6
\$\begingroup\$

I have a +5V/+12V dual output power supply. They share a common ground. When no load is connected to 5V, both outputs read 5V/12V by multimeter. As soon as I connect a load (LCD) to 5V line, the multimeter still reads 5V fine, but suddenly starts reading 14V instead of 12V on the other output. The voltage also visibly fluctuates between 14.1V ~ 14.5V. The 5V line does not fluctuate at all. The LCD connected to 5V probably draws less than 300mA at best, which is way below the 3A maximum given from the specification.

Is this behavior normal? Why does the 12V power supply suddenly jump to 14V?

\$\endgroup\$
12
\$\begingroup\$

It's because you are ignoring the datasheet and not attaching a minimum load to the outputs.

enter image description here

\$\endgroup\$
  • 1
    \$\begingroup\$ Wow that thought has never crossed my mind, I've never seen a power supply with minimum load. Thanks \$\endgroup\$ – Xiagua Oct 12 '16 at 2:18
  • 2
    \$\begingroup\$ Another question. The V2 (12V) terminal requires 20% of minimum load. That terminal is rated at 12V @ 1.3A, so am I right to assume that I need to throw in a load that draws at least 0.26A of current? Can I just hook up a 46 or smaller ohm resistor (12V / 0.26A = 46) to satisfy this requirement? \$\endgroup\$ – Xiagua Oct 12 '16 at 4:20
  • 2
    \$\begingroup\$ Yes, that's exactly right. That's more than 3W so you'll need something like a 5W resistor Nd it will get quite warm \$\endgroup\$ – Spehro Pefhany Oct 12 '16 at 4:32
  • 3
    \$\begingroup\$ Xiagua many non-laboratory, non-linear power supplies have a minimum load. \$\endgroup\$ – rackandboneman Oct 12 '16 at 11:42
  • 3
    \$\begingroup\$ Also true of DC-DC converters (especially the cheap ones) even single-output ones. And most PC supplies. Sometimes you can live with it going a bit out of regulation. \$\endgroup\$ – Spehro Pefhany Oct 12 '16 at 11:58
-2
\$\begingroup\$

This happens because the designers of the power supply took the cheaper option that causes more pain for their customers.

There are at least two good ways to make a dual output power supply. One is to build two seperate flyback converters (the input rectification and if-present power factor correction can be shared), the other is to build a power supply for the higher voltage and then add a buck converter for the lower voltage.

There is also a crappy but cheap way, building a converter that shares the transformer and control logic between the two outputs. This works OK under high load where output voltage is deterined mainly by duty cycle but under light load the load current plays a much greater role making it very difficult to keep both outputs in regulation.

This behaviour may be acceptable for systems where the power drawn is roughly constant but it sucks for modern digitial systems which can have a very wide variation between idle power draw and full load power draw.

Read the datasheets carefully, especially for multi-output supplies and steer clear of power supplies with high minimum load requirements.


Almost every computer PSU has minimum load.

And when haswell came along and introduced new lower-power sleep states the PSU vendors had to hastilly redesign their products.

\$\endgroup\$
  • 2
    \$\begingroup\$ steer clear of power supplies with minimum load requirements, and pay more when you don't have to? The documentation for the supply is pretty clear on the minimum load, so I don't see how you can call this "cutting corners". \$\endgroup\$ – pipe Oct 12 '16 at 13:15
  • 1
    \$\begingroup\$ Almost every computer PSU has minimum load. They work great for modern digital systems. \$\endgroup\$ – Agent_L Oct 12 '16 at 13:21

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.