Assuming you have accurate figures for guidescope focal length and guidecam pixel size, then trying to get beyond 0.5" regularly with HEQ5/EQ6 is pointless as they resolution of the steppers is the limiting factor.
My 'rule of thumb' is that a guiding RMS close to the imaging pixel size is the ideal, but even an RMS of twice that is hard to spot, especially if that's evenly spread between RA and DEC.
My 130P-DS/450D/coma corrector combo has a pixel scale of 1.83" per pixel, I have no trouble with it. The 150PL/450D is 0.89" per pixel, so I prefer to get sub-arcsecond guiding with it, although slightly worse doesn't seem to cause problems. I use a 183mm f/l guidescope with the 130P-DS and 66 ED, and a 220mm f/l guidescope with the 150PL, both with the 120MC camera.
The usual cause of trouble is when one of DEC or RA is showing spurious peaks once or twice per sub, even if the RMS is only a bit high, as this means bright stars come out as 'doubles'. Usually badly balanced DEC and sometimes solved with resist switch.