• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Why does the audio industry like linear transformers so much over SMPS?

ZeDestructor

Active Member
Joined
Jul 28, 2019
Messages
119
Likes
68
As the title says, why does the audio industry like linear transformers so much? Based on all the results we keep seeing with mains leakage and such, wouldn't it be much simpler and cheaper to just put in a decent SMPS? I mean, basically all the chart toppers (AHB2, THX 789/887, SP200, Purifi, Hypex, SMSL M500, for example) use an SMPS...

Edit: On top of that, linear PS also tend to be very sensitive to the incoming AC sine wave, which means needing power conditioners or some sort, and preventing the use of such gear with line-interactive UPSes.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
My two (three) cents:

1. Worrying about long term reliability.
2. The need for serious engineering skills if a custom solutions is required.
3. Angst (unfounded) about PSRR in linear post regulation.

Yes, simple voltage regulators don't have good high frequency rejection, so you have to design accordingly.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Reliability, I certainly think in the past smps did not have the best reliability but I'm not so sure now. Smps units are everywhere now, how many do you have in your home? I had a laptop one fail a year ago, but otherwise.
 
OP
ZeDestructor

ZeDestructor

Active Member
Joined
Jul 28, 2019
Messages
119
Likes
68
My two (three) cents:

1. Worrying about long term reliability.
2. The need for serious engineering skills if a custom solutions is required.
3. Angst (unfounded) about PSRR in linear post regulation.

Reliability, I certainly think in the past smps did not have the best reliability but I'm not so sure now. Smps units are everywhere now, how many do you have in your home? I had a laptop one fail a year ago, but otherwise.

Reliability-wise I'd say the claims are very unfounded. I have computers with over 15 years of use on and off with no issues on their original SMPS. There was a while when things were a bit unreliable though - the Capacitor Plague era where a stolen, incomplete electrolyte formulation caused a wave of short-life capacitors to come out of basically every capacitor manufacturer.

As for custom solutions.... eh? Looking at the teardown of some of Benchmark's stuff (like the DAC3), they're just using generic, unmodified industrial power supplies with the right rating and calling it a day.

If you really need some very custom solution, even then life is easy - companies like Artesyn and Delta have off the shelf modules you can buy and configure the voltage of (using control resistors). The rest is just a matter of filtering ripple using the right selection of capacitors.

Seems to me like the Audio industry believes in a lot of outdated FUD perhaps?
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,579
Likes
38,279
Location
Gold Coast, Queensland, Australia
Transformer based supplies are more reliable long term. They can take significant overloading without shutting down. It takes an awfully serious continuous overload to cook a decent transformer. All they need to protect them is a fuse (and if you're in the EU, a stupid thermal fuse buried in the primary).

They are simple, reliable and much more resilient to transient line condition aberrations, surges, lightning strikes etc. (we get a lot in Qld)

SMPS supplies can be excellent, tightly regulated and very lightweight and compact. They also can suffer from thermal issues and poor reliability when inadequately designed. I have no problem with them as long as they are silent, both from a radiated EMI perspective and from an acoustic perspective.

SMPSs are definitely the way forward, for better or for worse. Freight costs are based on size and weight and SMPSs clearly win.

When it comes to amplifiers, SMPS based supplies can offer more reliable protection for amplifier malfunction as the amplifier itself can shut the power supply off within a few cycles, whereas a linear supply has vast storage sitting in filter capacitors and it isn't "smart". Those joules want to go somewhere in the event of a failure or short and that can be rather spectacular.

There are compelling reasons for both, but as always, it comes down to the "fit for purpose" argument and that may vary from one implementation to the next.
 

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
There is also just so much "SMPSs are bad and only suitable for cheap crap" and "if it doesn't weigh at least as much as a boat anchor, it can't be High Quality" folklore and old wives tales still around. Manufacturers try to do what customers want....
 

Killingbeans

Major Contributor
Joined
Oct 23, 2018
Messages
4,089
Likes
7,546
Location
Bjerringbro, Denmark.
As for custom solutions.... eh? Looking at the teardown of some of Benchmark's stuff (like the DAC3), they're just using generic, unmodified industrial power supplies with the right rating and calling it a day.

Yes, but then have a look inside the AHB2 ;)

My point being that many of the designers, who's skill level barely covers trafo + diode bridge + big big honking capacitors, would run away screaming if asked to build a power amp SMPS from scratch.

Seems to me like the Audio industry believes in a lot of outdated FUD perhaps?

Absolutely. The list is endless. And those who don't believe it prospers on those who do.
 
OP
ZeDestructor

ZeDestructor

Active Member
Joined
Jul 28, 2019
Messages
119
Likes
68
Transformer based supplies are more reliable long term. They can take significant overloading without shutting down. It takes an awfully serious continuous overload to cook a decent transformer. All they need to protect them is a fuse (and if you're in the EU, a stupid thermal fuse buried in the primary).

They are simple, reliable and much more resilient to transient line condition aberrations, surges, lightning strikes etc. (we get a lot in Qld)

Good points. The counterpoint is that they really, *really* dislike having wonky, non-sinusoid waveforms and have gnarly power factors. This can become an issue if some other poorly-filtered SMPS else or a UPS is crapping out loads of RF and messing up the waveform.

In addition, most modern SMPS have some MOVs on the input to handle overvoltages of all sorts, while the main regulation will happily handle and shutoff when encountering a significant undervoltage situation. The downside ofc is that if a surge blows your MOVs, you need to do some soldering rather than just swapping a fuse, but the hardware after the PSU should be fine.

SMPS supplies can be excellent, tightly regulated and very lightweight and compact. They also can suffer from thermal issues and poor reliability when inadequately designed. I have no problem with them as long as they are silent, both from a radiated EMI perspective and from an acoustic perspective.

SMPSs are definitely the way forward, for better or for worse. Freight costs are based on size and weight and SMPSs clearly win.

When it comes to amplifiers, SMPS based supplies can offer more reliable protection for amplifier malfunction as the amplifier itself can shut the power supply off within a few cycles, whereas a linear supply has vast storage sitting in filter capacitors and it isn't "smart". Those joules want to go somewhere in the event of a failure or short and that can be rather spectacular.

There are compelling reasons for both, but as always, it comes down to the "fit for purpose" argument and that may vary from one implementation to the next.

Agreed.

Yes, but then have a look inside the AHB2 ;)

My point being that many of the designers, who's skill level barely covers trafo + diode bridge + big big honking capacitors, would run away screaming if asked to build a power amp SMPS from scratch.

Hehe, indeed. That said, I reckon than the 1.6kW monster that powers my desktop PC could give that AHB2 PS a good run for the money with the help of some more filtering capacitors on the output side (and a different, higher-voltage output rather than just +12V).
 

ajawamnet

Active Member
Joined
Aug 9, 2019
Messages
288
Likes
460
Reliability, I certainly think in the past smps did not have the best reliability but I'm not so sure now. Smps units are everywhere now, how many do you have in your home? I had a laptop one fail a year ago, but otherwise.
I just designed a Linear Tech/Analog Dev. part into an emergency vehicle system. Just amazingly low noise from 1Hz to 3GHz.
My bench:
IMG_20160114_140749.jpg


I was using some high end RF probes and was really amazed at how low they got the noise.

I do a lot of Mil gear and for some we have to be below 24dBuv in various chambers.
re102.png



One chamber is at PAX River:
http://thehowlandcompany.com/ISTF/ASIL.htm

P-8C PAX NAVAIR ANECHOIC CHAMBER.JPG


About two-three months ago I spent a silly amount of time there - trying to get mine and someone else's shit to get below that threshold.

And don't even get into conducted ... that just gets silly.


One issue is getting the output impedance down on SMPS . The topology - whether it's DC-DC (SEPIC, Switched Inductor, Switched Cap), the various Flyback types and whether they are buck, boost or buck/boost - makes a big difference how low the effective output resistance/impedance is of the supply. This can affect power delivery to the various analog stages. This is one of the reasons I see mentioned by the old timer analog mfgs... But I've seen some kickass SMPS designs that blow away poorly designed linear supplies.

Here's a buyers guide for non-electronics people - https://www.industrybuying.com/articles/different-types-of-switched-mode-power-supply-smps/


A bit more technical article : https://www.calex.com/pdf/3power_impedance.pdf
"Introduction
Noise in power supplies is not only caused by the power
supply itself, but also the load’s interaction with the power
supply (i.e. dynamic loads, switching, etc.). To lower load
induced noise, most uninitiated designers use the biggest,
lowest impedance capacitor that can be found, thinking that
this will absorb the dynamic load current. This method is often
hit and miss, and can even cause the power supply to oscillate
and make the situation worse."



Here's a thesis on SMPS output impedance:
https://pdfs.semanticscholar.org/101e/471955e3bfcca2c89f7b93fb9c1bc942b6e9.pdf

And another:
http://www.deltartp.com/dpel/dpelconferencepapers/33.3_10275.pdf
 
Last edited:

JJB70

Major Contributor
Forum Donor
Joined
Aug 17, 2018
Messages
2,905
Likes
6,148
Location
Singapore
Linear transformers are much bigger, heavier and are generally a lot more expensive therefore they are clearly better. Well, that's the logic (or maybe lack of) on planet audiophile.
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,579
Likes
38,279
Location
Gold Coast, Queensland, Australia
My bench:
index.php

Nice setup!

I can see the old Leader sign gen and LDM-170 with a docked laptop on top. Gotta love the MS optical mouse- they never die and fit man sized hands. I have two on my desk too. A couple of Rigols, and are they lights hanging off the roof above your scope? I see you like collecting tip cleaners... :)
 

ajawamnet

Active Member
Joined
Aug 9, 2019
Messages
288
Likes
460
Nice setup!

I can see the old Leader sign gen and LDM-170 with a docked laptop on top. Gotta love the MS optical mouse- they never die and fit man sized hands. I have two on my desk too. A couple of Rigols, and are they lights hanging off the roof above your scope? I see you like collecting tip cleaners... :)

Also got an old LSG-231 stereo generator there. The local PC parts shop (PME Micro - just a bad ass place to buy PC parts here in D.C.) gave it to me - his dad used to fix TV's and stereos at that location. Then PC's came around in the early 80's and they eventually evolved into that. Still had all his gear tho.

Yea - those are lights with extended beam shaping rings to keep the glare off of the scopes - The laptop is connected to a 8 port KVM that also does bench connections for repairing and the security camera DVR that I mod'ed to not phone home to China... I actually consoled into the thing and shut off all the silly botnet stoopid - It boots Busybox which I called Busybot on Kreb's site ...

The tip cleaners are for the various wattage irons. If you look close you'll see one that could be used as a cattle prod (according to the old Hakko site).

Love my wired optical mice... I do a lot of RF designs and the less RF BS flying around the better. There's a door in the wall behind the bench that goes to the next room (where the CNC gear is) that I used for some pre-compliance testing. Had to get a Siglent spectrum analyzer, since the Rigols at the time didn't do log X axis limit lines correctly...

I used to have a nixie tube freq counter too... That got lost during my whole IoT fiasco in the 1990's when we had three take over attempts:

http://www.ajawamnet.com/amnet/index.html
... once they figured out what an @ symbol meant (see the story)

Yep - I gots one o' da first patents on da IoT bullshit. That and 2 bucks will get me a coffee at Starfuq's... NOPE... need $3...

Oh well...

"I've failed at so many things so many times, I make Abraham Lincoln look like Bill Gates..."
~~~ me
 

rwortman

Addicted to Fun and Learning
Forum Donor
Joined
Jan 29, 2019
Messages
740
Likes
683
Linear power supplies are simple so just about any hobbiest can build one. Many high end designers are barely above hobbiests. One day some audiophile gurus said switch mode supplies were bad so now it’s hard to sell serious gear that has one. Same thing happened to tone controls and direct drive turntables, although DD is making a high end comeback. My latest amp has a switch mode supply feeding a switch mode amp. Funny, it doesn’t sound anywhere near as bad as it’s supposed to.
 

ElNino

Addicted to Fun and Learning
Joined
Sep 26, 2019
Messages
557
Likes
722
Linear power supplies are simple so just about any hobbiest can build one. Many high end designers are barely above hobbiests. One day some audiophile gurus said switch mode supplies were bad so now it’s hard to sell serious gear that has one.

This is true. Also, early switchers tended to have more noise in the audio band than linear supplies, and the "audio gurus" decided they were bad on that account. These days, a good switcher is usually cleaner than a basic linear supply.
 

gene_stl

Addicted to Fun and Learning
Forum Donor
Joined
Jun 14, 2019
Messages
867
Likes
1,198
Location
St.Louis , Missouri , U.S.A.
As restorer john said Linears are good for stopping or at least slowing down a lightning strike. At least seemingly better than the first switchers.

I had an Ophthamologist from Panama send me a Zeiss slit lamp power supply that had absorbed a lightning stroke. When I got it open I was surprised to see a switcher. It was way beyond repairability and though a switcher it was pretty old. The main brain was a custom and unavailable.


One thing I have seen too many times is that switchers are all a little different. The only guy that actually understands it is the guy that designed it and almost certainly HE doesn't work for the company anymore. For this reason I used to hate them. But just like happened with early transistor amps , they are much better now than they were at first.

I do still wonder about all that ultrasonic leakage. even though its only a few fractions of a watt. I would rather it didn't get dissipated in my tweeter voice coils. I wonder the same thing about DSD out of band noise.
 

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
I do still wonder about all that ultrasonic leakage. even though its only a few fractions of a watt. I would rather it didn't get dissipated in my tweeter voice coils. I wonder the same thing about DSD out of band noise.

You could just put a small capacitor in parallel with your tweeter.
 
Top Bottom