Look at a clothes line. It is anchored at each end. The line begins at an angle from one anchor point and ends at the same angle at the other anchor point.
Now, with the image of the clothes line in your mind, invert it. Now it looks more like a trajectory for a cannon shot.
Unless the rollers are angled in such a way to support the end-point angles, there is no way that the shape of the "hump" is going to look like a "...rope suspended between two supports."
It WILL have a definitive S-Curve shape. NO DOUBT ABOUT IT!
Now, why would anyone want to have a "hump" between the two rollers?
I think the answer to that is rather obvious... the purpose of the "hump" is to prevent any significant stretching in the material. Sure, there will be a little bit of surface tension on the outside of the three deflections as the material goes "through the hump". But, if properly controlled, that tension can be minimized.
How can stretching occur if the rollers are suppoed to be traveling at the same speed?
First, the original stipulation was that there were two sets of rollers. That is a Given! We have no choice... that is the spec that we are working under! The first set of rollers was... I think... horizontal, and the second set of rollers was vertical (again, I think. In any case, the two sets were oppositely oriented.).
In either case, again, the answer should be obvious. Controlling the Rotational Velocity of the rollers is a pretty damned straight-forward kinda deal. However, controlling the Linear Velocity of the rollers is not so straight-forward.
As time goes on, things happen... in this case, the "thing" would generally be a changing in the diameter of either the infeed rollers, or the outfeed rollers, or both... for whatever reason.
Oh, sure, some of you say... If the diameter(s) is(are) no longer the same as it was (they were), then it's simply time to clean or replace the rollers!
OK, but what if you are in the middle of a 3-mile run of this stuff? Are you going to let the hot stock just sit there while you clean or replace the roller? Not likely!
And if you let the system run, then the re-bar, which is supposed to meet very strict physical requirements, is being subjected to at least one of the very effects that the requirements are designed to prevent! Namely, "stretching" of the material!
Stretching is NOT the same as Compressing! Duh... Do ya think?
So... in order to keep the material from being stretched (between the two rollers), the Linear Velocities must be held equal. How can this be assured?
As long as the material remains at a horizontal level you can NOT be assured that the linear velocities are the same!
It is only by developing a "hump", and a method to monitor that "hump" that you can be assured that the "gozintas" and the "gozoutas" are equal! That is, that the two Linear Velocities are the same. If both Linear Velocities are the same, then there's absolutely no stretching!
We all know that we can NOT cause both roller sets to have exactly the same Linear Velocity... not exactly... no way... no how!
We can, however, develop a "buffer" that allows us to "fudge around the equal point".
By developing that "buffer" (that "hump") and then trimming the Outfeed Roller velocity, a bit faster or a bit slower, we can maintain the proper relationship between the Infeed Roller Velocity and the Outfeed Roller Velocity.
So... the purpose of the "hump" is to give us the ability to meet the Construction Engineering Specs (as specified by those that specify such things)!
Without the "hump"... we would have no way of knowing that we weren't developing a stretch (a little "Reverse-Logic" in that...)! At least, not a stretch caused by differences in the Linear Velocities of the rollers.
All of that was based on the idea that the Outfeed Roller Linear Velocity would be INCREASED because of an increasing Outfeed Roller diameter, thus causing stretching.
There is also the case where the diameter of the Infeed Roller increases, thus causing excess, and ever increasing, material to exist between the two roller sets.
This might make you think that all you need is to monitor for a "hump" developing because the Infeed Linear Velocity was faster than the Outfeed Linear Velocity... that would be wrong.
In that case, as long as the "hump" is not developed... all you know is that there is not excess material between the two rollers. You do NOT know whether, or not, stretching is occurring!
Having the "hump" is the only way to assure that the material is not being stretched!