Any machine discriminating out iron will "null" on even a silver coin once it reaches a certain point in depth. Reason being? The microscopic iron in the coil reaches a point where it is now stronger than the weak coin signal past a certain depth. At the very last fringe where a detection field can see the coin it'll drop down to the iron range before it can't even be seen as that anymore.
Also, a super deep coin can also null due to mineralization for the same reasons as iron. It reaches a point at depth where it now is regarded as part of the ground signal and it might not "null" as being iron, but it will null the threshold as a fast change in ground matrix.
Far as iron rejection, I used to be a believer that using no iron rejection was the only way to improve unmasking. I no longer believe that, after seeing/comparing a sharp DD line (12x10) using high built in iron rejection on a machine to one using very minimal iron rejection and seeing the same unmasking ability. Not saying either way is better to unmask, just saying I now believe there is more than one way to skin a cat.
Reason being that lower iron rejection isn't "seeing" a coin at the same time as iron better, but rather it's reducing recover time between "reject mode" of the iron and "accept mode" of the coin right up against it. If a detection field sees iron first, no way to see a deeper coin for the most part. Just the nature of detection fields. First target seen, even off to the side but still in the field, is the one that gets reported, with minor exceptions.
So if the reject/accept mode "lag" thing by reducing iron rejection is what is really going on, then how is there another way towards the same goal of unmasking? The other being that if you use a sharper DD line, you've more cleanly defined when you are seeing the iron, to not seeing it, and then to seeing the coin as the coil moves on. It's more cleanly transistioned, like turning a page on a book, so in effect you've increased the time between "reject" and "accept" in a way, just like lower iron rejection will do.
That's MO anyway and I'm sure others differ. Believe me, a few short years ago I thought otherwise, so I don't blame people for thinking otherwise as well by their experience if they don't agree. I can only go by what I've seen with my own eyes and ears on undug stuff in the field. I don't argue one is more important than the other, only that there are several paths to the same goal...
PS- Depending on the iron/mineral content, even shallow targets can null out. I've seen at one very bad beach that discrimination mode is nulling on coins at only 5 or 6" beach, while all metal pin point mode is easily banging on them hard much deeper. All metal has it's purpose and benefits for sure in some situations. The thing to keep in mind though is it's not seeing a coin through iron any better than discriminate. The laws that govern detection fields still apply- First target seen in the field and it's game over on seeing something deeper for the most part. Where I feel all metal can "unmask" better is the ability to sound off to stuff that the ground matrix is "masking. Not masking by iron like nails or such, but microscopic iron playing havok with the discrimination circuit and choking out targets due to that microscopic iron or other mineralization. I've seen even the best of VLF machines get the depth of a cheap radio shack detector in a bad matrix. That's were all metal or an all metal pin point mode can really shine.