I've always been told that DNA assembly can be tricky when using very small DNA parts due to low efficiency. I've also seen this when using 3A biobrick assembly to assemble promoters and RBSs together. However, there doesn't seem to be any information that I can find which definitively shows what the minimum DNA part size is that can be used before the efficiency becomes so low that assembly is unfeasible. The most I can find are warnings that small DNA parts are problematic (e.g. https://blog.addgene.org/plasmids-101-golden-gate-cloning).
What is the minimum DNA part size that can be used in DNA assembly before it becomes unfeasible? I'm specifically interested in Golden Gate assembly, but information for other digest-ligate assembly methods could also be helpful.