Understanding power amps and how they interact with speakers

[Thomas savage]

hi, having spent some time among audiophiles it seems to me there is a lot of confusion about the significance of rated power of a amp and its effects on what we hear.

Just what measured parameters should we look out for when pairing a amp with a speaker and what the hell is going on between the two?

I keep bumping into guys who think getting a huge amp to drive their tiny fragile stand mounts will give them more bass and is the most effective way to get the sound they want...

It's my intention to create a simple resource for audiophiles, to help them avoid pit holes and not waste money through total ignorance.


Please help me, and them. Keep in mind this is a resource and not a thread for idel imaginings and bizarre video links. Such unhelpful posts will simply be deleted.

Thank you

Consider me a moron and proceed accordingly

 

[Purité Audio]

I could be wrong ,and often am, I think you should be looking at the lowest impedance of your loudspeakers.
Manufacturers give a 'nominal ' impedance but it often does tell the whole story, very low impedance is difficult to drive, hence Krell and Apogee .
Essentially you need a power amplifier with enough capacitance to meet the requirements of the loudspeaker throughout its frequency range.
Conversely if you have a loudspeakers with a nice flat impedance 'curve' at high impedance ie 16 ohms that loudspeaker is going to be easy to drive and a lower power amp will be suitable.
There might be a Stereophile test of the Giyas showing their impedance plot.
Keith

 

[Thomas savage]

What measurments am I looking at regarding making sure capacitance of amp meets requirements of speaker throughout the frequency range?

I need more infomation, please members

Check this out! This question needs a definitive home.
http://www.hifiwigwam.com/showthread.php?135665-I-m-going-crazy!!!!&p=2330599#post2330599
Let's make it here

I would love to be able to post a link to here on threads like that but no point at the moment!

 

[Speedskater]

This is a topic the the old "Audio Critic" spent a lot of time writing about. It's not just low impedance loads that are a problem, it's also reactive loads that are a bigger problem (both capacitive and inductive). The Power Cube test instrument that Audio Critic used was excellent choice.
The Audio Critic archive:
http://www.biline.ca/audio_critic/audio_critic.htm

I sort of recall that maybe 5 years ago Stereophile had an article on the subject, probably written by Keith Howard.

 

[RayDunzl]

What measurments am I looking at regarding making sure capacitance of amp meets requirements of speaker throughout the frequency range?

General Old School Case Lesson:

The AC line "tops off" the capacitors as the voltage of the transformer secondary (if present) exceeds the stored voltage of the capacitor by .7V or so (to turn on the rectifier diode) and continues charging the cap until the voltage at the secondary of the transformer is no longer .7V (or so) above the voltage stored in the capacitor as the peak time of the waveform passes by.

The Capacitors have a chance to be briefly charged every 8.33ms (60HZ main) or every 10ms (50Hz main).

There are (at least) two capacitors, one charging from the positive peak, and one from the negative peak, for the + and - power rails.

So, you measure the voltage and 'ripple' on the rails to see how far the capacitors discharge between moments of charging, and the value of the voltage to give an indication of the limits to which the output transistors can 'swing' when applying that voltage to the speakers.

Ripple voltage from a full-wave rectifier, before and after the application of a smoothing capacitor

The impedance of the speaker in conjuction with the voltage applied to the speaker will determine the amount of current drawn from (and discharging) the capacitors. The capacitance of the capacitors will determine how many coulombs of electrons are stored.

Note that the nice voltage decrease seen above is indicative of a steady load. A musical signal would not represent a steady load, and the downward sloping ramp will show the increased drop on musical peaks in the waveform.

If the discharge of the capacitor impinges upon the voltage swing requested of the output transistors by the driver section, the output signal will be clipped. Here you can see the amplified signal exceeding the stored capacitor voltage, so those peaks would be clipped. (My laziness comes in here, still showing a 'steady load' ripple, it should irregularly drop a little more due to the current drawn by the applied signal)

Increased Capacitance would reduce the slope of the drop in voltage on the rails.

Switched power supplies (now common) have a higher recharge rate, and can use smaller capacitors.

Regulated Output Rails (uncommon, but mine are) should not show the same type of ripple.

 

[NorthSky]

Ω http://www.gcaudio.com/resources/howtos/spkramp.html

 

[Thomas savage]

♦ http://www.gcaudio.com/resources/howtos/spkramp.html

I don't agree with how he has determined the 'power of the speaker' and the idea you need to have a amp that's 50% more powerful than the power handling of your speakers...

Nonsensical?

 

[DonH56]

I wrote a piece on WBF showing the interaction of amplifier output impedance with speaker impedance. I have noticed that even some highly-regarded high-end speakers exhibit significant (nasty) impedance excursions that make them more sensitive to amp/speaker interaction.

The impact of power supply capacitance is tricky as it is a long-term thing and feedback mitigates its effect the vast majority of the time, at least until the rails sag too low.

How much power is "enough" is also tricky since we hear different frequencies differently (different sensitivity) and it depends upon the source. Chances are 50% distortion won't be heard if it is at 10 Hz, 20 kHz, or in the middle of an explosion. A quick clip is likely to be ignored, and is probably on something like a cymbal crash where we wouldn't notice anyway. Less than 1 % distortion can be reliably heard on single tones, especially in an ABX with somebody who knows what to listen for. Chances are we'd all hear 10% or more distortion on massed voices (like a choral) or with loud symphonic passage but I do not know other listener's sensitivity.

A lot of times a little extra distortion makes the sound "fuller" to people so they may not even know it's distortion. This is particularly true in the deep bass region where a 40 Hz tone may be hard to discern but the 120 Hz third harmonic distortion terms stands out even at a much lower level than the fundamental. Folk like the 120 Hz "bass" and feel the speaker (or whatever) that has that extra distortion sounds better, not knowing the "why".

Speaker handling is not well specified and any given guideline will likely quickly be disproven. Some speakers may handle 50% extra, many will not, but it depends upon how they specified their power rating. In addition to mechanical limits there is overheating and nonlinearity (distortion) to consider. I also posted an article on WBF showing how energy is distributed as you clip the amp. It can get a bit complicated.

IMO as I don't feel like digging up all the supporting references:
+1 dB is barely noticeable to most listeners on music or movies and requires 1.26x the power.
+3 dB is just louder to most of us, about what I do when someone asks me to bump the volume a hair, and takes twice (2x, 100% more) the power.
+10 dB in the midrange sounds twice as loud to most people and requires ten times the power.

So if you really do need more headroom then you probably need a lot more power... Gotta' love it when you go from 90 to 100 to 110 to 125 W/ch going up the line in AVRs, with the salesman touting the extra power at each step.

IME, IMO, FWIWFM, YMMV, my 0.000001 cent (microcent), etc. - Don

 

[Speedskater]

Found the Stereophile article:

Heavy Load: How Loudspeakers Torture Amplifiers
By Keith Howard • Posted: Jul 29, 2007

Read more at http://www.stereophile.com/reference/707heavy/index.html#Bgl3jugOXpycRKvb.99

 

[NorthSky]

I don't agree with how he has determined the 'power of the speaker' and the idea you need to have a amp that's 50% more powerful than the power handling of your speakers...

Nonsensical?

Does the documentation/specs that came with your hi-end speakers have a recommendation for amplifier power?

* Mine (main F/L speakers), recommended amplifier power is between 50 and 350 watts. I blew an integrated amp before, rated @ 40 watts per channel.
Usually it's preferable to go in the high range of that recommendation. My speakers dip to 4Ω in the critical regions of the audio spectrum; bass and highs.

We normally only use few watts (say from 2 to 10), but instantaneous peaks can be be in the thousands.
Klipsch speakers they don't need lots of power. Some stats need powerhouse monoblocks.
Sensitivity and impedance. Average speaker's sensitivity (all stereophile speaker's reviews) = approx. 87dB.

But Tom you know that stuff already.
_______

Some guys have speakers that dip lower than 1Ω! They don't buy a Denon receiver for them.
Your Vivid speakers Tom, do you give them the proper diet? ...Clean and adequate current...I'm sure you do.

 

[NorthSky]

Oh, and the majority of class A/B amps waste 50% of their efficiency in heat.

 

[RayDunzl]

Oh, and the majority of class A/B amps waste 50% of their efficiency in heat.

If you look at the amount of Acoustic Power actually radiated, virtually all of the electrical power supplied goes up in heat.

 

[Thomas savage]

Does the documentation/specs that came with your hi-end speakers have a recommendation for amplifier power?

* Mine (main F/L speakers), recommended amplifier power is between 50 and 350 watts. I blew an integrated amp before, rated @ 40 watts per channel.
Usually it's preferable to go in the high range of that recommendation. My speakers dip to 4Ω in the critical regions of the audio spectrum; bass and highs.

We normally only use few watts (say from 2 to 10), but instantaneous peaks can be be in the thousands.
Klipsch speakers they don't need lots of power. Some stats need powerhouse monoblocks.
Sensitivity and impedance. Average speaker's sensitivity (all stereophile speaker's reviews) = approx. 87dB.

But Tom you know that stuff already.
_______

Some guys have speakers that dip lower than 1Ω! They don't buy a Denon receiver for them.
Your Vivid speakers Tom, do you give them the proper diet? ...Clean and adequate current...I'm sure you do.

I drive them with this...,

Belles MB-200 Monoblock Amplifier specifications are as follows:

Monoblockpower ....................200wattsinto8ohms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .400 watts into 4 ohms

Distortion ..........................THDandIMlessthan0.005% DampingFactor .....................over2000
PeakCurrent .......................64amperes
Inputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Unbalanced only

Gain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 or 26.44 dB
InpuySensitivity .....................1.9voltsforratedpower
InputImpedance ....................100,000ohms
Hum&Noise .......................over100dB
FrequencyResponse .................0.2to125,000Hz
Dimensions ........................17"(432mm)Wx3.5"(88.9mm)Hx13"

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .(330 mm) D not including feet

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .and binding posts ShippingWeight .....................37pounds(16.8Kg) PowerRequirements .................120VAC50/60Hz6.3AFuse

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .240 VAC 50/60 Hz 3.15 A Fuse

 

[Thomas savage]

Then maybe best to get a class D amplifier with 1,000 RMS watts a side?
_________

• http://www.stereophile.com/content/matching-amps-and-loudspeakers-dummieslike-me-1
• http://www.stereophile.com/features/99/index.html#RoGRpP5WCL85RVgG.97
_________

Ray, your Martin Logan speakers they like power right? And you support that love with plenty right? ...500 Watts a side?
_________

Both that article and the designer of my speakers ( who is also a ee and has designed amps) approve of my amplifier choice

This is more about helping others..

 

[Fitzcaraldo215]

Oh, and the majority of class A/B amps waste 50% of their efficiency in heat.

It is far worse for class A. That is one reason why class D will continue on the ascendancy. Actually, I replaced some monster class A monoblocks several years ago with a class D. The drop in my electrical bill was quite obvious. Actually, it sounded better, too.

My ecological system of the future will use hamster wheel driven generators for power to class D amps. I understand that the hamsters deliver the best sound compared to lab rats or other small mammals. Although, this setup is still plagued by momentary dropouts, I understand. I need to get a good audiophile-approved whip to avoid that.

 

[Thomas savage]

It is far worse for class A. That is one reason why class D will continue on the ascendancy. Actually, I replaced some monster class A monoblocks several years ago with a class D. The drop in my electrical bill was quite obvious. Actually, it sounded better, too.

My ecological system of the future will use hamster wheel driven generators for power to class D amps. I understand that the hamsters deliver the best sound compared to lab rats or other small mammals. Although, this setup is still plagued by momentary dropouts, I understand. I need to get a good audiophile-approved whip to avoid that.

My bill dropped a bit when I went from krell FPB 300 to my class a/b Belles monos

 

[RayDunzl]

Ray, your Martin Logan speakers they like power right? And you support that love with plenty right? ...500 Watts a side?

Recommended Amplifier Power 20‒500 watts per channel

My amps are rated for 700@4 ohms, but the wall outlet's breaker will pop before those levels are ever seen here.

The amps claim 1700W maximum draw, they are both on a single 1800W circuit (15A x 120V) along with all the other AV gear and some of the kitchen.

If I listen while the kitchen overhead tubes (240W) and the PC (300W or so) are on and run the vacuum cleaner (750W) the breaker has popped.

I figure a big double handful of watts going to the speakers is about right for loud listening here.

(They aren't green, I know - 100W each at idle, and good for a couple of watts output, with, I think, seven levels of increasing bias, and corresponding increases in draw and output above that. I usually listen with occasional jumps into bias level 2, 3 is loud 4 is real loud. The highest draw I've seen is 500W each, and they get warm quickly)

 

[RayDunzl]

Doubling the power into 4Ω is the sign of a true serious amplifier. And doubling again into 2Ω, and again into 1Ω

I see that as a ratings trick, that is, under -rating the power available at the higher impedances.

I have an amp rated at 250/350 8/4. It does not "double down"

If I re-rate it at 175/350 8/4 it suddenly looks like it "doubles down"

Example: Krell FPB 600 - 600/1200/2400 - as tested it produced 945W into 8, but is under-rated to 600 so it can appear to double since that makes good ad copy.

In the case of my 250/350 amp, they were probably more concerned with the higher 8 ohm rating for their intended audience.

 

[RayDunzl]

doubling the power each time the impedance is halved, from measurements graphs, is serious power amplifier supplies.

Maybe at the limits, but it's not so spectacular at "normal" levels... If your listening spends 10W on an 8ohm speaker, then you will spend 20W if you replace that speaker with a 4 ohm job.

You can do that with most any reasonably competent amplifier, I would think.

 

[Thomas savage]

My amp..,