Switch mode versus traditional power supplies

[@mariofilippi3539]

After years of trying to understand the difference, your explanation of linear vs. switching power supplies was the most well thought out explanation. Thank you.

[@MichaelLivingston-me]

Well done, Paul. That's not an easy topic to discuss and keep it from becoming highly technical. Excellent explanation.

[@wilcalint]

A commendable job by Paul on this video. One thing that Paul briefly mentioned that I wanted to expand on. "Switching" Power Supplies have oscillators in them. And those oscillators are certainly over 10Khz. Paul said that Switching power supplies create "noise" and that is very true. And because of that, per the FCC ( Federal Communications Commission ), requires any consumer device with an oscillator over 10Khz MUST carry a testing for compliance certification number. Purchasing an off the shelf power supply that is already certified does not absolve you from getting your entire product retested and certified. All of that can be a pain in the ass. Been there, done that.

[@mrlazda]

And if you sell it in Europe, you must always certify entotr product no regardless did you use all certified modules or not. Europe basically only certifies whole product.

[@RogierYou]

Toroidal transformers are heavy and make your equipment feel more expensive 🤣

[@KlaweKlapki]

xD toroids are not heavy try ei core same power as toroid

[@scottholmes4388]

Torrids do not have a power line buzz and produce much less noise.

[@RogierYou]

@@scottholmes4388 yep as a ham radio operator I agree with this!

[@fabrized]

I've had buzzing toroidals but never ei

[@TPQ1235]

Paul, thanks. Many others have tried to explain this, took longer, used graphs and acronyms, and left me perplexed.

[@ivoverhoef1160]

Great ‘simple’ explanation Paul. Thanks!

[@d4sched567]

Switching power supplies are nothing new. Science discovered long long time ago that as the frequency went up, the size of the transformer went down to transfer the same amount of power. So, if you wanted a smaller power supply, you would take the incoming 50 or 60 Hz power, put it to an oscillator that would "switch" it at a much higher frequency, then use a much smaller transformer to make a conventional (linear) power supply. But now, you would also need some extra filters to clean up the higher frequency components brought on by the oscillator. That's what it is in a nutshell... In most cases, switching power supplies are chosen to eliminate an large, expensive transformer. Therefore, most switching supplies have the number one design goal to be "cheap" and, in line with this goal, the filtering I mentioned above is also not done properly. The lack of proper filtering and shielding makes most commercially available switching power supplies a poor choice for high end audio. Of course, a switching power supply can be designed and constructed in a (expensive) way to have an excellent noise filtering, which would make it suitable for high end audio use. However, almost nobody does that because you don't want to deal with such a complex design when you will likely have no cost or performance advantage over an equivalent conventional (linear) power supply.

[@tkjho]

I have a Technics mini-system that has a separate Pulse Power Supply Unit SH-C01 for the 50W RMS amp SE-C01. It was bought in 1979 and looks like it was 45 years ahead of its time.

[@L.Scott_Music]

You explained it well and it was less than two minutes more than the average video.

[@ThinkingBetter]

For a power amp, traditional transformers are not regulated and the voltage drops by the signal load and mains power fluctuations. Capacitors are used to help on that. A switch mode power supply can provide precise regulated voltage and also eliminate the need for an external power regenerator as it sort of does the same.

[@bronymusiclover8891]

It's true for a majority, but just want to mention there are power amps that use traditional big transformers and capacitors with regulated power supplies. As the example there is Technics SE-9600 which regulates voltage after capacitors from 67V to 50V - regulated linear power supply also powers main power transistors

[@maynellrodrigues9862]

I'll take the robustness of the transformer instead of a sensitive SMPS ✨

[@ScottsLifeOffGrid]

Paul, I really appreciate your videos, Thanks!

[@chebrubin]

SMPS in an Emotiva Reference Differential sounds amazing.

[@tpock63]

Better or cheaper 🤔 The bean counters always win

[@stephens2r338]

They are improving but they are still chosen purely due to cost savings. Sometimes up to 25% of the retail price. Most have the switching PS outside the box on the power cable to reduce radiated noise, case size and weight. Just like with all laptops. One of the leaders in this field are Lumin Audio owned by Pixel Magic. Their built in switching PS have lots of shielding. However their higher end models still use toroid or offer one as an upgrade. They just sound better.

[@hoobsgroove]

No because traditional power supply engineers can't build switch mode well enough

[@ramsaybolton9099]

@@hoobsgroovewhy?

[@hoobsgroove]

@@ramsaybolton9099 more complicated than a traditional power supply what is easy! You have to think about feedback circuit

[@stephens2r338]

@@hoobsgroove Thanks for your reply. So are you saying that a switching power supply is superior just that electrical engineers aren't clever enough to know how to make them.

[@hoobsgroove]

@@stephens2r338 more like traditional power supply engineers are set in their ways but yes you could say like that

[@gtrguyinaz]

Spectron first developed a switch mode power amp and it was very good in 1999… I owned 1 and it was great… combined with my audible illusions pre amp, it was as good as anything …

[@davidstevens7809]

Stewart pa1800 early excellent amp..it was better than carver cube amp.

[@KenTeel]

If I remember right, a switch mode supply uses a transistor that turns fully on and fully off very fast. This is called switching a transistor between saturation (fully on), and cut off (fully off.) The out put of the transistor is hooked to the primary windings of a transformer. The secondary winding (which is and isolated winding from the primary windings) has more windings than the primary windings. Therefore it's a stepup transformer. This output voltage, from the secondary windings, is fed into a capacitor (or several capacitors.) As the transistor switches on and off, alternately, the secondary windings charge up the capacitor. This capacitor is the "resovoir" of current and voltage that powers the amplifier section of an amp. Now, I'm going by memory on this description, so, if anyone sees flaws in this general description, chime in and correct what I've posted here. In my description I used a single transistor. In reality there can be more than one switching transistor.

[@jimeadie4415]

MOSFET rather than transistor typically. The duty cycle is continuously adjusted to maintain a set output voltage.

[@KenTeel]

@@jimeadie4415 A MOSFET IS a TRANSISTOR (metal-oxide-semiconductor field-effect TRANSISTOR ) I didn't specify what kind of transistor is used. On the old JBL 6233 amps (one of the first switch mode power supplied amps, called the "ice man") bi-polar transistors were used for switiching. Of course other types of transistors can be used (which is why I didn't specify WHICH TRANSISTOR type can be used.) And, the duty cycle is implied in what I said. The voltage across the capactor is charged, and of course the charge/drain, implies frequency. Please add to what I said, rather than just stating what I've already said.

[@PlatypusPerspective]

Switched mode can operate as step up or step down as required for the application. The transformer(s) in a computer PSU have considerably less secondary turns than the primary to deliver the 12V, 5V etc from the 160V to 320V DC rectified mains that the input side runs from through the switching transistor(s) and transformer primary. The amplifier SMPS in this context will be a step-down design if it is supplied directly by the power mains. The higher DC source (where I am in Australia with 230V AC mains it would be the approx 320V) needs to be converted to the amplifier supply rail voltage, say 60 to 70 volt, with two outputs required for positive and negative rails.

[@KenTeel]

@@PlatypusPerspective Thank you. That's good information. It seems very accurate to me. I hope that all will appreciate your words.

[@ScottBeattyProSound]

Thank you for this Paul.

[@rugerthedog396]

Good overview! However, Paul didn’t discuss all of the detail of the noise problems with switch-mode supplies. Noise problems occurs not only at the switch frequency and it’s harmonics, but also sneakily via ringing on the switch transition. Ringing can have a fundamental and harmonics from 10’s of MHz to beyond 100MHz. Ringing noise is particularly more prominent where the switching transistors are external to the controller. Both types of noise can travel not just via radiation, but conductively onto the power supply output voltage, and also back onto the power line. Part of the art Paul refers to is understanding the pitfalls and techniques of making the power supplies “good citizens” that don’t put out wide-band noise interference into the product itself and into the user’s environment.

[@ThinkingBetter]

@@Douglas_Blake Agree completely. The audiophile industry is still in the mode of the 1980s on many topics. Using SMPS for a power amp makes your power rails regulated and more immune to mains power fluctuations and noise. My first PC (Taiwanese IBM XT 4.77MHz Intel 8088 clone) in the 1980s had a switch mode power supply that was super noisy and it resulted in my neighbor being unable to watch some particular TV channel. My wired mouse acted as antenna and even at a distance of some 50-100 feet, his roof TV antenna captured the noise from my mouse wire. The issue was found out by some specialist with some equipment investigating the source of this radiation as our neighbor had swapped antenna and tried to solve it in several ways and in the end had to get some help from some inspector investigating RF emission violations. The guy came into my room of my parents home (in my teens) while I was alone in the house in the evening and didn't have the main door locked with his huge equipment and antenna; and he quickly found out where the noise came from (some first generation PC mouse with RS232 interface). After grounding the PC chassis, the noise disappeared.

[@ThinkingBetter]

@@Douglas_Blake Yes, for example, many audiophiles are still judging digital streaming audio with how it sounded 15+ years ago when 128kbps MP3 was the norm. Nowadays streaming audio can easily beat any disc media incl. SACD (DSD-64) with much superior data rate. Streaming FLAC 352.8kHz PCM can beat a DSD disc because any decent DSD music has gone through a mixer where the output typically was DXD and a LOSSY transcoding turned it into DSD. Still there is some religion about DSD being better than the master stream of the DXD PCM, which is complete nonsense. And then I find it quite hilarious that you can sell at some insane cost a CD player using some 15 year old Philips CD deck design, exclude the DAC and wrap it in a nice aluminum enclosure to then call it a “transport”. Transport my butt… any USB flash drive costing a few Dollars can save an entire CD collection more reliable than some US$20,000 transport.

[@somename8831]

Technics had an amp with battery powered power supply, I guess we don't do that anymore for a reason? Maybe it's good again with the new type of batteries?

[@samirpradhan7735]

Looks like, we all have missed one very important aspect with respect to bass production. It is normally believed that SMPS in hi-fi audio can't handle transient bass response like linear PS does. That may be one of the reasons why audiophiles do not like SMPS.

[@PlatypusPerspective]

There's no reason to expect switched mode power supplies have any problems with bass if they are correctly specified for the design centres of the amplifier. As a technician I would believe what is being referred to is the "reserve" an unregulated power supply has to provide short term power output greater than the continuous rated power of the amplifier (headroom). An amp with a SMPS has closely regulated supply rail voltage, and can output up to its rated power, say 150W, whether for 10 milliseconds or ten minutes. If an amplifier with conventional power supply can provide 150W continuous, then the supply rail voltages will be higher at lower average powers, and a transient may be able to drive the amplifier to, say 170W output briefly without clipping. An SMPS powered amplifier would clip attempting to do the same thing. I suspect the effect has probably become attributed to bass because an occurrence seeking such power is overwhelmingly likely to be bass oriented, such as an explosion in a movie soundtrack or a cannon in the 1812 overture. If the conventional amp was pushed too hard it would clip too, so I think it's just a matter of recognizing a slightly different characteristic of amplifiers.

[@davidstevens7809]

I found that out studying the carver cube amps..25 years ago..I ststed the fact and waa laughed at.. wade stewart quickly designed the stewart pa1800 and it ran faster and fixed the lag issues that triac controlled amps had.

[@marcusbrsp]

Let me get this straight. So, PS Audio will be switching to switch mode power supplies in a near future? I don't wanna get it switched up.

[@jimeadie4415]

The difference is a switched mode power supply uses an inductor rather than a transformer. With a switching frequency of 100KHz rather than 120 Hz for full wave rectification they switching power supplies don’t need as much filtering to remove audio frequency noise. Somewhat like a comparison of a conventional water heater vs tankless. Old approach vs higher efficiency.

[@mrlazda]

No, isolated switch mode power supply uses transformer in all topology except in flyback, which uses coupled inductors, consent it is only topology (flyback) that can't use full wave rectifiers.

[@milkman100001]

great vid paul. thanks

[@cletusberkeley9441]

Recently switched over (no pun intended) to a regulated linear power supply for my DAC, made quite a difference compared to the switch mode unit I was running for three years prior. They each have their place. Switching PSs are way cheaper to manufacture, smaller, lighter, way greater power density, etc., hence their ubiquitousness today. Properly designed they are excellent, but the run of the mill ones are quite horrible noise generators IMHO.

[@ConnorDylan]

Made quite a difference how? Is this one of those "I could hear a HUGE difference" things??

[@cletusberkeley9441]

@Connor Dylan I'm not into the snake oil thing at all (I'm very happy with electrical wire as my speaker cables). Noise floor for one thing went down.

[@cletusberkeley9441]

@Richard Wielgosz they're inexpensive, just audition them both for yourself and use the one you like best.

[@GrahamAtDesk]

I had a similar experience with my Pi streamer. The upper frequencies were painful to my ears, making me want to leave the room after a couple of minutes. Changing to another switching supply (a Google pixel charger) really cleaned things up. Adding two ferrite beads to the lead took it the extra mile. I tried all the 5v switching supplies I had in the house and saw a reduction in ear discomfort with all of them. Only the Google one was smooth enough for the ferrite beads to stop it sounding slightly harsh at times. I'd still like to try a linear PSU at some point to see if it makes it slightly more mellow.

[@rubenmejia4881]

Excellent explanation 👌 👏 👍

[@zeusapollo8688]

Less metal=more cheaper

[@chrisladouceur4093]

A good rule of thumb is that cheap switchers will be worse than any linear. But expensive purpose built switchers can be better than linears. A lot of people don’t want to believe that switchers could outperform linears (probably due to all the $$ they spent on them) but it’s true. Switchers have a greater potential, but are usually cheaply build crap

[@hoobsgroove]

A traditional power supply is switch mode as well If it has a regulator, the only difference is it works at a lower switch mode then a high mode switch mode what is better for audio quality. Well you can use a big transformer in a switch mobile power supply it's just not necessary generally Where you buying your smps Taiwan. What frequency are you running them at Paul. There's a American guy who worked for NASA building smps on the space station maybe should employer him?? You can build it yourself 400 w open source

[@paulstubbs7678]

So why wait until you can buy 'off the shelf'. why not design your own, then it will match exactly what you want. Kind of hard tweaking a pre-built supply, but when it's your own, no problems. Like that 'pulse locked supply' mentioned, where they are probably synchronising it's switching to the rest of the system to avoid having two clocks that beat against each other.

[@stefanegger]

Bringing a new PSU to the market would need a lot of certification and laws etc. By using already tested and certified products, it is easier and cheaper for them.

[@bf0189]

It's extremely hard and they would need to hire a separate PSU engineer and they ain't cheap. Plus lots of regulations etc. It's more practical for a more boutique firm like PSAudio.

[@paulstubbs7678]

@@bf0189 But if you keep doing this, then you kind of become just a re-seller, and eventually pointless. The customers will eventually learn where the good stuff your using is coming from and gravitate to a more direct approach to acquiring it. If you design it, you have the tech, the customers have to come to you.

[@JohnSmith-qi6co]

@@paulstubbs7678 All business is a balancing act. You are focusing on particular elements of that balance. You also have to consider that timing is everything. Sure, they could become "just a re-seller" of certain items, but be moving to others simultaneously. Like them or not, PS Audio is a successful business and someone there is working the balancing act.

[@jorgezuni2818]

Is there any difference in power savings with any one of these power supply ???

[@bartlevert9647]

@Paul, why are Class D amps typically fit with a SMPS, and Class AB with a toroidal transformer?

[@PlatypusPerspective]

Class D amplifiers are able to be compact and lightweight for their power output, not requiring large heat dissipation. Using a switched mode power supply makes sense in terms of also being compact and lightweight in comparison to a conventional transformer with very large reservoir capacitors.

[@ratman5727]

I am no expert, but the efficiency of a D class is certainly better than an AB. That is not to say that an AB couldn't use a SMPS, I think that most people that are in the market for an old tried and true AB would prefer the big transformer with large caps. The old technology works well, is quiet, and delivers good power at the expense of some extra weight (and heat). In my opinion, one just needs to look at some of the "no holds barred" designs like the McIntosh's and Krell's to realize that when money is no object, that is the design of choice. I have run many different amplifiers over the years, from NAD, Levinson, to Carver. They all have their pros and cons. It will be interesting to see where new technology takes us in the future.

[@hushpuppykl]

Many say SMPS is noisy. Benchmark amplifiers use SMPS. It measures extremely well. Question is whether manufactures want to use good quality SMPS or not. Are they cost saving?

[@adrianwilliams763]

Great explanation

[@kcr2365]

If a transformer is rated for a single frequency 60hz how can it also work with 50hz power supply

[@PlatypusPerspective]

A transformer needs to be operated at its design frequency in order to work properly, so the simple answer is that it can't. The turns per volt of the windings will be wrong for the core area. Depending on which way around the mismatch is, you could expect problems like overheating, poor regulation, noisy operation, incorrect output voltage.

[@ssgeek4515]

If I may say sir what about reliability. Too msny componants in a smps to go wrong.I have a 1980s psu with the big lump in it that has never failed even under heavy loads for fiddling? A smps will not last

[@JohnSmith-qi6co]

Does it have large electrolytic capacitors? They will likely be out of spec pretty soon!

[@dummerdum4714]

Still feel switch mode PS has a shorter life span and more prone to spikes

[@AhmedAddas]

How to get smps with ultra low emi

[@LeonFleisherFan]

Never say never. I'm curious as always. Thus far have only heard few products using SMP that I thought convincing, and even there, always in terms of their quality/price ratio (within the respective quality and price category, as a layperson would think they should cost as much or less, not more), and that included products in the 20+k price category where I still kept wondering if they might not sound even better if… But since SMP are what mostly everyone is using these days, there's upward potential designing better ones. One of the best external power supplies available today is a hybrid construction, for example. Nothing at all wrong with higher efficiency, ultimately, the proof is in the listening.

[@ConnorDylan]

"Convincing"? What does that mean?

[@LeonFleisherFan]

​@@ConnorDylanSonically. That's the sole purpose of an audiophile component. At least as far as I am concerned.

[@MrCatalysis101]

Very interesting thanks.

[@edmaster3147]

I thought PS was about the sound, but it is about the cost...?

[@JohnSmith-qi6co]

They can be "about the sound," but if they aren't also "about the cost," they will go out of business and then they won't be there to be "about the sound." That's the way everything works--put your oxygen mask on first then help others.

[@edmaster3147]

@@JohnSmith-qi6co I feel it's more a matter of choosing which quality to offer at which consumers. Being a small company in a small market expanding geographically is a good idea for PS, but to maintain such an organisation, you have to meet (legal) standards and keep sales at a level, which compromises sales.For one your products need to be fool and bullet proof. I bought speakers off a company that produces perhaps 2 sets a year and put all their heart and soul into it, besides their other products at a very small scale. But for them not many buyers and for me not many sellers. They indicated for instance, if you want a bit mor high, solder these resistors in. lol,

[@peddie1972]

What about linear power supply? I’m no expert, but from what I read and understand you get a lot better sound quality using that. Is it worth the extra money?

[@chrisladouceur4093]

Linear is what he’s comparing switchers against. Toroidal is linear. Linear will be surpassed by high end switch mode eventually. Some already do

[@PlatypusPerspective]

​@@chrisladouceur4093 I suspect the question might be about a linear regulated power supply, which is rare as they are bulky, inefficient and expensive. But deliver very clean supply rails.

[@chrisladouceur4093]

@@PlatypusPerspective same story though. Regulated or not, there is a lower limit that linears cannot get below due to the transformers

[@PlatypusPerspective]

@@chrisladouceur4093 A lower limit in what parameters?

[@Ineedtotakeabreak]

How long until the EU bans Class A space heaters also known as amps 😊

[@willjohnson2722]

I think the EU plan is to make electricity so expensive, nobody can afford to use them!

[@joeythedime1838]

So no more massively heavy power amps?

[@davidfromamerica1871]

Some day as nano technology becomes much more advanced. The way you will be listening to music will be totally different. Molecular structures storing terabytes of information sharing this information with stringed together other molecular structures. You won’t be living in a World with hardware scattered across your living room like you have now. Power cleaners, power this and power that. Audiophile cable polish at $500.00 an once😊etc. Yeah if there was such a product. You people would be telling everyone it makes a huge difference in the Bits. 😊

[@genejas]

It really is a stretch to call switch mode power supplies "new"... The Apple II had one in 1977, and you could argue the actual invention was done in 1959.

[@barney6888]

All I know about the power of moving electrons is to think if it like flowing water (as it's not really all that different) and to use the proper analogy between the two. Volts is the spray pressure coming out of the nozzle. Amps is how many electrons cross the exit point of the nozzle in whatever amount of time Ohms is the resistance and not your grandma Watts is the storage tank that's accessed when needed and is also on 2nd base because Who's on first and I probably have all of the above totally mixed up because I have a 500 watt brain with the wiring burned out causing me to not remember... and stuff. and I ramble. and when's Superman and Capt Kangaroo coming on?? I've been sitting here for 65 years, and no sign of it yet. Thank Goodness for the 3 Stooges and the Marx Bros, God Bless their souls indeed.

[@jamiemcneil9462]

Now PS Audio will be able to compete with high end companies like Emotiva. Paul wants to be the first to sell gear at the highest margin ever while using cheap garbage components. Good luck.

[@brianbob7514]

I think you say it more like ooru-y

[@antoniojoseandreomartinez9710]

Like.

[@Mikkelfrost]

Hey. 1 to seee

[@ClassifiedBrief]

Second.

[@NoEgg4u]

kzbin.info/www/bejne/mXmbmqlqo5eemdE

[@digggerrjones7345]

Saying "when I can't use real technical terms" is insulting to your viewers! In so many of the videos you suggest and insinuate that most people don't understand the industry jargon or the basic principles of electrical engineering. I think you're dead wrong in assuming that as by just reading the comments, it is clear there are a lot of very intelligent viewers who possess your level of knowledge, if not more!

[@geddylee501]

maybe he was referring to himself

[@jaredproudfoot9494]

Fernando, the viewer asking the question, requested that the answer be low in technical description.

[@finscreenname]

Or maybe he was talking to me. Not all of us are electrical engineers. Now that I know that "switch mode" is just the power supply being flipped on and off at super speeds to regulate the input instead using a transformer makes it worth no more than the other to me.

[@wty1313]

Listen to the first 40 seconds of the video. A non-technical explanation is what the writer of the letter requested.