Back a long time ago, the way it was explained to me was that the sag rods are placed (nearer the compression flange)such that they brace the compression flange for internal wind pressure. Plus they carry the dead load of the girts to the eave. The siding braces the compression flange for external wind pressure.
The sag rods are usually placed at the center of the girt. The sag rods and the siding work together to create some torsional restraint at the compression flange location (at the interior of the building) under wind suction. This is normally sufficient to assume a restraint at this location.
I agree with JedClampett and Jike. I also make sure the sag rod is capable carrying some compressive force (kl/r<200). Depending on what your girt spacing is this may be very difficult to accommodate.
Man I know we've talked about this a while back - but can't find it with the Search engine.
If you have a horizontal wall girt spanning between building columns, it will get both suction and positive pressure - either way - causing compression in either flange.
If you have one or more vertical sag rods connecting to it, then under the suction condition the sag rods would act as a lateral brace by developing a couple between the wall connection and the rod....ONLY....if the predisposition of the girt is to laterally buckle downward, away from the "tension only" rod.
However, if the girt wants to buckle upward, then the sag rod has very little stiffness and will not serve to create this couple to resist the lateral movement of the interior compression flange.
To ensure that the girt will tend to buckle downward, the girts can be specified to be installed with an initial downward sag. Any future compression in the interior flange will result in a downward lateral movement and the sag rods will adequately work.
Since most girts will naturally have this downward sag, the traditional sag rod system has tended to work well. However, some over-zealous contractor might over tighten the nuts on the threaded sag rod and create an unwanted upward camber.
It is one thing to say that sag rods will brace a flange but quite another to prove it. We're engineers and we should quantify these things when possible. It should be possible to find the strength and stiffness of the sag rod for both tension and compression. The AISC mannual provides us a means by which to compare the strength and stiffness of our desired bracing with a minimum necessary requirement.
Chances are, the stiffness in tension works and compression probably doesn't, as JAE wrote. So the solution is to make the compression work or remove the likelihood of the girt buckling such that it produces compression in the rod.
Read through the commentary in your girt manufacturers handbook. You may find that the capacities given in their design charts only apply if their proprietary bridging is used.
The difference is that the sag rod may provide lateral bracing to the flange, but the proprietary bridging provides torsional restraint as well.
