AVS Forum banner
1 - 20 of 21 Posts

·
Banned
Joined
·
14,420 Posts
Discussion Starter · #1 ·
Here's a real world data point for the controversy about power amps needing to be able to put out twice the power for half the impedance:


I ran some bench tests on what may be one of the most overbuilt power amps in the history of audio - the Threshold SA/4e from the early 1990s.


No visible effort was spared to build a robust power amp, as this article shows:

http://www.thresholdlovers.com/artic...?lng=en&pg=394


The Threshold SA/4e weighs about 120 pounds, according to my calibrated biceps. You really want two people to tote it around, and it facilitates this by having two comfortable handles on each end.


In contrast, in similar tests a Behringer A500 which weighs about 15 pounds puts out about the same amount of power at clipping at 8 ohms, but puts out about 144 watts at both 4 and 2 ohms. This is what I would expect from a typical contemporary mainstream amplifier.


Among other features, the Threshold SA/4e amplifier has extremely soft clipping. In the end I decided to measure its output power at about 0.1% THD @ 1KHz.


The actual measured power is as listed here:

http://cgi.audiogon.com/cgi-bin/auc....82027&auc&3&4&


Namely 112 Watts per channel @ 8 ohms with both channels driven @ 1000 hz, 196 Watts @ 4 ohms, 264 Watts @ 2 ohms.


This is still very far from doubling power as load impedance halves, particularly at 2 ohms.
 

·
Registered
Joined
·
1,508 Posts
Sorry, but what controversy?


An "ideal" amp doubling it's output into half the resistance is according to Ohm's Law. The most basic equation in electrical engineering.


Power = (volts*volts)/Resistance.


So if you halve the resistance, the power should double. This isn't a "need", it's an ideal. But like many things, the ideal and the reality are two different things.


As I've often said there are few amps that will do it into 4 ohms and very, very few that will do it into 2 ohms. Anyone who is expecting this to happen, is going to be disappointed.


The Threshold is indeed a monster, but it is also a pure class A amp, which are extremely inefficient. It's rated for 100 wpc, so it does essentially achieve doubling it's power into 4 ohms. I think they might be hitting limits of Class A technology and their design with their 2 ohm output. They might well be understating their 8 ohm rating, but I'm not going complain about them slightly understating a power rating. Though measurements @ 1khz are known for overstating actual power.
 

·
Banned
Joined
·
14,420 Posts
Discussion Starter · #3 ·

Quote:
Originally Posted by TPnBobcats /forum/post/17018515


Sorry, but what controversy?

The once that you clearly don't see:


Quote:
An "ideal" amp doubling it's output into half the resistance is according to Ohm's Law. The most basic equation in electrical engineering.


Power = (volts*volts)/Resistance.


So if you halve the resistance, the power should double.

Agreed.

Quote:
This isn't a "need", it's an ideal. But like many things, the ideal and the reality are two different things.

Not everybody seems to agree with your view. I tend to agree with you, but I'm probably not very agressive on this issue in the larger picture.

Quote:
As I've often said there are few amps that will do it into 4 ohms

I've said that on the test bench, there are none. Some may come close, but none can go all the way to the theoretical ideal. There are theoretical reasons why no simple power amp will go all the way.

Quote:
and very, very few that will do it into 2 ohms. Anyone who is expecting this to happen, is going to be disappointed.

Again, I tend to agree. In fact a real world power amp could be constructed that tested this way out on the test bench (not just specsmanship), but it would not be simple.

Quote:
The Threshold is indeed a monster, but it is also a pure class A amp, which are extremely inefficient. It's rated for 100 wpc, so it does essentially achieve doubling it's power into 4 ohms.

The Threshold SA/4e comes close, but no cigar. It falls apart trying to do 2 ohms.

Quote:
I think they might be hitting limits of Class A technology and their design with their 2 ohm output.

Class A does not necessarily cause this kind of a limit. The amp could smoothly transition from Class A to AB, and only us guys with meters and schematics would be the wiser.

Quote:
They might well be understating their 8 ohm rating, but I'm not going complain about them slightly understating a power rating.

Ordinarily, power amps put out more at clipping at 1 KHz than 20 Hz. Therefore, being somewhat underated at 1 KHz makes sense. I did some testing of this kind on the Threshold and it seemed to be very unusual. It seemed to deliver far more power at 20 Hz than at 1 KHz, and at all impedances tested. Incipient tripping of my ower line breaker stopped further investigations.

Quote:
Though measurements @ 1khz are known for overstating actual power.

That's because it is usual for the amp's power supply to have a little more ripple when powering a 20 Hz signal.
 

·
Registered
Joined
·
1,508 Posts

Quote:
Originally Posted by arnyk /forum/post/17018697


The once that you clearly don't see:

Apparently. Though I can't recall this being an issue in any amp thread I've ever seen.

Quote:
Originally Posted by arnyk /forum/post/17018697


Class A does not necessarily cause this kind of a limit. The amp could smoothly transition from Class A to AB, and only us guys with meters and schematics would be the wiser.

Not inherently, it doesn't. I was thinking more along the lines of because pure class A is so inefficient and produces so much heat, that may well impose it's own set of limitations.

Quote:
Originally Posted by arnyk /forum/post/17018697


I did some testing of this kind on the Threshold and it seemed to be very unusual. It seemed to deliver far more power at 20 Hz than at 1 KHz, and at all impedances tested. Incipient tripping of my ower line breaker stopped further investigations.

That is interesting, but my understanding of things was that the overstatement of power had more to do with it being easier to amplify a single frequency than a range of frequencies. It's not impossible that the difference you tested had to do with your "load". Since with speakers the impedance varies over the frequency range.


It is funny however, how little people understand that the line power is as great a limitation on their amps as anything else. You can't hit the rated power on a lot of the really big amps even with a dedicated 20 amp line.
 

·
Banned
Joined
·
14,420 Posts
Discussion Starter · #5 ·

Quote:
Originally Posted by TPnBobcats /forum/post/17018815


Apparently. Though I can't recall this being an issue in any amp thread I've ever seen.

Well even just AVS is a big place, and there's a far larger world out there. Certainly no fault of yours.


I've also seen people who did not think it was a requirement, but were impressed by it. Krell is the brand most frequently mentioned. By the first time I saw this kind of discussion praising Krell, I had previously read a report from a Krell owner of what its true bench performance was.

Quote:
Not inherently, it doesn't. I was thinking more along the lines of because pure class A is so inefficient and produces so much heat, that may well impose it's own set of limitations.

You're right about that. The heat can be managed with giant heat sinks and even forced air. What is harder to manage is the loss of SOA reserve. In the case of the big Threshold, they used over 20 OPTs per channel.


Quote:
my understanding of things was that the overstatement of power had more to do with it being easier to amplify a single frequency than a range of frequencies

Its more about the nature of that single frequency, 1 KHz. I KHz is high enough to be easy on the power supply, and low enough to be easy on the various sources of high frequency loss.

Quote:
It's not impossible that the difference you tested had to do with your "load". Since with speakers the impedance varies over the frequency range.

This load was composed of a bank of 8 8 ohm 250 watt 1% NI resistors. I've got a lot of experience with testing on both resistive loads and also loads that fairly closely duplicate some of the more challenging but still reasonable speakers out there. I don't think it is fair to test amps with loads that go under 1 ohm at 20 KHz, for example. 3 ohms, no problem.


At any rate, speakers are generally easier loads than resistors, as long as the amp has good performance into reactive loads. Reactive load performance has long been due to a secondary performance issue in power transistors which has been addressed increasingly well over the years.

Quote:
It is funny however, how little people understand that the line power is as great a limitation on their amps as anything else. You can't hit the rated power on a lot of the really big amps even with a dedicated 20 amp line.

Yes, and I'm in the midst of the installation of a 30 amp 120/208 line to the room where I do amp tests.
 

·
Registered
Joined
·
1,508 Posts
This is OT, but you seem to know your stuff. There's an argument raging in the emo thread with a "bomb thrower" about bi-wiring. He's claiming that figure 3 in this link here.

http://www.st-andrews.ac.uk/~www_pa/...ire/Page1.html


Is a "parallel" bi-wiring. Which is what it's described as.


Now my days with this sort of thing are long past, but from my dim memories of such things, while it might be described as "parallel" I don't think it really is. IE. it isn't a real parallel connection and thus won't halve the resistance, but don't feel like digging up my old EE books to make sure.


Am I correct in this?
 

·
Registered
Joined
·
18,923 Posts
I read the first post, and I am not sure what I am supposed to get from it? You are comparing a 120 pound amplifier that probably cost 1000's of dollars to a cheap pro amp?


We would expect the pro amp to fall short, right?
 

·
Registered
Joined
·
18,923 Posts

Quote:
Originally Posted by TPnBobcats /forum/post/17022737


This is OT, but you seem to know your stuff. There's an argument raging in the emo thread with a "bomb thrower" about bi-wiring. He's claiming that figure 3 in this link here.

http://www.st-andrews.ac.uk/~www_pa/...ire/Page1.html


Is a "parallel" bi-wiring. Which is what it's described as.


Now my days with this sort of thing are long past, but from my dim memories of such things, while it might be described as "parallel" I don't think it really is. IE. it isn't a real parallel connection and thus won't halve the resistance, but don't feel like digging up my old EE books to make sure.


Am I correct in this?

Far as I know, bi wiring is a parallel circuit. You know in EE, two wires are assumed to behave the same unless there's a good reason for them not to. I think EE's laugh at the concept of biwring improving audio. I am sure I have read a few mention that it makes no difference. Who do you trust? Talented EEs who have measured stuff, or golden eared audiophiles?
 

·
Registered
Joined
·
18,923 Posts
I remembered something. Some amplifier makers supposedly lower their power ratings just so that they can hit double power. This is according to a few articles. I don't know if it's true though.
 

·
Banned
Joined
·
14,420 Posts
Discussion Starter · #10 ·

This is OT, but you seem to know your stuff. There's an argument raging in the emo thread with a "bomb thrower" about bi-wiring. He's claiming that figure 3 in this link here.

http://www.st-andrews.ac.uk/~www_pa/...ire/Page1.html


Is a "parallel" bi-wiring. Which is what it's described as.

he can call it what he wants, but I call it no bi-wiring at all. Its using two paralleled cables as one.

Quote:
Now my days with this sort of thing are long past, but from my dim memories of such things, while it might be described as "parallel" I don't think it really is. IE. it isn't a real parallel connection and thus won't halve the resistance, but don't feel like digging up my old EE books to make sure.

I hate to disagree with you, but the gods of technology especially Ohm and Kirchoff say you are wrong.


No way is it bi-wiring. Its powering a speaker through two cables in parallel. The effective cable resistance is half of the resistance of either cable if they are identical.


In figure 3 the jumpers at the speaker end parallel the two cables, and also have the same basic effect as the shorting bars that most speakers set up for biwiring come with pre-installed.


BTW I know the author Jim Leserf, as he posts on the Usenet uk.rec.audio. He's very sharp theorectically, and his articles about skin effect are among the best I've seen.
 

·
Banned
Joined
·
14,420 Posts
Discussion Starter · #11 ·

Quote:
Originally Posted by MichaelJHuman /forum/post/17023183


I remembered something. Some amplifier makers supposedly lower their power ratings just so that they can hit double power. This is according to a few articles. I don't know if it's true though.

You just thought of this? I labor under the delusion that I said so almost explicitly. Well on rereading what I said, I was actually hinting pretty broadly.


But nomatter the banter, you're right. Any simple power amp that doubles its output power when you halve the load impedance does so by means of specsmanship, and can reasonably be expected to be unmaksed on the test bench by a savvy tech.


It might be possible to build an amp with an intelligent power supply or DSP that actually did double its power into halved loads, but there may be other questionable consequences. I could see someone like Crown who already have some pretty smart amplifiers, having a firmware update that actually implemented this idea.
 

·
Registered
Joined
·
18,923 Posts

Quote:
Originally Posted by arnyk /forum/post/17023452


You just thought of this? I labor under the delusion that I said so almost explicitly. Well on rereading what I said, I was actually hinting pretty broadly.


But nomatter the banter, you're right. Any simple power amp that doubles its output power when you halve the load impedance does so by means of specsmanship, and can reasonably be expected to be unmaksed on the test bench by a savvy tech.


It might be possible to build an amp with an intelligent power supply or DSP that actually did double its power into halved loads, but there may be other questionable consequences. I could see someone like Crown who already have some pretty smart amplifiers, having a firmware update that actually implemented this idea.

Sorry if I am slow



Your locale makes me think of the movie Grosse Pointe Blank
 

·
Registered
Joined
·
1,261 Posts

Quote:
Originally Posted by arnyk /forum/post/17018334


Here's a real world data point for the controversy about power amps needing to be able to put out twice the power for half the impedance:


I ran some bench tests on what may be one of the most overbuilt power amps in the history of audio - the Threshold SA/4e from the early 1990s.


No visible effort was spared to build a robust power amp, as this article shows:

http://www.thresholdlovers.com/artic...?lng=en&pg=394


The Threshold SA/4e weighs about 120 pounds, according to my calibrated biceps. You really want two people to tote it around, and it facilitates this by having two comfortable handles on each end.


In contrast, in similar tests a Behringer A500 which weighs about 15 pounds puts out about the same amount of power at clipping at 8 ohms, but puts out about 144 watts at both 4 and 2 ohms. This is what I would expect from a typical contemporary mainstream amplifier.


Among other features, the Threshold SA/4e amplifier has extremely soft clipping. In the end I decided to measure its output power at about 0.1% THD @ 1KHz.


The actual measured power is as listed here:

http://cgi.audiogon.com/cgi-bin/auc....82027&auc&3&4&


Namely 112 Watts per channel @ 8 ohms with both channels driven @ 1000 hz, 196 Watts @ 4 ohms, 264 Watts @ 2 ohms.


This is still very far from doubling power as load impedance halves, particularly at 2 ohms.

Stereophile has found a few who were capable of doubling but I honestly can't remember which.

The emotiva XPA-2 is not expensive and they claim doubles at 4 ohms but there is no spec for 2 ohms. Have you tested it?
 

·
Banned
Joined
·
14,420 Posts
Discussion Starter · #14 ·

Quote:
Originally Posted by MichaelJHuman /forum/post/17022738


I read the first post, and I am not sure what I am supposed to get from it? You are comparing a 120 pound amplifier that probably cost 1000's of dollars to a cheap pro amp?

Back in the early 1990s the Threshold ran just under $7k. Today, a new one would probably run about twice that.


In contrast, the A500 sells for under $200. Not only that, but it is far from being a robust example of a pro audio amp. Its performance at 2 ohms seems far poorer than typical for a pro amp.


On balance, the A500's performance is about the same as a typical receiver, except that it has larger heatsinks and is actually rated for 4 ohm operation.

Quote:
We would expect the pro amp to fall short, right?

Thing is, if you are using well-designed 8 ohm speakers that don't go below 8 ohms (pretty typical), there should be no practical difference between the two amps. They both put out about 110 watts into 8 ohm loads without clipping, and have under 0.1% THD while doing it.


If the Threshold remains available to me, I might try to set up some ABX tests, and publish the results.
 

·
Registered
Joined
·
18,923 Posts
I had that same thought before, actually. If two amps out out the same power into 8 ohms measured the same way, and the speakers were well behaved 8 ohms nominal speakers, would it do you much good if one of the amp had better 4 ohm performance?


Your opinion seems to be maybe not.
 

·
Banned
Joined
·
14,420 Posts
Discussion Starter · #16 ·

Quote:
Originally Posted by sergiohm /forum/post/17023503


Stereophile has found a few who were capable of doubling but I honestly can't remember which.

Interesting speculation, but I'd need something more exact to comment on. Got any actual product names?

Quote:
The emotiva XPA-2 is not expensive and they claim doubles at 4 ohms but there is no spec for 2 ohms.

That's not what their own test reports show. They show just over 200 wpc into 8 ohms and 300 wpc into 4 ohms.


Their spec sheet shows 300 watts into 8 ohms and 550 into 4. Close but no cigar on that power doubling claim. They use what I consider to be a bogus means of rating it - power at 1% THD. IMO, power amps should be rated by power at clipping, not power so deep into clipping.


By the same standards that I consider the A500 to be a 100 wpc amp, the XPA-2 is a 200 wpc amp. You can buy 4 A500s for the price of one XPA-2.


As a rule, you should get more watts per buck as the power goes up. Economises of scale and all that.


There's only a 3 dB difference in power, which you'd be hard put to actually hear unless you do a very close comparison.

Quote:
Have you tested it?

Looks kinda interesting, but I've already got way too many power amps. For the money, I'd rather have a Crown XTi-2000. If memory serves, its like a 400wpc amp as I see amp power. Costs far less, too. Not nearly so hard on the shelves in the equipment cabinet.
 

·
Banned
Joined
·
14,420 Posts
Discussion Starter · #17 ·

Quote:
Originally Posted by MichaelJHuman /forum/post/17024835


I had that same thought before, actually. If two amps out out the same power into 8 ohms measured the same way, and the speakers were well behaved 8 ohms nominal speakers, would it do you much good if one of the amp had better 4 ohm performance?


Your opinion seems to be maybe not.

My opinion is definately not.


It's like asking whether a car that could go 250 mph goes faster at 80 mph than a car that can *only* go 100. ;-)


Given that the 250 mph car is like the Koenigsegg at over $200,000 and the 100 mph car is any old V6 minivan... Well, the Koenigsegg is very cool and provides a lot of bragging rights, but for everyday hauling and long trips, I'd rather have the minivan.


I gotta do something with my big problem with too much moderation... ;-)
 

·
Banned
Joined
·
14,420 Posts
Discussion Starter · #18 ·

Quote:
Originally Posted by MichaelJHuman /forum/post/17023483


Sorry if I am slow



Your locale makes me think of the movie Grosse Pointe Blank

Yes, I live in one of the Grosse Pointes. Of course the movie was very bogus -there are absolutely no convenience stores in any of the Grosse Pointes.


Someone did slip up, and we do have a Wendys. But no other fast food shops, no McDonalds, etc.


The main street in the Grosse Pointes has about 400 banks on it within about a mile. ;-) I think the actual number is about 12. There are two Banks of America in that one stretch of road... Why????
 

·
Registered
Joined
·
1,508 Posts

Quote:
Originally Posted by arnyk /forum/post/17023422

Quote:
Originally Posted by TPnBobcats /forum/post/17022737


This is OT, but you seem to know your stuff. There's an argument raging in the emo thread with a "bomb thrower" about bi-wiring. He's claiming that figure 3 in this link here.

http://www.st-andrews.ac.uk/~www_pa/...ire/Page1.html


Is a "parallel" bi-wiring. Which is what it's described as.

he can call it what he wants, but I call it no bi-wiring at all. Its using two paralleled cables as one.

That is actually what "bi" wiring is, it's just that "bi" wiring doesn't actually accomplish anything, except in the highly suggestible audiophile crowd's imagination.

Quote:
Originally Posted by arnyk /forum/post/17023422


I hate to disagree with you, but the gods of technology especially Ohm and Kirchoff say you are wrong.


No way is it bi-wiring. Its powering a speaker through two cables in parallel. The effective cable resistance is half of the resistance of either cable if they are identical.


In figure 3 the jumpers at the speaker end parallel the two cables, and also have the same basic effect as the shorting bars that most speakers set up for biwiring come with pre-installed.

Oh, you aren't disagreeing with me. I didn't think this did anything or accomplished anything. I though it was identical to just using the shorting bar, it was that the 3rd diagram was confusing enough that I wasn't 100% sure it wasn't doing something funky I was unfamiliar with.

Quote:
Originally Posted by arnyk /forum/post/17023422


BTW I know the author Jim Leserf, as he posts on the Usenet uk.rec.audio. He's very sharp theorectically, and his articles about skin effect are among the best I've seen.

It seemed to be an excellent article and is referenced a number of places. I think one of the magazines reproduced it as part of their "audio myths" series.
 

·
Registered
Joined
·
18,923 Posts

Quote:
Originally Posted by arnyk /forum/post/17024875


Yes, I live in one of the Grosse Pointes. Of course the movie was very bogus -there are absolutely no convenience stores in any of the Grosse Pointes.


Someone did slip up, and we do have a Wendys. But no other fast food shops, no McDonalds, etc.


The main street in the Grosse Pointes has about 400 banks on it within about a mile. ;-) I think the actual number is about 12. There are two Banks of America in that one stretch of road... Why????

Well of course. The convenience store WAS there, until it was blown up
 
1 - 20 of 21 Posts
Top