OK- I know the answer - don't. Good old power supplies are better, easier to work with, and can deliver all the clean power you need - in line with the experience Nelson Pass convey . And I am not doubting that this works - I have had a class A amplifier that consumed 500 watts, and I had absolutely no complaint with the sound.
However, I am living in Europe, and we start taking energy consumption seriously. So my question: Has not one single company specialised in making decent switch mode power supply for hifi? Benchmark claims that it is possible with their power amplifier. I am not talking about only using them for Class D amplifiers - is it a dream to use such a thing for one of the smaller Pass single ended class A ( such as 20watt). With the efficiency they have, a dream could be to have the good sound with only the use of what an old light bulb might consume.
Is this just a dream? I am not talking about a DIY thing, but a power supply someone made for DIY'ers.
Please tell me if this is just a dream - but I would be even more interested in coming to know some solutions that might be - maybe not perfect, but acceptable.
However, I am living in Europe, and we start taking energy consumption seriously. So my question: Has not one single company specialised in making decent switch mode power supply for hifi? Benchmark claims that it is possible with their power amplifier. I am not talking about only using them for Class D amplifiers - is it a dream to use such a thing for one of the smaller Pass single ended class A ( such as 20watt). With the efficiency they have, a dream could be to have the good sound with only the use of what an old light bulb might consume.
Is this just a dream? I am not talking about a DIY thing, but a power supply someone made for DIY'ers.
Please tell me if this is just a dream - but I would be even more interested in coming to know some solutions that might be - maybe not perfect, but acceptable.
If I were you I would give it a try.
SMPS are not bad these days, things have moved on a lot.
They have mains filters and output filters.
LLC SMPS have lesser radiation than fly back as they provide the transformer with a sine wave. Fly back hit the transformer hard with a pulse.
A switching supply isn't going to overcome the inherent inefficiency of an SE classA amp. The major inefficiency there is in the amp itself, not in the power supply.
Hypex has a switching supply for hi-fi amps. There's also a vendor on Taobao selling switching supplies for audio amps. But don't expect superb efficiency overall unless you're willing to move over to classD from SE classA.
Hypex has a switching supply for hi-fi amps. There's also a vendor on Taobao selling switching supplies for audio amps. But don't expect superb efficiency overall unless you're willing to move over to classD from SE classA.
audiopower.it ran a thread about their 400W SMPS for audio amps about four years ago. Their site is still live and now sells a "nexus".
DPS-400 and DPS-500 power supplies are still listed on the site.
Whatever happend to connexelectronic? they used to be a diyaudio listed vendor of SMPS of 500W, now they are not. The trademark is now a US corporation selling some kind of industrial goods, not power supplies.
None of the above would be useful for a class A headphone amp. More for 200-300W class AB amps.
Whatever, both were a lot more expensive than transformers out of burnt up US made PA amps.
DPS-400 and DPS-500 power supplies are still listed on the site.
Whatever happend to connexelectronic? they used to be a diyaudio listed vendor of SMPS of 500W, now they are not. The trademark is now a US corporation selling some kind of industrial goods, not power supplies.
None of the above would be useful for a class A headphone amp. More for 200-300W class AB amps.
Whatever, both were a lot more expensive than transformers out of burnt up US made PA amps.
Last edited:
thank you for the responses - let me bring some more thoughts. The wonderful sound of class A se amps cannot be made by class D amps, or am I wrong in saying so? If we accept that, we also know that class D actually makes OK sound in the sub frequencies. We also know that most energy is in the lower frequencies. So - from an energy perspective still not selling out on the good sound, would it be possible to make a solution -maybe powered by switch mode all over - with a smaller wattage class A ( maybe down to 5-10 watt) with a class D sub - all dealt with with a clever DSP solution.
Is that a far out thought?
Is that a far out thought?
You clearly have not listened to the most recent models of a good ClassD amplifier. Yes, ClassA sounds different, but better is subjective. If one would compare an equally priced ClassA with a modern and well engineered ClassD amplifier, the second would easily win. But one would have to accept that technology advances and topologies that used to be superior, now become obsolete...
I only have listening experience from classA and classD amps side by side on headphones (not speakers) but based on that I'd suggest you're mistaken here. A full-digital amplifier (which is in effect a classD open loop DAC) sounds to my ears better than my best line-level DAC feeding a balanced SE classA amp. This difference could indeed be due to my 'best' DAC not being good enough so might not be an accurate reflection on the SE classA amp.
When selecting smps do not over-size as is common with linear supplies. The two main reasons are (1) efficiency is best near max but most importantly (2) if the smps is lightly loaded (if it is over-size) it will go into burst mode. Burst mode introduces unwanted lower frequency noise into the system. There is no way of knowing which frequency the smps bursts at and you may be unlucky and have the burst right there in the upper-midrange.
That said there may very well be smps today with controlled burst frequencies, or at least that burst > 20kHz. But for that we are probably talking mil-spec smps that costs more than they are worth.
(edit
Smps that are continously loaded by at least 70% of max rating will not burst, so they do seem best for class-A or at least hot biased class-AB amps where the current drawn is relatively constant.
That said there may very well be smps today with controlled burst frequencies, or at least that burst > 20kHz. But for that we are probably talking mil-spec smps that costs more than they are worth.
(edit
Smps that are continously loaded by at least 70% of max rating will not burst, so they do seem best for class-A or at least hot biased class-AB amps where the current drawn is relatively constant.
Last edited:
Worked fine for me +/-24V at 15A each!
I built a three channel 20W SE Rod Elliot class A amp for LCR tweeter duty. And as you know SE class A is the absolute worst efficiency. I found two NOS 24v 15a switchers so WTH! I could always build a cap multiplier linear supply later.
These power supplies are also open frame PC boards - not metal encased.
No audible noise and remember this is a tweeter amp. Just the white noise of the amp.
The amp pulls over 550w off the AC line even at idle so these supplies are quite loaded. Running about 2.7 amps per rail per amp no signal.
The only precautions I took was to wrap the B+ and B- leads once around a ferrite core. And I also installed a 20A Corcom PI filter in the AC line.
Also note that QSC amps are all switch mode power supplies, Class AB amp but with a switch mode power supply. These are used extensively in both commercial and mastering theaters. They are well regarded in the industry.
I'll try to post some pics.
I built a three channel 20W SE Rod Elliot class A amp for LCR tweeter duty. And as you know SE class A is the absolute worst efficiency. I found two NOS 24v 15a switchers so WTH! I could always build a cap multiplier linear supply later.
These power supplies are also open frame PC boards - not metal encased.
No audible noise and remember this is a tweeter amp. Just the white noise of the amp.
The amp pulls over 550w off the AC line even at idle so these supplies are quite loaded. Running about 2.7 amps per rail per amp no signal.
The only precautions I took was to wrap the B+ and B- leads once around a ferrite core. And I also installed a 20A Corcom PI filter in the AC line.
Also note that QSC amps are all switch mode power supplies, Class AB amp but with a switch mode power supply. These are used extensively in both commercial and mastering theaters. They are well regarded in the industry.
I'll try to post some pics.
Last edited:
I own a Peavey CS800s 260 W/ch @ 8 ohm amp with a switch mode rail supply. I used it at 1/8 to 2 W in my living room, didn't hear anything objectionable. Speakers are 101 db @1W 1m.
Last edited:
+1, that's it exactly.
Unregulated, a "linear" big iron and big cap supply (which is anything BUT linear, BTW!) will probably be more efficient in terms of power usage. But regulated (having much lower ripple at hard to remove frequencies near 100Hz), the switching supply will win hands down. People go on about how switching type anything is bad in audio, because anything invented after 1930 can't possibly sound good
. But keep in mind that the ripple left over from a high frequency switching supply will be above audio frequencies while the ripple left over from a big iron and big cap bank supply will be at pretty much the worst possible frequencies for an audio system. I run a 55W SE class A amp setup with switching supplies and have zero complaints about it (though I should add, I'm using a choke-fed topology which inherently rejects any high frequency ripple, kind of the best of both worlds in terms of power supply noise).Also, for non-class-A, when amplifier efficiency is high, a switching supply could be dropping in and out of irregular switching modes which will up the output noise including at lower in-audio-band frequencies.
Last edited:
Yeah well I guess Peavy either pulls enough current thru the smps at idle that it does burst, or the burst freq is > audible.
I use a smps that probably bursts in my car amp b/c it is over-sized, no bad sound there either, but if selecting a smps from the start, selecting one that is just right is better than selecting one that is over kill.
It is just a tip, I need more experience with smps before making hard claims, but most lightly loaded smps do burst, just the way it is.
Well for one thing, the audio information is not carried as an AM signal. It's more analogous to FM where the audio is encoded as pulse widths.
I'm sure they have some minimal idle current to prevent the smps from bursting. And/or it is possible to make the burst at > audible frequencies. You can also synchronize the smps switching with the amp's switching.
But for a diy'er buying a seperate pre-made smps it may not be easy to know the burst frequency. Which is why it is not good to over-size the smps. That's all Im saying, not that smps cannot be good for HiFi, I think it absolutely can. Over-sizing decresaes the efficiency as well, kinda debunking the reason to use them in the first place.
I may have been conservative in my earlier post where I say select one that is at least loaded to 70% of max rating. Depending on topology, some smps probably wont burst untill the load is < 25% or so.
The reason for burst (or 'skipping') is that as the output reaches it's design value, the switcching stops. If it doesnt stop the voltage will go beyond it's design value. As it stops, the load draws the output down. As the output reaches a threshold the smps starts switching again to get it's output back up. This goes on and on at a 'burst frequency'. The switching itself during operation may be hundreds of kilohertz, but the bursting may in some cases happen to be in the audible frequency range. Or if luck strucks, it burst way above, and in that case no prob.