Running torque, also known as prevailing torque, is defined by NASA as “the torque required to overcome kinetic friction of the mating threads plus the torque required to overcome any locking feature when 100% of the locking feature is engaged and the fastener is unseated”*.
In other words, running torque is the rundown torque required to overcome friction and/or any locking feature in a threaded fastening application, which produces no clamp load, or bolt stretch.
Even more simply, it’s the amount of torque needed to seat a screw.
This running or prevailing torque may be required to overcome intentional interference designed into the bolted joint, like the prevailing torque locking we described in a previous article, as well as unintentional interference that is due to friction (which will inevitably vary from joint to joint) and other environmental factors. But ultimately, it does not generate any preload.
At times, confusion can arise among manufacturers on whether the torque value specified for a fastener is inclusive of running torque or not.
When it comes to aerospace applications, NASA requires engineering documentation to clearly identify when the installation torque refers to the torque above running torque.
When engineering documentation does not explicitly specify “above running torque”, the indicated torque is understood to mean the final torque visualized by the torque instrument used for fastening or measuring.
In other applications, the torque value indicated in the specifications typically refers to the final torque inclusive of running torque.
For joints where running torque has a certain inherent variation, it may be more appropriate to focus the fastening and quality control processes on the “torque above running” value.
In aerospace applications, for instance, where quality control abides to extremely severe standards, even a very small variation of prevailing torque between joints can be considered to cause a significant-enough difference in clamp load between them. Therefore, it is often preferred to center the tightening process around clamping torque rather than final torque, since the latter would include variable amounts of prevailing torque and thus produce different amounts of preload on different joints.
In less demanding applications, however, variations in friction between joints are deemed to be small enough where it is acceptable to use a final torque value that is already inclusive of running torque. Nonetheless, it is often crucial to validate this low-variability assumption by measuring the running torque during each tightening and recording its value for quality control, or even failing the operation if the measured value is outside the acceptable range.
No matter what the application at hand, Kolver’s K-DUCER makes dealing with running torque easier than ever.
The “Prevailing/Running Torque” feature can help measure the amount of running torque encountered during rundown and, if needed, it can dynamically compensate for it by adding it to the total torque applied to the joint.
Let’s take a look at both of the aforementioned examples and how we can set up our K-DUCER tool to deal with each application.
RUNNING TORQUE - COMPENSATE
Let’s say we are working on an aerospace application with installation torque specified to be 21.5 Nm above running torque, where the running torque encountered can be between 1.1 and 6.8 Nm.
Specs:
Torque above running: 21.5 Nm
Running torque: 1.1-6.8 Nm
The goal in this case is to apply a total amount of torque that varies based on the amount of running torque detected, resulting in a consistent amount of clamping force on each assembly operation (as opposed to a consistent amount of total torque but a variable amount of clamping force).
The first thing to do is to run several tightenings of this joint near the maximum final torque allowed by this application, in this case our 21.5 Nm target plus the higher end of the running torque range, 6.8 Nm, for a total target of 28 Nm.
To do that, make sure any advanced torque option has been disabled and that the target torque in the main menu is set to 28 Nm.
We can now perform 10 or more tightening operations.
Because we want to visualize the torque/angle and torque/time graphs, we find it easier to work on a computer using Kolver’s free K-GRAPH software.
For each tightening operation, we want to note down the angle and/or time point at which point the screw becomes seated. This is easily recognizable as the torque graph suddenly starts increasing at a steeper angle until tightening is complete.
After 10 or so rundowns, you should have a consistent average value for that angle or time point.
In this example, we measured the seating point to be, on average, after 9500 degrees of rundown.
On the K-DUCER’s touch screen interface, let’s open the main menu and then select Advanced Torque.
Let's ignore the Max-Power Phase option, keeping it to OFF, and let's set the Running Torque feature to “compensate” mode.
For the window value, we are going to set 0 - 9000 degrees (slightly less than our 9500 average value, just to give it a little extra cushion).
Because the prevailing torque we measured, while oscillating, was trending relatively flat, we will choose the “average value” option.
And as the specifications indicate that running torque can be expected to be between 1 and 7 Nm, we will set the torque bounds to be those values +- 10%, i.e. min = 0.90 Nm and max = 7.7 Nm.
Let’s now go back to the main menu, open Torque&Angle settings and adjust the target torque.
As you recall, for our previous step, we had originally set the target torque to include the running torque; we now want to clear that and set our target torque to match our “above running” specification of 21.5 Nm, since the running torque will now be automatically detected and added to that value.
Now we can go back to the main screen and start working in running torque - compensate mode.
As we successfully complete our first tightening operation, notice how we successfully reach our “above running” target value and how the interface visualized the running torque we encountered and the total torque that was applied.
In this case, we can see the clamping torque applied was 21.5 Nm, right on target.
Right below, we see that the K-DUCER measured 5.6 Nm of PVT (Prevailing/Running Torque) which it then added to our target torque in order to calculate the total torque to apply. Aiming for a total target torque of 27 Nm, it shows to have been right on target at 27 Nm of total torque applied.
If we run a few more tests, we will observe how the measured prevailing torque varies slightly from tightening to tightening, yet the screwdriver compensates for it by varying the total torque applied, ensuring our target torque remains consistent throughout our series. 
In our final rundown, for instance, we can see that the measured PVT was 5.3 Nm, with the K-DUCER applying a total of 26.8 Nm in order to achieve our intended target clamping torque of 21.5 Nm.
RUNNING TORQUE - MONITOR
Now let’s take an application where the installation torque is specified as a total-torque value, already inclusive of running torque.
Specs:
Total Torque: 9.9 Nm
Running torque: Unknown
In our new example case, we have a screw that needs to be tightened to 9.9 Nm and while this figure already accounts for running torque, we still want to measure it to ensure that we don’t run into a higher-than-expected variability from joint to joint.
After setting the target torque to 9.9 Nm, and ensuring that all advanced torque options are turned off, let’s run a series of rundowns to identify the running torque window, as in the previous example.
After several tightenings, it’s clear that the seating point always happens after about 6300 degrees, so let’s go back to our K-DUCER’s menu, open our advanced options, turn on the PV/RunningTorque to “monitor” mode, and set our angle window to 0 - 6000 (again, I prefer to give a little cushion to the seating point, by ending the slightly shortening the running torque window). 
We also noticed that the running torque is consistently near 0.56 Nm, and we expect it to vary very little during production, so we let's set the min and max values to be 5.08 and 6.21 Nm respectively. This is important because if we later encounter an avg running torque value outside of these bounds, the tightening operation will (correctly) fail.
In other words, during production we are trusting a target value that's inclusive of running torque because we expect running torque to offer little variability, but we also want the assurance that the tightening operation will result in a NOK (fail) signal if the running torque encountered does vary more than expected, as that would result in an incorrect clamping torque.
Back to the main screen, we are now ready to perform our first tightening.
The screen now indicates that the screw was tightened to an acceptable 10.2 Nm and also notes, right below, that 0.56 Nm of running torque was measured during the rundown.
As we run a few more tightening operations, we are shown that there is, in fact, very little variability encountered in running torque, with the K-DUCER consistently showing a PVT MEASURED value of around 0.56 Nm.
While in this case running-torque compensation was not needed, the K-DUCER’s monitoring feature gave us the peace of mind of knowing that its variability was extremely low and unlikely to affect the clamping force on the joint.
Had we encountered a running torque of, say, 1 Nm, the tightening would have failed.
In our next article, we will look at a more complex use case where a multi-step torque strategy needs to be employed to account for a joint that has two different phases of running torque.
* NASA