Home |
Search |
Today's Posts |
#1
|
|||
|
|||
Null speaker test
Years ago I read about something called a null speaker test.
It was a way to detect differences between amps by connecting 2 amps to a single speaker. If there were differences between the 2 amps then the speaker would generate sound. Does anyone know of this test and can you give details? |
#2
|
|||
|
|||
In order to do this you better be sure that BOTH amps are of the old style
of 2 prong, floating ground with the speaker negative and the chassis ground common. Tie the amp chassis together with a ground strap. Lots of people use to create a fake center channel by connecting a speaker across the positive terminals L and R of their stereo amp. New amp designs have both negative and positive signals; voltage between both wires and chassis ground. You'll blow up an amp like this. James. ) " wrote in message link.net... Years ago I read about something called a null speaker test. It was a way to detect differences between amps by connecting 2 amps to a single speaker. If there were differences between the 2 amps then the speaker would generate sound. Does anyone know of this test and can you give details? |
#3
|
|||
|
|||
" wrote in message link.net... Years ago I read about something called a null speaker test. It was a way to detect differences between amps by connecting 2 amps to a single speaker. If there were differences between the 2 amps then the speaker would generate sound. Does anyone know of this test and can you give details? **Such a test is highly likely to destroy one or both amps. Signals should ONLY be combined at low signal levels, then listened through a power amp. There have been several methods in which this can be accomplished. Hafler proposed quite a good system. You have either been misled, or your memory is flawed. -- Trevor Wilson www.rageaudio.com.au |
#4
|
|||
|
|||
In .net, on 10/11/05
at 10:58 PM, " said: Years ago I read about something called a null speaker test. It was a way to detect differences between amps by connecting 2 amps to a single speaker. If there were differences between the 2 amps then the speaker would generate sound. Does anyone know of this test and can you give details? At first glance, this is a very clever trick, however, I don't think it is worth the (very minimal) trouble. Others have warned about things that could go wrong with this sort of adventure. I share some of these misgivings, especially if you attempt to do this with equipment designed for cars. Home audio amplifiers are a better risk, unless they are bridged or "dual Mono". In any case, I don't think this is a very sensitive test. What if there was a slight time delay between amplifiers? If so, the test speaker would be quite noisy, even if the amplifiers were otherwise identical. In a few cases one amplifier might be inverting, while the other is not. Sensitivity will also be tied to speaker efficiency, speaker frequency response, room response, and your own sensitivities. Assuming the above considerations are not show stoppers, which amplifier is better? All you know is that they are "different". In my opinion, the only interesting feature about this sort of test is that one can use music as the test signal. Music is much more complex than the usual test tones. ----------------------------------------------------------- spam: wordgame:123(abc):14 9 20 5 2 9 18 4 at 22 15 9 3 5 14 5 20 dot 3 15 13 (Barry Mann) [sorry about the puzzle, spammers are ruining my mailbox] ----------------------------------------------------------- |
#6
|
|||
|
|||
On Tue, 11 Oct 2005 22:58:48 GMT, "
wrote: Years ago I read about something called a null speaker test. It was a way to detect differences between amps by connecting 2 amps to a single speaker. If there were differences between the 2 amps then the speaker would generate sound. Does anyone know of this test and can you give details? That's extremely likely to destroy *both* amplifiers. The classic null test uses an attenuator at the output of the amp, exactly matched to the gain and phase characteristics of the amp, and the output of the attenuator is then summed with the input signal. the output from this summing junction will be any amplifier artifacts. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#7
|
|||
|
|||
|
#8
|
|||
|
|||
In article .net, " wrote:
Years ago I read about something called a null speaker test. It was a way to detect differences between amps by connecting 2 amps to a single speaker. If there were differences between the 2 amps then the speaker would generate sound. Does anyone know of this test and can you give details? I think it sounds like an interesting idea. Having a wide range speaker with two separate voice coils would do it. Since the fields null, the speaker impedance will be flatter. I would suggest lower drive levels. Having low level nulling transformation sampling the amps, might be better. Amplifiers would tend to shwo differences at the higher frequencies. In any case, I would like to hear more about this sort of testing vs humanoid ABX comparison. greg |
#9
|
|||
|
|||
"Barry Mann" wrote in message
om In any case, I don't think this is a very sensitive test. As Peter Irwin mentioned Peter Walker of Quad advocated this test as a demonstration that the best amplifiers have distortion products which are inaudible even in the absence of signal. David Hafler revisited this type of test as alluded to here, and many other places http://www.audioasylum.com/audio/cab...es/100288.html http://www.snippets.org/pipermail/di...ry/009563.html What if there was a slight time delay between amplifiers? Unlikely in the narrow sense, but in the broader sense of phase shift, then this is a serious problem that Hafler ran into and tried to manage. If so, the test speaker would be quite noisy, even if the amplifiers were otherwise identical. In a few cases one amplifier might be inverting, while the other is not. That's the basic problem - there are so many variables that lead to sound that survives simplistic nulling that has nothing to do with distortion. When I tried to come up with a way to increase the audibility of errors in power amps, I came up with the process of sending the music through the amp again and again, to allow the errors to build up. Here are some real-world examples that people can simply download and listen to: http://www.pcabx.com/product/amplifiers/index.htm |
#10
|
|||
|
|||
"Trevor Wilson" wrote in message ... " wrote in message link.net... Years ago I read about something called a null speaker test. It was a way to detect differences between amps by connecting 2 amps to a single speaker. If there were differences between the 2 amps then the speaker would generate sound. Does anyone know of this test and can you give details? **Such a test is highly likely to destroy one or both amps. Signals should ONLY be combined at low signal levels, then listened through a power amp. There have been several methods in which this can be accomplished. Hafler proposed quite a good system. You have either been misled, or your memory is flawed. -- It's been at least a dcade since I read the article that descibed the set up, soo it's entirely possible my memory is flawed, but not about the basic premise. |
#11
|
|||
|
|||
"Don Pearce" wrote in message . .. On Tue, 11 Oct 2005 22:58:48 GMT, wrote: Years ago I read about something called a null speaker test. It was a way to detect differences between amps by connecting 2 amps to a single speaker. If there were differences between the 2 amps then the speaker would generate sound. Does anyone know of this test and can you give details? Despite what the others have told you, there is a way to do this without danger. I'm presuming the amplifiers are niminally similar, or you wouldn't even be trying this. The way you do this is to connect the amplifers in a bridged formation. Ignore the ground terminals at the amplifier outputs, and connect the speaker between the two live terminals. Now connect the audio source to both amplifier inputs in parallel. There should be no net audio on the output if the amps are the same. Use an oscilloscope on one of the outputs to see just what signal level is present while you are doing this. I'm not really sure I'll try it in any case, I just thought it would be agood topic of discussion, particularly if my memory of it only producing sound when there is a difference is correct. The output will not be zero. In a good amplifier, the gain is set by a pair of resistors - the feedback network. If good 1% resistors are used for these, you can expect to hear a signal perhaps 40dB below that from a normally wired amplifier. You can trim the level to one amplifier (the one with the higher gain) slightly to null out the remaining residue of signal. What remains now will be slight frequency response differences and distortions. d Hopefully I can find the issue of the magazine where I saw this setup mentioned, and reveal the whole thing to one and all. I thought it might offer some alternative to those who get their panties in a twist at the metion of ABX. |
#12
|
|||
|
|||
In k.net, on
10/13/05 at 12:15 AM, " said: [ ... ] I thought it might offer some alternative to those who get their panties in a twist at the metion of ABX. I'm not sure if the null speaker test could satisfy many listeners. Generally, one wants to discover which is best "A" or "B". In the null test one can only establish that there is a difference. The null test might be an interesting way to compare a production reference amplifier to a newly manufactured unit as part of a comprehensive final test procedure. ----------------------------------------------------------- spam: wordgame:123(abc):14 9 20 5 2 9 18 4 at 22 15 9 3 5 14 5 20 dot 3 15 13 (Barry Mann) [sorry about the puzzle, spammers are ruining my mailbox] ----------------------------------------------------------- |
#13
|
|||
|
|||
It's like bridging the amps but not inverting one of them. If one has
higher gain than the other you would get some sound. But this test does not make much sense to do. DaveL " wrote in message hlink.net... "Don Pearce" wrote in message . .. On Tue, 11 Oct 2005 22:58:48 GMT, wrote: Years ago I read about something called a null speaker test. It was a way to detect differences between amps by connecting 2 amps to a single speaker. If there were differences between the 2 amps then the speaker would generate sound. Does anyone know of this test and can you give details? Despite what the others have told you, there is a way to do this without danger. I'm presuming the amplifiers are niminally similar, or you wouldn't even be trying this. The way you do this is to connect the amplifers in a bridged formation. Ignore the ground terminals at the amplifier outputs, and connect the speaker between the two live terminals. Now connect the audio source to both amplifier inputs in parallel. There should be no net audio on the output if the amps are the same. Use an oscilloscope on one of the outputs to see just what signal level is present while you are doing this. I'm not really sure I'll try it in any case, I just thought it would be agood topic of discussion, particularly if my memory of it only producing sound when there is a difference is correct. The output will not be zero. In a good amplifier, the gain is set by a pair of resistors - the feedback network. If good 1% resistors are used for these, you can expect to hear a signal perhaps 40dB below that from a normally wired amplifier. You can trim the level to one amplifier (the one with the higher gain) slightly to null out the remaining residue of signal. What remains now will be slight frequency response differences and distortions. d Hopefully I can find the issue of the magazine where I saw this setup mentioned, and reveal the whole thing to one and all. I thought it might offer some alternative to those who get their panties in a twist at the metion of ABX. |
#14
|
|||
|
|||
On Thu, 13 Oct 2005 00:15:32 GMT, wrote:
"Don Pearce" wrote in message . .. On Tue, 11 Oct 2005 22:58:48 GMT, wrote: Years ago I read about something called a null speaker test. It was a way to detect differences between amps by connecting 2 amps to a single speaker. If there were differences between the 2 amps then the speaker would generate sound. Does anyone know of this test and can you give details? Despite what the others have told you, there is a way to do this without danger. I'm presuming the amplifiers are niminally similar, or you wouldn't even be trying this. The way you do this is to connect the amplifers in a bridged formation. Ignore the ground terminals at the amplifier outputs, and connect the speaker between the two live terminals. Now connect the audio source to both amplifier inputs in parallel. There should be no net audio on the output if the amps are the same. Use an oscilloscope on one of the outputs to see just what signal level is present while you are doing this. I'm not really sure I'll try it in any case, I just thought it would be agood topic of discussion, particularly if my memory of it only producing sound when there is a difference is correct. The output will not be zero. In a good amplifier, the gain is set by a pair of resistors - the feedback network. If good 1% resistors are used for these, you can expect to hear a signal perhaps 40dB below that from a normally wired amplifier. You can trim the level to one amplifier (the one with the higher gain) slightly to null out the remaining residue of signal. What remains now will be slight frequency response differences and distortions. d Hopefully I can find the issue of the magazine where I saw this setup mentioned, and reveal the whole thing to one and all. I thought it might offer some alternative to those who get their panties in a twist at the metion of ABX. The method is no good for this. It will reveal all sorts of differences that heave nothing whatever to do with audible differences between amplifiers. These would be to do with small phase and amplitude differences between two perfectly competent amplifiers. d |
#15
|
|||
|
|||
"Don Pearce" wrote in message
On Thu, 13 Oct 2005 00:15:32 GMT, wrote: I thought it might offer some alternative to those who get their panties in a twist at the metion of ABX. The method is no good for this. It will reveal all sorts of differences that heave nothing whatever to do with audible differences between amplifiers. These would be to do with small phase and amplitude differences between two perfectly competent amplifiers. Agreed. Note that much of this stuff about nulling amps was known *before* we developed ABX. We knew quite a bit about it and had been there and done that. It's an interesting experiment, but sheds not that much light. The best methodology that I've seen for increasing the audibility of differences between amps is the approach used at www.pcabx.com. It is really just a 21st century retread of some of Peter Walker's other work of long ago. |
#16
|
|||
|
|||
"Barry Mann" wrote in message om... In k.net, on 10/13/05 at 12:15 AM, " said: [ ... ] I thought it might offer some alternative to those who get their panties in a twist at the mention of ABX. I'm not sure if the null speaker test could satisfy many listeners. Generally, one wants to discover which is best "A" or "B". In the null test one can only establish that there is a difference. But, if there's no difference, ther's nothing to prefer other than features. Since there seems to be preicous little in the way of differnce as revealed by ABX, and since there is a vocal and very nasty cadre of people who scream that it HIDES difference, it seems to me that a way that doesn't require blinding or several trials, a simple, objective comparison like a null test, ought to have some quieting effect on those in denial. The null test might be an interesting way to compare a production reference amplifier to a newly manufactured unit as part of a comprehensive final test procedure. I'm hoping for a way to drive home the point that the "incredible" differences that get reported, for the most part simply don't exist, due to the state of current technology making it simple and inexpenisive to build amps that are sonically perfect. IOW they pass a signal that is not only indstinguishable from another reference, but in fact identical. ----------------------------------------------------------- |
#17
|
|||
|
|||
something like page 7 of this?
http://www.hafler.com/techsupport/pd...80_amp_man.pdf "Barry Mann" wrote in message om... In .net, on 10/11/05 at 10:58 PM, " said: Years ago I read about something called a null speaker test. It was a way to detect differences between amps by connecting 2 amps to a single speaker. If there were differences between the 2 amps then the speaker would generate sound. Does anyone know of this test and can you give details? At first glance, this is a very clever trick, however, I don't think it is worth the (very minimal) trouble. Others have warned about things that could go wrong with this sort of adventure. I share some of these misgivings, especially if you attempt to do this with equipment designed for cars. Home audio amplifiers are a better risk, unless they are bridged or "dual Mono". In any case, I don't think this is a very sensitive test. What if there was a slight time delay between amplifiers? If so, the test speaker would be quite noisy, even if the amplifiers were otherwise identical. In a few cases one amplifier might be inverting, while the other is not. Sensitivity will also be tied to speaker efficiency, speaker frequency response, room response, and your own sensitivities. Assuming the above considerations are not show stoppers, which amplifier is better? All you know is that they are "different". In my opinion, the only interesting feature about this sort of test is that one can use music as the test signal. Music is much more complex than the usual test tones. ----------------------------------------------------------- spam: wordgame:123(abc):14 9 20 5 2 9 18 4 at 22 15 9 3 5 14 5 20 dot 3 15 13 (Barry Mann) [sorry about the puzzle, spammers are ruining my mailbox] ----------------------------------------------------------- |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
A model of the brain, & quick-switch | High End Audio | |||
Comments regarding: Cables, Hearing, Stuff!! | High End Audio | |||
rec.audio.car FAQ (Part 4/5) | Car Audio | |||
Bose 901 Review | Vacuum Tubes | |||
A question for the cognoscenti autoformers | Vacuum Tubes |