Hi,
Yes, the zener diode is a good idea but the linear will also track the input noise to some degree, so a little filtering is in order here too. A resistor in series with the zener (not too high of a value) followed by some capacitance to ground will probably do it, and the zener voltage high enough to surpass any noise on the input. Maybe 2v would be better.
However, this is NOT the correct way to build a switching pre regulator power supply.
Normally the way this is done is the switcher is made to track the linear, not the other way around. This way when the linear is adjusted the switcher tracks it. That's how the circuit i posted back a while works and that's the typical way of doing it.
The problem with making the linear track the switcher (this new circuit) is that the output is no longer regulated with load. As the load increases the switcher output loads more and more which causes a drop in voltage at the output of the switcher. This in turn makes the output of the linear drop too, as much as the switcher plus some more because of its own change in internal impedance. A load that draws a pulsed current for example will cause the output voltage of the linear to bob up and down. It's not as bad as no regulation, but it's not as good as the other way around.
When the switcher tracks the linear (as more typical of this kind of dual regulator) the switcher keeps up with the linear so the output stays constant with load, and the linear is the sole regulator so the output is well regulated.
The way to simulate this new circuit would be to use a DC input voltage with a small series resistance and use a spice model of the linear regulator and use the zener or whatever. Load the regulator WITHOUT changing the input DC voltage and see what happens. The change in internal impedance of the linear regulator will cause a change in output voltage which is more than if it was really allowed to regulate.
There is probably a way to correct this by using the linear regulator as usual but bias it differently, but what's the point if there is a simpler way to do it with a transistor and zener to make the switcher track the linear.
For example, for a fixed output a zener from the output of the switcher to the bottom of the lower resistor of the linear regulator might do it. That way the linear still regulates, but it's still not as good.