Computing stuff tied to the physical world

Meet the Manson NSP 3630

In Hardware on Apr 30, 2013 at 00:01

It’s always nice to explore more equipment, to see how it behaves in the lab. I already have a very nice dual-voltage lab power supply, but this one is interesting due to its relatively low cost (€90 + shipping), fairly high power, and very convenient small size:


It’s the Manson NSP 3630, available from Reichelt in Germany by mail-order. At 27 x 15 x 7 cm, it really should easily fit in most home labs. With many thanks to David Menting for letting me play with it a bit and report some first impressions.

The display is quite simple, yet still nicely readable:


Power consumption, without any load is 0.07W when turned off with the power switch on the back of the unit, and about 2.0W when turned on, without load.

There are two (optical encoder) rotary knobs to adjust the maximum voltage and current, respecively, over a range of 0.7 .. 36V and 0 .. 3A – i.e. plenty for most situations.

A drawback with this particular supply, is that it’s a bit of a hack to pre-set the current limit. You can turn it down during use to reduce the limit to the point where the voltage starts dropping, but if you want to set it up ahead of time, then the way to do it is to short out the power supply while adjusting the knob, to see the value it displays.

This is a switched-mode power supply, which explains why it is so small and requires only a small fan, but it does lead to some residual noise on the output. Hooking it up to the scope, still under no-load conditions, you can see that the output varies between about +10 mV and -20 mV of “ripple” above and under the preset value, respectively (5V in this case):


To produce this image, the scope was set to “peak-detect” mode, which captures the high and low value at each point in time, and the trace is drawn in “envelope” mode, which is a variation of persistent display showing the most extreme values ever reached as long as the scope is kept on. The last trace is the one shown in the middle, while the top and bottom lines are the largest variations ever reached – I think I left it running for a minute or so.

It’s a pretty good result, actually, for a switching power supply. Variations of this kind should not cause any problems for most digital circuits, which usually can tolerate fairly large variations on the power supply lines.

  1. Is there a fixed relationship between the current setting and the knob rotation? Could you just print out a label for the current knob (after suitable calibration), or is it multiturn, or an optical encoder with no end-stop?

  2. The latter. Many turns are required to traverse the entire range.

  3. Very interesting, I wanted to buy such a device yesterday from Conrad under the label Voltcraft 1-20V 5 A (cost 150 €) I decided against, because it was too expensive.

    When you measured the noise, did you reduce the bandwidth ? Often the power supply noise is measured with a bandwidth limitation of 20 Mhz and the signal looks way better then.

    • Yes, bw-limited, see the “BW” tag in the lower left corner on the scope screenshot.

    • Sorry for asking I should have known better.

      I really fell in love with the portability of this little power supplies. I hope we still have one in the school lab so I can test it tonight. I bought an overload from itead, so I can test it under load.

      Cheers Rubi

  4. I also could get my hands on a supply to test it, I uploaded some screenshots to my flicker account.

    Interestingly there is little noise difference when you hook up a load to the supply.

    Cheers Rubi

Comments are closed.