The h-topology introduced by Voevodsky on the category $Sch/K$ of separated schemes of finite type over a field $K$ is the Grothendieck topology associated with the pretopology whose coverings are of the form $\{p_i\colon U_i\to X\}$ where $\{p_i\}$ is a finite family of morphisms of finite type such that the morphism $\coprod p_i\colon:\coprod U_i\to X$ is an universal topological epimorphism.
Why is the h-topology not subcanonical? I.e. what is an explicit example of a scheme $X\in Sch/K$ such that the presheaf $\operatorname{Hom}_{Sch/K}(-,X)$ is not a sheaf in the h-topology?
I will prove that presheaf $F$, represented by $X = Spec(\mathbb{k} [x] / (x^2))$ is not a sheaf.
Let $pt$ be $Spec( \mathbb {k} )$
$F(X)$ has more than one element.
Assume that $F$ is a sheaf in h topology.
Consider ${ pt }$ as a covering of $X$. Notice, there is only one map from $pt$ to $X$. Then $F(pt)$ is one element set. So $F(X)$ has one or zero elements. Сontradiction.