Getting 7v from 5v and 12v power supplies

[ratatax]

This is quite a simple question, I'm wondering if doing this is safe:

I have a 12V DC power supply that feeds into a 5V DC/DC converter, with a common ground. I want to power a fan with 7V, so I connect its positive terminal to the +12V and its negative terminal to the +5V output of the converter.

Could this damage or put increased strain on the DC/DC converter or the 12V power supply ?

This question was often asked by PC builders who wants to undervolt their fans - never getting precise anwser except that its "not recommended but it will work".  What do you think?

[Manul]

It is ok, if DC/DC converter can sink current. Some of them can. Or if it is guaranteed, that the current on 5v rail is higher then 7v current at all times.

If none of these two conditions are true, then there is a problem, 5v output will rise up with possible consequencies like destroying electronics powered by this 5v rail.

[magic]

I suppose most ATX PSUs don't sink current so the trick relies on 5V loads sinking all the current from the fan(s).

This typically works, but if anything goes wrong such as too many fans and not enough power draw by the motherboard or one of the fans develops a short circuit, then the motherboard is kinda screwed.

[InsertUsernameHere]

Would a 1W resistor or potentiometer ( to adjust fan speed) not be safer?

Example (https://www.blackfiveservices.co.uk/fanspeed.shtml):

Original fan voltage: 12V

Original fan current: 0.2A (Read from sticker on fan)

    V=IR (Ohm's Law), so 12 = 0.2R   =>   R = 12 / 0.2   =   60

Fan resistance: 60 Ohms

Target voltage: 7V

    V=IR (Ohm's Law), so 7 = 60I   =>   I = 7 / 60  =  0.12 (2 d.p.)

Target current: 0.12A

Voltage drop across resistor: 12V - 7V  =  5V

    V=IR (Ohm's Law), so 5 = 0.12R    =>    R = 5 / 0.12   =  42

Target resistance:  42 Ohms

Power dissipated by resistor:  5V * 0.12A  =  0.6W

[Manul]

Voltage drop resistor is a poor choice to control motor speed. It works to some extent, but basicaly it reduces torque, not RPM. If you lower torque too much it may not be able to start at all from a stand still. Yet if you give a kick start, it will still develop quite high rpm. And its RPM will be affected strongly by friction losses, wear and load.

[ratatax]

I'm seriously considering the resistor option, the load on the 5V rail will be way more than the fan's consumption but at startup i'm not sure it may take a few ms to start drawing current (it's a raspberry Pi on the 5V rail).

The fan is only rated 60mA at 12V so the resistor doesn't have to be very big !

Voltage drop resistor is a poor choice to control motor speed. It works to some extent, but basicaly it reduces torque, not RPM. If you lower torque too much it may not be able to start at all from a stand still. Yet if you give a kick start, it will still develop quite high rpm. And its RPM will be affected strongly by friction losses, wear and load.

It's one of those CPU fans, with integrated brushless control circuitry, RPM always seemed to be proportional to voltage from what I've seen but I haven't tested lowering the voltage with a resistor...

[Manul]

The problem is, that voltage drop over resistor is proportional to current. Yes, it works to some extent, everyone is free to do it. But there is a common misunderstanding about this method, which I tried to explain.

[madires]

If the fan has a 4-pin connector use the PWM input to control the speed.

[schmitt trigger]

Hopefully tomorrow I will be able to take some fan current waveform.

From what I remember it is not constant but increases and decreases as the fan turns.
The plate current is the average of the peaks and valleys.

First I have to find a fan on my junk box 🙂

[schmitt trigger]

I found a pair of fans.

The first one is 24 volt, about 8 cm square
See the attached current waveform. The average current consumption is 60.5 mA, but it has peaks reaching 120 mA. But it has narrow peaks of -76 mA.

The second one is 12 volt, about 4 cm square, and rotating faster.
You can see that the average current is 76.3 mA, but it has a maximum of 98 mA and peaks of -26 mA.

In both instances, if I slowed down the fan with my finger, the period would increase. Meaning as I had mentioned previously, that the current changes with the rotation.