Switching Power Supplies in AV Receivers

[Johnny2Bad]

Did you look at the performance? Output noise of 4 micro-Volts. Broadband. Show me a linear PS with that low a level of ripple.

You won't find many linear PS's that meet those values, and those that do will entail a higher cost.

The people at Benchmark have a few words on the subject.
https://benchmarkmedia.com/blogs/application_notes/inside-the-dac2-part-3-power-supplies

Plus, a linear PS is not trivial to design. Sure, we all know the block diagram, but are your snubbers the correct value? What are you going to do to mitigate ringing?

I don't see your "massive increase in points of failure". What, the resistors are unreliable? The presence of an Xtal bother you?

Quartz crystals are amongst the most reliable components in all of electronics. I have never, personally, heard of a single failure.

If SMPS's were as unreliable as you suggest, we'd all be replacing Wall Warts on a regular basis. I don't know about you, but I have Rubbermaid Tubs of the things, all in working order and all having outlasted whatever device they originally powered.

The original question was are SMPS's suitable for audio. Clearly they are, with careful design.

As for which approach is best with regard to a specific application, that, as always, depends on the application. But to summarily dismiss one or the other reveals a poor regimen of design.

 

[slipperybidness]

Did you look at the performance? Output noise of 4 micro-Volts. Broadband. Show me a linear PS with that low a level of ripple.

You won't find many linear PS's that meet those values, and those that do will entail a higher cost.

This argument is a moot point, simply because a POWER AMP does NOT require these stellar ripple numbers!

In other words, this would be solving a problem that does not exist for a Power Amp Power supply.

I can provide references for these statements if needed.

Going by memory, I believe 10uF ceramic caps are the default choice for snubber caps. And, I do agree that this is a detail that TLS did not mention, and according to the books that I have read, it is very important to use these on your linear PS.

 

[PENG]

I still don't prefer SMPS but whether I like them or not I believe they will eventually become much more popular for reasons mentioned already. I suspect the potential replacements for my Bryston (still have more than 10 years warranty left) or Halo power amp will have some sort of SMPS in them, hope not though. I don't like CVTs, in fact not even automatic transmission, but I have been forced to go with automatic transmission since ages ago, unless I can afford to have two cars and two SUVs. Actually even then I would still be severely limited in my potential choices, such as having to give up the leather package.

 

[slipperybidness]

I still don't prefer SMPS but whether I like them or not I believe they will eventually become much more popular for reasons mentioned already. I suspect the potential replacements for my Bryston (still have more than 10 years warranty left) or Halo power amp will have some sort of SMPS in them, hope not though. I don't like CVTs, in fact not even automatic transmission, but I have been forced to go with automatic transmission since ages ago, unless I can afford to have two cars and two SUVs. Actually even then I would still be severely limited in my potential choices, such as having to give up the leather package.

The main driving force for moving to SMPS in audio is cost savings. The more $ saved on production, the more profits for the manufacturer.

As TLS likes to say, "it's a race to the bottom".

My new vehicle has CVT and they did a very good job on this one! I don't see any downsides at all on the 2014 Nissan Murano.

Now, my 1985 300zx is a 5-speed manual, and I wouldn't have it any other way!

 

[TLS Guy]

This argument is a moot point, simply because a POWER AMP does NOT require these stellar ripple numbers!

In other words, this would be solving a problem that does not exist for a Power Amp Power supply.

I can provide references for these statements if needed.

Going by memory, I believe 10uF ceramic caps are the default choice for snubber caps. And, I do agree that this is a detail that TLS did not mention, and according to the books that I have read, it is very important to use these on your linear PS.

My over sight, I should have included the snubber caps.

I still like simplicity. Unless space and weight are at an extreme premium, I say that the simpler, and I maintain the more reliable, unregulated supply is best. I don't believe the complex supplies will bring any audio benefit.

 

[gzubeck]

The NAD had the same rated continuous output of 100 watts/ch but, as far as I can tell, it didn't have the bass punch which the SR5010 has.

How would you explain then the important reduction in weight if the receivers still use the conventional power supplies. To my knowledge, the standard power supply requires a massive transformer to provide the necessary current at 50 or 60 Hz to drive high wattage outputs. So, how would they manage to use smaller transformers to deliver high dynamic power with a conventional power supply?

By the way, QSC have been manufacturing their DCA (Digital Cinema Amplifier) line which use the SMPS technology, with either Class AB or 2-tier Class H output circuitry depending on maximum power output. They have been used in many theatres for several years. If they had not been reliable, I am sure they would have been removed from cinemas.

In the Montreal area, we have 6 Cineplex theatres fully equipped with QSC sound systems. I had the opportunity to watch several Met Opera direct satellite broadcasts in two of those theatres and they perform pretty well.

I think more capacitance in the power supply design.

 

[gzubeck]

Setting up DB tets of power amps is not easy. The only one I have done that with is Quad and Perreaux some years ago, with B & W speakers and also Thors. There was a definite preference for Quad especially driving the Thors which are more revealing.

I really am suspicious of receivers and Far Eastern Amp designs in general. They are far too complex with huge component counts in the power supplies and amps, as well as having a lot of internal adjustments. To me they sound woolly in the bass and not smooth in the HF with poor string sound.

I have just stayed with my brother in the UK for the second time in six months. He has NHT VT 2.4s.

He is driving then from a Yamaha receiver. I heard these some years ago at the friends house he bought them from. He was powering them form a solid amp, that I don't recall. I do remember I was impressed with them and the bass was pretty punchy. From the receiver they sound mediocre to poor. The bass is flabby and speech is not crisp.

I have a Denon at Eagan driving a pair of my JW full rangers that I know well. They are a nice speaker. I have listened to them quite a bit lately, but I seem to get listening fatigue. The top end I think is not a smooth as it should be. I will have to bring a Quad set up down and compare.

I'm pretty sure that in terms of the power amp sections of receivers, and probably a lot of amps, the sound is not top notch.

I have always enjoyed my Quad amps. They never seem to put a foot wrong, and above all are very reliable due to no internal adjustments and very low part count.

One thing I do suspect is that a lot of receivers and amps may start out OK, but diverge from new performance over time.

Peter Walker was obsessed in his designs, with not only good sound, but reliability and the ability of the design to not downgrade as the components aged. In other words, he made sure his designs were tolerant of components have a pretty wide spec range without affecting performance.

So yes, I believe that there a lot of differences between power amps. I have a suspicion that receivers in terms of their power amps, do not deliver state of the art performance. I have strong preference for separates. I also am a strong advocate of Quad amplifiers, due to superb sound which is maintained over time and high reliability over time. I have had a little over 60 years experience with Quad electronics. I admit it could be I'm used to their sound, but I doubt it. In addition I use speakers of my design exclusively and so there is the possibility that my designs optimally match Quad electronics. However I hate trouble, and the Quads are less prone to bring it to my door step than other stables. In my book they are well worth the cost.

I think if there is a hell they should make all those japanese executives try and power their premium JBL 4350s with cheap japanese receivers. If its good enough for americans its good enough for the japanese....couldn't resist.

 

[gzubeck]

I too am suspicious of Switch Mode Power Supplies, but it has to be said that in the last year or two, designs and supporting chipsets that are oriented towards Audio have become available. Previously the typical SMPS was a model oriented toward computers, or other non-Audio applications. There has been tremendous improvements in that area recently, and the mere presence of a SMPS in an Audio application isn't necessarily a negative.

It should also be said that the design of the audio electronics themselves can mitigate many SMPS issues. If the ripple rejection and high frequency filtering is adequate, it doesn't really matter which type is utilized. As a SMPS is less expensive to implement, consumers and HiFi enthusiasts alike can benefit from lower overall costs.

That doesn't mean the SMPS gets an automatic pass, but only that careful evaluation may lead to a quality manufactured or DIY build that doesn't suffer audibly simply because a conventional power supply wasn't used.

They might not care about longevity anymore. Emotiva switched to smps on their new amps and their warrantee period is now down to 3 years. Good luck when your amp goes down after five years and your linear power supply lasts 20 years.

 

[gzubeck]

Here is the problem though.

An unregulated power supply works really well. There is a transformer, a rectifier, and two large caps. Only three points of failure, and low risk failure for all those components. If the caps ares sized properly there is gobs of peak power instantly avialble.

Now lets look at what you are promoting.

A massive increase in points of failure and a bunch of chips with a high failure risk.

I see all downsides, with no real advantage over the simpler approach.

And another thing, I have always found that elegant amps with low part counts invariably sound the best. Amps with a bunch of stuffed and cramped circuit boards, always seem wanting to me.

I think we have a bug on the site. I can see the image in my post and the edit, but it disappears on posting.

I think we're dealing with Bic disposible products now below a certain price point. What I'd like to know is who's got good power caps now that panasonic is no longer manufacturing quality electrolytic power caps. suggestions.

 

[TLS Guy]

I think we're dealing with Bic disposible products now below a certain price point. What I'd like to know is who's got good power caps now that panasonic is no longer manufacturing quality electrolytic power caps. suggestions.

When did that happen? I have relied for years on Panasonic caps for replacement.

This is the trouble now .Few people want to pay for quality. Now I suppose we will have just Chinese junk or worse. They just shouldn't bother to get out of bed.

 

[gzubeck]

When did that happen? I have relied for years on Panasonic caps for replacement.

This is the trouble now .Few people want to pay for quality. Now I suppose we will have just Chinese junk or worse. They just shouldn't bother to get out of bed.

I just did a search on digikey and the Panasonic T-ha are listed as obsolete. For the money they we're the best...I found out about it 3 months ago when I did a search and then inquired directly from customer support. Panasonic still produces capacitors just not large Power caps...I think its just cdc, united chemicon, or the screw type caps from multiple manufacturers....

https://www.digikey.com/products/en/capacitors/aluminum-capacitors/58?FV=34003f,fc01ac,ffe0003a&mnonly=0&newproducts=0&ColumnSort=0&page=1&stock=0&pbfree=0&rohs=0&quantity=&ptm=0&fid=0&pageSize=25

 

[TLS Guy]

I just did a search on digikey and the Panasonic T-ha are listed as obsolete. For the money they we're the best...I found out about it 3 months ago when I did a search and then inquired directly from customer support. Panasonic still produces capacitors just not large Power caps...I think its just cdc, united chemicon, or the screw type caps from multiple manufacturers....

https://www.digikey.com/products/en/capacitors/aluminum-capacitors/58?FV=34003f,fc01ac,ffe0003a&mnonly=0&newproducts=0&ColumnSort=0&page=1&stock=0&pbfree=0&rohs=0&quantity=&ptm=0&fid=0&pageSize=25

I guess Nichicon would be the best bet, but lead time is nearly 6 months! I guess I had better order ahead.

 

[gzubeck]

I guess Nichicon would be the best bet, but lead time is nearly 6 months! I guess I had better order ahead.

They do have quite a bit of leftover 80v panasonics and they do carry 80v nichicons if you can live with that power rating...

 

[PENG]

They might not care about longevity anymore. Emotiva switched to smps on their new amps and their warrantee period is now down to 3 years. Good luck when your amp goes down after five years and your linear power supply lasts 20 years.

Marantz $8000 flag ship integrated amp switched to smtp too, but then Denon's equivalent still has the linear conventional PS, for now I guess.

 

[PENG]

The main driving force for moving to SMPS in audio is cost savings. The more $ saved on production, the more profits for the manufacturer.

As TLS likes to say, "it's a race to the bottom".

My new vehicle has CVT and they did a very good job on this one! I don't see any downsides at all on the 2014 Nissan Murano.

Now, my 1985 300zx is a 5-speed manual, and I wouldn't have it any other way!

My Subaru's CVT works practically as the regular AT too, just don't feel the same, but at least it has the paddle shift thing to play with. Like SMPS, CVT/AT are more complex and also more expensive to repair, it still eventually basically replaces the stick/clutch. Many of us don't like SMTP and may never change our mind, but they do perform, and are improving all the time to the point some work really well in both performance and reliability. More and more amps are going that way and it is a matter of time before it gets more popular then the still much more popular huge transformer, and caps approach. It did take AT and CVT many years to take over, especially in Europe and Asia, so let's see..

 

[Johnny2Bad]

They might not care about longevity anymore. Emotiva switched to smps on their new amps and their warrantee period is now down to 3 years. Good luck when your amp goes down after five years and your linear power supply lasts 20 years.

Warranties have little to do with longevity when it comes to electronics.

Most failures happen in the first 24 hours, a significant number of those within seconds of first turn-on. The balance is largely skewed to the first 30 days. After that, if it still works, it probably will work for years. The long term failure rate (say, over 10 years) isn't perfect, but it's close to perfect compared to the first month.

Warranty length is not related to reliability, as even a six month warranty covers more than 99% of failures; it's part marketing, but with any digital products, it's the inability to guarantee parts availability past a short period of time. Device manufacturers retire digital chips after a short product cycle. Three years reflects that more than anything else, and it's not even related to the choice of Power Supply, as the parts in the SMPS are available over a longer term.

Just maybe Emotiva (which admittedly has a marginal reliability record so far) switched to SMPS for it's advantages?

Lower Cost to manufacture
Lower Shipping costs due to weight reduction
A huge reduction in radiated magnetic interference, allowing closer parts placement, smaller enclosures, better channel to channel consistency, eliminating the need for steel shielding on the much larger Linear Supply transformer, superior rejection of 60Hz and 120Hz hum, superior broadband noise performance, and the overall adequate performance for the job.

A better argument against SMPS than reliability would be the question asking whether the lower noise is really needed, as most amplifiers have sufficient Power Supply Rejection Ratio (PSRR) to deal with the higher noise levels of Linear Supplies. You know, a real evaluation, not a Straw Man argument.

They can also inject noise into the AC line that may affect other audio components on the same feed, but that has to be tempered with the abundance of devices in the average home that do the same thing. An LED lamp, for example. It can be addressed with an SMPS built for audio applications with careful design.

You know, I currently am building some amplifiers and I chose Linear Supplies for them. They cost me 4x more than the SMPS options I had, but I wanted consistency with other amplifier topologies for comparison purposes. I may end up building another set that differ only in the Power Supply, but that is not the plan as of now.

I certainly don't have any inherent prejudices with regard to one technology vs another. I use what i use for reasons after considering the options and the performance goals. Again, getting back to the OP's query ... are they suitable for audio? They are.

Some people dislike this topology or that, and I have zero problem with that ... make yourself happy, is my way of approaching audio ... but if it then becomes an indication of a belief system versus a performance goal, it's OK to admit it. Why lie to yourself, of all people?

 

[PENG]

Use the following integrated amps as examples, both by D&M

Marantz PM-10 ¥600,000 (USD 5,520) 200W 8 ohms, 0.05%THD, 400 W 4 ohms, 0.1% THD, 47 lbs
Denon PMA-SX1 ¥580,000 (USD 5,336) 50W 8 ohms, THD 0.1%, 100W 4 ohms, THD 0.7%, 67 lbs

The Marantz uses SMPS, hence the power and weight advantage over the Denon that still uses conventional linear power supply with heavy toroidal transformers and heat sinks (not so heavy but still..). For about the same price, everything else being equal, if you must choose between the two, would you not take the Marantz, for 6 dB more output at much lower THD?

I know there are other specs that matter, but I would think either one would be well beyond the point of diminishing return such that the decision would boil down to just power output, aside from the physical features.

 

[Johnny2Bad]

I believe examples like the Marantz you referenced, PENG, will become rare in five years. For those that prefer them, think about buying sooner rather than later.

 

[gzubeck]

Warranties have little to do with longevity when it comes to electronics.

Most failures happen in the first 24 hours, a significant number of those within seconds of first turn-on. The balance is largely skewed to the first 30 days. After that, if it still works, it probably will work for years. The long term failure rate (say, over 10 years) isn't perfect, but it's close to perfect compared to the first month.

Warranty length is not related to reliability, as even a six month warranty covers more than 99% of failures; it's part marketing, but with any digital products, it's the inability to guarantee parts availability past a short period of time. Device manufacturers retire digital chips after a short product cycle. Three years reflects that more than anything else, and it's not even related to the choice of Power Supply, as the parts in the SMPS are available over a longer term.

Just maybe Emotiva (which admittedly has a marginal reliability record so far) switched to SMPS for it's advantages?

Lower Cost to manufacture
Lower Shipping costs due to weight reduction
A huge reduction in radiated magnetic interference, allowing closer parts placement, smaller enclosures, better channel to channel consistency, eliminating the need for steel shielding on the much larger Linear Supply transformer, superior rejection of 60Hz and 120Hz hum, superior broadband noise performance, and the overall adequate performance for the job.

A better argument against SMPS than reliability would be the question asking whether the lower noise is really needed, as most amplifiers have sufficient Power Supply Rejection Ratio (PSRR) to deal with the higher noise levels of Linear Supplies. You know, a real evaluation, not a Straw Man argument.

They can also inject noise into the AC line that may affect other audio components on the same feed, but that has to be tempered with the abundance of devices in the average home that do the same thing. An LED lamp, for example. It can be addressed with an SMPS built for audio applications with careful design.

You know, I currently am building some amplifiers and I chose Linear Supplies for them. They cost me 4x more than the SMPS options I had, but I wanted consistency with other amplifier topologies for comparison purposes. I may end up building another set that differ only in the Power Supply, but that is not the plan as of now.

I certainly don't have any inherent prejudices with regard to one technology vs another. I use what i use for reasons after considering the options and the performance goals. Again, getting back to the OP's query ... are they suitable for audio? They are.

Some people dislike this topology or that, and I have zero problem with that ... make yourself happy, is my way of approaching audio ... but if it then becomes an indication of a belief system versus a performance goal, it's OK to admit it. Why lie to yourself, of all people?

If the smps can last 10+ years with heavy use then you may be right...what percentage of linear power supplies have failed vs. smps supplies over a ten year period? Is there a statistic? Maybe they've figured out how to get smps supplies to last 10+ years by over engineering them then kudos to those engineers as I'm not one to stand in the way of progress...look at how other appliances (washing machines, refrigerators, dishwashers etc.) have gone in the last twenty years...they now fail after 3 years vs. lasting 15-20 years like they used to and they're more expensive than ever. Progress is now measured in designed obsolescence vs. longevity because you can now just sell them a new amp in 5-7 years as it will be just new profits.

 

[<eargiant]

Great topic.

Anyone have an opinion on the link below?
https://benchmarkmedia.com/blogs/application_notes/152143111-audio-myth-switching-power-supplies-are-noisy