|
|
Post by mgbpuff on Jul 15, 2017 7:12:01 GMT -5
The one thing that isn't measured and can't easily be measured is the characteristc impedance match between amplifier and any possible speaker. To keep things somewhat equal amps are tested into a resistive load. Speakers are not purely resistive, some have a lot of inductance and some a lot of capacitance. Someone should come up with a load substitute testing machine that covers the usual reactive range of speakers. This is one reason why a PROPERLY designed active speaker can outperform a separate amplifier / any speaker combination.
|
|
|
|
Post by leonski on Jul 15, 2017 11:54:21 GMT -5
2 things: you can test an amplifier into a load other than a resistor. ANY search on Google will show dozens of 'Dummy Load' proposals for this purpose. As near as I can tell, NO standard exists since everyone seems happy with the BS that is power into a resistor. A company called Audiograph makes what they call the Power Cube Measurement system. The connected amp is stressed with +-0 / 30' / 60' signals into a 8 / 4 / 2 ohm load. The resultant graph clearly shows an amps capability into OTHER than a resistive load. To make such a measurement as MGB proposes would require some kind of STANDARD be adopted so everyone was on the same page. Right. As if that could ever happen. I'm glad to agree with MGBs last post. I've been harping on that very point for quite a while. The capactivie and inductive shifts are generally called 'Reactance'. Going at it from the SPEAKER end, reading Stereophile test results of Speakers shows some to be really wacky loads with huge phase angles sometimes at sub-2ohm impedances. This is a recipe for an amp killer of a speaker. IMO, the bottom line is that amp designers and speaker designers don't talk much. As for powered speakers? Sure. Not only can you use an active crossover but Passive, as well. A Passive Line Level XO (PLLXO) is simple to make using readily available parts. The only special part of it is you should know the input impedance of the amp and output impedance of the PREamp. The length of cable that can be driven is limited. You don't need OPAMPS or a power supply, either. You can easily build such a crossover into an Altoids Tin. Example of Power Cube output. Easy to 'see' what's going on without muss or fuss www.bostonaudiosociety.org/meetings/2014/14-11_meeting.htm |
|
|
|
Post by Casey Leedom on Jul 20, 2017 12:51:17 GMT -5
|
|
|
|
Post by leonski on Jul 27, 2017 1:10:27 GMT -5
In reading the description of the Power Supply in the Stereophile test, I can't help think about Bob Carver and his 'tracking' power supplies, starting with his 'CARVER' amps in the 80s.
My m400t basically turned the power line on and off in time to the music, tha amp storing very little energy. At full juice into low sensitivity Magnepans, I could get the house lights to flicker
in time to the music.
Bob worked this issue and refined the circuitry or simply started with a clean sheet of paper an unknown number of times.
Did Bob have any input or consultation with EMO on the PS or the amp in general?
The article makes NO mention of an input filter to the test equipment. That would get 'rid' of switching noise and you'd never see it. And one other thing? Magnepan .7s a 'bad load'?
I need to see the reactance data for that speaker. Low sensitivity like all the REST in the line? Sure. But a bad load? Show me.
|
|
|
|
Post by mgbpuff on Jul 27, 2017 8:01:28 GMT -5
2 things: you can test an amplifier into a load other than a resistor. ANY search on Google will show dozens of 'Dummy Load' proposals for this purpose. As near as I can tell, NO standard exists since everyone seems happy with the BS that is power into a resistor. A company called Audiograph makes what they call the Power Cube Measurement system. The connected amp is stressed with +-0 / 30' / 60' signals into a 8 / 4 / 2 ohm load. The resultant graph clearly shows an amps capability into OTHER than a resistive load. To make such a measurement as MGB proposes would require some kind of STANDARD be adopted so everyone was on the same page. Right. As if that could ever happen. I'm glad to agree with MGBs last post. I've been harping on that very point for quite a while. The capactivie and inductive shifts are generally called 'Reactance'. Going at it from the SPEAKER end, reading Stereophile test results of Speakers shows some to be really wacky loads with huge phase angles sometimes at sub-2ohm impedances. This is a recipe for an amp killer of a speaker. IMO, the bottom line is that amp designers and speaker designers don't talk much. As for powered speakers? Sure. Not only can you use an active crossover but Passive, as well. A Passive Line Level XO (PLLXO) is simple to make using readily available parts. The only special part of it is you should know the input impedance of the amp and output impedance of the PREamp. The length of cable that can be driven is limited. You don't need OPAMPS or a power supply, either. You can easily build such a crossover into an Altoids Tin. Example of Power Cube output. Easy to 'see' what's going on without muss or fuss www.bostonaudiosociety.org/meetings/2014/14-11_meeting.htmThanks for that link! I have seen these power block curves before, but I forgot about them. They certainly tell more about amplifier performance than other tests, but still they are only done at one frequency at at time, in this case they only did 1khz. They should cover 20hz, 1kz, 10khz, and 20khz as a standard. And speaker manufactures should provide the resistance and reactance phase for each of these points also. Then and only then do we have a chance of matching amps to speakers without trial and error. |
|
|
|
Post by pedrocols on Jul 27, 2017 8:03:51 GMT -5
2 things: you can test an amplifier into a load other than a resistor. ANY search on Google will show dozens of 'Dummy Load' proposals for this purpose. As near as I can tell, NO standard exists since everyone seems happy with the BS that is power into a resistor. A company called Audiograph makes what they call the Power Cube Measurement system. The connected amp is stressed with +-0 / 30' / 60' signals into a 8 / 4 / 2 ohm load. The resultant graph clearly shows an amps capability into OTHER than a resistive load. To make such a measurement as MGB proposes would require some kind of STANDARD be adopted so everyone was on the same page. Right. As if that could ever happen. I'm glad to agree with MGBs last post. I've been harping on that very point for quite a while. The capactivie and inductive shifts are generally called 'Reactance'. Going at it from the SPEAKER end, reading Stereophile test results of Speakers shows some to be really wacky loads with huge phase angles sometimes at sub-2ohm impedances. This is a recipe for an amp killer of a speaker. IMO, the bottom line is that amp designers and speaker designers don't talk much. As for powered speakers? Sure. Not only can you use an active crossover but Passive, as well. A Passive Line Level XO (PLLXO) is simple to make using readily available parts. The only special part of it is you should know the input impedance of the amp and output impedance of the PREamp. The length of cable that can be driven is limited. You don't need OPAMPS or a power supply, either. You can easily build such a crossover into an Altoids Tin. Example of Power Cube output. Easy to 'see' what's going on without muss or fuss www.bostonaudiosociety.org/meetings/2014/14-11_meeting.htmThanks for that link! I have seen these power block curves before, but I forgot about them. They certainly tell more about amplifier performance than other tests, but still they are only done at one frequency at at time, in this case they only did 1khz. They should cover 20hz, 1kz, 10khz, and 20khz as a standard. And speaker manufactures should provide the resistance and reactance phase for each of these points also. Then and only then do we have a chance of matching amps to speakers without trial and error. And then you add the room. |
|
|
|
Post by mgbpuff on Jul 27, 2017 8:19:03 GMT -5
I don't think the room would affect these electrical measurements, but of course the room does affect what gets to your ears.
|
|
KeithL
Administrator 
Posts: 10,261
|
Post by KeithL on Jul 27, 2017 10:20:37 GMT -5
You've pinned the problem..... There's only one "8 Ohm resistive impedance load", but there are an infinite number of possible reactive load combinations, and every speaker is different. Therefore, beyond "proving that an amplifier performs consistently with a variety of different impedances", there's no way to test one with all possible combinations, or even close, so it's a losing battle. Even if you tested all of the values in that cube, they're still all "simple reactances" - so they don't cover everything that a speaker might ask an amp to do by a long shot anyway. 2 things: you can test an amplifier into a load other than a resistor. ANY search on Google will show dozens of 'Dummy Load' proposals for this purpose. As near as I can tell, NO standard exists since everyone seems happy with the BS that is power into a resistor. A company called Audiograph makes what they call the Power Cube Measurement system. The connected amp is stressed with +-0 / 30' / 60' signals into a 8 / 4 / 2 ohm load. The resultant graph clearly shows an amps capability into OTHER than a resistive load. To make such a measurement as MGB proposes would require some kind of STANDARD be adopted so everyone was on the same page. Right. As if that could ever happen. I'm glad to agree with MGBs last post. I've been harping on that very point for quite a while. The capactivie and inductive shifts are generally called 'Reactance'. Going at it from the SPEAKER end, reading Stereophile test results of Speakers shows some to be really wacky loads with huge phase angles sometimes at sub-2ohm impedances. This is a recipe for an amp killer of a speaker. IMO, the bottom line is that amp designers and speaker designers don't talk much. As for powered speakers? Sure. Not only can you use an active crossover but Passive, as well. A Passive Line Level XO (PLLXO) is simple to make using readily available parts. The only special part of it is you should know the input impedance of the amp and output impedance of the PREamp. The length of cable that can be driven is limited. You don't need OPAMPS or a power supply, either. You can easily build such a crossover into an Altoids Tin. Example of Power Cube output. Easy to 'see' what's going on without muss or fuss www.bostonaudiosociety.org/meetings/2014/14-11_meeting.htmThanks for that link! I have seen these power block curves before, but I forgot about them. They certainly tell more about amplifier performance than other tests, but still they are only done at one frequency at at time, in this case they only did 1khz. They should cover 20hz, 1kz, 10khz, and 20khz as a standard. And speaker manufactures should provide the resistance and reactance phase for each of these points also. Then and only then do we have a chance of matching amps to speakers without trial and error. |
|
KeithL
Administrator
Posts: 10,261
|
Post by KeithL on Jul 27, 2017 10:23:55 GMT -5
Just one correction..... The new BasX amps DO NOT have the new switch mode power supply; the BasX amps use regular linear supplies, with toroidal transformers, just like our older amps.  Gen 1's and 2's don't have switching power supplies, they have a huge toroidal transformer and a linear power supply with a monster bank of capacitors. Cheers Gary And that's what makes them weigh so much compared to the Gen. 3s. garbulky , being a 'heavy iron' man, definitely prefers the Gen 1 & 2 XPA products for that reason. I'm as yet unconvinced, but I must say that the Gar's Gen 2 XPA-1s do sound mighty. The only Emotiva amps I've heard with the "new" switching power supplies are the BasX A-300 and A-500. To my ears, they both sound excellent. As excellent as the earlier XPA models? Not having them side-by-side, I couldn't say. But I can say that I think they're awfully close. But I still think that Gar's Gen 2 XPA-1s sound better than either of the BasX amps. YMMV As to the Stereophile review, I think it was honest. That's all you can ask for. The reviewer isn't an audio auricle - only someone giving their opinion. If one reviewer likes (or dislikes) a specific product, then it's just another opinion. If a dozen reviewers all like (or dislike) a product, then it's less likely that they're all wrong - but it's still just opinions. Even measurements may be suspect - What was the plus or minus tolerance of the test equipment? Was the test equipment properly calibrated? Were the measurements in different reviews actually apples-to-apples comparisons with identical test parameters? But I digress... The good thing about reviewers is that they should be relatively consistent. If Robert E. Greene, for example, likes a smooth power response from speakers, and another reviewer thinks that power response is academic, then you'd expect for the two to have different opinions of a speaker. But it is humorous that some reviewers will do their best to damn a component with faint praise. I recently saw a review of a Yamaha integrated amplifier that I thought did exactly this. I thought that the review was dancing on a pinhead to avoid saying that the amp was really good (due to its big-Japanese-manufacturer source). But maybe that's just my impression. Cheers - Boom |
|
KeithL
Administrator
Posts: 10,261
|
Post by KeithL on Jul 27, 2017 10:29:02 GMT -5
I've heard that the new "medical" stuff is pretty good  Just read the review. I've heard a few different Zu Audio speakers at shows and I think they can be sub-optimal unless paired with the right amp. At two shows, the sound from the Zu room drove us out of the room with it's hardness; I'm talking headache. The third time I heard a Zu setup, I didn't want to leave. It was gorgeously smooth but had substantial upper frequency extension. Here's a video hosted by Herb interviewing the owner of High Water Sound in Brooklyn. Eccentric, esoteric and entertaining in my mind and it might give you an idea what he loves in terms of sound. I just want to TRY what these guys had to smoke!! 😲🎉🎶😇 |
|
|
|
Post by leonski on Jul 27, 2017 11:50:36 GMT -5
Keith Said:
:You've pinned the problem.....
There's only one "8 Ohm resistive impedance load", but there are an infinite number of possible reactive load combinations, and every speaker is different. Therefore, beyond "proving that an amplifier performs consistently with a variety of different impedances", there's no way to test one with all possible combinations, or even close, so it's a losing battle. Even if you tested all of the values in that cube, they're still all "simple reactances" - so they don't cover everything that a speaker might ask an amp to do by a long shot anyway.
I'll agree that the power resistor is the EASIEST load with which to test and amp. But it also represents ZERO speakers. None. Getting amp manufacturers to AGREE on a reactive dummy load is also a non-starter. Everybody wants to look good and many will simply look awful into a reactive load. I like the simplicity of the Power Cube. Even though it is also not representive of a real speaker. In some sense, it is worst-case.
But ya gotta start SOMEWHERE and people deserve to know. I guarantee some surprising results as some highly regarded amps suddenly don't do so well.
IMO, Stereophile has done HALF A JOB. They test and measure speakers for both reactance and frequency response. And while It would be better if they presented the graphic part of the data as a Smith Chart, they leave amps ENTIRELY alone content to test only into a resistor. This is a disconnect. Speaker and amp don't match! Even if they decided on an available speaker or published the 'curve' of their 'real speaker' dummy load, that would help.
When you read about a certain speaker, they always make a few general comments about 'the partnering amp'.
|
|
|
|
Post by mgbpuff on Jul 27, 2017 12:01:42 GMT -5
On the other hand most of the audience for amplifier/speaker systems can't even handle the DC version of 'Ohms Law' let alone the universe of AC's imaginary math. That doesn't mean I wouldn't like to see more information, it just means we are highly unlikely to ever get it!
|
|
|
|
Post by leonski on Jul 27, 2017 12:16:37 GMT -5
On the other hand most of the audience for amplifier/speaker systems can't even handle the DC version of 'Ohms Law' let alone the universe of AC's imaginary math. That doesn't mean I wouldn't like to see more information, it just means we are highly unlikely to ever get it! Too bad you're about 99% correct. I see so much 'fuzzy science' related to this sport that it makes me laugh sometimes. My favorite is the misuse of 'efficiency' when what is really meant is 'sensitivity'. And lots of persons think they have speakers which are a 'bad load' when really, they aren't THAT difficult to drive. Problem is the lack of real, measured data and comparison standards. How it sounds is up to YOU! |
|
KeithL
Administrator
Posts: 10,261
|
Post by KeithL on Jul 27, 2017 15:36:00 GMT -5
I should mention that there is another side to this question. A well-designed amplifier should be very close to being a "voltage source". What this means is that the ideal amplifier simply puts out a voltage that is the same as the voltage you feed into it multiplied by some constant. This can be expressed in a number of different ways, but the end result is that the output voltage is the same REGARDLESS OF THE LOAD IMPEDANCE. Of course, nothing is perfect, and everything has limits. But, instead of providing a whole slew of measurements, and graphing them all in a variety of attractive colors, all you have to do is specify the outer limits of the area where your amplifier operates properly. The Power Cube is sort of a way of "feeling around" for those outer limits in a few specific directions.... for example, most amplifiers find purely capacitive loads to be difficult to drive, so one limit is probably going to be the biggest capacitor that it's going to be able to drive. Of course, as you mentioned, one of the drawbacks is that this will in fact make a lot of equipment look somewhat bad (or limited). And another drawback is that it really isn't all that representative of what the amplifier will normally be called on to do in real life. (This is real life. Would you really be willing to pay $5k more for an amplifier that could deliver full voltage into a 20 uF capacitor - even though you're never going to need it to do that?) The other big catch, specific to audiophiles, is that a lot of audiophile equipment is what I would call "subjectively designed". For example, tube power amplifiers have a VERY HIGH output impedance. And, since the basic design criteria for "an ideal voltage amplifier" include "high input impedance and very low output impedance", as an attempt at an ideal voltage amplifier, a tube power amplifier is a resounding FAIL. Yes; really. If you run one of those Power Cube tests on a tube power amp, the results will be bad; comically bad; really REALLY bad. So, shall we reserve the easy side of the cube for tube amps, or shall we just skip it altogether? Now, what would be really cool would be to do a big project, and collate EVERYTHING. For example, you could actually use the full cube to measure everything. Then you could measure every speaker, and LOCATE THEM IN THE CUBE. Now you could tell how the amplifier would perform with any load by looking at the cube, and how it would work with a particular speaker by looking at HOW WELL THE AMPLIFIER PERFORMED IN THE AREA OF THE CUBE WHERE THAT PARTICULAR SPEAKER FIT. Of course, this is never going to happen, because it would cost a lot of money, and nobody wants to do it THAT badly. You're talking about serious standardization - in an industry where most companies don't especially benefit from being concisely compared with everyone else. Here at Emotiva, we'd love to see that sort of comparison.... because you'd get to see how well our amps stack up against competitors that cost 5x - or maybe 50x - as much as ours. (But, what do you say when the $50k amp ,whose manufacturer just bought a full page ad in this month's issue, gets beaten by one that costs $2k?) Keith Said: :You've pinned the problem..... There's only one "8 Ohm resistive impedance load", but there are an infinite number of possible reactive load combinations, and every speaker is different. Therefore, beyond "proving that an amplifier performs consistently with a variety of different impedances", there's no way to test one with all possible combinations, or even close, so it's a losing battle. Even if you tested all of the values in that cube, they're still all "simple reactances" - so they don't cover everything that a speaker might ask an amp to do by a long shot anyway. I'll agree that the power resistor is the EASIEST load with which to test and amp. But it also represents ZERO speakers. None. Getting amp manufacturers to AGREE on a reactive dummy load is also a non-starter. Everybody wants to look good and many will simply look awful into a reactive load. I like the simplicity of the Power Cube. Even though it is also not representive of a real speaker. In some sense, it is worst-case. But ya gotta start SOMEWHERE and people deserve to know. I guarantee some surprising results as some highly regarded amps suddenly don't do so well. IMO, Stereophile has done HALF A JOB. They test and measure speakers for both reactance and frequency response. And while It would be better if they presented the graphic part of the data as a Smith Chart, they leave amps ENTIRELY alone content to test only into a resistor. This is a disconnect. Speaker and amp don't match! Even if they decided on an available speaker or published the 'curve' of their 'real speaker' dummy load, that would help. When you read about a certain speaker, they always make a few general comments about 'the partnering amp'. |
|
hemster
Global Moderator
Particle Manufacturer
...still listening... still watching
Posts: 51,951
|
Post by hemster on Jul 27, 2017 18:52:53 GMT -5
Here at Emotiva, we'd love to see that sort of comparison.... because you'd get to see how well our amps stack up against competitors that cost 5x - or maybe 50x - as much as ours. ( But, what do you say when the $50k amp ,whose manufacturer just bought a full page ad in this month's issue, gets beaten by one that costs $2k?) ^This. |
|
|
|
Post by leonski on Jul 27, 2017 20:08:08 GMT -5
Tube amps with high output impedance are generally considered a non-ideal CURRENT source amp.
SS is a non-ideal Voltage source amp.
Fact is, the best speaker for tubes and SS are electrically Different. Overlap? Sure, why not? Magnepans work well with most good quality amps. Even TUBE!
Carver, for example, included a resistor in series with ONE set of speaker outputs to be used with some speakers. Turned it into sort of a Current Source amp.
Lots of your single driver, full range speakers seem to like some series resistance. Other designs, like the LS3/5 series from a Bunch of makers (starting as a BBC project) also enjoy
tube amps, even though they are medium to low sensitivity.
What DO you say when a 3k$ amp bests a 50k$ amp in a certain test?
'Sure, but how's it SOUND?'
Horses for Courses
|
|
|
|
Post by mgbpuff on Jul 28, 2017 3:19:41 GMT -5
Well, I seem to remember from my college days that tubes are essentially voltage devices and transistors are current devices. Tubes operate at voltages too high for speakers therefore must have output transformers to get the right voltages but come at the cost of high impedance (1/10th to 1/20 of the load). Since this is appreciable with respect to the load impedance, output accuracy and particularly damping factor is affected. Transistors run at voltages that a speaker can live with so they can be directly tied to speaker loads and the output impedance of a transistor is very low so the output characteristic impedance changes less with load variations and the damping factor is very high. Also transistors are easy to add negative feedback and that reduces harmonic distortion to a very low number compared to what is possible with tubes. Push pull output eliminates even harmonics and thus reduces harmonic distortion. Tubes can be designed push pull but many are single ended. Transistors must be push pull design. Thus, the old thought that tubes produce even harmonic distortion and solid state does not. Tube amps have generally higher levels of harmonic distortion because output negative feedback is difficult to obtain ( a special feedback winding on the output transformer is possible). High power tube amps either run high voltage on thr tube or parallel tubes. Transistors are parallel for power. Multiple output devices cause accuracy problems with either, but remember negative feedback is easier with solid state.
|
|
|
|
Post by audiobill on Jul 28, 2017 7:56:18 GMT -5
So, given the difficulties Keith describes, ultimately "How Does it Sound" is the relevant question.
And, for many of us, tubes just sound better.
|
|
|
|
Post by vneal on Jul 28, 2017 8:13:38 GMT -5
Not an expert by any means. But the speakers I have owned (at least 20 different models) had a couple 0f common man things to take notice of. The lower efficiency models were usually 4 ohms and wanted lots of power. Whereas the more efficient were 8 ohms. Most amps lets say 20 years ago would have a harder time with the 4 ohm speakers and I would usually compensate with a higher power amp. I have had Klipschorn and LaScallas (102db) where a McIntosh 30 watt tubed amp could blow you out of the room. I believe a 1000 watt mono amp would be a waste with these models. Whereas my old Thiel CS 2.4 (4ohm) wanted power 100-200-300-400 watts. were not too much.
Flash forward to current technology and I think the Emo amps are hard to beat. Emotive please come out with a new Mono Reference amp for 2 channel or frnt theater as a reference standard
|
|
|
|
Post by leonski on Jul 28, 2017 13:22:32 GMT -5
vneal, Some. repeat SOME Thiel models are an awful load. Stuff like a crossover board which wouldn't fit in a Monopoly Box tells the story At that point, even a capable amp might only deliver 1/2 or less its rated power at some frequencies. en.wikipedia.org/wiki/Power_factorThe first paragraph tells most of what a stereo guy would need to know. If we were powering a factory, other stuff would than become more important. |
|
