Hi,
The so called "instability" is relative to the application.
In other words, any jitter would be a small percentage of the
total period and this would only affect the application slightly.
Whether or not that jitter would ruin the application is something
that would have to be assessed of course, but one example
where the jitter is small relative to the time period is with the
control of a triac trigger, where the one shot produces a
delay after the zero crossing so the gate can be fired at
some time later, which regulates the output power.
This works out extremely well for regulating the power to
an ac operated device like a motor or lamp. The motor runs
very smooth and the speed goes from full down to almost
zero. The frequency is rather low however, around 60Hz,
so the jitter affects it probably less than 0.1 percent. At
higher frequencies it might affect the phase shift more so
you'd have to try it.
Other ways of adjusting phase are possible, for example
using a phase shift network and then detect the zero
crossing, but this isnt as easy to get the full 0 to 360
degree (or nearly that) phase shift, and after all nothing
is going to be perfect.
A PLL is another idea.
Generating a ramp and detecting a particular voltage level
is another way to do this.
Depending on the frequency and what kind of phase step
increment you need, you might also use divider chips or
of course a microcontroller.
It might be a good idea if you tell us the min and max
frequency you will be working with. The lower the
better actually. Also, what the min step in phase you
need, such as 1 degree. The bigger steps are easier
to generate.