51 responses Add your response
Yes you do need a good quiet power (toriodal) supply that keeps noise away from the circuitry - and big capicitance like on a big power amp is one way to do that. Of enormous criticality is how power supply is kept away from ground and how balanced the signal lines are designed with respect to ground.However, it can't be simplified to just that.
It is just as important to have features other than just a good power supply - balanced, high output levels with great channel separation and low output impedance and high input imepdance and of course low S/N. High quality tone controls with a "defeat" option for refernce. Volume control needs to be well implemented like with a discrete resistor network steps (like a crystal CS3310 or Burr Brown PGA2310) rather than just a cheap pot and wiper with continuous volume control, IMHO.
Hi Chris . . . it's a mistake to so superficially "analyze" both the engineering decisions and marketing motivations behind any consumer product . . . it's true that there are many questionable design and engineering practices behind many specialist audio products - but simply looking at the size of the "power supply" (what does that include?) is akin to starting a thread with a title of "What's the proper diameter of a woofer?" . . .
. . . and if it's a mains transformer and filter capacitors you're looking at -- let's just say that dramatically oversizing these components is a pretty ineffective way of ripping off the consumer.
Kirkus, I'm not saying that an oversized power supply is all you need. I was trying to find out why some experts keep telling that a big power supply is essential, even with preamplifiers. Why do you think manufacturers are making two (or four) chassis preamplifiers with the entire power section occupying an entire chassis? Spectral preamplifiers for example have a power section that is bigger than some power amplifiers! You can't deny that this is way overkill. So what are the benefits of such a design?
I see what you're asking.
Power-supplies are commonly separate from the main chassis usually either for UL approval (when pre-fabricated wall-warts are used) or to reduce magnetic coupling from the power transformer into the audio circuitry. In the latter case it may be actually cheaper to produce a separate chassis than to apply the correct parts and engineering time to eliminate the interference when these parts are in close proximity.
Truly low-noise circuit designs usually require low signal impedances, which means more signal current, which means more power-supply current, which can easily lead to a bigger supply . . . and everything runs warmer, so it has to be spread out more for heat dissapation. Also frequently, internal grounding mistakes can cause supply ripple currents to be injected as hum into signal grounds . . . and if the cause of the hum isn't well understood, then it may be solved simply by heaping on more power-supply capacitance and reducing the ripple current.
There are also standard core sizes that transformers are made in . . . and other electrical characteristics (like load regulation, EM field output, and leakage reactances) that are related to the size of the core and the winding methodology. Again, it simply may occur that to acheive other engineering goals . . . bigger ends up being the cheaper/simpler way to get better.
You might also run into issues where these products are sold in countries that have different power tolerances (50Hz runs hotter and requires more capacitance), or tropical climates where a lower temperature rise is requires.
So, in the end, it's these and a thousand other factors. In the end, just like underwear . . . it all Depends.
Elizabeth is just flat out wrong on this and apparently doesn't understand the importance of a component's power supply.
Some of the finest line level amplifiers use large, even massive power supplies to ensure dynamics are reproduced accurately.
They include Blue Circle, Plinius, Marantz, Tom Evans, Pass, Atma-Sphere, First Sound, and on and on.
I think first you have to seperate the tube amps from the transistor units as they have very different requirements.First they need a stable B+ high voltage supply usualy upwards of 240volts for a preamp and a seperate low voltage supply for the heater circuit some using voltage regulation.Many designers prefer to incorperate in line inductors and by the time you get recifiers and storage caps for both high and low voltage circuits.Its getting both costly and heavy.Then you have to seperate the transistor guys into those that use switch mode supplys and those that like the high current transformer and massive energy storage supplys some also using inductors.Linn uses a lot of switch mode designs they sound good when effectively used but surely savings in cost as well as size,weight and efficiency are factors.I like to think the power supply is a critical part of the circuit and not the area to cut cost.
Hi Bill, you forgot to include Spectral. Still it is not very clear to me if a big bad MF of a power supply will ensure that you get: killer bass, bold dynamics, authority and spaciousness.
Btw, some of the fellow Audiogoners use passive preamps (resistor or transformer based) with NO power supply at all --> do they have poor dynamics and bass response?
Oversize is a relative term. If a preamp has a high-power amp's supply, then yes, based on power it would be overkill.
As an EE that specializes in power supply design, my personal opinion is that a large power supply is a good thing. The reason is that large transformers have a low internal impedance which allows them to pass current very quickly to the circuit. This speed and availability is what makes them better than a small transformer with thin wire.
Sometimes, sounding "dynamic" is due to a lack of bass. So yes, a small power supply can sound more "dynamic" than a large one - but that is simply because tonal balance is relative to itself and not because of power supply design. The two should not be confused.
I have an AudioPrism Mantissa preamp. It's getting a bit long in the tooth now, but still a very nice product overall. It has an external power supply. There was also an upgraded version that had a beefier power supply available.
Now AudioPrism has long had a reputation as understanding power supplies and power filtering, so I kind of trust them on this one. Anyway, the difference between the original power supply and the upgraded one, in terms of sonics, is quite easily heard. More dynamics and ectension. An overall sense of increased ease and blacker backgrounds.
It's more than just boosting capacitance or the size of the transformer though. There's probably more discrete filtering in there and probably additional isolation between the sections feeding power to the analogue circiutry also.
Beefier supplies also tends to mean better design and implementation. Or at least more complex design and implementation that should lead to better sonics, for the reasons I cited above.
Btw, some of the fellow Audiogoners use passive preamps (resistor or transformer based) with NO power supply at all --> do they have poor dynamics and bass response?
This is a terrible mistake. You will not achieve high input impedance and low output impedance with a device like this. In the worst cases the sound will change character audibly with volume as the circuit will act like a filter when it couples with the source and power amp. Remember a preamp is also a "buffer" - it allows the source to see high impedance (making the source have an easy life) while driving the power amp faithfully with whatever it needs - irrespective of small load variations. Buffering and high S/N cannot be achieved without active circuitry. Volume adjustments require around 120 or more decibels of control - it is not a simple task if you want to do it with high channel separation and high S/N and low distortion/transparency.
Chris, that is a question many have asked. A larger transformer and power supply has helped every pre amp I have ever tried it on from the Japanese mfgrs all the way up to the high end gear from Jeff Rowland, Threshold and Mark Levinson. A nice 600 watt toroidal transformer will cost 80-120 dollars plus an outboard case to hold it, since almost no manufacturers use a power supply of that size.
After 30+ years of listening I now feel that the size of a preamp power supply has no more to do with the sound of the music than the size of the woofer in a speaker.
This might be music to your ears but it is probably even more music to a manufacturer's ears.
Are you really suggesting (as a generalization) that a manufacturer can now use a small cheaper power supply and a small cheaper woofer with small cheaper motor and small cheaper voice coil (often no bigger than a tweeter): And in general it sounds better or just as good (in general no benefit from using better parts)?
I agree there is such as thing as diminishing returns and many reasonably sized but well designed power supplies are adequate for the job but please....this is audiophile heresey! LOL - Just Kidding ;-)
Actually, no, that is not what I'm suggesting at all.
We agree then on that aspect - good.
Do not mistake quantity for quality, they both start with the letter "Q", but are very different.
Sorry, my bad.
Don't confuse smaller with cheaper.
How silly of me.
Are you suggesting that all you can eat buffets offer the best quality food?
Huh? Are you offering to buy me lunch? ;-)
I am a firm believer in over-specified power supplies in audio, for several reasons:
Large transformers tend to have high primary inductance, this acts as a RF filter, no garbage in, no garbage out.
An over-specified transformer will not distort or limit its voltage going to the PS caps, due to the HUGE current spikes generated by the rectifiers (more than 10 times the average current).
Large capacitors generally have low ESR (equivalent series resistance). PS caps are effectively in series with the output signal, the less ESR the more transparent and "fast" the sound quality. This seems couter-intuitive because conventionally circuit diagrams are drawn with caps in paralell to ground and transistors or tubes in series with the output, however, if we mentally re-draw the circuit, it makes sense.
There are three main "loops" in a typical amplifier circuit:
1- The power transformer secondary + rectifiers form an AC loop that gets rectified and feeds>
2- The PS capacitors, which feeds DC to the control devices (tubes or transistors), which modulate the PS with an AC fac-simile of the input signal (music is variable-frequency/intensity AC), which generates>
3 - The output AC loop, formed by electrons flowing from PS ground to the PS capacitor (+) terminal, then through the tubes and transistors, then to the output. The output AC loop is closed through the load (speakers or the next component input impedance.
In other words:
A preamp or power amp uses tubes or transistors to AC-modulate (to the tune of the music) voltage coming from a power supply, which is the source of all power delivered to the load.
Transistors or tubes act as "control valves", allowing more or less of the PS voltage to pass to the output, hence the term "valve" adopted in the UK for tubes. Also, a FET's control element is called gate for the same reason.
A good listen to any Naim CD player with the Super Cap optional power supplies will convert the non-believer. Or a Musical Fidelity heaphone amp with the high-current PS.
Another example: ARC preamplifier's power supplies are over-specified by at least a factor of ten. ARC's tube preamps are known to sound as good or better than many transistor preamps in the bass.
P.S. I am not associated with Naim, MF or ARC...but I used to own an ASR Emmitter II amp with 1,000,000 uF (yes, one Farad) in the double-mono power supplies. Its bass quality, soundstaging and dead-quiet background are unsurpassed in my experience.
Dazzdax, just for the sake of friendly discussion, the purpose of a preamp is to present a proper load to the source(s), switch inputs, adjust volume and drive properly the amplifier.
A resistive passive preamp is basically made of a 10 K ohms potentiometer used as a voltage divider. The potentiometer attenuates the incoming signal.
As a load for the preceding component, it performs poorly, the 10 K amplifier input impedance in paralell with the passive preamp's internal resistance add up to an equivalent ~5-6K ohms load, too low for most CD players and even worse for phono stages.
As a driver, its average output impedance is 5 k ohms, plainly unsuitable for driving a 10 K ohms power amp* input impedance and incompatible with long or high capacitance cables.
* Most engineerring textbooks and many audiophiles agree that the preamp output impedance must be lower than one tenth of the driven component's input impedance, for optimum transfer of signal and no loss of bass/treble/dynamics. 10 K amp needs a 1 K source (or lower).
The symptoms of low input impedance (poor load) and high output impedance(poor driver) are lack of punch, dynamics and weak bass.
Also, if the ICs are high-capacitance ones or longer than 1 meter, the 5K internal impedance will form an RC low-pass filter that will attenuate treble.
All of the above is measurable and not subjective at all.
Of course there is one advantage to a resistive passive preamp: lack of active circuits, therefore one hears purer midrange and zero noise.
However, IMHO, passive resistive preamps are a flawed design and do not play music (specially rhythm) well.
Because rhythm and dynamics are so important to the kind of music that I like, I use an active preamp.
TVCs are a different ball game, they present a high input impedance to the source and a low output impedance to the load, because they are transformers.
The role of a transformer is to step up or down AC voltages (great as an attenuator) and to match impedances (great as a sorce AND a driver).
Provided that a TVC is well designed and has enough inductance, it will have no measurable frequency response roll-off from ~20 Hz to ~50 Khz. You can read the data sheets at the Steve & Billington, Sowther and Bent Audio sites.
All IMHO and IME, take my comments for what they are worth.
Casouza, thx for your precious input. It leads certainly to a greater understanding of this matter. Why do some digital front end manufacturers like Wadia advocate the use of passive (digital) built in preamplifier? If transformer based preamplifier is the answer, why are most preamplifiers still built in the conventional way?
Chris, my comments were regarding resistive passive preamps versus transformer passive preamps. Between those, I prefer TVCs.
Most preamplifiers are built the traditional way because people need gain, specially audiophiles with phono front ends, low gain amps and low efficiency speakers. A preamp must be (ideally) able to cope with all kinds of sources, amps and speakers, no easy task.
As for Wadia's integrated preamp, it is active, not passive (I have seen the opamps inside a Wadia). Its attenuation is partly analog, partly digital. It sounds great and it replaces handsomely a preamp + PC + IC, as long as one adjusts the internal analog gain DIP switches in order to use a MINIMUM of digital attenuation. If one goes deep into digital attenuation, there is a clear loss of resolution/detail/soundstaging.
As for active preamps, there are excellent ones, some with gain, some with unity-gain buffers. I am not familiar with the Pass buffered preamp.
In summary, there is no clear-cut BEST solution.
In general, I follow these rules-of-thumb:
-If your system lacks dynamics and punch or you need to use the preamp above 3 in the dial, you need more gain, therefore no passive preamp will do the trick. Buy an active preamp that drives your system to VERY LOUD levels and still has about 6 db spare gain for those quiet recordings.
-If your system has too much gain ( volume pot stays around 8 in the dial, "jumpy" volume control near 9) or you can hear noise/tube hiss, you need a preamp with less gain or a TVC preamp IMHO.
-If you want a minimalist solution, a Wadia or Audio Aero CD player connected directly to the power amp will sound great IME.
Audio Aero has an internal active analog, tube-based preamp followed by a buffer to lower output impedance.
Opus 21 also has an internal solid state active analog preamp (SS) and it sounds great for the money.
There may be other brands of CDPs with internal preamps.
Anyway, my original point about the cap-charging current peaks in a preamp DC power supply reaching 10 times the average current remains (and is validated by engineering textbooks). A transformer that is specified for 10 times the preamp continuous power draw is NOt overkill, it is just sensible design.
I hope this helps
I can think of hundreds of scenarios where an active linestage is necessary (as I can think of just as many scenarios were SETs will sound terrible). But the question remains, if you have plenty of gain, low capacitance cables, 100kohm input impedance on the amps with high sensitity - that is a source and amp/speaker combo where there absolutely no need for gain or buffering to match impedances, under though circumstaces how and it what way could an active linestage better a passive resistor preamp? I have had a CAT, Lamm, ARC, Joule ME, Dodd, and even a Bent TAP and now I use a Lighspeed Attenuator to drive the Music Reference RM10 and RM9 SE amps - I don't hear loss of dynamics, soundstage, texture, PRAT, bass, or anything else, except noise. Though I do use a Atma combo becuase the the passive approach just does not work for me with my Atma amps. I would not draw conclusions about passive preamps unless they are in the right system context, as I would not judge whether 1.5 watt SETs are any good by trying them on a pair of Thiel or B&Ws. And if I needed buffering, I'm sure the TVCs or AVCs would sound better than the resistor-based passives. Now this may all work becuase the power supplies in the EMM Labs gear is pretty beefy.
Casouza - yes, transformer should be oversized since mentioned current spikes, while driving recifier/capacitors, heat up the windings (RMS much higher than Average) and their high frequency content heats up the core but 10x is perhaps overkill. Imagine 40% efficient class AB 200W amp with 5kW transformer. Large amount of capacitors is beneficial providing better filtering but when you think of it - linear power supply is really a bad case of SMPS working at 120Hz. SMPS working at 100kHz will do the same with 10x smaller toroidal transformer, is line and load regulated and is quiet since higher frequency is easier to filter out, non-audible and switching is done (in modern SMPS) in zero voltage/zero current. For that reason Jeff Rowland uses SMPS in Capri preamp (where efficiency is not important). SMPS got such bad rap from cheap ones used in computers that people don't want them even for class D amps (that are SMPS) and designers do what sells.
Preamps need some oversizing perhaps 2x-3x to cover losses related to operation with rectifier and losses in voltage regulators that have to work in 90-132V mains range. Filtering here is done mostly by the regulators and current is pretty much constant (class A) with transitional current supplied from local caps.
Benchmark DAC1 supply is not an overkill - it takes from mains "16W peak" of average power but because of mentioned much higher RMS value of current spikes it needs to be about 32W and it looks like 30W toroidal transformer to me. It is toroidal because outside field of evenly woven toroidal transformer is zero and Benchmark achieves measured 140dB S/N.
You are right about power supply caps being in series with the speaker. People don't realize how important supply caps are. Cheap caps not only have high ESR "eating out" dynamics that amp could provide but also have high inductance filtering out fast transitions. Placing film caps in parallel is a band-aid and often a bad one since pure capacitance in parallel to main caps inductance creates parallel resonant circuit that rings (cap is in series with the speaker). There are low inductance caps (like slit foil caps) but very expensive. I would rather go with fast responding SMPS and that's what Jeff Rowland did (AFAIK) in latest high power class AB amps.
There is another thread talking about a custom made tube preamp that sports an oversized power supply so much so that the unit weighs 75 lbs (34 kg). The owners of these particular devices have only accolades for the sound that this special preamp creates. Of course their descriptions oft include the size of the power supply as though any preamp worth of hifi should have a power supply capable of driving speakers. I have seen that there is some science to putting an oversized power supply into a preamp for improved noise performance and voltage stability. Beyond that, I do not see the point. Unless I am mistaken, the power output (audio power) of a preamp is, at most, about 50 milliwatts. That is assuming a 20k ohm impedance looking into the amplifier. I have seen some preamps boasting 250 VA power supplies. All that to deliver 50 milliwatts? Sure, there are the tube heater circuits to power and other ancillary circuits, but that is not the case with SS preamps. It this just a case where we as audiophiles expect our preamps to be big heavy boxes with gobs of reserve power and so the hifi builders deliver to those expectations? Preamps have had oversized power supplies for decades now. It just seems like a type of arms race that each generation of preamps must have even bigger power supplies to show some type of advancement in technology. I remember when very expensive SS preamps could be left on all of the time because they consumed little more power than a night light.
Could someone please simplify this discussion for me back to the basics? To quote Jeremy Irons in Margin Call, "speak to me as though I were a young child...or a golden retriever."
Why is a separate power supply necessary at all? In the last hi-fi system I had (Mac C22, MC275, Marantz 10B), the wall plug in the NYC apartment I had at the time seemed to be a pretty good power supply. I'm now getting back into hi-fi and recently acquired a Levinson 25S phono preamp with a separate PLS 226 power supply. I haven't yet used it since I'm still in the process of acquiring the other necessary stuff (signal source, preamp etc.)
I gather that the 25S (or 26S preamp) wouldn't be of much use if it could be plugged directly into a wall socket. Why, in very simple terms, is that?
The whole idea of a power supply in audio electronics is to convert the alternating current (A/C) from the wall to direct current (DC). DC current is required for anything regarding audio (it just cannot run from straight A/C like some motors can).
The reasoning behind creating a separate power supply will vary depending on manufacturer. In a nutshell, a separate power supply will shield the core audio circuits from the electromagnetic noise caused by the transformer in the power supply. Generally it is not really a problem, but it can cause a slight buzz in the audio if the transformer is too close. The second reason could be that the manufacturer just did not have the room to build in the power supply they wanted within the size of the preamp chassis. A good example of this is the Krell 202 preamp:
You can see from the pictures that the power supply is enormous with two large transformers and a lot of transistor heatsinks for their shunt regulated DC power supply.
The alternating current AC from the wall comes in waves that run at 60 hz (about 60 waves per second) where half the time the voltage is "pushing" and the other half it is "pulling". DC requires a 100% constant "push" in voltage that is one direction only. The power supply tries to "leak" the current pressure from the AC and also store as much voltage in the power supply capacitors so that it can continue to provide a 100% DC current even when the AC is not pushing current at all (which is between the waves).
The size of the power supply can also dictate how good of a sound quality you get. With regards to extremely large power supplies, it can help further smooth out the sound. My research and experimentation has shown that conventional circuits and op amps do not really require a huge power supply. However, when you start using Class A type audio circuits, then the power supply size becomes hugely important. The Class A circuits require a large amount of constant current and you can actually reduce the sound quality by providing a "too small" power supply. You are actually starving the audio circuit and the end result could be things such as a thin/bright/harsh type of sound. A prime example of this is the older Krell KAV-280p preamp. This is also similar to the Krell Phantom III preamp. They both have such a small power supply with limited capacitance (4 x 3300uf in the 280p) that the sound is bright/thin. I found that doubling that to 4 x 6800uf significantly changed the sound and it was no longer bright/thin - had a tremendous amount of bass and midbass strength. When you get into the larger Krell preamps, such as Phanton II or Phanton I, you will get the larger power supplies for very similar audio circuits.
Power supplies should be very robust. They should provide an unshakable foundation for absolute control of the signal to the amp, plus a high degree of resistance to perturbations created anywhere in the system, including from the amp(s). No one can deny one has more control with an ‘oversize’ power supply. Just like driving a car with a standard transmission and more than adequate torque provides more refined control of the drive.
Of course, all those who have never experienced the control power gives can never ‘get’ it. However, once experienced never forgotten. The unfortunate aspect of these forums is that most contributors simply don’t have access, have never experienced, reference grade components. So their comments are only correct and of value to those who have similar limited performance units. Nothing new here. (But today I find it annoying that shills and charletans make great “to do’s” about the utterly minutes differences between red black or blue wall plugs. As if they have power supplies and systems that can identify the virtually, no, non- existent differences. Crass purveyors and promoters of these ridiculous cash grabs should post some technical support for their fanciful, or should I say, Fantasy claims. Off topic, but, after all, this is just a forum :)
Anthony Cordesman is a classy audiophile gentleman who I've had the pleasure of listening with as he used and reviewed my products many moons ago. A shout out to Anthony!
Now about transformers. My answer would be somewhat different than those who go with bigger is better only. Innovation in audio is being advanced at a more progressive pace over the last 15 or so years and I see and hear design improvements leaning toward smaller transformers.
I look at transformers the same way I look at speaker crossovers, more and bigger parts is usually a result of something not working up to par somewhere in the chain.
Interesting comments. In repairing and building gear I have found that just because something may seem "over built" to some, the manufacturer has designed it that way to work with his component. Think back to Counterpoint. The basic preamps were build with really nothing power supplies so that they were simply built to a price point. As you went up the line, separate power supplies came into play. One simple reason to move the transformer out of the main chassis was noise. The separate power supplies also had better parts to complement the preamp that made the sound better. Tube regulation, tube rectification, etc., were added to improve the sound. You can build a power supply to regulate the AC with a resistor and many cheaper preamps have just that.
My power supplies are about 30lbs for my DAC and preamps. Probably more than the amps most of you are using. It is not that they are overbuilt, but necessary to get the sound I wanted each component to reproduce, plus noise, etc. Parts quality are also part of the equation. Dueland capacitors for example, are much larger than other standard capacitors, they also weigh more so not so much over build but in there for sound which requires more space. Each part has a sonic impact. That was why Counterpoint used Plitron transformers as a recommended upgrade. Expensive and larger (and quieter) that the stock ones Counterpoint originally used. Does all this make a difference in sound quality, to me YES, to others, I cannot answer for them. But if anyone used Counterpoint gear and went up the ladder, they heard the differences. So I do not think that this means that something is not working up to par somewhere in the chain, this is part of the design implementation. Speakers may need to be of a more simple design with quality parts, but when filtering AC and making the sound quiet, dynamic, micro/macro swings, control bass, clarity extension of high frequencies, etc., it may be what makes one component sound better that the other. Solid State versus tube designs also require different implementation. So my point is, try doing that with a resistor to control the AC.
How many manufacturers use point-to-point wiring? Do you think that you could hear the difference in your components if they were point-to-point wired? Is that overkill/overbuilt or a better design implementation?