So the SRAM guys sent me out a PF30 BB to play with (see the first post on this topic here) and I have to say: so far, I don't see that there's much point (at least for mountain bikes).
Note that I am decidedly NOT just a hater of all things new. I think the new direct mount front derailleurs kick ass. I like through axles even for XC bikes and am working on getting set up to build around tapered steerer tubes. I like it when stuff gets better, because it means that mountain biking can be even more fun for all of us (the question of stuff always getting more expensive is another story).
To whit: the claimed advantages of the system are 1) Weight and 2) Stiffness. There's also some stuff about improved bearing life, but given that you can buy an external cup King BB with a 10 year warranty for around $120 retail, I don't see the point in worrying about longevity comparisons. Existing systems last quite well.
Now I can't speak to the stiffness claim, but I submit that *crank/BB systems are plenty stiff already*. Sure, you can make something deflect less in the lab by making it bigger, but we reached the point long ago where no human being can really detect crank flex unless they're a champion track sprinter or something. And heck, some of the track champs probably don't notice or care either. I can't personally even detect significant flex in a square taper setup, for heaven's sake, and I'm a (supposedly) pro cyclist.
As an aside, I'd like to see stiffness reported in absolute numbers - ie, we attach the cranks to some rigid test structure, subject them to typical cycling forces, and measure how much the crankarms and chainrings deflect - in millimeters, or fractions thereof, from their intended positions. It's not that useful to say something is "50% stiffer!" when in fact that means it defects .001mm instead of .002mm - there's a threshold at which nobody cares anymore. But instead every year we hear how much stiffer everything has gotten - without any concrete deflection numbers to make the information meaningful.
There's also a claimed improvement in ankle clearance for the PF30/BB30 (though not stance width, which they call "q factor" in their literature), but as a *massively* duckfooted cyclist (I wear the paint off chainstays) I don't find that ankle/crank clearance has ever been a big problem, so I can't imagine it matters to more than a percent or two of the cycling population, and even to them not a great deal.
So really, since the PF30 system is significantly more expensive and a moderate pain the butt to build with (though certainly not as bad as BB30, which is just ridiculous), there had better be a significant weight advantage. But I just don't see it, after inspecting and weighing a few bits and pieces.
The BB itself is 88g. Not too shabby, but a generic SLX-level Shimano external bearing BB is around 90g (it depends a bit on how many spacers you need, but it's basically 90g). So 2g. But the PF30 shell weighs 125g, as opposed to the 100g of a plain-jane english threaded shell from Paragon. So thus far, the PF30 setup is 23g behind. And that's not to mention that if we used a nice King BB, we'd save another 20g for the traditional setup, albeit at significant cost.
Yes, you'd miter all the tubes a tiny bit (about 1/4") shorter with the PF30 shell, which helps a bit, but it's not that significant. About 15-20g worth of tubing, if you're building with steel, and depending a bit on what kind of tubes you use. So call it dead even so far.
I don't have a set of PF30/BB30 cranks in my grubby paws to weigh against their standard counterparts, and the best info I can find is this article in Bicycling, which seems to indicate that the BB30 cranks are a bit *heavier*. I'd be hard pressed to figure out a way that the PF30 crank (assuming the chainrings and bolts are the same) could shave off more than a few grams, since the only difference is in the spindle, though SRAM claims their XX crankset is 60g lighter (including BB) for PF30.
So there's a bit of confusion, but my impression is that when all is said and done, the weight is a wash. At best, you might save a couple of grams. At worst, you've actually got a heavier system than a conventional one.
And of course the PF30 is a bunch of extra work - you've got to have a setup to adapt it to the fixture, you've got to have a 2" hole saw (turned down a tiny bit on the lathe), and you've got to get set up to machine the inside of the shell to size after you're done building the frame. It doesn't require the same crazy tolerances as the BB30 standard, but it's still a lot of extra work.
Bottom line: I don't see the point. So I will probably not even bother to get a shell and build anything. Memo to SRAM: if it ain't broke, don't fix it. You guys have done lots of kickass stuff over the years, this BB30 and PF30 stuff isn't, at least so far.
Of course, I could certainly be wrong about a lot of what I've said. Maybe the stiffness improvement is night and day, and when I finally ride a PF30 bike, I'll refuse to go back. But I doubt it.