On the relevancy of ACD tests, Power Supplies, Amplifier topologies

[jeannot]

I would like this thread to start informational, and hopefully turn into a discussion where technical opinions are exchanged for the common good. People who are sensitive to over-engineering should hit Back right now.

First a quick background on Power Amplifiers. We all know that a power amplifier is made of a power supply and circuitry, and both of these halves have to be perfect is perfection is the target.

Kinda like the engine AND the suspension being necessary for a fast car. The engine has more to do with acceleration (akin to dynamics in audio, or ease at high volumes), and the suspension to the handling (musicality, softness and presence).

A perfect power supply delivers a constant voltage at any current demand between 0 and infinity. The perfectly constant voltage part is not possible without using expensive regulated power supplies (à la Naim, Quad 303, some old Yamahas-Carver, etc), but a near-constant voltage can be achieved by over-sizing the transformer and filtering (also acting as reserve) caps. That is expensive, and in direct relation with two things: 1-The price of the amplifier, 2-The pride the manufacturer puts in its products. Some manufacturers cleverly use high quality regulated power supply for the sensitive input stages, and a non-regulated power supply for the output stages.

It is difficult and impractical to evaluate the power supply on its own, but amp power tests can give us a window into its quality. A perfect power supply will allow an amplifier to give the same power whether it drives 1, 2 or ACD. And to almost double its power in 4 ohms. You will notice generally the more expensive (and generally the heavier) an amp/receiver is, the more these numbers are approached.

A perfect amplifier is, as an old British magazine once coined it, a "straight wire with gain". Amplifying a signal is no small task, especially amplifying voltage with the non-linear current amplifying devices we call bipolar transistors. That is a much more complicated topic than power supplies (I can start another thread if there is demand), but suffice it to say that we can tell how serious a manufacturer is by looking at the design in a shop manual. Starting with the ubiquitous 7-transistors topology, as manufacturers get more serious, you start seeing current regulators, current mirrors, multi-stage differential inputs, symmetrical configurations and parallel output transistors.

Case in point, my Denon 3808CI:
Power Supply: The receiver tests at 186 watts per channel at 8 ohms 2 channel driven, and 113 watts ACD. That translates into a drop from 54.5V to 42.5V, not something to be proud of. It would likely sound much better even at high volumes and cost MORE to produce if it gave 115 watts in stereo and 113 watts ACD.

Amplifier topology: The receiver uses 9 transistors per channel, some Mark Levinson and Harman Kardon power amps use over 35 in the name of open-loop linearity, low-feedback, power supply rejection and low output impedance.

The importance of these on the day-to-day performance can certainly be argued, a sure sign being that the 3808CI is considered like a pretty good receiver by most.

 

[PENG]

Case in point, my Denon 3808CI:
Power Supply: The receiver tests at 186 watts per channel at 8 ohms 2 channel driven, and 113 watts ACD. That translates into a drop from 54.5V to 42.5V, not something to be proud of. It would likely sound much better even at high volumes and cost MORE to produce if it gave 115 watts in stereo and 113 watts ACD.

No, I disagree as it defies logic. It would likely cost less to produce 115W in stereo and 113 in ACD than to produce 186W in 2 channel and 113W in ACD. All Denon needs to do is put less powerful amplifiers in the receiver to achieve the same 2 channel and 7 channel output of 115W. Instead, Denon wisely put the money where the mouth is, by beefing up the amp section, leaving less money for the PS so the PS can drive the amp to its rated output level in 2 channel but not into 7 channels simultaneously.

I much prefer this approach, than Harman Kardon's that would have invested more on the power supply, leaving less for the amp section so they are forced to put in weaker amplifiers. At the end of the day, it is the dollar per watt that counts. Everything else (THD, SN, CT, features) etc. being equal, in terms of $ per W, not too many AVR can beat the 3808. End of story.

 

[Alex2507]

No, I disagree as it defies logic. It would likely cost less to produce 115W in stereo and 113 in ACD than to produce 186W in 2 channel and 113W in ACD. All Denon needs to do is put less powerful amplifiers in the receiver to achieve the same 2 channel and 7 channel output of 115W. Instead, Denon wisely put the money where the mouth is, by beefing up the amp section, leaving less money for the PS so the PS can drive the amp to its rated output level in 2 channel but not into 7 channels simultaneously.

I much prefer this approach, than Harman Kardon's that would have invested more on the power supply, leaving less for the amp section so they are forced to put in weaker amplifiers. At the end of the day, it is the dollar per watt that counts. Everything else (THD, SN, CT, features) etc. being equal, in terms of $ per W, not too many AVR can beat the 3808. End of story.

Harman watts are better.



















j/k lol

 

[PENG]

Harman watts are better.

j/k lol

Absolutely,sound much better too, Engineer James would have thanked them for the new definitions!!

 

[3db]

No, I disagree as it defies logic. It would likely cost less to produce 115W in stereo and 113 in ACD than to produce 186W in 2 channel and 113W in ACD. All Denon needs to do is put less powerful amplifiers in the receiver to achieve the same 2 channel and 7 channel output of 115W. Instead, Denon wisely put the money where the mouth is, by beefing up the amp section, leaving less money for the PS so the PS can drive the amp to its rated output level in 2 channel but not into 7 channels simultaneously.

I much prefer this approach, than Harman Kardon's that would have invested more on the power supply, leaving less for the amp section so they are forced to put in weaker amplifiers. At the end of the day, it is the dollar per watt that counts. Everything else (THD, SN, CT, features) etc. being equal, in terms of $ per W, not too many AVR can beat the 3808. End of story.

I guess this boils down to the use of the receiver. If one's listening habits is more to all channels such as multichannel audio, then a more stout supply would be more advantageous. On the other hand if two channel and home theater is your thing, then putting stronger amps in the main channels would be more advantageous. The whole arguement is academic anyway because even the HK along with the Denons and the Yamahas will drive speakers into ear damagaing levels without clipping for two channels.

 

[Alex2507]

The whole argument is academic anyway ...

I'm not so sure about that. Although the main part of a piece of music can be played overly loud by a 65 watt H/K (like mine) I feel like you miss out on getting the full effects of the transient peaks as I have heard cranked out by ParadigmDawg's Rotel amp. Then there is also the issue of how they behave with lower impedance loads as with my buddy's 75 watt H/K. I was cranking Money For Nothing on that loud enough for it to shut itself off before my ears were even scared.

Even though I like teasing PENG about H/K because I think it drives him crazy I completely agree with his position on this. You never need all channels driven equally loud but you always need power on tap to handle the loudest screams in opera.

Bring on the 4310 ...

 

[jeannot]

No, I disagree as it defies logic. It would likely cost less to produce 115W in stereo and 113 in ACD than to produce 186W in 2 channel and 113W in ACD.

I do agree that when the rubber hits the road, the 3808 is a solid amp leaves a lot of the others in the dust. I would like to explain why for the same ACD power, 170 watts stereo costs less than 115 watts stereo.

In order to maintain a more stable supply voltage, a transformer must have higher gauge wiring, both primary and secondary. That costs a lot more. By using a thinner gauge wiring, all they have to do is 10 or 15 more turns of the thinner gauge on the secondary to achieve higher voltage, allowing meeting bench tests for a lot less money. But the thinner gauge has more resistance, hence the caving under high current demands.

The disadvantage of a power supply that caves under pressure is not about power, it is about the amplifier circuitry seeing constantly fluctuating supply voltage, the current regulators and feedback circuitry having to try to compensate for it all the time.

It's like reading under fluctuating lights. You still read your book, but it won't stop you from wishing they were stable.

This article is not about putting down the only electronics we can afford, just about explaining the signs of engineering compromises.

I drive a Crown Vic and I love it, but give me a Maybach and I'd send the Crown Vic to the Environment faster than you can say tow truck.

 

[PENG]

I do agree that when the rubber hits the road, the 3808 is a solid amp leaves a lot of the others in the dust. I would like to explain why for the same ACD power, 170 watts stereo costs less than 115 watts stereo.

In order to maintain a more stable supply voltage, a transformer must have higher gauge wiring, both primary and secondary. That costs a lot more. By using a thinner gauge wiring, all they have to do is 10 or 15 more turns of the thinner gauge on the secondary to achieve higher voltage, allowing meeting bench tests for a lot less money. But the thinner gauge has more resistance, hence the caving under high current demands.

The disadvantage of a power supply that caves under pressure is not about power, it is about the amplifier circuitry seeing constantly fluctuating supply voltage, the current regulators and feedback circuitry having to try to compensate for it all the time.

It's like reading under fluctuating lights. You still read your book, but it won't stop you from wishing they were stable.

This article is not about putting down the only electronics we can afford, just about explaining the signs of engineering compromises.

I drive a Crown Vic and I love it, but give me a Maybach and I'd send the Crown Vic to the Environment faster than you can say tow truck.

This time you have got the theory right, so yes I agree with what you said about the higher costs in larger transformers and yes, electronics are generally cheaper, even when factoring in the not so cheap output devices you may still be marginally correct.

However, I do have trouble with you logic. You are saying something like the

3805 - 38 lbs that tested (approx only):

132.2X2 8 ohms, 218.4X2 4 ohms, 114.9X5 8 ohms

HK 630 - 41 lbs, tested

84.6X2 8 ohms, 146.9X2 4 ohms, 79.7X5 8 ohms

HK has a more expensive/better transformer than the Denon because it's 2 channel output is closer to it's 5 channel outputs? You know your basic electrical theory right? I know you do, based on some of your other posts.

I had posted the above examples years ago to make a point about facts and figures vs hearsays. The two models were reviewed by the same person in Aug 2004 and their prices at launch time were comparable.

 

[jeannot]

However, I do have trouble with you logic. You are saying something like the

3805 - 38 lbs that tested (approx only):

132.2X2 8 ohms, 218.4X2 4 ohms, 114.9X5 8 ohms

HK 630 - 41 lbs, tested

84.6X2 8 ohms, 146.9X2 4 ohms, 79.7X5 8 ohms

HK has a more expensive/better transformer than the Denon because it's 2 channel output is closer to it's 5 channel outputs?

What I wrote is that for the same ACD power, the receiver with the lower stereo output has a better and more expensive power supply.

You do have trouble with my logic, you're right. Or, you have it out for me.

 

[mtrycrafts]

Case in point, my Denon 3808CI:
Power Supply: The receiver tests at 186 watts per channel at 8 ohms 2 channel driven, and 113 watts ACD. That translates into a drop from 54.5V to 42.5V, not something to be proud of. It would likely sound much better even at high volumes and cost MORE to produce if it gave 115 watts in stereo and 113 watts ACD.
...

What nonsense Of course that 3808 delivers 113 watts to two channels and 113 watts to all channels driven. You have your perfect amp

 

[mtrycrafts]

I guess this boils down to the use of the receiver. If one's listening habits is more to all channels such as multichannel audio, then a more stout supply would be more advantageous. ....

Multi channel audio doesn't drive the amps to full power at the same instant as some would like to believe it occurs all the time. To do that, you'd have to convert the signal to mono and send it to all the channels. Then, it could happen if it is timed correctly, no time delays.

 

[PENG]

What I wrote is that for the same ACD power, the receiver with the lower stereo output has a better and more expensive power supply.

You do have trouble with my logic, you're right. Or, you have it out for me.

For simplicity, consider amp A & B are tested for a pure resistive load of 8 ohms:

Amplifier A

ACD - 100W
2 ch - 130W

Amplifier B

ACD - 100W
2 ch - 110W

If I understand correctly you may be saying that amplifier B has a better and more expensive power supply because its voltage does not fluctuate under load, or I misunderstood?

It seems to me you are assuming that Amplifier B has less 2 ch output (110W in my example) because the power supply is holding the voltage more or less constant regardless of the load current. I would ask you to please read the HTM lab measurements such as the following, and pay attention to the graphs as well as the stereo output to ACD output ratio that you talked about:

http://www.hometheater.com/content/nad-t-747-av-receiver-ht-labs-measures

http://www.hometheater.com/content/harmankardon-avr-630-receiver-ht-labs-measures

You will notice from the graphs that the NAD and HK both have pretty much the same high 2 ch output to ACD output ratio like comparable Denon units. The reason why HTM rated their 2 ch output levels lower had to do with the 0.1% THD criteria. In other words, even the NAD models that typically have stout power supplies are capable of producing twice as much power in 2 ch (stereo) than in ACD, just that at that high output level their distortions are much higher than 0.1% that HTM uses as a criteria. So please do not be fixated on the power supply as a limiting factor, the amplifier section/output devices are also a major limiting factor. That is the point I have been trying to make, not that I disagree with your points on power supplies.

 

[3db]

I'm not so sure about that. Although the main part of a piece of music can be played overly loud by a 65 watt H/K (like mine) I feel like you miss out on getting the full effects of the transient peaks as I have heard cranked out by ParadigmDawg's Rotel amp. Then there is also the issue of how they behave with lower impedance loads as with my buddy's 75 watt H/K. I was cranking Money For Nothing on that loud enough for it to shut itself off before my ears were even scared.

Even though I like teasing PENG about H/K because I think it drives him crazy I completely agree with his position on this. You never need all channels driven equally loud but you always need power on tap to handle the loudest screams in opera.

Bring on the 4310 ...

That's the caveat of celan power. One doesn't how loud one is listening until one goes to bed and wonder why one's ears are ringing!!

 

[jeannot]

Your analysis is absolutely correct.

If I understand correctly you may be saying that amplifier B has a better and more expensive power supply because its voltage does not fluctuate under load, or I misunderstood?

Yes, that is what I am saying.

It seems to me you are assuming that Amplifier B has less 2 ch output (110W in my example) because the power supply is holding the voltage more or less constant regardless of the load current.

Yep. For better regulation, a transformer needs bigger wiring, and/or better core material (or better config like a toroid). Better regulation cost more, hence to save money you cut on regulation, raising the idle voltage and giving more stereo power.

You will notice from the graphs that the NAD and HK both have pretty much the same high 2 ch output to ACD output ratio like comparable Denon units.

I come up with an ACD/2ch ratio of 82% for the NAD, and 94% for the HK.

So please do not be fixated on the power supply as a limiting factor, the amplifier section/output devices are also a major limiting factor.
That is the point I have been trying to make, not that I disagree with your points on power supplies.

From my original post: "We all know that a power amplifier is made of a power supply and circuitry, and both of these halves have to be perfect is perfection is the target. "

We are on the same page, and may have been all along.

You do raise an interesting observation though. The 0.1% arbitrary criteria that HTM chooses will make the amp that goes out of its comfort zone appear less powerful than another one. So you are correct that they skew the readings if we only look at power. That 0.1% limit has the fortunate side effect of making a poorly designed amplifier look less powerful.

 

[M Code]

For simplicity, consider amp A & B are tested for a pure resistive load of 8 ohms:

Amplifier A

ACD - 100W
2 ch - 130W

Amplifier B

ACD - 100W
2 ch - 110W

If I understand correctly you may be saying that amplifier B has a better and more expensive power supply because its voltage does not fluctuate under load, or I misunderstood?

It seems to me you are assuming that Amplifier B has less 2 ch output (110W in my example) because the power supply is holding the voltage more or less constant regardless of the load current. I would ask you to please read the HTM lab measurements such as the following, and pay attention to the graphs as well as the stereo output to ACD output ratio that you talked about:

http://www.hometheater.com/content/nad-t-747-av-receiver-ht-labs-measures

http://www.hometheater.com/content/harmankardon-avr-630-receiver-ht-labs-measures

You will notice from the graphs that the NAD and HK both have pretty much the same high 2 ch output to ACD output ratio like comparable Denon units. The reason why HTM rated their 2 ch output levels lower had to do with the 0.1% THD criteria. In other words, even the NAD models that typically have stout power supplies are capable of producing twice as much power in 2 ch (stereo) than in ACD, just that at that high output level their distortions are much higher than 0.1% that HTM uses as a criteria. So please do not be fixated on the power supply as a limiting factor, the amplifier section/output devices are also a major limiting factor. That is the point I have been trying to make, not that I disagree with your points on power supplies.

Ok lets clear the air...
Within an AVR the most costly single component is the power transformer, that is why its specs are frequently decreased... Also the power supply capacitors enter into the cost formula, the lower the capacity and rated voltage the cheaper...

Regarding the amplifier components there are price grades for various output devices depending upon their voltage and wattage capability. For example, the HK 630 uses Toshiba output devices 2SC5358 (NPN) and 2SA1986R (PNP) in matched pairs so actually 14 pcs total are used in each AVR. HK has chosen to use these Toshiba output devices even in lower model AVRs so they can drive up the qty for lower component pricing. Since HK builds >250K AVRs means that they purchase from Toshiba 250,000 x 14 pcs which equals 3.5 million pcs. You can be assured HK gets a very competitive price for each output device.

Conclusions
The capability of an AVR's amplifier to output reliable, high power & current is dependent upon multiple factors including the power transformer, capacitors, grade of output devices and size of the amplifier's heat sink area inches. But the most significant point is their cost..
If their electric performance specs are decreased significantly, then the AVR has little chance to perform well.

Also for an HD AVR the market is expecting its selling cost to be decreasing yet in other factors such as IP royalties have gone up 2.5 times over an SD AVR. Plus the significant additional costs for video, HDMI and connectivity ICs as well..


Just my $0.02...

 

[jeannot]

The capability of an AVR's amplifier to output reliable, high power & current is dependent upon multiple factors including the power transformer, capacitors, grade of output devices and size of the amplifier's heat sink area inches. But the most significant point is their cost......Also for an HD AVR the market is expecting its selling cost to be decreasing yet in other factors such as IP royalties have gone up 2.5 times over an SD AVR. Plus the significant additional costs for video, HDMI and connectivity ICs as well..

Completely agreed (with your complete post). Your quote up here is why I will be going to separates from now on. The quality of my power amp won't go down anymore, because I will only send receivers with a pre-out to the Environment.

Looking at how the features (Audyssey, analog upconversion, networking for example) work their way through the cheaper receivers though, it appears to me that we are getting more pre/processor power for less money every year. And that may be at the cost of the amplifier section, which is what i think you are saying.

 

[jeannot]

Ok lets clear the air...

I forgot to thank you for that, it is important.
Sometimes we get carried away, a lot of us work very hard and get tired, words don't come out right. I'm certainly guilty of that sometimes.

 

[GranteedEV]

when in doubt get seven krell monoblocks.

 

[M Code]

I forgot to thank you for that, it is important.
Sometimes we get carried away, a lot of us work very hard and get tired, words don't come out right. I'm certainly guilty of that sometimes.

You are welcome..
The majority of my posts are targeted to help the audience better certain technical areas and/or features, if they did I am satisfied..
Thanks.

Just my $0.02..

 

[3db]

when in doubt get seven krell monoblocks.

WIMP!!!! I'm getting 14 mono blocks so I can bi-amp!! Oh yeah BABY!!! and open a welding shop on the side.