I don't use a portable XRF, but am familiar with the theory (we have a benchtop XRF for soil analysis in our lab).
The degree to which XRF penetrates various substance may be the issue here. As a general rule, the further "down" the periodic table you go, the shallower the x-rays penetrate. In the noble metals group on the periodic table, copper is on top, then silver in the middle, then gold at the bottom. So x-rays will pass right through even a thick piece of copper, while they will be blocked by just the few millimetres surface layer of pure gold. Silver is halfway between - for a thick ingot, the x-rays shouldn't make it all the way through, but it does depend on both the thickness of the piece of silver and the intensity of the x-rays, which in turn derives from the power level of the machine.
What I suspect is happening, with your "98% silver" reading of a silver ingot, is that a tiny percentage of the x-rays are passing right through the silver ingot, and the XRF is picking up "contamination" from whatever is underneath the ingot - the table, benchtop, floor or whatever it's sitting on when you run the test.
I think it's safe to conclude that the "98% silver" item is in fact pure silver, if for no other reason that nobody who is making fake silver ingots would make them out of 98% silver - there's no profit for them in doing that.
Don't say "infinitely" when you mean "very"; otherwise, you'll have no word left when you want to talk about something really infinite. - C. S. Lewis