Power amp question

[DonH56]

If you run the room correction program in the AVR it should compensate for the gain difference. Otherwise, adding an amp may or may not change the volume depending upon the gain of the amp compared to the gain of the internal amp.

Most people can reliably detect well under 1 dB when switching between components. But bumping the volume on a given component takes about 1 dB to be noticeable.

I have six or seven AVRs around the house (yes, pales to the component count of John and others). One, a Sony ES, really struggled driving a 5.1 system, and its multichannel rating was way below the 2 ch rating. All the others did fine with nominal 8-ohm'ish speakers and didn't noticeably degrade with multiple channels driven. Sighted testing, no measurements. I only noticed because the Sony was really poor compared to the rest.

Others have already discussed the potential benefits of adding an amp. IME expectation bias is very strong and I suspect most folk buying an external amp for their AVR do not need it. But this is the audio world wherein the exception makes the rules so everyone must have an external amp to play... And it is the easiest "upgrade" for many people whether it helps or not. I would guess most of those folk would be better off adding room treatments or upgrading their speakers.

 

[Sancus]

The most likely thing you'd notice adding a power amp is having more power, especially since most AVRs are not really designed to drive 4 ohm speakers. I regularly overheated my Denon X4000 before adding an external power amp, for example. There is no audible quality difference when playing at middling volumes, though.

 

[noobie1]

I have 4ohm speakers that need a lot of current. I went from a Denon AVR that has been measured to output at about 130W per channel in stereo mode at 4 ohms (I believe 1% THD). When I first plugged in my 400W per channel Classe amp, the sound that came out of the speakers was night and day different from the AVR. Certain passages of music require more transient power than the conventional averaged numbers. I think Amir has said this before but people typically underestimate the amount of wattage they need for music listening.

 

[Dimitrov]

I have 4ohm speakers that need a lot of current. I went from a Denon AVR that has been measured to output at about 130W per channel in stereo mode at 4 ohms (I believe 1% THD). When I first plugged in my 400W per channel Classe amp, the sound that came out of the speakers was night and day different from the AVR. Certain passages of music require more transient power than the conventional averaged numbers. I think Amir has said this before but people typically underestimate the amount of wattage they need for music listening.

But you are my case in point. Night and day difference, immediately. Simply adding more power shouldn't make a night and day difference unless you exceeded the limits of your previous amp.

Something else is going on here.

If you don't use the extra power it shouldn't make any difference at all. Its just unused. So my suspicion is that there was gain difference, which you interpreted as a sound quality difference.

 

[Blumlein 88]

I have 4ohm speakers that need a lot of current. I went from a Denon AVR that has been measured to output at about 130W per channel in stereo mode at 4 ohms (I believe 1% THD). When I first plugged in my 400W per channel Classe amp, the sound that came out of the speakers was night and day different from the AVR. Certain passages of music require more transient power than the conventional averaged numbers. I think Amir has said this before but people typically underestimate the amount of wattage they need for music listening.

If you have a multimeter with peak hold, it might be nice to put it across the speaker terminals and play a few songs. See what the peak voltage asked of your amp is. That technique isn't optimum, but it is something to work with.

 

[Krunok]

If you have a multimeter with peak hold, it might be nice to put it across the speaker terminals and play a few songs. See what the peak voltage asked of your amp is. That technique isn't optimum, but it is something to work with.

Aren't multimeters set to measure voltage correctly only when frequency is 50Hz?

 

[DonH56]

Aren't multimeters set to measure voltage correctly only when frequency is 50Hz?

Many of the cheapest meters target power-line AC and so have a bandwidth of only 100 to 150 Hz or so. But most higher-quality (esp. true RMS) meters will cover the audio range and beyond. The midlevelish Fluke 179 goes to 100 kHz, for example. The problem is many DMMs do not specify their frequency range.

 

[Sangram]

If you have a multimeter with peak hold, it might be nice to put it across the speaker terminals and play a few songs. See what the peak voltage asked of your amp is. That technique isn't optimum, but it is something to work with.

The current into the speaker is a far better measure of output stage requirements than voltage, as the impedance varies wrt frequency. True RMS meters, which can measure high frequencies, are not optimal either because they use zero-to-zero averages and have problems when the sample-to-sample variance is too high (of course given we are talking about music signals, not sinusoid/etc). A music signal is usually not either sinusoid, triangle, square or easily deducible waveshape to begin with, and the zero crossings may not be completed for a single frequency before something else modulates it.

In actual fact an appropriately attenuated oscilloscope is probably the best bet, with both current and voltage being measured and the entire waveform recorded (eg using Picoscope). The reactance of the speaker may cause current draw to be different from (i.e. more than) what Ohm's Law will mean, because some lag effects can be observed when measuring speakers even with sine waves. Negative impedance phase (transition to capacitative) may cause large levels of current and inductive phase may see some voltage peaking. With a poorly designed crossover this can have some profound and unpredictable effects on the measuring instruments as well as the driving amplifier.

The other issue is dynamic headroom - to produce your rated SPL @listening position, let's say you need 8W, a reasonable enough figure (though little high, in practice 6W is usually enough for most speakers). Once you add a 20dB peak you are at 80W, and most receivers will be gasping for breath already. Even some dedicated power amplifiers will not be capable of sustained operation here. Mercifully for movies peaks tend to be short in duration, but this is not a worst case scenario anyway.

 

[GGroch]

Dimitrov, you started an interesting topic here...but it is really more than one topic, and I think this has caused some confusion. Perhaps breaking it down a little would help.

1. Does substituting separate power amps for the built in amplification really improve performance?

The answer to this is "it depends"...but in most cases No. It depends mostly on the quality of the receiver, and whether the interaction of speakers, room, and listening levels result in audible clipping. If you choose an AV receiver that matches your gear needs, the power amp is unlikely to make any real difference.

RE: Quality of receiver, Amir has not measured many multi-channel AV receivers, but did did measure the Denon AVR-4306 This is a 13 year old mid-line receiver available used for less than $200. The results, tested only in 2 Channel Stereo using "Direct" mode (no processing) are good enough to be considered almost totally transparent....certainly with any speakers that it is likely to be paired with. You cannot hear .01% SNAD & 79dB+ S/N in normal environments, and 183 WPC is far more than sufficient in most living rooms. The major variable is that the WPC will be significantly lower using full surround mode...but I still think it is very unlikely to clip in appropriate use.

2. Why do so many people hear huge improvements when substituting separate power amps for built in amplification?

This is far easier to explain. Many people hear huge differences in DACS, Speaker Cables, and even Power cords when very little and in most cases NO difference actually exists. Major industries have flourished based on expectation/ confirmation bias...and there are a lot of threads on ASR discussing this.

The effects of confirmation bias is strongest where quick & valid A/B/X testing is hard to achieve. Substituting a power amp takes a least a few minutes, and because volume re-calibration will almost always be required, often takes more than an hour. Our memory for audio lasts no where near that long.

I do not see how SPL meters have any impact on this bias. Certainly a setting of "6" on the volume control will nearly always result in a different volume when amps are changed...but whether it is louder or softer would depend on the input gain of the internal vs external amps. It is just as likely to be softer than louder. Plus, the hour or so it takes to make the switch would eliminate the louder is better bias you hear when say, auditioning speakers at an audio store through a switcher.

Confirmation/Expectation bias for our listening experiences is incredibly strong, and impacts all of us...it cannot be avoided. It does not prove we are gullible, just that we are human.

 

[Dimitrov]

The other issue is dynamic headroom - to produce your rated SPL @listening position, let's say you need 8W, a reasonable enough figure (though little high, in practice 6W is usually enough for most speakers). Once you add a 20dB peak you are at 80W, and most receivers will be gasping for breath already. Even some dedicated power amplifiers will not be capable of sustained operation here. Mercifully for movies peaks tend to be short in duration, but this is not a worst case scenario anyway.

Using bass management and subwoofers can help tremendously. It can often mean the difference between the amp clipping like crazy and coasting along without a worry in the world.

 

[Blumlein 88]

Aren't multimeters set to measure voltage correctly only when frequency is 50Hz?

It depends. From what I've seen most any will do 400 hz (because some military and aviation gear uses 400 hz AC or did in the past). Plenty will do a couple khz flat. As most power is in the lower few hundred hz one that works to 400 hz should do for this purpose. I've tested a few recent cheapish meters off Amazon. All were good to at least 1000 hz.

If you have doubts you could check it by feeding DAC output to it with test tones. See what frequency begins to drop off on the meter.

 

[Blumlein 88]

The current into the speaker is a far better measure of output stage requirements than voltage, as the impedance varies wrt frequency. True RMS meters, which can measure high frequencies, are not optimal either because they use zero-to-zero averages and have problems when the sample-to-sample variance is too high (of course given we are talking about music signals, not sinusoid/etc). A music signal is usually not either sinusoid, triangle, square or easily deducible waveshape to begin with, and the zero crossings may not be completed for a single frequency before something else modulates it.

In actual fact an appropriately attenuated oscilloscope is probably the best bet, with both current and voltage being measured and the entire waveform recorded (eg using Picoscope). The reactance of the speaker may cause current draw to be different from (i.e. more than) what Ohm's Law will mean, because some lag effects can be observed when measuring speakers even with sine waves. Negative impedance phase (transition to capacitative) may cause large levels of current and inductive phase may see some voltage peaking. With a poorly designed crossover this can have some profound and unpredictable effects on the measuring instruments as well as the driving amplifier.

The other issue is dynamic headroom - to produce your rated SPL @listening position, let's say you need 8W, a reasonable enough figure (though little high, in practice 6W is usually enough for most speakers). Once you add a 20dB peak you are at 80W, and most receivers will be gasping for breath already. Even some dedicated power amplifiers will not be capable of sustained operation here. Mercifully for movies peaks tend to be short in duration, but this is not a worst case scenario anyway.

I agree with all this, but not many people have an O-scope around anymore. Lots of people have a multi-meter. Current would be a better gauge and current and voltage better still. I was simply pointing out one method to get some idea. That is why I described the technique as not optimum.

 

[DonH56]

Since most amps act more like voltage sources than current sources getting an average power reading using pink noise and a DMM is not unreasonable IME/IMO. No, it won't provide an accurate power reading, but if you know your amp's power rating into 8 ohms or whatever you can figure out the voltage limits to see how close you are to (voltage) clipping.

The last time I tried to do this "right" I used a current sensor (can't recall if basic EM clamp or a Hall effect device) and voltage probe to sense the output voltage and current without having to add a series R and get current from that (challenging given the currents and impedances involved for high-power systems like audio). I then did math in the 'scope to calculate peak and average power from the peak and RMS current and voltage.

 

[rwortman]

I am doing this now. I have a very nice Yamaha RX-A1070 AVR. Before decided to put together my first surround system(two AVR's ago), I was listening to two channel stereo using a preamp and an NAD 150wpc power amp. I decide to include the amp in the surround system for two reasons. One is what has previously been mentioned that AVR power outputs are rated with two channels driven and fall off rather drastically when more channels are added. Two was that I had it, liked it and wanted to continue to use it and the front L/R speakers with two channel gear that was not connected to the AVR. A simple two way switchbox connected to the amp input let's me do this. I used the calibration functions of the AVR so there isn't a level difference from using an external amp. There is a largish available power difference but I am not sure I am using it. I have not attempted at A/B comparison because I don't care enough to spend an evening swapping speaker cables around. I suspect any difference if any would not be of the night and day variety. Most AVR's that have pre-outs are powerful enough that this shouldn't be the case. It is possible that certain models are anemic enough with all channels driven that in very loud situations there is a significant difference. I an not even certain that my dedicated two channel set up sounds any different than the AVR in stereo mode. I just like the idea of playing my turntable using a traditional two channel setup and since I have it anyway, why not add a CD player and a streamer/DAC. Audiophiles like toys that play music.

 

[Dimitrov]

I know some of you have said that folks will use the autocalibration setting to account for level differences but how many do that really? How many after just buying their new power amp?

I think it's highly doubtful people who report hearing these (often) "night and day" differences have even done a calibration. I don't think it even crosses their mind.

 

[Dimitrov]

I did some homework looking at the bench test data here :

https://www.soundandvision.com/category/power-amplifier-reviews

"Krell Chorus 7200, $9.5K: "An input level of 148.0 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +25.66 decibels using the RCA input. When using the XLR input, a level of 147.7 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +25.68 decibels." NOTE: RCA and XLR inputs are slightly different in some amps and in others very different such as the Marantz below.

Monoprice Monolith 7 $1.5K: "An input level of 103.3 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +28.75 decibels. "

Emotiva A-500 amp, ~$500 [part of their review of package system "Emotiva BasX System": "An input level of 97.90 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +29.24 dB. "

Sherbourn PA 7-350, $2.8k: "An input level of 52.2 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +34.70 decibels using the RCA input. When using the XLR input, a level of 52.5 mV was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +34.67 dB. "

Marantz MM8077 $2.4K:"An input level of 100.8 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +28.99 decibels using the RCA input. When using the XLR input, a level of 195.0 mv was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +23.28 dB. "

If we limit our examination to RCA connections of these random examples, they vary tremendously from a low of +25.66 dB to a high of +34.70 dB, a huge and clearly audible 9dB swing. That's crazy. That could easily account for folks hearing differences (I mean improvements).

This ignores expectation bias and memory, but I think the overriding factor by a large margin is that the levels are (based on the examples above) usually large enough to easily result in sound quality differences and NOT necessarily because the amps have better quality power/or more headroom, simply because of gain differences.

 

[rwortman]

I know some of you have said that folks will use the autocalibration setting to account for level differences but how many do that really? How many after just buying their new power amp?

I think it's highly doubtful people who report hearing these (often) "night and day" differences have even done a calibration. I don't think it even crosses their mind.

Most people couldn't program a VCR. It's unlikely that those same people are calibrating their AVR's. I helped one guy who told me he bought his second AVR after having left channel problems with the first one and now the second one was doing the same thing. I asked him how it was connected. Analog outs from the TV into stereo analog inputs on the AVR. WHAT! I told him to connect the HDMR from the cable box into his AVR and then HDMI from the AVR video out to the TV. He was skeptical but the next day reported how awesome it worked. He had purchased a full 5.1 surround system and never heard anything but the stereo output from the DAC in his TV.

 

[noobie1]

Dimitrov, you started an interesting topic here...but it is really more than one topic, and I think this has caused some confusion. Perhaps breaking it down a little would help.

1. Does substituting separate power amps for the built in amplification really improve performance?

The answer to this is "it depends"...but in most cases No. It depends mostly on the quality of the receiver, and whether the interaction of speakers, room, and listening levels result in audible clipping. If you choose an AV receiver that matches your gear needs, the power amp is unlikely to make any real difference.

RE: Quality of receiver, Amir has not measured many multi-channel AV receivers, but did did measure the Denon AVR-4306 This is a 13 year old mid-line receiver available used for less than $200. The results, tested only in 2 Channel Stereo using "Direct" mode (no processing) are good enough to be considered almost totally transparent....certainly with any speakers that it is likely to be paired with. You cannot hear .01% SNAD & 79dB+ S/N in normal environments, and 183 WPC is far more than sufficient in most living rooms. The major variable is that the WPC will be significantly lower using full surround mode...but I still think it is very unlikely to clip in appropriate use.

2. Why do so many people hear huge improvements when substituting separate power amps for built in amplification?

This is far easier to explain. Many people hear huge differences in DACS, Speaker Cables, and even Power cords when very little and in most cases NO difference actually exists. Major industries have flourished based on expectation/ confirmation bias...and there are a lot of threads on ASR discussing this.

The effects of confirmation bias is strongest where quick & valid A/B/X testing is hard to achieve. Substituting a power amp takes a least a few minutes, and because volume re-calibration will almost always be required, often takes more than an hour. Our memory for audio lasts no where near that long.

I do not see how SPL meters have any impact on this bias. Certainly a setting of "6" on the volume control will nearly always result in a different volume when amps are changed...but whether it is louder or softer would depend on the input gain of the internal vs external amps. It is just as likely to be softer than louder. Plus, the hour or so it takes to make the switch would eliminate the louder is better bias you hear when say, auditioning speakers at an audio store through a switcher.

Confirmation/Expectation bias for our listening experiences is incredibly strong, and impacts all of us...it cannot be avoided. It does not prove we are gullible, just that we are human.

I’m just a single data point. I have a bunch DACs of various shapes and sizes. From $50 to $800. Every DAC above $100 sounded very similar to me. I really had to squint to hear differences.

This is not the sort of sonic difference I perceived when going from a 130W receiver to a 400W amp. It was night and day. Like an anvil hitting you on the forehead.

 

[noobie1]

I did some homework looking at the bench test data here :

https://www.soundandvision.com/category/power-amplifier-reviews

"Krell Chorus 7200, $9.5K: "An input level of 148.0 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +25.66 decibels using the RCA input. When using the XLR input, a level of 147.7 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +25.68 decibels." NOTE: RCA and XLR inputs are slightly different in some amps and in others very different such as the Marantz below.

Monoprice Monolith 7 $1.5K: "An input level of 103.3 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +28.75 decibels. "

Emotiva A-500 amp, ~$500 [part of their review of package system "Emotiva BasX System": "An input level of 97.90 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +29.24 dB. "

Sherbourn PA 7-350, $2.8k: "An input level of 52.2 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +34.70 decibels using the RCA input. When using the XLR input, a level of 52.5 mV was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +34.67 dB. "

Marantz MM8077 $2.4K:"An input level of 100.8 millivolts was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +28.99 decibels using the RCA input. When using the XLR input, a level of 195.0 mv was required to produce an output of 2.83 volts into an 8-ohm load, indicating an overall gain of +23.28 dB. "

If we limit our examination to RCA connections of these random examples, they vary tremendously from a low of +25.66 dB to a high of +34.70 dB, a huge and clearly audible 9dB swing. That's crazy. That could easily account for folks hearing differences (I mean improvements).

This ignores expectation bias and memory, but I think the overriding factor by a large margin is that the levels are (based on the examples above) usually large enough to easily result in sound quality differences and NOT necessarily because the amps have better quality power/or more headroom, simply because of gain differences.

I frequently use an iPhone app to check decibel levels. It’s prob not the most ideal way to go about things but you can have a ballpark idea whether the volume is level matched.

 

[Jaxer]

I'm new here, but, just my two cents,

Hi, I have owned a custom T-Amp - https://www.amazon.com/Tripath-Stereo-Subwoofer-Amplifier-Digital/dp/B00VAC7M7G since 2006-2007 I bought the one that had all the rave reviews back then and then gutted it like many did and built my own chassis, PSU, matched resistors instead of a volume pot, etc.

They sound amazing and only have 1 drawback. Low power. For computer speakers they are PERFECT!!! I power 4Ohm Vienna Acoustics Haydn's which require a ton of power and 4Ohms is very hard on stereo equipment.

According to amazon, that thing has a higher wattage than my chip but... that doesn't really mean anything since watts don't exist without a @THD rating so who actually knows what the wattage is with it having the THD rating behind it.

Either way at the distance you will have these speakers I believe that amp will be AMAZING and you will be shocked. Mine creates clear crisp deep bass but when the speakers are only a few feet from your ears it doesn't take a lot of power.

Also, the power supply and caps are part of the bass response. If you are good at soldering you can tear that thing apart in a few years and upgrade it if you want to.

I do not recommend the D-Class amp though no offense to Evilsizer I am sure he knows what he is talking about a trip at t class chip is going to sound much better. If you were looking to power a subwoofer then I would say have at it.

T-Class amps are d-class but there are differences that make them much better. There are amazing D-Class amps out there and I have no idea whether the one Evilsizer recommended is good or not.

What I can tell you is that there are plenty of reviews of TP2020 based T-Amps - https://stereodamage.com/top-5-power-amp-for-axe-fx-tested-and-reviews/ that are done by audiophile magazines that have plenty of good to say about them. Like I said I think it is the same chip that started all the rave and craze when they first came out. They disappeared off the market for a while. I bought mine right when tripath stopped making or slowed the production of the chips. Since then they have started making them again.

Yesterday I found an excellent video on this topic on YouTube, maybe this will help someone in the future