I know I keep banging on about it, but also don't forget about the possibility of intersample variations. e.g. I bought 2 pairs of HD600 within a couple of weeks of each other from Amazon, I sent the second pair back for a refund because it had harsh and 'shouty' mids. Much later I bought a used pair of HD600 for my dad. This pair has poorer bass extension than my pair but I actually prefer the sound of his pair because it has more "air" in the high treble and therefore sounds a little more open than my HD600.
In the last couple of years I added the HD660S, HD650 and HD58X to my collection, all bought new. According to my measurements, my HD600 is almost identical to the HD650 with respect to bass extension to below 20Hz, but the HD650 has more bass from 20Hz to 300Hz when a perfect seal is made. My HD660S has the weakest bass extension, and my HD58X has the strongest bass extension.
With regards to my HD600 vs HD650, I prefer the 600 because the 650 sounds more muddy/bloated in the mid-bass/upper-bass and slightly thicker and more shouty in the upper-mids. The fact that my HD650 also has less treble than my HD600 further increases my preference for the HD600. I can however make my HD650 sound almost identical to my HD600 with a few simple EQ adjustments. With this EQ applied you'd find it very difficult indeed to distinguish between these two headphones in a blind A/B test.