Audioholics Amplifier Measurement Testing Methodology

[admin]

There has been much discussion on many of the audio forums lately regarding mid priced A/V receivers and multi channel amplifier power capabilities as well as their abilities to drive low impedance loads. Much of the controversy steams around the infamous " All Channels Driven Test " which simulates a best case test load in a worst case environment and by all intents and purposes an unrealistic real world scenario. As a result, we have come up with a standard for measuring amplifier performance that deals with various types of power tests as well as other metrics that directly affect sonic performance and how the amplifier can sound powering various types of speaker loads.

Discuss "Basic Amplifier Measurement Techniques" here. Read the article.

 

[Shiitaki]

Thanx

When I was doing all of my critical listening in a car, being able to drain the battery on the freeway while pounding bass was the key measurement. The manufacturer power ratings on amplifiers really are useless. The Audioholics review of my receiver, along with other reviews led me to a brand I'd not taken seriously before. The measurements are very important to me, not because of a specific watt measurement, but the results as a whole. Any difficulties, or details of operation that are not what would be expected are also of high importance. For instance, a input that limited bandwidth or misapplied bass management. Surprises aren't always fun.

Thank you for taking the measurements seriously, I want to say thank you to everyone for doing a great job. Consumers lack the skill and equipment to fully evaluate things like amplifiers; independent analysis keeps firms honest and provides buyers peace of mind. It's nice to make an informed decision.

Thanks again, Allan

 

[highfigh]

When I was doing all of my critical listening in a car, being able to drain the battery on the freeway while pounding bass was the key measurement. The manufacturer power ratings on amplifiers really are useless. The Audioholics review of my receiver, along with other reviews led me to a brand I'd not taken seriously before. The measurements are very important to me, not because of a specific watt measurement, but the results as a whole. Any difficulties, or details of operation that are not what would be expected are also of high importance. For instance, a input that limited bandwidth or misapplied bass management. Surprises aren't always fun.

Thank you for taking the measurements seriously, I want to say thank you to everyone for doing a great job. Consumers lack the skill and equipment to fully evaluate things like amplifiers; independent analysis keeps firms honest and provides buyers peace of mind. It's nice to make an informed decision.

Thanks again, Allan

Back in the early '70s when the FTC mandated that certain information be included in amplifier specs, car audio included a few speakers, cassette or 8 track players, FM adapters and a few other items. Power amplifiers for cars didn't even exist and at best, they offered power boosters, which we all know were crap. Then, when the marketing departments got involved and competition for market share increased, power ratings that were known in the car audio biz included "100 Sony Watts", "100 Watts, JBF" or "100 Watts, WLS" were thrown around when we had serious doubts about their accuracy. JBF stands for "Just Before Fire" and WLS stands for "When Lightning Strikes".

When a manufacturer would rate their amp at 100W and it was 5" x 5" x 2" with one 5A fuse, we were in the position of telling people that what they bought at the big sale put on by a liquidator was garbage and couldn't put out that much power if it was hit by a solar flare.

 

[KEW]

Gene,

Thanks for providing this explanation/education.
It is good to have an unbiased explanation of what matters and what doesn't relative to what we hear as well as insight into relative measures between different test methods used by various publications.

Perhaps the most important thing Audioholics is doing is providing a well thought out "real world" set of tests. If a test is not realistic, manufacturers are tempted to expend resources "building for the test" rather than building for audio quality.


In case anyone thinks this is a stretch, I'll offer two cases where this has been done:

Computer graphics card manufacturers incorporated a graphics sequence commonly used for testing into their board so as to by-pass actually having to process those graphics. These boards inexpensively returned exceptional test scores (on that particular test), but their real-world performance was not exceptional.

Honda (and likely other auto manufacturers - Honda is the one I know about) incorporated a design modifications to improve NHTSA side impact test results; however, if you were to relocate the point of the impact 6" forwards or backwards, the result would not reflect the car's safety rating.

Thanks,
Kurt

 

[KEW]

The graph showing the spike in the Class D Axiom A1400-8's frequency response piqued my interest.
Is there (or will there be) a review on this amp (my search came up empty) with a more detailed explanation on this test result, or was this simply an isolated measurement to demonstrate the behavior of a Class D amp?

Thanks,
Kurt

 

[MinusTheBear]

The graph showing the spike in the Class D Axiom A1400-8's frequency response piqued my interest.
Is there (or will there be) a review on this amp (my search came up empty) with a more detailed explanation on this test result, or was this simply an isolated measurement to demonstrate the behavior of a Class D amp?

Thanks,
Kurt

Genes in-depth review of this product is right here.

http://www.audioholics.com/reviews/amplifiers/axiom-audio-a1400-8

 

[KEW]

Genes in-depth review of this product is right here.

http://www.audioholics.com/reviews/amplifiers/axiom-audio-a1400-8

Doh!

What am I doing wrong?!!!

I put "Axiom Audio A1400-8" into the search field and get nothing!!

Thanks,
Kurt

 

[MinusTheBear]

Doh!

What am I doing wrong?!!!

I put "Axiom Audio A1400-8" into the search field and get nothing!!

Thanks,
Kurt

I used google. Didn't try the AH search function.

 

[gene]

Doh!

What am I doing wrong?!!!

I put "Axiom Audio A1400-8" into the search field and get nothing!!

Thanks,
Kurt

That is bizarre. I have to look into why that is the case. Thanks.

 

[3db]

Thanks for the write up. You've excluded the all channels test becuase you say that this is a non event in real life. But so are just straight resistive loads of 4 ohms. I may have missed it in reading through this but does your equipment test for reactive loads as well? Any receiver can drive a 4 ohm resisitive load....but throw a complex impedance at it and it changes the whole game. If you do test for that you would probably be one of the first in the industry.

The other thing I'm curious about is that manufacturers using variacs during their tests to maintain line voltage. This could be easily eliminated if you provide a non standard 60/100 amp circuit. I know its non standard but I also can't help but think that people with big home theater setups would have dedicated feeds of higher amperage circuits.

 

[gene]

Thanks for the write up. You've excluded the all channels test becuase you say that this is a non event in real life. But so are just straight resistive loads of 4 ohms. I may have missed it in reading through this but does your equipment test for reactive loads as well? Any receiver can drive a 4 ohm resisitive load....but throw a complex impedance at it and it changes the whole game. If you do test for that you would probably be one of the first in the industry.

The other thing I'm curious about is that manufacturers using variacs during their tests to maintain line voltage. This could be easily eliminated if you provide a non standard 60/100 amp circuit. I know its non standard but I also can't help but think that people with big home theater setups would have dedicated feeds of higher amperage circuits.

I did NOT exclude the ACD test. We do this test. I just question its relevancy in real world. Also this is really an instantaneous 1kHz Power Sweep vs distortion, not a full bandwidth continuous power test like some magazines and manufacturers would have you believe it is.

I would like to do reactive load testing. I am trying to get a manufacturer to build a dummy speaker load for me that pros in the industry can all agree represents a typical loudspeaker. This will eventually be incorporated into my testing.

I use a 120V/20A line. I am sure there are people that use 220V/20A but I can't imagine anyone using higher amperage lines than that. You will certainly blow a fuse in the amp if you go that high. I prefer using a typical line and NOT holding it constant as its more representative of real world. I never see much appreciable line sag especially when doing Psweep tests. In those cases, it typically only affects my measurements by a few watts per channel. I can live with that.

 

[3db]

I did NOT exclude the ACD test. We do this test. I just question its relevancy in real world. ..

MY apologies; misread that.

Also this is really an instantaneous 1kHz Power Sweep vs distortion, not a full bandwidth continuous power test like some magazines and manufacturers would have you believe it is

Some companies such as NAD are moving to all channels full frequency type testing. However, I don't how long the signal is being applied for. ..

I would like to do reactive load testing. I am trying to get a manufacturer to build a dummy speaker load for me that pros in the industry can all agree represents a typical loudspeaker. This will eventually be incorporated into my testing.

I use a 120V/20A line. I am sure there are people that use 220V/20A but I can't imagine anyone using higher amperage lines than that. You will certainly blow a fuse in the amp if you go that high. I prefer using a typical line and NOT holding it constant as its more representative of real world. I never see much appreciable line sag especially when doing Psweep tests. In those cases, it typically only affects my measurements by a few watts per channel. I can live with that.

Thanks for your replies. I hope your endeavour will be successful in getting the reactive load built. Its clear to me that your tests are very thorough and comprehensive.

 

[gene]

Some companies such as NAD are moving to all channels full frequency type testing. However, I don't how long the signal is being applied for. ..

They may be doing that for up to two channels but I can assure you, them like everyone else that claims ACD are doing 1kHz Psweeps like I stated in my article.

 

[GlocksRock]

I drive all 7 channels of my amp simultaneously when I play guitar hero, I set my AVR to 7 ch. stereo and crank it up loud when I rock out. That is about the only real life situation I can think of for ACD, except for someone just playing music with all the speakers turned on, or my fiancee's mother and sister who keep their avr set on 5 ch. stereo for everything.

 

[3db]

They may be doing that for up to two channels but I can assure you, them like everyone else that claims ACD are doing 1kHz Psweeps like I stated in my article.

Based on this;

http://nadelectronics.com/audio-topics/NAD-Full-Disclosure-Power

They claim to run full frequency test in their all channels driven. Are you saying that they are NOT doing this?

 

[gene]

Based on this;

http://nadelectronics.com/audio-topi...sclosure-Power

They claim to run full frequency test in their all channels driven. Are you saying that they are NOT doing this?

I'd like to see their test conditions (IE. Line Voltage / Current), how they are running the sweep, if they are really running 4 ohm loads for all channels simultaneously, etc.

NAD is sending me one of their receivers to review early next year so trust me I will be testing it ACD to their specification.

The reality may be their receiver can deliver full rated power into 4 ohms with all channels driven, but the consumer can never reach these levels using a 120A/15A line which is what most people run. It doesn't really matter since ACD is NOT a real world scenario to begin with

 

[3db]

I'd like to see their test conditions (IE. Line Voltage / Current), how they are running the sweep, if they are really running 4 ohm loads for all channels simultaneously, etc.

NAD is sending me one of their receivers to review early next year so trust me I will be testing it ACD to their specification.

The reality may be their receiver can deliver full rated power into 4 ohms with all channels driven, but the consumer can never reach these levels using a 120A/15A line which is what most people run. It doesn't really matter since ACD is NOT a real world scenario to begin with

To tell you the truth, I was hoping you would be able to test a unit. Just do me the favour of using a 120V/60A line.

 

[GranteedEV]

I would like to do reactive load testing. I am trying to get a manufacturer to build a dummy speaker load for me that pros in the industry can all agree represents a typical loudspeaker.

Yeah, but a typical loudspeaker is probably more like an 8 ohm with the ocassional 6 ohm dip.

I'm more interested in knowing if an amp is going into shutdown trying to drive something like a magnepan IE not only low impedance but with occasional very low dips. THis is because I think most amps can in fact drive a typical 8 or 6 ohm speaker load and while 4 ohms continues is nice to see as well; it's the atypical loads that seem to create issues in the first place.

 

[gene]

THis is because I think most amps can in fact drive a typical 8 or 6 ohm speaker load and while 4 ohms continues is nice to see as well; it's the atypical loads that seem to create issues in the first place.

4 ohm loads can be had for receivers to drive, especially if they employ the Class D Ice Module. I do want to make a reactive load bank one day however that flips to low impedance at low F and low impedance at high F.

 

[gene]

Looking at NAD specs of their receivers:
http://nadelectronics.com/products/av-receivers/T-785-A/V-Surround-Sound-Receiver

They don't specify load impedance for their power measurements. I suspect their ratings are for 8 and 4 ohm loads with current limiting employed for the latter hence why their power ratings are the same under both load conditions.