You cannot have a clock with 0% tolerance even when using an X-tal there will be a precision of some ppm.
So what I read from your post is, that you need a precise clock. Full stop.
The precision of the IMO is given in the PSoC datasheet and is about 5%.
If that is not enough you HAVE to use an X-tal as reference, preferrably a 24 MHz one. Now the precision of your (any) clock derived from this source depends on the precision of that equipment.
Since all clocks are derived from either the IMO (that's where the "M" stands for) or from (the VERY unprecise) ILO you'll have no other choice.
A singal with +-0% ?
I think even crystal still specify Xppm which is still not 0%.
There are new MEMS clock source that are low ppm but not sure if you can find 0%.
There is a method of achieving very high precision measurements, like T, V, I, Freq.
Attached is a simple diagram of the hardware used.
Basically hi accuracy data communicated to processor, and processor measures deviation from actual,
building tables of correction factors. This is done at production test, and can include T and V variations.
You cannot achieve 0%, all signal generation has some noise and uncertainty associated with it. But
you can get sub ppm performance with this method.
The only drawback is component value changes over time, unless those are well understood they
compromise total accuracy over time that can be achieved.
Hi Accuracy Measurements.jpg 115.2 K