1
\$\begingroup\$

I have an APC baseboard that runs off 24v.

The two fans (2 x 24v @ 0.16A = 7.7W) were a noisy and power-hungry problem. I've got a couple of Noctua fans of the same size, with "almost the same" or "plenty good enough" static pressure and airflow, that take just 12v @ 0.06A (1.4W) and are almost silent. The fan models are AD0824HS-A71GL and Noctua NF-R8 Redux-1800-PWM if it helps, and the case is in a cool environment under low use, where heat is less of an issue than in a datacentre.

Naively I could put the fans in series (2 x 12v = 24v), but as answers to several other questions point out, this is a mistaken approach as they are inductive in nature.

Given that the PCB and the new fans are both well engineered, I'm looking for the cheapest and simplest way to power these fans from the original 24v pins. It doesn't need to be perfect, but I'd like to avoid a constant 1.4W wasted power drain of using 2 x 200R series resistor, one on each, (or 1 x 100R common to both), or constant loss through a zener diode.

What is my simplest/cheapest solution?

\$\endgroup\$
1
\$\begingroup\$

If the two fans have the same Power ratings they will split the voltage equally.

ADDA AD0824HS-A71GL 24V 0.16A 3.84W

If not equal then the lower power one will run at higher V and RPM until it’s incremental impedance matches the other.

Miniature BLDC Fan motors are designed more like controlled voltage/current drivers wto look like a resistor. They have very low starting torque as a result but torque rises with current and Power out matches the DC power square law consumption. They use precision Hall sensors to trigger each phase of the commutation and use PWM to reduce the average current as a ratio of the applied DC voltage.

i.e. I is proportional to V.

Hence a BLDC fan motor behaves like a resistor.

The current noise is LC filtered to reduce it somewhat. So the input Voltage splitting or divider ratio depends on the fans being matched for power or voltage and current specs.

enter image description here enter image description here

Note the linear slope of current with horizontal voltage and the f(x) quadratic equation has a 0 linear term for x.


I have done this successfully with 2 matched 24V fans on 48V supply with temp controlled variable LM317 current limiter using transistor to Vadj. This was for high volume production.

I found that the linear current speed control method was more stable than using PWM for speed control here due to aliasing.

Here are some other examples. enter image description here

\$\endgroup\$
  • \$\begingroup\$ To make the fans equalize better (the first to spin up has a larger voltage drop, delaying the second), I'd add a resistor of ~500 Ω across each fan. It would waste a bit of power as extra heat, though. \$\endgroup\$ – DrMoishe Pippik Jul 23 '18 at 21:02
  • \$\begingroup\$ If you have a proper plenum design, the air flow guarantees if one starts the other starts at the same time as stiction is overcome by forced air from the other. Worked in production very well. 1k systems/mo \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Aug 23 '18 at 6:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.