We have a number of demo data sets on the wiki and, while some of these have a reasonable set of patients with thousands of observations for >100 different concepts, I don’t know how representative they are.
Is anyone sitting on a medium-to-large OpenMRS dataset that is anonymized? Or that could be relatively easily anonymized? In return, you could get your data converted into a format for analysis for free. 
There are a few people working on updating the ETL mappings for OpenMRS → OMOP’s Common Data Model (CDM) both to provide some good clinical demo data for the OHDSI community and to help expose the OpenMRS Community to the power of OHDSI tools.
Cheers,
-Burke 