Will the number of documents affect the topic distribution of LDA?

Question

I am building a LDA model with 840 documents

def compute_coherence_values(dictionary, corpus, texts, limit, start=2, step=3):
"""
Compute c_v c