How much can you 'laymanize' a PhD presentation?

First, it is important to recognize that every scientist is effectively a layman for almost all other science.

For example, I am a computer scientist, and I have only a vague layman-level understanding that when you say "non-local waveform collapse" it's probably about quantum mechanics and Bell's theorem, and the math you put up on a slide will certainly be rather alien and fairly incomprehensible to me. If I put up a slide on the semantics of field calculus, it would probably be likewise alien to you, and if a biologist of my acquaintance put up a slide on the chemical mechanisms of chromatin remodeling it would be alien to both you and I. Think of a layman, then, not as some sort of idiot to be condescended to, but as an intelligent and interested person whose expertise happens to be in a different field.

Furthermore, our relative ignorance of one another's work is greatly amplified while listening to a talk: when I am puzzled and need to think about what you've just said, I miss the next things being said and have to try to catch up! In effect, only people who are extremely close to your own work, even within your field, will be able to follow deep technical details in a talk. Every audience except the smallest and most intimate, then, is effectively composed mostly of laymen---but likely with a few experts as well.

So, how do you give a talk to heterogeneous audiences? Personally, I find it effective to think of a talk as an advertisement, the goal of which is to make people interested enough and convinced enough by your work that they want talk with you about it afterward or to read the manuscripts that present it fully. Start with a discussion of the goal of your work, how the work relates to other work and why its goals might be interesting, narrow down to a few key details (which might only be understood in detail by the few experts in the audience), present evidence for why your insights are correct, and widen back out to explain why what you just presented satisfies the goals you laid out in your introduction. Such a structure can then "laymanize" even the most complex topic, by explaining how it fits into the larger world, while at the same time respecting the intelligence of the audience and not sacrificing communication to experts.

For the details of how to accomplish this effectively, I personally have found Patrick Winston's "How to Speak" lecture to be an excellent resource for heuristics and tips, and recommend it to others as well.

In my experience the best approach in these situations is the use of analogy.

The reason for this is, if you can find an analogy which is applicable it can be accessed by most members of the public as well as your peers. There is an added bonus that an analogy is far less condescending than a talk during which you treat everyone as if they were 5 years old :)

Analogies can be scaled to the audience. For example, I once saw a talk at a conference where kinetochores were likened to couples dancing. It was effective and I feel would have been accessible to anyone. While speaking to an academic audience terminology was used which made it clear that the 'couples dancing' analogy was not fully suitable however it was a good 'base' on which to build the questions and hypothesis.

Another advantage of this technique is that if you are clever about it you can make your talk humorous in a natural way.

A good rule of thumb when speaking to lay people (or scientists from other fields) is to talk mostly about why you research a particular topic, and only a little bit about how you actually solve the problem. A general audience isn't going to understand the details anyway, but you should be able to explain to them in broad strokes what it is that you're interested in and why anyone should care. Indeed, if you can't explain these last two points, then you probably did not understand them well enough yourself.