COMMUNITY : FORUMS : TECH

Are high power resistors really neccessary on a benchtop PSU conversion?

Was looking through the site and I was just wondering--why all the huge resistors on the bench-top power supply conversions of PC Power supplies? From personal experience this seems like a waste of perfectly good electricity.

I know that they require some current to just stay on however in designing a project for my school's Engineering Technology department I found that the heat generated by such a small resistance (Around 10 ohms) was unacceptably high. Originally I was looking at Instructables and this sitethis site for inspiration but all the cooling measures taken to prevent the high power resistor from becoming a hazard seemed rather silly.

A few calculations and experiments later with the 250 watt power supply and I determined that 160 ohm1 watt resistors and 1K 1/2 watt resistors were perfectly acceptable for the purpose of keeping the PSU awake and functioning. I connected one of each between each voltage and ground. According to calculations I can get away with dissipating a grand total of two watts or less spread across multiple resistors.

The current divider rule dictates that if you add resistances in parallel, the resulting resistance will be smaller meaning more current will flow through the overall circuit. However this increased current will divide itself across the parallel resistances according to the rule Ix= RtIT/(Rx+Rt).

The current through and power dissipated by the resistor you've soldered into the PSU will not change enough to be significant no matter how large or small a resistance you attach in parallel with it--with the exception of an effective short and what in God's name are you doing intentionally shorting the terminals of your bench-top PSU?

Now several months later, the PSU is still operating happily and powering multiple micro-controller projects on a display board. Therefore I can reliably conclude that the high-power 10 ohm resistors in many computer power supply conversions are probably a gratuitous waste of wattage. You can get away with using a higher resistance and a resistor that dissipates much less current.

sort by: active | newest | oldest
gschoppe says: 5 years ago
westfw is right, most newer power supplies regulate very well... you might want one just in case, but it's not normally necessary.
westfw says: 5 years ago
The theory is that some significant load on the 5V supply is needed to maintain regulation on the 12V outputs. I've also heard that while this was true on older power supplies, it is no longer true on most newer supplies. I haven't checked. 10 ohms across your 5V is only 2.5 Watts; it shouldn't get THAT hot (compare to 4W night-light bulb) (but you do need to use at least a 5W resistor!) (But I'm also not sure this is enough of a load for the power supplies that need it.)
gmoon says: 5 years ago
tech-king in reply to gmoon5 years ago
if waste is a problem, you could also use an inductive ballast, as opposed to resistive. i have a 1 amp min current desktop supply. so yeah, the load does vary a lot.
gmoon in reply to tech-king5 years ago
No good, in this case. An inductor doesn't have any reactance in a DC circuit (once it's saturated at startup.) Without AC, it would work as a resistive load, since it's just a coil of wire. Depending on the resistive properties of the wire coil, it could get hot, melt, or just be a constant resistive load.
tech-king in reply to gmoon5 years ago
ops. ac/dc. my bad. i forgot about that. this make me wonder though: do wire wound resistors work on dc? i know they are not for hv applications and all, im just wondering.
gmoon in reply to tech-king5 years ago
I'm thinking most power (low value) resistors are wound-- 'cause they can use beefier wire to handle the current, and wire resistance is more accurate than formed resistance. It sort depends on the use....

If you're thinking AC (inductance and noise from the wire being a coil), the really high-wattage resistors are probably bifilar wound--two wires wrapped as a pair and connected at one end, which cancels out any inductance...
Psickattus (author) says: 5 years ago
If you've already built it, I wouldn't worry about changing anything. There are many ways to make a resistive load. My different approach was, now that I think more about it, due to 1) Me not wanting to crack open the PC power supply and 2) Not having the 10 ohm 10 watt resistor that seems to be the standard for these conversions. In an EE supply closet with some 1223 odd drawers there were no single resistors in the recommended ranges for the conversions. Either their resistance was much too low for their high wattage, or it was much too high to draw a significant amount of current. Not wanting anything to get too hot, I opted for a different approach. And as gmoon says, your PSU's minimum load(s) may vary. In my case I just split it across different voltages.
guyfrom7up says: 5 years ago
Maybe because people don't know that? I was just following tutorials and I attached the power resistor, lol. I'm not taking it off because it's glued to my case...