To put it short, it was not really worth the trouble, especially now in 2010 when you can have quiet 180mm PC fans from Silverstone and similar.
But as I went to all the trouble to get the fans and test them, here's the longer report.
First of all, 172/50mm industrial fans come in three basic classes, around 205, 240 and 280 CFM, with RPM of 2850/3400/4000 for PAPST. I tried to hunt the lowest speed ones, but the middle pack ones were the only ones I could get cheaply. Buying new ones you could naturally get whatever, but they cost over hundred euros per unit, so I did not even consider them.
Now, after some hunting I ended up with 2 Comair Rotrons and 3 PAPSTs. Comair Rotrons are just too loud, period. The PAPSTs are sorta promising, but not ultimate solution for a silencer.
Here's a photo showing the label of the tested unit, along with your standard Nexus 120mm fan:
And here you can see that the PAPST is indeed twice as thick:
The PAPSTs respond very well to undervolting, and once the fan had started, I could turn the dial all the way down to about 4.2 volts and it still ran just fine. This is not bad at all for a fan that has nominal voltage of 24 and that's rated to withstand voltages up to 28.
Because I lack fancy SPL measurement tools, I just put the two fans side by side, Nexus at 12V and tuned the PAPST beast voltage until their level of noise was equal. It was very much a judgment call, because the noise signatures were totally different. While Nexus exhibited mostly broadband turbulence noise, the PAPST unit was like an smooth engine idling a block away.. chunk-chunk-chunk, as the motor hit the next coil on it's way round. Throughout the whole useful range (4-12V) for quiet computing motor noise formed the main part of the acoustic signature.
I ended up calling 4.66V the point of equality, but as said, this was very hard to determine, due to differences in signatures. At this speed the PAPST did not seem to move much air; Nexus at 12V and similar noise level most likely beat it. It's hard to say for sure without measuring it, because 172mm naturally has a lot more area to move the air through.
So where and why would you consider using a fan like the PAPST tested here? The first advantage of the PAPST is a lot higher CFM rating, even at half the voltage, 12V, and it's still not totally ear-drum breaking at that speed. With intelligent fan controller you could move some air with low noise and then get huge amount of airflow if you need it.
The another advantage is pressure. Because PAPST is twice as deep, I would assume the CFM drops a lot slower than for the Nexus, so if you have a system with high back pressure (tightly packed case, radiator or especially two radiators sandwitched or similar) the actual CFM figure per noise level can be competitive.
The third advantage is these fans are, well, industrial. They are build to last, even under extreme conditions. They are designed to cool airplanes, industrial lasers, rack-mount chassis and so on. PAPST promises 4 year lifespan for a new fan even under the maximum operating temperature of 72C.
So all in all, it was interesting, but not really something everyone should rush to do. However, the results are not totally crushing, so I will later test the fans on large radiator and compare cooling results at each noise point, because the proof is in the results. Silverstone FN181s will hopefully be in the mix by then, too.
And oh, the last advantage.. these monsters are damn cool! They are not just designed to be solid, they also feel and look very heavy-duty. Any PC fan feels just like plastic toy after handling these beauties.