This is the thing that frustrates me on the issue of "CNPS". Passive cooling solutions with heatsinks and even heatpipes do not work as well (or at all) without airflow. TEC cooling solutions do not work well without airflow and good case exhaust. Even with water-cooling, you still need fans for the radiator, whether external or internal.
My sister-in-law's Dell Pentium III 900 system is whisper-quiet. We built her a P-4 2.8C system in a midtower case, I confess this was one of my earlier "designs", and I visited periodically to replace smaller fans with bigger fans, and add ducting solutions. She has pointed out to me how the Dell is "whisper-quiet", but I have been "inside" the Dell, and by today's standards it is a sweat-box. The P-III systems could handle a higher case-interior ambient. So in addition to the CPU HSF, there were maybe two 80mm fans spinning at speeds that could not have been above 1,500 rpm.
My own system is still due for modification, with an option of replacing two 92mm fans with either one or two 120's. Or, I can leave the 92's, but I want to try a low-volume-high-airflow solution across my motherboard, and either re-orienting the 92mm exhaust fans or replacing with 120's will make the ducting design and construction easier.
Even so, I've got rid of almost all motor whine. Although I can tune up my intake and exhaust fans to higher speeds and CFMs where there is either some slight bearing rattle or slight motor whine, it does not improve cooling to any significant degree. The system is slightly noisier than the sister-in-law's Dell, but my load CPU temps never go over 40 or 41C at room temps between 70 and 73F.
Yet the noise-haters among various users may fail to grasp the practical trade-offs in the cooling-vs-noise issue, the fans-with-water issue, the TEC with HSF issue.
I'm not going to hate myself because someone in the noiseless-computer crowd has determined that my computer goes "shshshshshhhhh" and sounds like an air-conditioning vent with the AC fan on "Low" or even "Medium".