Earl, the fittings shown in the above image would not be terribly expensive to have machined. In fact I can make that sort of thing in my own shop. All I would need is some aluminum or other raw stock, a lathe and a milling machine + time. I have everything right now other than time. But I can get the parts made by a machinist at a decent price. (possible group buy?)
The pipe looks like standard aluminum pipe - PVC might be smoother, but may have some internal ripple. It is possible to polish and smooth the interior of metal (or other metal, like copper) pipe using a rig like a cylinder hone (like those used on engines) and a suitable rotational device - electric drill, and the proper compound and abrasives...
A question that arises in my mind is the effect of the "adapter"? It would seem to me that it needs to be pretty well set for the specific exit angle of the driver's throat area?
You mentioned "inverse taper"? That's unclear to me... narrowing the diameter vs. length??
Ok, so now we have the "bouncing off the walls" information at various distances down the tube. Neglecting the specifics of the "simple algebra", what are we going to do to this 'other than the plane wave' signal to reveal useful information??
Also, what length does one want or need the tube WRT going lower in frequency?
The PWT that WE used on the 555 driver's tests was so long they ran it out a window...
_-_-bear
The pipe looks like standard aluminum pipe - PVC might be smoother, but may have some internal ripple. It is possible to polish and smooth the interior of metal (or other metal, like copper) pipe using a rig like a cylinder hone (like those used on engines) and a suitable rotational device - electric drill, and the proper compound and abrasives...
A question that arises in my mind is the effect of the "adapter"? It would seem to me that it needs to be pretty well set for the specific exit angle of the driver's throat area?
You mentioned "inverse taper"? That's unclear to me... narrowing the diameter vs. length??
Ok, so now we have the "bouncing off the walls" information at various distances down the tube. Neglecting the specifics of the "simple algebra", what are we going to do to this 'other than the plane wave' signal to reveal useful information??
Also, what length does one want or need the tube WRT going lower in frequency?
The PWT that WE used on the 555 driver's tests was so long they ran it out a window...
_-_-bear
Last edited:
I believe most terminated tubes are sealed ended. This means they will have the usual 1/2 wave, 3/2 wave etc. resonances based on their length. The damping wedge will damp this to a high degree and usually there will just be some remnant of the first resonance (at least in the curves I've seen). Length is somewht irrelevant if the termination is good.
I'm not sure that we need to take a cylinder hone to the inside of one.
David
I'm not sure that we need to take a cylinder hone to the inside of one.
David
1audiohack
If you are serious, here is the test:
Test the driver, rotate it 90 degrees (if four hole mount, 120 if three hole), test again, rotate again, test agian, rotate again, test again. Plot all four result on a single plot. If they are identical then there is no non-axi-symmetric issues and we can probably ignore that aspect. But, if, as I suspect, they are different at the HFs, then non-axisymmetric behavior is present and the results of any single angle will not be valid.
If you are serious, here is the test:
Test the driver, rotate it 90 degrees (if four hole mount, 120 if three hole), test again, rotate again, test agian, rotate again, test again. Plot all four result on a single plot. If they are identical then there is no non-axi-symmetric issues and we can probably ignore that aspect. But, if, as I suspect, they are different at the HFs, then non-axisymmetric behavior is present and the results of any single angle will not be valid.
OK answers in order as best I can:
This tube was designed based from information gleaned from the AES paper 1id-1991(r2003).
Post 10, your right that the point is to "shrink your universe" in order to remove as many variables as possible. It's a two edge knife however as now you don't know what it sounds like so it becomes an objective tool in a visual only format.
Post 11, good gating has ben available since the late sixties, read Richard Heyser / TEF.
Your point 1, flys in the face of everything I have seen or read, please elaborate.
Your point 4, this puzzles me, please elaborate.
Post 13, the HF limit is determined by the diameter, the LF by the length.
Post 14, thank you Earl, I find that very interesting.
Post 16, you will never get a "set of protocols to obtain usefull results that could be compared with the results of other" using a $40. USD mic.
Post 17, I made it in one night after work. The mic slips into a rubber sleeve, an automotive seal actually. The black JBL 2327 1" to 2" adapter has been internally re-taper sleeved to a 1.5" exit. While experimenting with the dampening I wanted to use the lightest driver I had as it was R&R'd probably 50+ times.
Post 18, "Simplicity?" Depends on ones resources. "No one could afford one like this?" Im in mine less than $150 USD. "No one will build it?" Are you sure, I did.
Post 19, I designed it. I use TEF as a measurement platform as did JBL on their 2435, 2450, 2451, 2452, 475Nd etc. drivers.
Gotcha's, you bet, getting a six foot wedge of fiberglass or foam pulled through the tube is one, if you wish to obtain realistic distortion measurments you will need at least a 170dB mic, drivers are pretty clean when the drive is in microvolts. Also with a standard 145-150 dB measurment mic the mic's pre amp won't swing the voltage and the SPL line just gets flatter and flatter with increased drive, it's not reality.
Post 21, If you guys really want to make some of these to some kind of standard we can certainly do it, we have a CNC lathe and two CNC mills.
The tube pictured is 1.550" ID, thats the as finished ID of schedule 40 extruded aluminum, you were kidding about honing the ID weren't you?
The flange adapter has a 500Hz taper from 1.480" to the 1.550", thats matches the expansion rate of the 1.5" JBL drivers phasing plug as near as I can tell. I made a tapered sleeve to fit the into throats of the 2" drivers since thats the phase plug exit diameter anyway so I can test these drivers as well but the 1.5" throatless drivers are what this tube was really built for.
Post 22, properly dampened it makes no difference if the end is terminated with a cap or not.
Post 23, If I'm serious? Does not the fact that I built this hold the answer to that question? ;^) I did that, I remember far less than 1 dB SPL but more than 0.25 dB SPL. I guess I never understood what this test would illustrate. Brush Wellman mentions it in their white paper on Be diaphragms though I didn't think they showed any appreciable variation. I can't imagine a well designed properly assembled driver exhibiting substantial variations.
In the end these are comparative tools with a very high degree of test repeatability. After I was satisfied with the results at the shop, I broke the setup down, driver off, mic in the bag etc. and brought it home where I reset it up several days later, test repeatability of less than 0.25 dB SPL, that's about what you get just unbolting the driver and remounting it and resweeping it. Try that with a mic in space pointed at any horn.
At that point removing the driver cover and comparing diaphragm materials and coatings became meaningful for me.
Hope that helps a bit.
All the best,
Barry.
This tube was designed based from information gleaned from the AES paper 1id-1991(r2003).
Post 10, your right that the point is to "shrink your universe" in order to remove as many variables as possible. It's a two edge knife however as now you don't know what it sounds like so it becomes an objective tool in a visual only format.
Post 11, good gating has ben available since the late sixties, read Richard Heyser / TEF.
Your point 1, flys in the face of everything I have seen or read, please elaborate.
Your point 4, this puzzles me, please elaborate.
Post 13, the HF limit is determined by the diameter, the LF by the length.
Post 14, thank you Earl, I find that very interesting.
Post 16, you will never get a "set of protocols to obtain usefull results that could be compared with the results of other" using a $40. USD mic.
Post 17, I made it in one night after work. The mic slips into a rubber sleeve, an automotive seal actually. The black JBL 2327 1" to 2" adapter has been internally re-taper sleeved to a 1.5" exit. While experimenting with the dampening I wanted to use the lightest driver I had as it was R&R'd probably 50+ times.
Post 18, "Simplicity?" Depends on ones resources. "No one could afford one like this?" Im in mine less than $150 USD. "No one will build it?" Are you sure, I did.
Post 19, I designed it. I use TEF as a measurement platform as did JBL on their 2435, 2450, 2451, 2452, 475Nd etc. drivers.
Gotcha's, you bet, getting a six foot wedge of fiberglass or foam pulled through the tube is one, if you wish to obtain realistic distortion measurments you will need at least a 170dB mic, drivers are pretty clean when the drive is in microvolts. Also with a standard 145-150 dB measurment mic the mic's pre amp won't swing the voltage and the SPL line just gets flatter and flatter with increased drive, it's not reality.
Post 21, If you guys really want to make some of these to some kind of standard we can certainly do it, we have a CNC lathe and two CNC mills.
The tube pictured is 1.550" ID, thats the as finished ID of schedule 40 extruded aluminum, you were kidding about honing the ID weren't you?
The flange adapter has a 500Hz taper from 1.480" to the 1.550", thats matches the expansion rate of the 1.5" JBL drivers phasing plug as near as I can tell. I made a tapered sleeve to fit the into throats of the 2" drivers since thats the phase plug exit diameter anyway so I can test these drivers as well but the 1.5" throatless drivers are what this tube was really built for.
Post 22, properly dampened it makes no difference if the end is terminated with a cap or not.
Post 23, If I'm serious? Does not the fact that I built this hold the answer to that question? ;^) I did that, I remember far less than 1 dB SPL but more than 0.25 dB SPL. I guess I never understood what this test would illustrate. Brush Wellman mentions it in their white paper on Be diaphragms though I didn't think they showed any appreciable variation. I can't imagine a well designed properly assembled driver exhibiting substantial variations.
In the end these are comparative tools with a very high degree of test repeatability. After I was satisfied with the results at the shop, I broke the setup down, driver off, mic in the bag etc. and brought it home where I reset it up several days later, test repeatability of less than 0.25 dB SPL, that's about what you get just unbolting the driver and remounting it and resweeping it. Try that with a mic in space pointed at any horn.
At that point removing the driver cover and comparing diaphragm materials and coatings became meaningful for me.
Hope that helps a bit.
All the best,
Barry.
Hi Barry
I don't think that many (any?) of the rest of us have that kind of resources. So what will you sell us those tubes for? $150?
If you have "good gating" then why do you need the tube dampened? TEF is very old technology, there are much better systems avaialble these days for free - like HolmImpulse for example.
Point 1
If there are only plane waves in the tube then the distance that we measure along the tube will not matter, its all the same, right? But there is a slope discontinuity at the junction between the driver and the tube and this will generate Higher Order Modes (HOM). These modes will propagate with a wavenumber that is complex below the cut-in frequency and real above that frequency. In a complex exponential wave this means that below cut-in the waves are damped exponentials that decay expoenetially, but complex exponentials above cut-in which are true propagating waves. The damped expoentials are called evanescent waves and they decay and eventually disapate as they propagate. If you measure too close to the junction you will be measureing these waves and you don't want to do that.
As to the non-axisymmetric performance, I have no data on that other than Don Keele said that he measure substantrial differences when he did it. The issue with non-axisymmetric waves is that they have much lower cut-in frequencies than the axial ones. So they affect the measurements at a much lower frequency.
Why don't you run the test again, maybe with two different brands of drivers and post the data for all of us to see? 1 dB in the tube is actually a fairly substantial non-axi motion when you think about it in linear terms as a percentage.
Finally - all work that I know of for PWT assume plane waves. This all falls apart when you understand that its not all plane waves, HOM can and will be present and if you do not account for them then you don't really know what you have. Accounting for them, not only improves the measurement, but yields other equally as important data along the way.
For example, if you averaged all four rotated measurements then you would have excluded any non-axial symmetric modes and the measurement would be more accurate. Diaphragm do have rocking modes and these will be non-axisymmetric - and a serious problem in some cases.
I am not sure that I agree with your comments on the mic. I think that the mic will be good where the measurements are good, both will become tentative at higher frequencies. And these modern mics are actually pretty good. I used to design microphones and there is no reason that a good mic has to be expensive anymore. 30 years ago you needed a B&K mic to get anything reliable. Today that's not the case.
and
I don't think that many (any?) of the rest of us have that kind of resources. So what will you sell us those tubes for? $150?
If you have "good gating" then why do you need the tube dampened? TEF is very old technology, there are much better systems avaialble these days for free - like HolmImpulse for example.
Point 1
If there are only plane waves in the tube then the distance that we measure along the tube will not matter, its all the same, right? But there is a slope discontinuity at the junction between the driver and the tube and this will generate Higher Order Modes (HOM). These modes will propagate with a wavenumber that is complex below the cut-in frequency and real above that frequency. In a complex exponential wave this means that below cut-in the waves are damped exponentials that decay expoenetially, but complex exponentials above cut-in which are true propagating waves. The damped expoentials are called evanescent waves and they decay and eventually disapate as they propagate. If you measure too close to the junction you will be measureing these waves and you don't want to do that.
As to the non-axisymmetric performance, I have no data on that other than Don Keele said that he measure substantrial differences when he did it. The issue with non-axisymmetric waves is that they have much lower cut-in frequencies than the axial ones. So they affect the measurements at a much lower frequency.
Why don't you run the test again, maybe with two different brands of drivers and post the data for all of us to see? 1 dB in the tube is actually a fairly substantial non-axi motion when you think about it in linear terms as a percentage.
Finally - all work that I know of for PWT assume plane waves. This all falls apart when you understand that its not all plane waves, HOM can and will be present and if you do not account for them then you don't really know what you have. Accounting for them, not only improves the measurement, but yields other equally as important data along the way.
For example, if you averaged all four rotated measurements then you would have excluded any non-axial symmetric modes and the measurement would be more accurate. Diaphragm do have rocking modes and these will be non-axisymmetric - and a serious problem in some cases.
I am not sure that I agree with your comments on the mic. I think that the mic will be good where the measurements are good, both will become tentative at higher frequencies. And these modern mics are actually pretty good. I used to design microphones and there is no reason that a good mic has to be expensive anymore. 30 years ago you needed a B&K mic to get anything reliable. Today that's not the case.
Last edited:
Hi Earl;
I will see what I can build them for and post. I built mine on hand wheel machines since I only made the one. I made mine 1'5" diameter as that was the driver family of my interest. Since diameter limits the HF, I personally would not use adapters.
Gating is great for the measurement no doubt, the dampening is to terminate the tube and present a constant impedance to the driver over a wide frequency band.
So are phonographs, I would say the topic is still open for debate. I don't want to turn this into a measurement platform debate. I was asked what I used and answered, the one I use the most. I also have LMS, MLSSA, the full version of ARTA and have used SMAART and have mucked around with some of the freeware stuff including Holm, all have weaknesses and limitations, given a single choice however for me it would be TEF.
I will dust off the tube and measure the axisymmetric performance of my reference driver which is a JBL 2451 with a Brush Wellman Be diaphragm and post it, in a day or two. The only other brand drivers I currently have is a pair of Rankest Hind's 1" drivers a friend sent me to measure, since I don't have a dedicated 1" PWT I don't think a comparison would be relevant, it certainly wouldn't be favorable.
About the low cost mic comment, I didn't mean that one couldn't use one for comparative purposes, simply that using such a mic would leave the measurement test repeatability between multiple users suspect at best.
All the best,
Barry.
I will see what I can build them for and post. I built mine on hand wheel machines since I only made the one. I made mine 1'5" diameter as that was the driver family of my interest. Since diameter limits the HF, I personally would not use adapters.
Gating is great for the measurement no doubt, the dampening is to terminate the tube and present a constant impedance to the driver over a wide frequency band.
So are phonographs, I would say the topic is still open for debate. I don't want to turn this into a measurement platform debate. I was asked what I used and answered, the one I use the most. I also have LMS, MLSSA, the full version of ARTA and have used SMAART and have mucked around with some of the freeware stuff including Holm, all have weaknesses and limitations, given a single choice however for me it would be TEF.
I will dust off the tube and measure the axisymmetric performance of my reference driver which is a JBL 2451 with a Brush Wellman Be diaphragm and post it, in a day or two. The only other brand drivers I currently have is a pair of Rankest Hind's 1" drivers a friend sent me to measure, since I don't have a dedicated 1" PWT I don't think a comparison would be relevant, it certainly wouldn't be favorable.
About the low cost mic comment, I didn't mean that one couldn't use one for comparative purposes, simply that using such a mic would leave the measurement test repeatability between multiple users suspect at best.
All the best,
Barry.
Low cost mics can have problem with stability of absolute levels, i.e. sensitivity, but for frequencies below about 8 kHz even the cheapest mics are dead flat. Its so easy to do, there would be no reason that it wouldn't be flat. Above 8 kHz and even the tube can be suspect as I said.
There may be pro's and cons of any particular piece of software, thats true, but as to the technique, there was an AES paper several years ago that put that subject to rest. It's only open to debate for people who didn't read that paper, otherwise log swept sine is the clear winner.
There may be pro's and cons of any particular piece of software, thats true, but as to the technique, there was an AES paper several years ago that put that subject to rest. It's only open to debate for people who didn't read that paper, otherwise log swept sine is the clear winner.
The paper that I was refeering to was not about plane waves, but about a comparison of all the know techniques for measurement. They looked at sensitivity to noise and nonlinearities as well as power input. They concluded that swept sine had the lowest energy input for a given output (hence high SNR), but that log swept sine had the added advantage of insentivity to nonlinearities. TEF is not a log sweep, but a linear sweep. And in Heysers original work he used quadrature to determine the amplituude at a given frequency. The more effective approach is to use deconvolution with the input signal (not possible in Heysers day) to find the impulse response directly and then the frequency response from the impulse response (Holm is the ideal implimentation of this technique, clumsey software, but ideal technique).
See AES Article
Title: Building a Plane-Wave Tube: Experimental and Theoretical Aspects
Author: Roberto Magalotti
Author: Carlo Zuccatti
Author: Paolo Pasini
Publication: AES-J, Vol. 47, No. 7/8, p. 596, Jul/Aug-1999
Abstract: The primary purpose of this report is to describe the building and testing of a plane-wave tube for measuring compression drivers in an audio laboratory. A new method for testing the tube for undesired reflections was devised, with and without sound absorbing materials. The experimental results show good agreement with the predicted behavior outlined in the AES document on plane-wave tubes, and confirm the predictions on usable bandwidth limits.
I have attached a copy here for your convenience.
It also references the relavent AES Standard:
Title: AES Information Document — Plane wave tubes: design and practice
Author:
Publication: AES-1id-1991 (r2003)
Abstract: Specification of Loudspeaker Components used in Professional Audio and Sound Reinforcement, in section 2.2.1, calls for the use of plane-wave tube measurement of horn drivers. Because many variations and results are possible, depending on the details of construction of plane-wave tubes, this document discusses those variations for the purpose of encouraging further experimentation.
Regards,
WHG
(copyrighted material removed)
Title: Building a Plane-Wave Tube: Experimental and Theoretical Aspects
Author: Roberto Magalotti
Author: Carlo Zuccatti
Author: Paolo Pasini
Publication: AES-J, Vol. 47, No. 7/8, p. 596, Jul/Aug-1999
Abstract: The primary purpose of this report is to describe the building and testing of a plane-wave tube for measuring compression drivers in an audio laboratory. A new method for testing the tube for undesired reflections was devised, with and without sound absorbing materials. The experimental results show good agreement with the predicted behavior outlined in the AES document on plane-wave tubes, and confirm the predictions on usable bandwidth limits.
I have attached a copy here for your convenience.
It also references the relavent AES Standard:
Title: AES Information Document — Plane wave tubes: design and practice
Author:
Publication: AES-1id-1991 (r2003)
Abstract: Specification of Loudspeaker Components used in Professional Audio and Sound Reinforcement, in section 2.2.1, calls for the use of plane-wave tube measurement of horn drivers. Because many variations and results are possible, depending on the details of construction of plane-wave tubes, this document discusses those variations for the purpose of encouraging further experimentation.
Regards,
WHG
(copyrighted material removed)
Non-axi-symmetric behavior would be due to the diaphragm rocking more than anything else. There is not much to constrain this (these) mode(s) because of the way a compression driver is supported. there should not be any forces to excite these modes however, so its kind of a balancing act. One that some companies do well and others not so well.
Good paper William. The authors show pretty good agreement between impulse techniques and damped tubes of both open or closed ends types. I think the referenced AES working group paper was pretty good too. Fancher Murray, who did all the JBL compression driver work when I was there was instrumental in it.
I think the OP was interested in woofers on a terminated tube. Has anyone seen any references on that?
David
I think the OP was interested in woofers on a terminated tube. Has anyone seen any references on that?
David
You may want to have a look at Fahy's article (attached).
Regards,
WHG
(copyrighted material removed)
Hi Earl and all;
This is how I normally view multiple lines of data, a curve is selected as reference,(the top curve) and becomes the straight line on the bottom and all additional data is compared against it.
I neglected to show cursers but the variations are larger than I remember, about 1.1dB below 8kHz and increasing above. I took five measurements at 0°, 90°, 180°, 270° and 360°, the fifth obviously being a return to the 0° position and the curve is absolutely line on line indicating that the variations are within the driver assembly. Interesting to say the least, something I want to chase down an answer as to what is actually at work here.
I can of course display this probably any way you would like to view it.
This is how I normally view multiple lines of data, a curve is selected as reference,(the top curve) and becomes the straight line on the bottom and all additional data is compared against it.
I neglected to show cursers but the variations are larger than I remember, about 1.1dB below 8kHz and increasing above. I took five measurements at 0°, 90°, 180°, 270° and 360°, the fifth obviously being a return to the 0° position and the curve is absolutely line on line indicating that the variations are within the driver assembly. Interesting to say the least, something I want to chase down an answer as to what is actually at work here.
I can of course display this probably any way you would like to view it.
Attachments
I would prefer to see a scale of say 1 dB per division, 10 dB is way too hard to read.
I think that this proves my point. The non-symetric modes are cutting in at about 5 kHz, which is well below the predicted axisymmetric mode cut-in frequency for this tube.
Clearly, above 8 kHz the results will vary by several dB depending on how the driver is mounted. And these are just the non-axisymmetric modes - there are even more errors when you also consider the symmetric ones. To sort those out you have to vary the mic location along the tube.
I think that this proves my point. The non-symetric modes are cutting in at about 5 kHz, which is well below the predicted axisymmetric mode cut-in frequency for this tube.
Clearly, above 8 kHz the results will vary by several dB depending on how the driver is mounted. And these are just the non-axisymmetric modes - there are even more errors when you also consider the symmetric ones. To sort those out you have to vary the mic location along the tube.
Sorry, I had to remove the AES papers- they're copyrighted and we can't have them posted here without permission.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.