Got a new 6450-48P and am looking into fan swaps. Thermals should reflect its light duty (<1Gbps most of the time, maybe 15Gbps peak; 20-30W PoE; racked in a well ventilated closet).
Hello, i received a 6450-48p some weeks ago and finally got to open it up while working on reading this thread from the beginning as well (now on p. 115)
While testing my Pabst 412/2 12V fans with a voltage range of 10 - 14V (5,9cfm, 18dB(A), 0.8W, 6000rpm),
and they stopped spinning at 4.5V as expected, i was thinking, isn't this what we want for a homelab, a completely silent switch?
As
@fohdeesha notes, most of our usage is so lightweight, thermals are most of the time (like 99%) not even close to the fan speed2 threshold.
So why not intentionally use 12V fans with a narrow voltage range, and so achieve a silent passively cooled switch that will spin up it's fans automatically if thermals reach fan speed2 threshold, but is mostly completely silent.
Is there any downside to this, do the non-spinning fans use any power if connected, but unable to spin?
mike
edit: the components are also designed to handle much higher temps sustained, so no no real issue there, and this method would save money because no need to pay extra for silent fans like noctua's, noisy high RPM enterprise grade fans would also do as long as their voltage range is around typical 10V - 14V.
edit 2: logs and terminal buffers filled with fan error warnings would be one downside?
edit 3: users who wants a constant small airflow could do with using one noctua or similar fan, and two high RPM 12V fans with 10-14V range, achieving the best of both worlds for a lower price, almost completely silent with automatic thermal protection.
edit 4: Regarding the fan error msg spam, one could create a container with support for the 2048bit ssh keys, plus some trick to strip the msg spam at bash level from reaching console and also strip them from logs, or maybe running wireshark and strip them at link level if using telnet?