There have been a few studies in which supposedly young materials were dated by the K-Ar method as millions of years old. It turns out that these materials were actually a mix of young material and much older material, so the conclusions are obviously invalid.
The Grand Canyon study was not done on rocks from the top of the canyon and from the bottom. It was on samples from four separate lava flows on top of the canyon, and on one phenocryst from one of the flows. The theory of isochrons predicts that such a study will
not produce the age of any of the lava flows… and it didn't. So the theory is supported rather than undermined by this result. But, in fact,
the determined age is almost certainly correct!! The isochron method measures the age at which the samples were
isotopically homogeneneous, which for the samples from separate flows was not when they erupted (at differnt times) but was when they were all together in the mantle under the GC. 1.34 billion years ago, as determined by Austin's "study".
In at least some of these cases the investigator knew (and in all of them he should have known) that the claimed errors were not errors at all, but were instead the product of deliberate and fraudulent sample selection and misrepresentation.
Finally, all tests are subject to an error rate. None of the problems you cited are valid. Even if they were, a few bad results, even a few hundred bad results, are a minuscule fraction of the many results consilient with each other and with non-radiometric methods. A few failures means nothing.
Edited by JonF, : No reason given.