I'm looking for the most adequate power supply that can go between ~0-30v and ~0-5A. These are going to be used for software development for automotive infotainment systems and the operation specs for the current one that we are developing goes from 7-18v at 1.4 A but could vary slightly in the future.

Beyond those specs, what other specs are good comparison points to tell how good is a power supply? (i.e. stable & clean power). Do brands matter?, if so, why and how?.

  • \$\begingroup\$ "The best" is not wxactly the most useful indication... \$\endgroup\$ – clabacchio Mar 26 '12 at 15:25
  • 1
    \$\begingroup\$ Shopping questions are off-topic. \$\endgroup\$ – stevenvh Mar 26 '12 at 15:59
  • \$\begingroup\$ Should I go to SuperUser then?, can my question be moved?, or any recomendations of where/how to ask about power supplies? \$\endgroup\$ – dukeofgaming Mar 26 '12 at 16:17
  • \$\begingroup\$ @dukeofgaming, Sorry to say shopping is off-topic on all sites, but that does not mean the definition does not vary from site to site. The question can be moved if it is a good question for the site, but I dont believe this one is. Asking for brand advice and such is inherently going to vary with time and not provide lasting useful information for the populace. If you are just wanting to buy a desktop power supply, you might find someone that makes test equipment does a good job for you. Agilent or such. \$\endgroup\$ – Kortuk Mar 26 '12 at 16:28
  • 1
    \$\begingroup\$ What could be an on-topic question is asking about what to look for when shopping for a power supply. See this blog post for more info. \$\endgroup\$ – Kellenjb Mar 26 '12 at 16:47

First there's the obvious, like good line and load regulation. Line regulation tells you how the output voltage varies with varying input voltage. Load regulation shows how the output voltage varies is you vary the load. If you use an integrated regulator like the typical LM317 you should have nothing to worry about.
Ripple rejection tells you how input ripple (like from the smoothed rectifier voltage) passes through the regulator and shows at the output.

Do you need to go as low as 0V? Most regulators start at a bit higher voltage, like 1.2V for the LM317. The datasheet shows an application to start from 0V, but notes that full output current are not available at high input-output voltages.

Then there's power management. If you build a linear regulator like with the LM317 internal dissipation can be high. A 0-30V supply with 0-5A output dissipates at least 150W if you need the 5A at very low voltages. One way to circumvent this is to use a transformer with several taps, where the control logic switches to a lower transformer output voltage is less output is required.
Related is short-circuit current. A classical power supply will just limit the current to the 5A, giving the high dissipation I mentioned. This high current may damage your circuit too. Solution is a foldback current limiter, which decreases the output current when a short-circuit has occurred.

enter image description here

Switchers (SMPS, short for Switched-Mode Power Supply) are much more efficient than linear regulators and are more and more used in circuit, now that advanced controllers have simplified their design. But switchers are best at a given input and output voltage and output current. Their design for wide input voltage/output current range is more complicated. You can find them commercially, however, and you can recognize them by their light weight.

  • \$\begingroup\$ I think the OP is asking for a benchtop supply, not an integrated one...but I may be wrong \$\endgroup\$ – clabacchio Mar 27 '12 at 7:09
  • \$\begingroup\$ @clabacchio - His original question was a shopping question. Even if he's still not interested in designing a benchtop power supply himself (which would make it a perfect question for this site) the parameters are still relevant. \$\endgroup\$ – stevenvh Mar 27 '12 at 7:15
  • \$\begingroup\$ Yes, but how are they evaluated it's different, because I think that benchtop supplies don't have only a LDO inside..anyway I appreciate the content, just suggesting to think about it in a benchtop form...also Kortuk intended it that way \$\endgroup\$ – clabacchio Mar 27 '12 at 7:19

Just a small contribution: since you have to test software for automotive, I expect that you don't have analog signals to process. So you can save in money and weight relaxing a bit the accuracy and the noise filtering.

If you want to automate a bit your measurement system, consider if you want a interface such GPIB (old), RS-232, USB or Ethernet that gives you the ability to control dinamically the output voltage. For instance, you may want to test your circuit over a range of voltages to check that it works properly in all situations.

And consider this: 30 V * 5 A = 150 W. If you don't plan to have both the maximum ratings at the same time, you may go for a lower power supply. For instance, your max expected consumption now is 18 V * 1.4 A = 25.2 W. If you don't expect big variations, a 50 W supply may be more than enough.


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.