# How efficient are lab bench power supplies? (affected by output voltage?)

As how I understand it, there are two ways of converting DC. One is by using a voltage regulator (zener diode, feedback combination etc), and the other is by using a buck boost converter. Firstly, if there are other methods, please tell me!

How do lab bench power supplies work? I've heard that buck-boost supplies are noisy (for precise lab requirements) and I know that voltage regulators basically drop the remaining voltage as wasted energy.

Now to my question: I have only seen a single transformer in any lab supply, so obviously the output voltage of it is fixed. If say the bench supply's max voltage output is 24V, then I'm assuming the output of the transformer would be ~24+V.

If this is going (after AC-DC conversion) to a voltage regulator, and say I pick 1V fixed o/p voltage and connect the bench supply to a load that draws 1A current. If I understand it correctly) the lab supply will waste at least (P=VI=[24-1]*1) 23W and have an efficiency of just 4%.

Is this true? If so, the heat output would increase dramatically as selected output voltage is lowered (and current increased). So much energy would be wasted? Not to mention heat sinking etc.

What am I missing? Is there a better way? Or is this a necessary sacrifice for getting a precise power supply?

• Most lab power supplies are switch-mode power supplies (which is the general category buck- and boost-converters fall into). Another example of a switch-mode supply is a flyback converter. Look into SMPS (Switch-Mode Power Supplies) for more information. – Shamtam Nov 6 '13 at 3:53
• There are lots of arrangements, a couple of mine are switching and another is linear - while that only has a single transformer a relay switches the secondary at about half the voltage range. – PeterJ Nov 6 '13 at 4:00

Many supplies nowadays use both a switching regulator and a linear regulator together. The two are set up to track each other, so the voltage dropped by the linear regulator is never excessive.
So for example, say you set the supply to 5V, you might have a 24V input which the switching regulator drops down to 6-7V, then the linear regulator drops that down to the desired 5V. This way you get the benefit of the switchers efficiency with the quiet output of the linear regulator.

However, as mentioned by Passerby, efficiency and weight is not always so much of a concern compared to reliability and quietness, so many simply use a purely linear supply and deal with the losses involved.
To keep from being ridiculously inefficient however, generally many linear supplies will have more than one tap on the transformer which are switched between at appropriate points in the range (e.g. instead of dropping from 24V to 1V, it would switch to, say, a 5V tap so a lot less power is dissipated) You can usually hear the clicks of the relays at certain points in the voltage range.

Efficiency of a lab power supply is found by reading the spec sheet for said lab power supply. And efficiency isn't too much of a concern in a multi-purpose device like a lab supply, because it is not a consumer or battery operated product where such efficiency is important to cut costs or produce something marketable. In a lab supply, even extended use, necessary cooling can be determined ahead of time because the variables are known. X amps, Y voltage range, means they know the maximum heat dissipation to expect.

This is too open ended otherwise.

It really depends on the power supply. Lab supplies are generally designed to have very good regulation on current and voltage as well as low output noise. Generally efficiency is not much of a concern unless it's a high output power supply. In that case, low output voltages at high currents have the potential to dissipate a huge amount of power in the regulators, which is something that should be avoided as much as possible. Some power supplies will have selectable ranges, so you can decide whether you need a higher current with a lower voltage or a higher voltage with a lower current.

I have a couple of Agilent power supplies that have a low range of 8v and 20a and a high range of 20v and 10a. They have a gigantic transformer with a pair of secondaries. It might be one tapped secondary; it's been a while since I had the cover off. The supply can put the two secondaries in parallel for the low voltage range and in series for the high voltage range. Then the regulator has two stages - a PWM'd preregulator and a linear postregulator. So it's sort of a hybrid switching and linear power supply.

This likely won't be the case for all power supplies, though. Agilent supplies will be highly engineered for the best performance (and you pay for that) while el cheapo made in china supplies will likely be very simple and purely linear. And then there will be a whole range in between.

• Most cheapo lab supplies that aren't SMPS based use a tap-switching topology, with a bunch of relays. The mains transformer has output taps every ~5V, and they switch taps so you're only ever dropping <5V in the linear regulator. It's a clever design: The low-noise performance of a purely-linear PS, and the actual dissipation is still kept pretty low. – Connor Wolf Nov 6 '13 at 9:02
• I think your definition of cheapo and my definition of cheapo do not agree. The cheapo ($70) supply I have has no relays in it at all. It's a no-name chinese brand, single output, 30 volt, 5 amp unit. I also have a BK precision triple output supply ($300) that does produce a relay click for transformer tap selection when the voltage selection exceeds a certain point. I can't remember what point offhand, though. I also used a piece of crap Agilent linear supply that didn't have any switching either. Probably becuase the current limit was like 150 mA. – alex.forencich Nov 7 '13 at 4:24
• Beats me, but my "literally cheapest model available on amazon" $199 dual-output bench supply has something like 4-5 output taps internally. It's one of those Mastech HY3005F-3 dual 0-30V, 0-5A dealies. You can really hear the relays click over as you adjust the output. – Connor Wolf Nov 7 '13 at 5:55 • I really needed the dual outputs, so It's not quite in the same ~$70 ball-park, but the single-channel version of that supply also has switched taps, and is ~\$100.. – Connor Wolf Nov 7 '13 at 5:57