I spent the last few days in Vermont doing a photo shoot, riding more miles than my legs could handle, and getting a lot of testing done in ways that are very applicable in the real world. Three guys spent a significant amount of time on Rails for the first time, and it was nice to hear their responses.
In any case, my mission yesterday was to find the steepest, twistiest, meanest piece of a descent that I could find. Fortunately, the "Switchbacks" segment at the top of Smuggler's Notch was nearby, and was close to a few great coffee joints. After the previous 3 days, I needed gallons of the stuff.
The mission was simple: ride down the segment the worst way I know how, stop at a predetermined spot, and immediately take the temperature of the rims. Then, I would ride back up partway, accelerate to a given speed, and then come to a full stop at the line and remeasure the rim temperature. These two test techniques would mimic a likely way that a very timid descender would approach a section like this, and replicate a situation like coming to a stop sign at the bottom of a steep downhill pitch. The kicker? Having done that with my own bodyweight (which on the morning of the test was 158 pounds), I would then add first an additional 25 pounds and retest, and then an extra 15 pounds on top of that and retest. The kicker to the kicker? I had to get back up the segment under my own steam each time.
The short answer to the testing is that it takes a heck of a lot worse than this test dished out to overheat a Rail. The segment in question is .9 km long, with an average gradient of 11%. The maximum rim temperature I recorded was 196*f, after the descent with 25 additional pounds. The temperature after the descent with 40 additional pounds was a few degrees lower, which indicates some amount of test error, which is actually great, since the conditions in which we ride aren't that well controlled. This is why we want to have nice buffer zones. When I rode down with just my own bodyweight, the rim temp at the bottom was 155*f. As physics would have it, weight does indeed make a significant difference.
The protocol of this test was simply to maintain 10mph, under conditions when the natural inclination of the bike would have been to go 40+ mph. This was a big, but not conclusive, part of the answer to the "how close to a Rail's limit would you be under certain extreme braking scenarios." The lab-tested limit of a Rail is just shy of 350*f. So, the answer from yesterday's test was "not close at all." Even factoring in some amount of testing error given that maybe I didn't scan the absolute hottest part of the rim (which I'd have absolutely no way to identify), and given the speed at which the rims cooled down, if you tack 40* onto each scan, you're still not close. It took me about 3 seconds to measure the rim heat after stopping, while the average additional heat loss in 5 seconds of being stopped (I did an immediate scan and a 5 second delay scan each time) was over 15* - and that's with no air cooling.
Significantly, these tests would have destroyed most previous-generation rims on the market, and would have killed plenty of current rims on the market. Certainly, it would have at least been right there with Kenny Loggins in the danger zone for a broad swath of rims that haven't taken advantage of (and suffered the additional cost of) the latest resin tech.
But the biggest and most immediate thing I learned was how devastatingly difficult that much additional weight makes climbing hills on a bike. What was a reasonably challenging experience at my bodyweight turned into a notable struggle with 25 extra pounds, and turned into a walk with 40 extra pounds.
Please note that this is not "the authoritative blog about brake heat under real world testing conditions." It's a big, and very reassuring part of it, but it's not it. Please also note that this doesn't mean that there isn't some other fuse that blows before your rims do - tires being the main suspect. Learn good descending and braking technique, and force yourself to use it.