Jump to content

AR9 Driver Diagnosis


kkantor

Recommended Posts

Hi.

I made a short soundfile to help a friend test all the drivers on an AR9. Here's how it works:

1- Download it onto some kind of player attached to your stereo.

2- The file is mono to save space. I recommend testing only one speaker at a time. Disconnect the other speaker during each test.

3- For the first 10 seconds, there's a wide-band noise signal. Use this to set the playback to a comfortable level.

4- Of course, remember to start with the volume very low, then edge the volume up. The playback level is not critical, but should be modest enough to let you listen with your ear close to the drivers.

5- After the first 10 seconds, the sound flips back and forth between full range and a driver-specific tone. It does this four times for each range; each flip lasts for one second.

F-U-L-L-WOOF-FULL-WOOF-FULL-WOOF-FULL-WOOF-FULL-LM-FULL-LM-FULL-LM-FULL-LM-FULL-UM-UM.... etc.

6- Should be fairly easy to spot a driver than isn't working, although this is tricky for the tweeters. The closer you are to the speaker, the easier it will be.

Important Notes-

- Since the tweeter in the 9 operates over an unusually high frequency range, the last test tone contains almost no energy below 8KHz. This makes it very, very difficult for most people to hear; totally impossible for the vast majority of male adults. If you don't hear it immediately, DO NOT attempt to turn up the volume. At these high frequencies, human hearing is pretty much all or nothing. So, if you can't hear it the first time through, turning up the levels will only risk damaging your tweeters and will not improve audibility. Most listeners are shocked when they hear how high in pitch 8KHz is.

- If you must prove this to yourself, please do it with headphones, and be careful. The top range can come in and out of audibility just be moving the headphones a tiny bit.

- If you need to diagnose the tweeter purely by ear, you will need to find someone with intact hearing above 10KHz. Try Craig's list.

-k

ar9_08.mp3

Link to comment
Share on other sites

Hi.

I made a short soundfile to help a friend test all the drivers on an AR9. Here's how it works:

1- Download it onto some kind of player attached to your stereo.

2- The file is mono to save space. I recommend testing only one speaker at a time. Disconnect the other speaker during each test.

3- For the first 10 seconds, there's a wide-band noise signal. Use this to set the playback to a comfortable level.

4- Of course, remember to start with the volume very low, then edge the volume up. The playback level is not critical, but should be modest enough to let you listen with your ear close to the drivers.

5- After the first 10 seconds, the sound flips back and forth between full range and a driver-specific tone. It does this four times for each range; each flip lasts for one second.

F-U-L-L-WOOF-FULL-WOOF-FULL-WOOF-FULL-WOOF-FULL-LM-FULL-LM-FULL-LM-FULL-LM-FULL-UM-UM.... etc.

6- Should be fairly easy to spot a driver than isn't working, although this is tricky for the tweeters. The closer you are to the speaker, the easier it will be.

Important Notes-

- Since the tweeter in the 9 operates over an unusually high frequency range, the last test tone contains almost no energy below 8KHz. This makes it very, very difficult for most people to hear; totally impossible for the vast majority of male adults. If you don't hear it immediately, DO NOT attempt to turn up the volume. At these high frequencies, human hearing is pretty much all or nothing. So, if you can't hear it the first time through, turning up the levels will only risk damaging your tweeters and will not improve audibility. Most listeners are shocked when they hear how high in pitch 8KHz is.

- If you must prove this to yourself, please do it with headphones, and be careful. The top range can come in and out of audibility just be moving the headphones a tiny bit.

- If you need to diagnose the tweeter purely by ear, you will need to find someone with intact hearing above 10KHz. Try Craig's list.

-k

It seems to me if you can't hear above 8khz, it won't matter whether or not the tweeters work, you won't hear much from them anyway since the crossover point is 7khz. If you can, putting your ear up to it with any program material will make it easily audible.

If either the UMR or LMR drivers are inoperative, they will leave a gaping hole in the FR of each system. Listening to them one channel at a time, you'd easily tell in seconds, especially listening to a piano because of its wide frequency range.

Telling if one or both woofers are operating should also be easy. If only one is operating, connecting the output of a 9V battery applied across the input on the back would make one woofer move in one direction and the other in the opposite direction, that is one will move in while the other moves out because the dead one will only respond to the change in internal air pressure by the operating one. If both are connected and working, they will both move in or out at the same time depending on the polarity.

The program switches conveniently on the front also make it easy to tell about the UMR, LMR, and Tweeter. Adjusting these when your ear is next to the driver in question will make their effect by switching their levels easily audible if they are working. If you can't hear a 3db change to any of the drivers under those circumstances, I'd sell the AR9s and trade them in for a Bose Wave Radio. Given your lack of hearing accuity, that would be more than sufficient for your needs. :rolleyes:

Link to comment
Share on other sites

Hi.

I made a short soundfile to help a friend test all the drivers on an AR9. Here's how it works:

1- Download it onto some kind of player attached to your stereo.

2- The file is mono to save space. I recommend testing only one speaker at a time. Disconnect the other speaker during each test.

3- For the first 10 seconds, there's a wide-band noise signal. Use this to set the playback to a comfortable level.

4- Of course, remember to start with the volume very low, then edge the volume up. The playback level is not critical, but should be modest enough to let you listen with your ear close to the drivers.

5- After the first 10 seconds, the sound flips back and forth between full range and a driver-specific tone. It does this four times for each range; each flip lasts for one second.

F-U-L-L-WOOF-FULL-WOOF-FULL-WOOF-FULL-WOOF-FULL-LM-FULL-LM-FULL-LM-FULL-LM-FULL-UM-UM.... etc.

6- Should be fairly easy to spot a driver than isn't working, although this is tricky for the tweeters. The closer you are to the speaker, the easier it will be.

Important Notes-

- Since the tweeter in the 9 operates over an unusually high frequency range, the last test tone contains almost no energy below 8KHz. This makes it very, very difficult for most people to hear; totally impossible for the vast majority of male adults. If you don't hear it immediately, DO NOT attempt to turn up the volume. At these high frequencies, human hearing is pretty much all or nothing. So, if you can't hear it the first time through, turning up the levels will only risk damaging your tweeters and will not improve audibility. Most listeners are shocked when they hear how high in pitch 8KHz is.

- If you must prove this to yourself, please do it with headphones, and be careful. The top range can come in and out of audibility just be moving the headphones a tiny bit.

- If you need to diagnose the tweeter purely by ear, you will need to find someone with intact hearing above 10KHz. Try Craig's list.

-k

Ken,

Great idea on a quick way to test the AR-9. The only caveat here would be the great risk of burning-out a tweeter by not realizing the amount of energy going into the tweeter. The "modest-enough" level might be much higher for some people than others, and that could result in a burned-out tweeter. In any event, maybe I'm missing something here, but couldn't one play music at a moderately high level and put an ear to each driver to be sure that everything is working? I have a stethoscope that I have used on occasion for that purpose, also. If in doubt beyond that, disconnect the driver and briefly touch the contacts of a 1.5 volt dc flashlight battery (or meter for continuity) to the driver terminals to listen for the distinct crackle.

--Tom Tyson

Link to comment
Share on other sites

Ken,

Great idea on a quick way to test the AR-9. The only caveat here would be the great risk of burning-out a tweeter by not realizing the amount of energy going into the tweeter. The "modest-enough" level might be much higher for some people than others, and that could result in a burned-out tweeter. In any event, maybe I'm missing something here, but couldn't one play music at a moderately high level and put an ear to each driver to be sure that everything is working? I have a stethoscope that I have used on occasion for that purpose, also. If in doubt beyond that, disconnect the driver and briefly touch the contacts of a 1.5 volt dc flashlight battery (or meter for continuity) to the driver terminals to listen for the distinct crackle.

--Tom Tyson

Tom, one reason I toggle back and forth with the wide-band noise is that it makes it more difficult to turn the level up to one that might damage the tweeter.

This file was made in response to difficulties the person was having using trying to differentiate between the sound of the UM and the HR. It's just designed to be a quick diagnosis, not to replace meters, microphones, etc. The origin of the approach are files I have made for on-line QC situations.

Try it, and let me know if you think it is useful or not.

-k

Link to comment
Share on other sites

Tom, one reason I toggle back and forth with the wide-band noise is that it makes it more difficult to turn the level up to one that might damage the tweeter.

This file was made in response to difficulties the person was having using trying to differentiate between the sound of the UM and the HR. It's just designed to be a quick diagnosis, not to replace meters, microphones, etc. The origin of the approach are files I have made for on-line QC situations.

Try it, and let me know if you think it is useful or not.

-k

"Tom, one reason I toggle back and forth with the wide-band noise is that it makes it more difficult to turn the level up to one that might damage the tweeter."

On the other hand, if you've added 11 more outboard tweeters per channel like I have and turn the program level switch for the AR tweeter to -6db, burning it up is not that easy to do...with a 60 wpc amplifier. :D

BTW, is it as easy to burn them up on AR9s as it was on AR3as and LSTs? Where do you go to get replacements? Tiffany? :rolleyes:

Link to comment
Share on other sites

Tom, one reason I toggle back and forth with the wide-band noise is that it makes it more difficult to turn the level up to one that might damage the tweeter.

This file was made in response to difficulties the person was having using trying to differentiate between the sound of the UM and the HR. It's just designed to be a quick diagnosis, not to replace meters, microphones, etc. The origin of the approach are files I have made for on-line QC situations.

Try it, and let me know if you think it is useful or not.

-k

Ken,

I did try it -- first with headphones to see what was going on -- and then downloaded it onto a CDR for playback. It does work! It still scares me a bit (anything with test tones), but I think you are right about wide-band noise level helping to prevent turning the output level too high. I still can't imagine someone not able to put an ear close to the tweeter and upper midrange to tell if these are working -- unless hearing acuity is so low at those frequencies that any such test would be useless. Yeah, the grill will have to come off, and the music will have to have a lot of high-frequency content, but it should still be fairly easy to make that assessment. Nevertheless, this test-tone download would be much faster for most circumstances.

--Tom Tyson

Link to comment
Share on other sites

Ken,

I did try it -- first with headphones to see what was going on -- and then downloaded it onto a CDR for playback. It does work! It still scares me a bit (anything with test tones), but I think you are right about wide-band noise level helping to prevent turning the output level too high. I still can't imagine someone not able to put an ear close to the tweeter and upper midrange to tell if these are working -- unless hearing acuity is so low at those frequencies that any such test would be useless. Yeah, the grill will have to come off, and the music will have to have a lot of high-frequency content, but it should still be fairly easy to make that assessment. Nevertheless, this test-tone download would be much faster for most circumstances.

--Tom Tyson

I think that many inexperienced listeners have never learned to relate words or numbers to perceived sound. They have no confidence about knowing if they are hearing "midrange" or "treble." So, they don't know what they are listening for why you ask them to see if the tweeter is making sound. If they hear any sound coming from the upper mid when they are trying to listen closely to the tweeter, they get confused. Even if you tell them to put their hand over over the UM, and use the toilet paper tube trick on the tweeter, they are still confused.

This general issue is a problem on production line testing, too. New line operators have no idea how to identify gross defects in the sound that more experienced operators can pick up on from across the room.

-k

Link to comment
Share on other sites

On the other hand, if you've added 11 more outboard tweeters per channel like I have and turn the program level switch for the AR tweeter to -6db, burning it up is not that easy to do...with a 60 wpc amplifier. :D

BTW, is it as easy to burn them up on AR9s as it was on AR3as and LSTs? Where do you go to get replacements? Tiffany? :rolleyes:

Huh?

Link to comment
Share on other sites

Huh?

Why the surprise? If you've followed my postings, you know I've been re-engineering all of my speaker systems starting almost 20 years ago. The AR9 tweeter in my system only produces around 5% of the total system energy output above 7 khz. The rest is directed at the walls, ceiling, and arrives through reflections. Their FR is also different to compensate for differential absorption as a function of frequency by the room boundaries. The effect is entirely different from a direct radiating speaker, even different from a bi-polar speaker. The closest approach to mine I've seen in a commercially manufactured speaker was the dbx soundfield speaker as I recall it but even that didn't go far enough for me. Bose 901 was re-engineered the same way. The indirect radiation at high frequencies in that system is even a higher percentage than at middle and low frequencies in that system now. There have been radical changes to the spectral balance of these systems as well. Why? They sound much more accurate to my ears. In fact when properly equalized for each individual recording as well, often very accurate. They sound to me like instruments would sound if they were in my room. On their own in a 2 channel sound system, I think that's the best they can be expected to do. Also as I've said in my postings, I am not interested in what is referred to somewhat ambiguously as "imaging" although on many recordings that seems to improve too but probably not in the sense audiophiles mean it. For example, a real piano is not a point source of sound and a recording of one shouldn't sound like it is.

Link to comment
Share on other sites

Why the surprise? If you've followed my postings, you know I've been re-engineering all of my speaker systems starting almost 20 years ago. The AR9 tweeter in my system only produces around 5% of the total system energy output above 7 khz. The rest is directed at the walls, ceiling, and arrives through reflections. Their FR is also different to compensate for differential absorption as a function of frequency by the room boundaries. The effect is entirely different from a direct radiating speaker, even different from a bi-polar speaker. The closest approach to mine I've seen in a commercially manufactured speaker was the dbx soundfield speaker as I recall it but even that didn't go far enough for me. Bose 901 was re-engineered the same way. The indirect radiation at high frequencies in that system is even a higher percentage than at middle and low frequencies in that system now. There have been radical changes to the spectral balance of these systems as well. Why? They sound much more accurate to my ears. In fact when properly equalized for each individual recording as well, often very accurate. They sound to me like instruments would sound if they were in my room. On their own in a 2 channel sound system, I think that's the best they can be expected to do. Also as I've said in my postings, I am not interested in what is referred to somewhat ambiguously as "imaging" although on many recordings that seems to improve too but probably not in the sense audiophiles mean it. For example, a real piano is not a point source of sound and a recording of one shouldn't sound like it is.

I thought you had simply put a tweeter array on top of your old Bose 901s. But now I'm with you....

Link to comment
Share on other sites

I thought you had simply put a tweeter array on top of your old Bose 901s. But now I'm with you....

My Bose 901s have 1 forward firing tweeter, two upward firing tweeters, and three backwards firing tweeters at different angles. The front tweeter does not have the same spectral balance as the others and all do not play at the same loudness. The equalization and relative loudnesses have been engineered as appropriate the acoustics of the exact spot in the room it's installed in.

As I see it, for a sound reproduction system to reproduce the sound of musical instruments with the same timbre as they would be heard in your room, the sound system must have a flat frequency response. Sounds simple until you consider that this means not just the on axis frequency response of the speakers but all of the reflections that reach your ears from the same general direction as the apparent source and it must be flat all the way back right through to the recording microphones. Lots of variables that have to be taken into consideration and compensated for on each recording and in each sound system installation separately to achieve that goal.

To reproduce the timbre of the instruments as they would be heard in a concert hall, the effect that the hall itself has on the sound reaching your ears has to be incorporated in the chain as well. As I've pointed out before, one way to understand the difference of RT as a function of frequency in a concert hall is to realize that reverberant sound fields are dynamic, not static phenomena. While the reflections can be considered in aggregate as reverberation, each individual reflection has its own time and direction of arrival, loudness, and spectral balance independently of the others and of the direct sound. Without reproducing the change to the spectrum of sound of the reflections as they die out as you would hear them live, you can't reproduce the timbre you hear there either. A simple filter mimicking the steady state spectral transfer from a test source on the stage to a measuring instrument at a point in the audience doesn't cut it. It doesn't sound the same. This in my view was a mistaken notion incorporated in the speaker design of AR3 by its designers. The lack of sharp transient attacks and clarity characteristic of the HF rich direct and early reflections are filtered out to an audible degree while the mellowness the HF poor late reflections imparts is lost also. Experienced concert goers who are accutely sensitive to these sounds and know what instruments sound like even if they aren't analytical about them know this to be the case and lament the fact that recordings can't capture that sound.

The frustrations of trying to solve these problems has its amusing aspects. It also has its triumphs.

Link to comment
Share on other sites

(Back to the orriginal topic)

Hi Ken,

Seems like a very useful file. What program did you use for creating it? I would like to make some sine and intermod tests to burn to CD. Random phase stereo noise too.

A test we did at KEF was to place a pair of systems face to face, then connect both, with one system out of phase with the first. With a vertical alligned array such as the AR9 you can get very complete cancelation from lowest to highest frequency when applying pink noise or your test signal. Rocking one speaker backwards would dramatically increase the (HF) sound level.

This is also good for checking the balance of one system vs another. For example, if wirewound pots are used to set levels you will get a precise null at some setting when the pot of either unit is varied. Note that this only works well if the units can be placed so that every driver is opposed by the same driver in the other system and no more than about an inch away. Not so good with non mirrored arrays like an AR3, 3a etc.

David

Link to comment
Share on other sites

Hi.

I made a short soundfile to help a friend test all the drivers on an AR9. Here's how it works:

1- Download it onto some kind of player attached to your stereo.

2- The file is mono to save space. I recommend testing only one speaker at a time. Disconnect the other speaker during each test.

3- For the first 10 seconds, there's a wide-band noise signal. Use this to set the playback to a comfortable level.

4- Of course, remember to start with the volume very low, then edge the volume up. The playback level is not critical, but should be modest enough to let you listen with your ear close to the drivers.

5- After the first 10 seconds, the sound flips back and forth between full range and a driver-specific tone. It does this four times for each range; each flip lasts for one second.

F-U-L-L-WOOF-FULL-WOOF-FULL-WOOF-FULL-WOOF-FULL-LM-FULL-LM-FULL-LM-FULL-LM-FULL-UM-UM.... etc.

6- Should be fairly easy to spot a driver than isn't working, although this is tricky for the tweeters. The closer you are to the speaker, the easier it will be.

Important Notes-

- Since the tweeter in the 9 operates over an unusually high frequency range, the last test tone contains almost no energy below 8KHz. This makes it very, very difficult for most people to hear; totally impossible for the vast majority of male adults. If you don't hear it immediately, DO NOT attempt to turn up the volume. At these high frequencies, human hearing is pretty much all or nothing. So, if you can't hear it the first time through, turning up the levels will only risk damaging your tweeters and will not improve audibility. Most listeners are shocked when they hear how high in pitch 8KHz is.

- If you must prove this to yourself, please do it with headphones, and be careful. The top range can come in and out of audibility just be moving the headphones a tiny bit.

- If you need to diagnose the tweeter purely by ear, you will need to find someone with intact hearing above 10KHz. Try Craig's list.

-k

Ken how can I burn this to disk my computer is not even in the same building as my system or can some one send it to me

Thanks Jim

Link to comment
Share on other sites

Ken how can I burn this to disk my computer is not even in the same building as my system or can some one send it to me

Thanks Jim

I'm not sure how to answer this, because there are so many different operating systems, and so many different kinds of software for burning CD's. But, the basic idea is to save the file to your computer, then burn a CD-R with it. Once you have a CD-R, you should be able to play it in your audio system the usual way.

-k

Link to comment
Share on other sites

(Back to the orriginal topic)

Hi Ken,

Seems like a very useful file. What program did you use for creating it? I would like to make some sine and intermod tests to burn to CD. Random phase stereo noise too.

A test we did at KEF was to place a pair of systems face to face, then connect both, with one system out of phase with the first. With a vertical alligned array such as the AR9 you can get very complete cancelation from lowest to highest frequency when applying pink noise or your test signal. Rocking one speaker backwards would dramatically increase the (HF) sound level.

This is also good for checking the balance of one system vs another. For example, if wirewound pots are used to set levels you will get a precise null at some setting when the pot of either unit is varied. Note that this only works well if the units can be placed so that every driver is opposed by the same driver in the other system and no more than about an inch away. Not so good with non mirrored arrays like an AR3, 3a etc.

David

David,

That's really interesting. I've used the out-of-phase trick to try and reduce room noise when breaking in speakers, but I have never thought of using it for QC. It's a testament to the quality of your manufacturing process control that you could null tweeters!

The file was made using Adobe Audition. After some false-starts, I wound up creating a broad-band noise signal, then filtering sections of it individually. There's an, (oddly named), "Scientific Filter" menu in Audition that lets you create very high-order filters. I'm also a big fan of the open-source (eg- free) software called, Audacity. It isn't quite as slick, but you can do a lot with it, especially for the money!

http://audacity.sourceforge.net/

-k

Link to comment
Share on other sites

David,

That's really interesting. I've used the out-of-phase trick to try and reduce room noise when breaking in speakers, but I have never thought of using it for QC. It's a testament to the quality of your manufacturing process control that you could null tweeters!

The file was made using Adobe Audition. After some false-starts, I wound up creating a broad-band noise signal, then filtering sections of it individually. There's an, (oddly named), "Scientific Filter" menu in Audition that lets you create very high-order filters. I'm also a big fan of the open-source (eg- free) software called, Audacity. It isn't quite as slick, but you can do a lot with it, especially for the money!

http://audacity.sourceforge.net/

-k

"Null" is a relative term I suppose, but my recollection is that it is pretty distinct. It certainly shows up if a driver is dead or if one unit of two systems is incorrectly phased.

Thanks for the other info. I have Cool Edit Pro (an earlier version of Audition) and will see if I can make some similar test signals.

Regards,

David

Link to comment
Share on other sites

"Null" is a relative term I suppose, but my recollection is that it is pretty distinct. It certainly shows up if a driver is dead or if one unit of two systems is incorrectly phased.

Thanks for the other info. I have Cool Edit Pro (an earlier version of Audition) and will see if I can make some similar test signals.

Regards,

David

David,

CoolEdit is a great program. I learned much just listening to some of the algorithms those guys came up with. I know that the basic, "2000" program required one to purchase the "Scientific Filters" as a seperate plug in, while the (more expensive) Audition includes them. I don't know if Cool Edit Pro has them as standard. These filters are very different than the stock filters in my non-pro version in that they let one specify very high slopes, explicit cutoffs, etc. If you need help locating the Scientific Filters upgrade package, just drop me an email.

-k

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...