Toroidal Linear vs Switch Mode Power Supply

Discussion in 'Amps [BG]' started by illusha, Jan 31, 2011.


  1. illusha

    illusha

    Joined:
    Jan 18, 2011
    Location:
    Orygun
    ok, I still haven't purchased the new amps because I'm still stuck on this whole "Toroidal vs SMPS" dilemma... we're talking about a large PA sound system, 8 hornloaded subs, 8 dual-15s and 4 single-15's, running at full blast for 8+ hrs at a time...

    built for DJ and EDM music to be played at indoor and outdoor venues, sometimes dirty warehouse power, sometimes huge genset split up through spider boxes, sometimes a nice club with proper wiring, gotta be ready for everything...

    the big question (still) is - toroidal vs switching power supplies - and how do they perform under sagging line conditions... there are several schools of thought, some people believe in one technology, some swear by the other technology, and some are on the fence like me and just don't know which way to jump... wouldn't mind lightening up my racks, but don't want to do that at the expense of sound quality or volume...

    I realize that amps don't work at 100% duty-cycle at all times, in fact, they don't at most times... however, what happens when there is that 30-second bass drop that the DJ eq'ed up ??? is it better to have a heavy-ass ol-school toroidal transformer that is storing all that current just waiting to be released - OR - is it better to have a light-weight new-age switcher efficiently drawing consistent power from the source ???

    any other thoughts on the matter or useful links to other discussions?
  2. greenboy

    greenboy

    Joined:
    Dec 18, 2000
    Location:
    remote mountain cabin Montana
    Disclosures:
    greenboy designs: fEARful, bassic, dually, crazy88 etc
    It comes down to specific models/designs, and not any one "school of thought", which more often than not turns out to be a "school of unthought" anyway.

    Ferinstance, my horrible experiences with multiple Stewart 1.2 would have soured me on lightweight amps had it not been for the fact that QSC Powerlight amps weren't in the same system pulverizing the joint.
  3. Balog

    Balog Supporting Member

    Joined:
    Mar 19, 2009
    Location:
    Bothell, WA
    Transformers do not store current.

    Pink noise is only a %50 duty cycle, I believe.

    Perhaps try a rental of new amps to try smps out.
  4. seamonkey

    seamonkey

    Joined:
    Aug 6, 2004
    Look for Universal Regulated Power Supplies that can work on a wide range of powerline voltages. Look for power supplied that don't have a huge power drain when turned on. Look for lightweight and low heat - cause when you get a rack full they will be heavy and hot. You're going to find SMPS fits your need. They are nothing new. Used in computers and medical gear for years.
  5. Register to disable this ad
  6. Jerrold Tiers

    Jerrold Tiers

    Joined:
    Nov 14, 2003
    Location:
    St Louis
    Computers and medical gear are a different sort of load.......... usually pretty predictable, although computers have a wide swing in loading. An amp may have a 500:1 change of load, at a relatively slow rate, a computer probably isn't quite that much.....but changes faster, with tighter specs.....

    That said, what works well at 250W of computer may not be as good at 1800W of amp....

    SMPS *may* be pretty stable in output over line variations, or they *may* be "linked" so that they fall as the line does.... SMPS may be set up to draw good sine wave power, or they may not. it just depends.

    A transformer and rectifier type supply (or the supply in some SMPS) is a horrible load on the mains.... huge current peaks, low "power factor", meaning you can actually get LESS power from the line than with a good SMPS.

    I expect most companies have it fairly well together with SMPS by now...... early ones had "issues".
  7. rodl2005

    rodl2005

    Joined:
    Jun 1, 2005
    Location:
    Tasmania, Australia
    Were the 'issues' really gutless sounding PA's?? Thats all I've heard re SMPS or even digital amp type PA's my old early early 90's Dynacord cruds all over damn near anything in it's NEAR power range... & sound quality.

    Show me A good SOUNDING Lite weight PA/amp system & I'll gladly eat my words... I have to carry it every week!!
  8. Bob Lee (QSC)

    Bob Lee (QSC) In case you missed it, I work for QSC Audio!

    Joined:
    Jul 3, 2001
    Location:
    Costa Mesa, Calif.
    Disclosures:
    Applications Engineer, QSC Audio
    You mean "toroidal non-linear" vs. switch-mode (also non-linear) power supply.

    Most power amp supplies are unregulated, so sagging line conditions mean that the power supply rail voltages sag by about the same proportion. For example, a 10% line sag would mean the supply rails sag about 10%, the point of clipping drops about 10%, and so the maximum output power before clipping would drop about 19%.

    And as Balog said, transformers don't store current.
  9. illusha

    illusha

    Joined:
    Jan 18, 2011
    Location:
    Orygun
    I know that transformers don't store current, what I meant was I've been under the impression that capacitors (that do store current) are differently designed in toroidal vs switching power supplies... and as far as "linear" vs "non-linear", I guess I didn't quite get the difference, care to enlighten me?

    ^^^ Jerrold, could you elaborate on that a little bit further? all 'dem *may's*...

    thanks for your input guys... so with SMPS, would it be much help to add a voltage regulator (such as TrippLite LCR2400) ??? any recommendations for SMPS amps that are known good performers?
  10. Bob Lee (QSC)

    Bob Lee (QSC) In case you missed it, I work for QSC Audio!

    Joined:
    Jul 3, 2001
    Location:
    Costa Mesa, Calif.
    Disclosures:
    Applications Engineer, QSC Audio
    OK, neither conventional line-frequency power supplies (with toroidal, EI, or other forms of transformer) nor switch-mode (high-frequency) power supplies are linear.
  11. Bob Lee (QSC)

    Bob Lee (QSC) In case you missed it, I work for QSC Audio!

    Joined:
    Jul 3, 2001
    Location:
    Costa Mesa, Calif.
    Disclosures:
    Applications Engineer, QSC Audio
    A capacitor stores energy (not current; current is the rate of electron flow, which would be a concept opposite from storage) as electrical charge. The amount of energy stored is proportional to the capacitance and to the square of the voltage. In power supplies, they're used as reservoirs to smooth out pulsed DC (from rectified AC) into fairly stable DC voltages.

    In a line-frequency power supply, the transformer operates at the frequency of the AC mains (either 50 or 60 Hz in nearly all cases), and the secondary voltages are rectified and fed into capacitor reservoirs on each supply rail. Typically, the caps are charged at 2× the transformer frequency, so the caps have to be sized to maintain the voltage reasonably while handling the energy demand and being recharged only every 8 to 10 milliseconds.

    A high-frequency power supply handles (at least) two frequencies: the AC mains and the transformer frequency. The incoming AC is rectified into DC and stored in a capacitive reservoir, which is switched at the higher freq for the transformer, and then the secondary voltages are rectified and stored in reservoir caps for the supply rails. The first reservoir is typically recharged at 2× the mains frequency, so the caps are sized to hold their voltage reasonably for the energy draw demanded of them while being recharged every 8 to 10 milliseconds. The DC energy is switched through the primary of a power transformer at a higher frequency (often anywhere from 50 to 150 kHz, depending on the design). Just like in the line-frequency supply, the reservoir caps on the rectified secondaries have to be sized and selected to maintain voltage taking into account the energy demanded and the recharge rate. But in the SMPS that recharge interval is measured in microseconds, not milliseconds.
  12. illusha

    illusha

    Joined:
    Jan 18, 2011
    Location:
    Orygun
    wow, BIG THANKS going out to Bob Lee for such a detailed explanation - I appreciate it... that's EXACTLY the kind of scientific info I was looking for - trying to understand the real-world differences between the two PS designs... would you care to comment regarding advantages and disadvantages of each of those 2 designs?

    and what is your opinion on the new-age SMPS amps? are they going to perform as expected across all frequencies? is there any advantage at all to the toroidal transformers?

    * side note - I gotta work on getting my terminology straigt =)
  13. 12bass

    12bass

    Joined:
    Jan 2, 2003
    Location:
    Saskatoon, Canada
    Cool thread!
  14. christw

    christw Get low!

    Joined:
    May 11, 2008
    Location:
    Dayton OH
    Disclosures:
    Endorsing: J Worrell Bass
    +1

    The engineer in me enjoyed this thread.
  15. Jerrold Tiers

    Jerrold Tiers

    Joined:
    Nov 14, 2003
    Location:
    St Louis
    Actually, I'm going to defend the "stores current" idea....... if you can stand a bit of techie-talk

    BOTH a transformer (and definitely a purpose-made inductor) and a capacitor store energy..... but they do it in different ways.

    A 'current" is moving electrical charge.... electrons on the move.

    An inductor DOES store current....... you "charge it up" by applying a voltage to it, and current through it increases.....If you then disconnect from teh voltage, and connect it to a load, the current continues to flow. Initially the same current flows, then it decreases with time as the energy is depleted by the load. The inductance "stores current" in that rather direct sense.

    Phase linear made amplifiers with a low frequency power supply based on that idea, and almost every SMPS ever made uses the principle.

    In fact, a so-called "flyback" SMPS uses a transformer that actually does "store current" in this way. However, the "ideal" transformer does not. In most cases other than flyback SMPS, any storage in the transformer is regarded as bad.

    The stored current in an inductor is energy in proportion to the current squared x inductance it is flowing in. Actually (I^2 * L)/2

    A capacitor also can be said to "store current", because it stores "charge"..... actual electrons. The electrons get there as a "current", moving electrons, so a capacitor can also be said to store current in that sense, very much as the lake behind a dam stores the flow of the river.

    In a capacitor the energy is in proportion to the capacitance x voltage, (V^2 * C)/2

    Both the inductor and the capacitor give back their stored energy in the form of current.
  16. Calaverasgrande

    Calaverasgrande

    Joined:
    Apr 22, 2006
    Location:
    East Oakland, California
    I am curious about SMPS vis a vis the switch frequency. Is this liable to something like the nyquist theorem? Such that the freq has to be twice the highest freq you want to reproduce? If so is there a lowpass filter as well?
  17. illusha

    illusha

    Joined:
    Jan 18, 2011
    Location:
    Orygun
  18. Jerrold Tiers

    Jerrold Tiers

    Joined:
    Nov 14, 2003
    Location:
    St Louis
    To answer both "calaveras" and "illusha"....

    "SMPS" refers to the power supply.... the actual amplifier is a different part. These days, an SMPS can be quite good at supplying the power for an amp, and most companies doing that have it pretty well together.

    As always, play it first. The IPR I have no direct knowledge of.... but they are supposed to be good power amps. Clever design, a bit scary, but workable.

    As for nyquist, no, not in the power supply. Amps have been reproducing frequencies way above the 50 or 60 Hz 'pulse frequency" in a regular transformer supply for many years.... There may be interactions, but not due to that cause. The power supply is not modulated by the audio signal, and isn't subject to nyquist etc in the way that the switching frequency of a class-D amp is.
  19. agedhorse

    agedhorse Development Engineer-Mesa, Product Support-Genz Supporting Member

    Joined:
    Feb 12, 2006
    To clarify a bit on SMPS, the value of the rectified line frequency storage caps may seem small if you compare them with a typical L.V. DC rail cap on a conventional amp but recalling Bob's comment that energy storage is proportional to the SQUARE of the voltage, the voltage on the line frequency caps is much higher and the square of this higher voltage means the value can be lower. The voltage can be 350 volts on an SMPS line side whereas maybe only 80 or 90 volts on a typical DC rail of the amp side. That's a 20x difference in energy storage, so whereas a typical L.V. DC rail cap may be 10,000uf... for the same energy storage at 350 volts the cap would only need to be 500uF.
  20. Jerrold Tiers

    Jerrold Tiers

    Joined:
    Nov 14, 2003
    Location:
    St Louis
    The energy storage capacity deal is true, but in reality it may be that the energy is not AVAILABLE in the same way as with a traditional supply.

    it is not uncommon to hear that "my SMPS amp just does not sound right in the lows".

    We are not talking here about "sine wave power", most amplifiers will satisfy their paper specifications. But most of us have noticed that even with virtually identical specs, different amps can have drastically different "feel". This is due to performance that is "not documented" in the specs. (we tried to document that low freqency transient performance with our 40 Hz single cycle power spec, but then lots of people yelled at us for having "fake peak-power specs")

    The reason for the difference is often that the SMPS comes with a "current limit"..... A traditional supply has no such thing.

    if your big lead-sled power amp wants a very large current for a low frequency transient, it just draws that current, as much as wanted, from the power supply capacitors. How much it can draw is simply a function of how big the capacitors are. "Good" amps had large capacitors. And capacitors have no current limit, they can supply pretty much whatever you want, limited only by their "capacity".

    Notice that this is NOT affected by the fuse, or the allowable line current, etc.... it is a short term pulse, and pretty much all the power comes from the "local storage".

    Now, with that SAME total energy storage at high voltage in an SMPS, it is "on the wrong side of" a solid-state device or devices.....the SMPS switching IGBTs are between it and the amplifier. Often the amount of storage on the "amplifier side" of the SMPS is much lower than normal for a traditional power supply

    ALL the power that the amp wants must come through those solid state switching parts. But those parts have a maximum current that they can conduct. if that is exceeded they will be overheated or even instantly destroyed.

    The SMPS control circuit always has a maximum current limit. SMPS parts are expensive, and it is not usual to drastically oversize them. So the current limit may be set to protect the parts, while allowing the sine wave power.

    BUT, with that current limit BETWEEN the "energy storage" and the "energy consumer", it is like a 'throttle valve", set low enough that the transient power may be choked off so that it is little more than the sine wave power.

    Inherently this is different from a traditional amp, which as mentioned has NO such "throttle valve"...

    You will hear the result in the low end, and I think that is one big reason for complaints about the low end on some SMPS amplifiers.

    Many manufacturers, but not necessarily all, have found ways to avoid this problem. It is inherent with an SMPS unless the SMPS is designed specifically for an amplifier application, and not just made like any other standard SMPS.
  21. 12bass

    12bass

    Joined:
    Jan 2, 2003
    Location:
    Saskatoon, Canada
    Excellent explanation, Jerrold!

Share This Page