One day several years ago, I was called to an Engineering meeting in progress where a telephone conference call was being conducted. An excited voice on the far end was describing a problem they were having that shut down the assembly line.
An assembly that had not given problems before had suddenly started breaking bolts during the tightening sequence. I was assured that the fastening equipment had been checked and was performing correctly. They were also using the correct part numbers in the assembly. I returned to the Fastening Lab and proceeded to order the parts and get the test equipment together to try to duplicate the failures. When the test fixtures were finished I was ready to try to duplicate what the assembly plant had experienced. I ran several samples with no failures and called the test Engineer and informed him of my results. We talked for a while and agreed that something had changed in the assembly or in the tooling to cause such a drastic event.

We set up another conference call and asked to have the equipment operators present. We asked if there was anything they could think of that may have gotten changed to cause the breakage. One operator said that he thought the bolts looked different than they had the last time they ran the assembly. When asked what they looked like he replied they were grey instead of black. He had checked the part number to print and the part number matched the print part number. The required part was M12 x 1.5 x 3.5” long with a phosphate/oil finish for retarding corrosion during service.

Somebody had changed the finish to zinc for better corrosion resistance and did not change the part number to indicate the change. The area in which the assembly was being made was subject to a FINE OIL MIST FROM THE AIR DRIVEN TOOLING EXHAUST.

The phosphate/oil finish was not affected by the oil mist during rundown and tightening and no one thought to test the new finish under the same conditions with the oil mist present and with the part number being the same there was no red flag that caused concern. What happened to the finish on the new bolt was a disaster. The oil mist, however fine, got onto the zinc finish and the result was a slimy coat that affected the friction adversely to the point of consistent bolt breakage at the same tightening torque applied. With this new information it was immediately apparent that the tightening strategy needed to be changed. The reason my testing did not show a problem was that the parts I ordered from our stock was the phosphate/oil finish and resulted in the proper result at the specified tightening strategy. We requested that they send some of the new bolts to our Lab so we could retest and come up with a new tightening strategy until they could get the older style bolts ordered. The bolts showed up the next day and I was ready to begin testing.

First I ran a failure mode test to determine the maximum torque that could be used with an oil mist on the zinc finish. Surprisingly the maximum torque required to break the bolts was less than half of the specified torque used in the original system. No wonder they were in a frenzy!!

The testing that followed uncovered another problem that resulted from various amounts of oil mist on the zinc. The friction range was so great that the clampload varied from low to breakage of the bolt!

It was difficult to make a determination of an absolute torque +/- a variance, due to the wide range of friction. I then decided to try to determine a tightening strategy that would hold the joint together and not break the bolt. The joint was one that needed a minimum clampload to eliminate slippage, so to get the clampload it was necessary to get the torque to a narrow range for repeatability. This would require the same amount of oil mist to be on each bolt (haha) to get the friction to be consistant!! But because friction is so non repeatable this appeared a conundrum! I decided to try a torque/angle answer but this could also be a problem because the assembly plant was not equiped with state of the art tooling.

There were approximately 2000 joints worth of bolts (4000 bolts) that were affected before the proper bolts could be delivered and manufacturing was not going to shut down for even 2 days to wait for the correct bolts.
I got the OK to go ahead and develop the torque/angle strategy and they would implement it by hand for the remainder of the bad batch of finish. To make a long and drawn out story shorter due to the various amounts of oil to be sprayed on the test bolts I finally determined that 15 Nm + 90 degrees of rotation would give the amount of clampload required to keep the joint together and not allow slippage! The workers quickly adapted and with a torque wrench set to click at 15 Nm and eyeballing 90 degrees +/- roughly 5 degrees the joints were sucessfuly assembled and production resumed, a little slower than before but enough to get the product out safely. This was a plus derived from a minus that helped to sell manufacturing production on the idea of torque plus an angle of rotation to assemble certain joints that required a minimum clampload to survive heavy service!! It is very difficult to change from just torque to a more certain strategy due to the type of tooling. The better computer controlled torque + angle tightening tooling is expensive and a time sensitive program. The added cost is really offset by the improved reliability and less returns for retightening the joints.

The following figures show what happens when certain finishes are subject to contamination by an external influence. The first figure shows the original joint with the phosphate/oil finish, torque control and the resultant clampload and variation. The second figure shows the wide variation in clampload at the same torque control, with various amounts of oil mist on the zinc plated bolts. The third figure shows the relationship of angle turn and the resultant clampload.