Torque Tooling Requirements

s.starr

Registered
I'm the first Quality Engineer at my company and fairly new to the discipline (under 2 years overall experience) and have been tasked with defining torque tooling requirements for all current and future torque wrenches.

In our manufacturing process, we use calibrated click-torque wrenches that can hit 10-100 Ft-Lbs., are priced anywhere from $50-$300, and all state to be 3-4% accurate. In general, thru-bolt applications are used (hex and carriage bolts with washers, and a nut), and specified torque values are not typically listed in our Engineering drawings or Production Work Instructions, but the most common nominal torque value used during assembly is 24 ft-lbs. The bolts are torqued 2-3 times during the manufacturing process: the first time to the minimum value (23 ft-lbs) in Production, the second time to the nominal value (24 ft-lbs) in QC, and optionally a third time to the maximum value (25 ft-lbs) in Test. If the torque value is not known, we have a QMS document called the "Mechanical Fastener Guide" which has a table listing the tightening torque values (per fastener grade, TPI, etc.) of every hardware piece we would use in our manufacturing. You would refer to this document to take the tightening torque value listed, then take -10% of that as your min value and -5% of that as your nom value, with the originally listed tightening torque value functioning as a max value in our process. Finally, we also mark the bolts with a paint pen to visually indicate the bolts have been torqued. This mark is on both sides if it's a hex nut, or just on the thread/nut if it's a carriage bolt. For the mark on the thread/nut side of the thru-bolt application, a continuous mark is made from the surface of the fastener onto the adjacent surface. The idea is that the break in the mark can be used as a visual inspection of torqued hardware. Each department has their own respective colored-mark.

We've been having a lot of issues with undertorqued bolts from Production (who's torquing to min). Since torque wrenches are generative equipment, not measurement, the paint mark is the visual indicator QC uses to decide what is "undertorqued." This has historically been anything that twists 90 degrees or more from the original Production torque mark. Ideally, if the bolts were torqued properly to the min value, the torque wrench should barely twist to hit the nom value. From some testing, this 90-degree turn is the equivalent of torquing to nom from hand-tightening the hardware, or using half of the nominal value. I've also done tests with torquing to 18 ft-lbs (25% of 24 ft-lbs.) to see visually how much twist the bolt moves, and that was always less than 90 degrees, despite being way below the min value, so anything QC receives where the bolt twists 90 degrees or more means the hardware is severely loose.

Recently, I conducted an experiment where I used 3 QC technicians to torque 10 bolts each on 2 pieces of our product: 1 with all hardware torqued to min with the same torque wrench, 1 with all hand-tightened hardware. The purpose was to see if the 1 piece with all correctly torqued hardware would pass inspection across the different techs using our torque process. In the end, of the 30 sample pieces from the correctly torqued hardware, 2 of the 3 techs falsely failed a combined 3 bolts, resulting in a 90% pass rate when it should have been 100%. This leads me to believe the tooling we are using is not consistent enough to hit our product's torque requirements. I also noticed the 3 failed rejections were all on hex bolts; (the 2 products were split between a combination of hex and carriage bolts).

Here's what I need help with: when torque wrenches are listed to 4% or whatever % accurate, does that mean the tool can hit the value set on the wrench within 4% of that value (ie. 24 ft-lbs +/- 0.96 ft-lbs for 4% accuracy), or does that mean 4 out of every 100 clicks are going to be something other than 24 ft-lbs (accuracy vs. precision)? Should our current tooling be hitting our product requirements already? If not, to what accuracy should the torque wrenches be to meet the product requirements consistently? Is a destructive MSA the right test for evaluating our current tooling, assuming it's calibrated properly or are torque testers an adequate means of testing torque wrenches? We have two digital torque testers at our plant, but neither are listed to evaluate anything higher than 21 ft-lbs, yet can still display up to 30 ft-lbs., should we be testing the wrenches with the testers at a reading outside of the listed tester value, or at all? Are click-style torque wrenches the way to go, or are alternatives like digital or cam-over wrenches better/more consistent? Should I use AQL?

I know this is a long post, but I wanted to provide as much detail as possible, any help is greatly appreciated, I'm looking forward to learning more!
 

Golfman25

Trusted Information Resource
So just a couple of thoughts, spit balling here.

I also would doubt your off the shelf wrenches could hit and maintain the torque accuracy you're looking at -- +/- 1 lb. You could do a test trying to verify them. You man need to upgrade your tools?

Second, why can't you just torque them down to a min value which will be more than you need. So if you need 24 lbs. min, set the wrenches higher than the min. -- say 25 to 30, or whatever works.
 

John Predmore

Trusted Information Resource
Here's what I need help with: when torque wrenches are listed to 4% or whatever % accurate, does that mean the tool can hit the value set on the wrench within 4% of that value
I am not a torque expert, but I did a bit of research before I bought torque wrenches earlier this year. The short answer is yes, the manufacturer claims with proper use, the wrench will click at a value within 4% of the setpoint.

Proper use of a click torque wrench means to only use it as static verification of an assembled joint. Avoid using the click wrench to assemble a threaded joint, because the working surfaces inside will wear with use. Proper use requires that the spring preset load inside the torque wrench be loosened to zero between uses, to prolong reliable function. Torque wrenches need to be re-calibrated often to have confidence in their accuracy, maybe more often than once a year. Inexpensive torque wrenches will need to be replaced more often than other hand gages.

@s.starr mentioned this as a “historical” problem, which makes me wonder: how was the torque spec originally developed? You mentioned the Mechanical Fastener Guide. Handbook values are legitimate initial values for fastener design, but those are based on optimal conditions. Was there any product or process validation done on the 24 ft-lb? Rough surface finish of male or female threads creates friction during assembly which steals torsional force which would theoretically go into tightening torque. Plated threads will behave differently in a joint than unplated, for example. If you are compressing a gasket or washer in the joint, that will also divert some force that would otherwise go into tightening torque. Components in any joint will stretch or otherwise deform under load over time, leading to a phenomenon called relaxation. Vibration or side-loading of surfaces in the joint can contribute to loosening of a joint.

Here is one source I found on the Internet with an extensive discussion of the technology of torque. I have no affiliation with this company
 
Top Bottom