Cross regulation vs Transient Response...SMPS for Audio ?

[grigson]

Regarding Full or half bridge SMPS for audio, (with split +/- output rails) Does any reader know :-

1. What percentage of the output inductors should be coupled.?

2. How to get good transient response , in spite of having coupled inductors.?
(-because coupled inductors improves cross-regulation, but worsens transient response)


The entire reason that the Audio world is slow to adopt SMPS power supplies, is due to the mystery surrounding the coupling of the output inductors in the split rail audio supply rails.

For example, a full-bridge, offline smps would have dual +/- outputs, and in order to get good cross-regulation, one couples the two output inductors.

....but there is a trade off..becuase if 100% coupling is chosen, then transient response worsens.

So, does any reader know what percentage of the output inductors should be coupled?

This is the gaurded secret that will unlock the floodgates for mass adoption of SMPS for audio.... the moment it is out!

 

[VEBill]

...The entire reason that the Audio world is slow to adopt SMPS power supplies...

Forgive my possible ignorance...

Back in the 1970s, home audio stereo receivers and amplifiers in the 100w RMS x2 class were typically very heavy due to the large AC-input power transformer. By that time they were all OTL (output transformerless) circuits so they were light weight except for the PS. Some were designed with the transformer in one corner making them extremely imbalanced and very awkward to carry.

For the last couple of decades, you can typically walk up to a multi-channel receiver (700+w RMS total) and pick it up easily. They're obviously not using linear power supply circuits. I had assumed that they'd switched to light weight switching power supplies. Is my assumption incorrect?

 

[grigson]

Up to about 500W, Mains transformers are much cheaper than SMPS's for audio use.

An undersized mains transformer can manage 500W audio by simply being overloaded for the short intervals when peak power occurs.....SMPS is not nearly so capable of being overloaded in order to handle short high-power bursts.

Also, many Audio amplififer supply rails are split rails (eg +/-50V).

A mains transformer will easily , via a split output winding, be able to provide good cross-regulation between the split rails.

Another point is that the Green Brigade haven't hit the audio world too hard......
partly because even high peak power audio systems consume very little on average.

-but there's also the fact that Class A amplififers are still in use.......which draw maximum power all the time when quiet........so theres little point in saying how inefficient mains transformers are when the amplififer is doing that, so the audio world can "get away" with using mains transformers.

 

[VEBill]

This is the gaurded secret that will unlock the floodgates for mass adoption of SMPS for audio.... the moment it is out!

Up to about 500W, Mains transformers are much cheaper than SMPS's for audio use.

Your statements are mutually contradictory.

 

[ScottyUK]

VE1BLL,

A lot of the bigger amplifiers for professional use direct switching from the rectified mains (or rectified & PFC-boosted mains, just to piss off grigson a bit further ) and the output transformer is a ferrite type. They are very compact for their output and are usually difficult to repair.

As for grigson's question, I'm not at all sure as it assumes a separate regulated power supply and amplifier.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[grigson]

I apologise for ambiguity i caused on that subject.....

What i meant was, that since the means to achieve good cross-regulation of the split rails in SMPS is so little known, it works out cheaper to do it with mains transformers instead......

...because at the moment, most audio companies have engineers who dont know how to get good cross-regulation with SMPS.....and they can't afford the wages of a specialist SMPS consultancy to do it for them.....so at the moment, smps is more expensive, due to this development problem ..

-once the cross-regulation problem in SMPS has been solved, and is out there in the common domain, SMPS will be cheaper.

At the moment, companies are often getting good regulation of the split outputs by literally implementing two smps's...one for each half of the split rails...which is very expensive.

-The mystery of the coupled inductor technique has not been widely solved....in relation to the trade-off between cross-regulation and transient response.

 

[VEBill]

"...coupled inductor technique..."

An active solution would be better, faster, cheaper.

 

[grigson]

By "Active" i take it that you are referring to cross-regulation techniques such as in Magnetic Amplifiers and Secondary Side Post Rectifiers?

 

[VEBill]

Magnetic amplifiers? Good grief...

No. I meant using one of these... ;-)

If this subject is a real concern, then just add some series regulators and have them communicate with each other via an intelligent control system. Amplifier Classes G & H prove that controlling power supply rails is completely practical and reasonably cost effective. More modern schemes might use switch mode and FETs, instead of linear and bipolar. High speed uC can make the control system do practically anything you want. Modern supercaps can fill in the gaps. Power supply perfection should be easier than ever before.

Have you gone down to your local High Street and tried lifting various home theater amplifiers? I know for a fact that some (not all) weigh next to nothing, and they certainly violate the rule "about 2.7Kgs per 100W" for the main transformer since the entire 700 watt RMS output system weighs only about 10kg. It seems most likely that they're not using a mains transformer, except perhaps a small one for the ancillary power supplies.

To drive the point home, there obviously aren't any mains transformer in the high power car stereos. Some of those are many hundreds of watts, and a few hoodlums are running thousands of watts. Even my car is supposed to have 420 watts in the stock system (Harmon Kardon). I've not had to cancel any road trips due to issues with poorly implemented cross regulation vs transient response tradeoffs.

It's worth noting that not all customers listen to light orchestral music. If the customer listens to (for example) Bachman-Turner Overdrive with the system cranked up to "11", it's going to be putting out a lot of power by any standard. Undersized mains transformers based on assumed peak-to-average ratios aren't going to survive the first day.

What exactly are your goals? It's been noted that you're posting up a storm on various forums.

 

[MacGyverS2000]

I see a common thread to your posts, grig, but I fail to see a clear point to it all...

Dan - Owner

http://www.Hi-TecDesigns.com

 

[LionelHutz]

If it really is a guarded secret, the guardians are likely the engineers from companies that have figured out the answers to your questions. Or more accurately, the engineers who have come up with what they feel is the best solution in regards to your questions.

 

[IRstuff]

"good cross-regulation of the split rails in SMPS is so little known"

Really? Didja think that audio designers are what, Neanderthal EEs that never crack open a textbook or a decent trade journal?

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[Skogsgurra]

If Grig is typical for the breed - that is what one is temted to think..

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[grigson]

My understanding of coupled inductors in dual outputs of a full-bridge SMPS is that current flow is obviosly not like in a transformer, where the flux of one coil cancels that of the other...........

-Rather, current flow in either coil puts the flux in the same direction.

-Neither coil "knows" that the other coil is there......it just "sees" this extra flux-per-ampere of current that it flows.......and it doesnt "know" that this flux comes from the other coil........it simply "thinks" that its inductance is greater.......and indeed it is....by mutual inductance theorem.

......the thing is, when one coil is flowing current and the other is not, then there is no "extra" flux.......so, in fact, the inductance really is less.

-So in coupled inductors like this, we have widely varying inductances.

-The output inductor is a crucial part of the full bridge smps small signal feedback transfer function.......

....and therefore , we have an extremely complex control loop at hand with coupled inductors.

It is the mastery of this which is evading the fraternity.

 

[Skogsgurra]

Lots of "thinking", "feeling", "seeing" and "knowing" to describe something that can easily be put in an equation. Why do you think all those words are necessary? Are audio equipment designers artists instead of engineers? Is it really that bad?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[IRstuff]

The usage of coupled inductors is hardly a mysterious concept known only to select acolytes. Slobodan Cuk was teaching coupled inductor power supplies in the late 70s, albeit he was coupling input//output. Nonetheless, an EE that graduated since the 1980s is likely to have been exposed to those types topologies in school.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize