I am not sure, if I understand you correctly.

PD4ML implements its own HTML/CSS rendering engine. The engine starts page layouting only after all resources (images, stylesheets etc) are completely loaded. So it is not possible, that page layout changes itself in 1-2 seconds.

So I am afraid, if a PDF conversion result does not match its original HTML page layout, there are probably other reasons (failed CSS parsing, unsupported HTML tags etc). An HTML/CSS samples would help to analyze the problem reason.

The above is relevant for the original PD4ML library. There is also PD4ML Web (based on WebKit native HTML renderer) – the version may suffer from network/rendering asynchrony or delays.