Yeah, the signal strength also increased, but I don't remember the exact numbers. In general, I prefer using SpeedFlex throughput to measure real-world throughput as a metric for the mesh link strength. Signal strength (RSSI) is a static heuristic that's meant to estimate strength, and I found that less meaningful in terms of real world customer impact compared to a Speedflex.
I would actually first recommend replacing the R500 mesh node. I have noticed that even compared to the R600, the R500 is less capable of establishing strong mesh links with good throughput across challenging distances. This is likely explained by a combination of 2x2:2 vs 3x3:3 *and* the fewer BeamFlex antenna patterns.
Perhaps an arrangement to try is the R710 as root and move one of your root R600's to replace the R500?