I only chronographed one round of 52 grains on a hot Brisbane summer day and I got 2475 fps in my R8, considerably less than both yours and Vaughans quickload data.
It's obvious, I know, but it is worth remembering QL is calculating aid to reloading, a very good one I believe, but just a tool and the pressures and velocities are based on equations not exhaustive physical measurements for all calibers and combinations.
I find it very useful for selecting starting powder and projectile combinations and figuring out initial loading ranges etc. It rarely comes in on the money with respect to MV. The exception there seems to be my 6br, for which it seems to calculate MV with surprising accuracy. For my 9,3, not so good (always a good deal higher than actual).
I also find it to be fairly conservative. My old 308 loads based on Hodgdon data and working up etc should have detonated my R93 if QL and the internet was to be believed. After buying QL from Pigeon I backed those loads off a bit even though I did not believe they were dangerous in my rifle. I figured it didn't hurt to be careful. QL seems to match up OK with Norma factory data in the 6br and 9,3.... This drivel being a long winded way of saying I'm not too surprised by those differences.
One thing that you might not think of (apologies if its obvious) - if you are getting up toward max pressures be very careful with messing about with seating depth. If you're near max pressure and you shorten your COAL you can push your load over safe limits fairly easily.