Very nice review MikeC as well as great to see that the new chamber is working nicelyâ€¦
But there is a bit of an issue. When doing thermal testing -- like PSU load testing -- the chamber gets warmer. A full cycle test -- from 20W to max load -- takes some 3~4 hours for a 700W model. In the last 30~40 minutes, some 500~1000W of heat is being pumped into the room. The resulting rise in ambient air temperature can affect the noise-to-power curve, making it steeper than it might be otherwise.
This rise in the ambient temp around the test loader also happened in the previous kitchen lab, but because there were open doorways to three other rooms/areas, the ventilation was good enough to keep the rise to maybe 2C max from start to finish, which seemed perfectly acceptable to me. Some real world applications would actually see much high temp rise than that if a system was pushed that hard -- imagine the temp under your desk after several hours of 3D gaming with PC containing a pair of 4870x2 + uber-quadcore .
The chamber is more enclosed with just a single door. And tho it's left wide open except when acoustic measurements or recordings are made, the heat buildup in there is more acute. During the DA700 testing, the rise was higher than 3C by the 400W load testing. The testing had to be stopped to pull the air temp back down to keep the test conditions stable (and comparative to previous tests). I think the exposed blue batting may also absorb and retain more heat than ordinary drywall, because even with a large box fan blowing hard from the doorway, it took some time to drop the temps by 2C.
I'm in the midst of another PSU review now (yeah doing them in a batch), and I'm playing around with hanging the box fan at the top of the doorway and keeping it blowing out at the lowest setting, starting at the 150W power test level. The position and airflow direction would take advantage of convection; hopefully cooler air from the kitchen lab would flow in through the bottom of the doorway. The main thing is to keep the heat rise in the room to 2C or lower.
FYI, a new automated PC-based PSU load tester has been in the works for a year. This was initiated as a project last fall by electrical/computer engineering students at Univ of BC. It counts as a course or two. The mechanical design is mine, the rest they came up with based on my requirements; I am their "client". They were supposed to have finished before the summer.... but there were parts supply issues and they're still testing / validating now.
When that thing is done, the testing procedure will be become much more streamlined, with all the electrical data collection done automatically. My guess is that it will cut the testing time by at least half. That might make it worthwhile to add another twist to the testing: Simulation of operation in a case that has a separate PSU chamber -- like the P180 series, Fusion -- and a handful of others. This would be only for the acoustics / thermals and hopefully, go quick. I'm itching to have that tester in here!