 ksi; 2300 lb for 220 ksi. (3) Minimum load per specification 2600 lb.

TABLE 4B Stress Rupture Test 1200F 1600 lb(1) 1. 49 hours broke in shank (pitch dia.) 2. 114 hours broke in thread 3. 35 hours broke in thread 4. 41 hours broke in thread 5. 64 hours broke in thread (1) Required load per specification is 1260 lb (27% overload).

-8-

SAE

AIR5065

TABLE 5 - Test Results on Beta C Titanium 200 ksi 7/16-20 Bolt Tensile Load Pounds 26,500 26,400 26,800 27,000 2 deg. underhead washer 24,900 25,400 25,500 24,700 24,700 Tensile UTS, Failure ksi 206 thread 205 thread 208 thread 210 thread 2 deg. underhead washer 193 thread 197 thread 198 thread 192 thread 192 thread Fatigue (R=.1) Cycles at Load 11,600 lb(1) 119,000K thread 30,000K thread 9,100K thread 10,400K thread 23,000K thread

(1) This is the required load for 200 ksi steel bolts. Two shear tests were run and the results were 109 and 111 ksi.

TABLE 6 - Test Results on MS21250, 180 ksi AISI 8740 1/2-20 Bolts Tensile Load Pounds 37,080 37,170 37,160 37,170 37,090 5 deg. underhead washer 37,320 37,110 37,120 36,880 36,900 Tensile UTS, Failure ksi 216 thread 216 thread 216 thread 216 thread 216 thread 5 deg. underhead washer 217 thread 216 thread 216 thread 215 thread 215 thread Fatigue (R=.1) Cycles at Load 13,900 lb 130K nf 129K thread 50K thread 65K thread 105K thread 105K thread Fatigue (R=.1) Cycles at Load 15,500 lb 55K thread(1)

(1) This part was a thread roll set up part and was tested unplated.

-9-

SAE

AIR5065

4. (Continued): The results obtained on the A286 10-32 engine bolts, presented in Table 4, are especially interesting because the load required to break the bolts was so high. Because these bolts have a reduced shank, the load applied at the head during the test was at a greater distance from the shank thereby increasing the possibility of head failure. The heads were subjected to tensile loads 42% greater than the minimum requirement. The stress rupture load was 27% above the required load. These bolts were not tested with a 5-degree washer because of the difficulty of keeping the bolt in the center of the washer (undersize shank). These bolts were in the solution treated and aged condition as required by the specification. The fact that these bolts had such excellent stress rupture properties is probably due to the high strength of these bolts. Apparently the high strength condition provides superior elevated temperature properties, at least up to 1200 F. Both the high strength A286 (Table 2) and the IN718 (Table 3) were tested in the unrecrystallized condition. This represents the extreme condition with respect to "grain flow" and cut "end grain". The "end grain" worry is probably a carry over from problems of exfoliation in aluminum structures. With respect to titanium, Ti-6Al-4V bolts are not subject to the requirement that there be no excessive cutting of flow lines. Many parts used in commercial aircraft are machined directly from bar stock. The Boeing specification, BPS-F-69, exempts Ti-6Al-4V from the requirement that too many flow lines not be cut. Possibly this is due to the fact that titanium is vacuum melted and extremely clean. However, many other materials such as IN718, A286, MP35, and PH13-8Mo are also vacuum melted, so it is difficult to understand why this requirement still exists. There are a number of 13-8Mo bolts on McDonnell Douglas prints for the F-18 which can be completely machined from the bar. All aerospace companies are concerned with head failures and will reject any lot of bolts in which a bolt head comes off during fatigue testing. Head failures during fatigue testing occur randomly. From Petersons Stress Concentration Design Factors, John Wiley & Sons 1966, it can be seen that for most sizes of aircraft fasteners, the stress concentration in tension at the fillet area is approximately two. Slight bending can significantly increase this. This compares with a stress concentration between three and five for the thread area. However, the threads undergo severe cold work compared to the head fillet area and are left with significantly higher protective compressive stresses. On the other hand, the stress area is smaller than at the fillet. The balance is such, that most bolts will fail in the thread. As fatigue tests are discontinued after 130,000 cycles, the users do not know if their fasteners would experience head failures if tested to destruction. A more meaningful test of the integrity of the fastener head would be to increase the fatigue load after the required minimum 65,000 cycles of most specifications, and test to failure. Any head failure would require that the entire lot be re-fillet rolled and retested. Such a requirement would be more discriminating than evaluating for the amount of cut flow lines under the head or any other head "anomaly". This could not be used for shear type heads. 5. CONCLUSIONS AND RECOMMENDATIONS: Grain flow of forged aircraft quality bolts has no effect on the integrity of the bolts. Not only will machined parts (all flow lines cut) meet the minimum specification requirements but during this program the heads were subjected to extremely high loads, well above the minimum requirements. No
- 10 -

SAE

AIR5065

failures occurred in the head of any of the bolts. The head failures which occur on a random basis in acceptance testing or in service are not due to "too many" flow lines being cut during fabrication. It is recommended that requirements concerning cut flow lines be deleted from aerospace fastener specifications. These requirements are subjective in that they are not numerically or dimensionally defined and provide no added integrity to the hardware. Requirements to start with clean stock and upset the heads should be retained. In place of the requirements for limited grain flow cutting, a requirement to fatigue test to failure would be more meaningful. When running the fatigue test to failure, the load should be increased 20% after 65,000 cycles and no head failures would be acceptable.

PREPARED UNDER THE JURISDICTION OF AMS COMMITTEE "AMEC"

- 11 -