Plane-wave tube (PWT) measurements can give you a lot of good information about a driver, but since they usually involve custom-built tubes of very long length filled with a carefully arranged gradient of sound-absorbing material, they have been out of reach for all but the most fanatical DIYers.
Thinking about this, it occurred to me that, with the advent of time-gated measurement software, PWTs don't have to be so complicated anymore. In fact, it seems to me that all you need is a cheap section of appropriate-diameter PVC pipe from your local hardware store--no stuffing or termination necessary.
The PWT will function correctly until the first reflection from the open end of the tube returns to the test microphone, which is positioned near the driver. So the length of your tube determines the necessary time gating.
For example, a typical 10-foot length of pipe would give good results down to about 65Hz--plenty low for most drivers.
It's a simple concept, but I haven't heard of anyone doing it this way.
Any thoughts?
Thinking about this, it occurred to me that, with the advent of time-gated measurement software, PWTs don't have to be so complicated anymore. In fact, it seems to me that all you need is a cheap section of appropriate-diameter PVC pipe from your local hardware store--no stuffing or termination necessary.
The PWT will function correctly until the first reflection from the open end of the tube returns to the test microphone, which is positioned near the driver. So the length of your tube determines the necessary time gating.
For example, a typical 10-foot length of pipe would give good results down to about 65Hz--plenty low for most drivers.
It's a simple concept, but I haven't heard of anyone doing it this way.
Any thoughts?
Last edited:
I played with the idea of using a tube to measure driver SPL profiles some years back. I built a tube using a 12" sonotube like those used for pouring concrete pillars. In my case I positioned the mic at the opposite end of the tube from the driver and lined the tube with 2" thick foam bumpy matress pad. It didn't work because what I basically constructed was a kundt tube.
What you have is a puzzlement to me. Why are you locating the mic close to the driver and having the tube end open?
What you have is a puzzlement to me. Why are you locating the mic close to the driver and having the tube end open?
Locating the mic close to where the driver mounts to the PWT allows you to get a clean measurement of the driver's power response while the wave propagates down the PWT, and before its reflection returns from the far end. (Whether the far end is open or closed, it will reflect the wave.)
Measuring at the far end of the PWT means you cannot gate out the reflection, because it would cross the mic nearly simultaneous to the wave's arrival.
Measuring at the far end of the PWT means you cannot gate out the reflection, because it would cross the mic nearly simultaneous to the wave's arrival.
I didn't realize plain wave tubes were so complicated. I'm familiar with the compression driver types and at JBL these were just clear Lexan tubes with a long sliver of fiberglass. They didn't give perfect response but the nulls were pretty well damped and were a constant in the measurement (might allow for some curve subtraction for correction).
Time gating can certainly work. Is it necessary?
David S.
Time gating can certainly work. Is it necessary?
David S.
I worked at Shure in the '70s . I remember plane wave tube measurements. I recall the measurement mikes were small ( 0.5 or 0.25 inch diameter omni condensers ). It seems they were mounted very close to the driver and at right angles. I think the tubes had a long tapered filler of lambs wool? and were open at the far end.
We also made similar measurements in the throat of various horns.
Don
We also made similar measurements in the throat of various horns.
Don
Last edited:
I have never seen a PWT in person, but my impression was that they had to be constructed with a fair amount of care to control the rate of absorption along the length and make sure any modes were well damped. That's why they're found mainly in manufacturer R&D labs.
So it was interesting to me to realize that Joe Hobbyist could get the same job done with nothing more than a freeware measurement app and $20 worth of water pipe.
And if the PWTs of old were often somewhat compromised, then gated measurements on open pipe would likely tend to be more accurate and repeatable.
For those who like to test drivers and build databases, it seems to me that this kind of thing might get them all excited. Many people are very interested in drivers' real-world power responses. It would be easy and cheap to measure drivers up to 4". Larger diameter pipes of 6", 8", and more get a bit more expensive and cumbersome.
Last edited:
The mics are being used in a pressure mode rather than a free field mode. In the free field the directivity of the mic gives flatter response on axis. In the tube itself you don't get that benefit, so a small mic diameter raises the rolloff frequency. I think JBL used the 1/4" unit also.
There was an AES workgroup on best practices for this. Might be good to read what they said if anybody has a copy.
David S.
There was an AES workgroup on best practices for this. Might be good to read what they said if anybody has a copy.
David S.
Bill,
I think you should get some sonotube and have a go at this. I think it could be very interesting, but I'm not exactly sure what the curves will look like. You are replacing the normal air load with a frequency independent resistive load (somewhat like an infinite horn). Based on this the units will give a curve unlike what we are familiar for.
I'm sure it is totally applicable for drivers meant to be used in horns, but not so sure what it means for conventional use.
David S.
I think you should get some sonotube and have a go at this. I think it could be very interesting, but I'm not exactly sure what the curves will look like. You are replacing the normal air load with a frequency independent resistive load (somewhat like an infinite horn). Based on this the units will give a curve unlike what we are familiar for.
I'm sure it is totally applicable for drivers meant to be used in horns, but not so sure what it means for conventional use.
David S.
Thanks for the description. That's how I have seen them depicted. I wonder if the distance between the driver and the mic is important. It seems to me that it wouldn't be as long as the mic is in front of the absorption filler and there are no standing waves in the tube...
How were the horn throat measurements used? That seems like it should be a gated measurement, too, to remove the horn mouth reflections.
Dave,
I've only been thinking of PVC so far--do you think sections of sonotube taped together would be stiff enough?
Yes, PWT measurements would be most applicable to horn loading, but it seems to me that a PWT's clean (no room reflections) power response measurement could be helpful in designing direct-radiator systems for smoothest off-axis response, too. Especially if you don't have access to an anechoic chamber...
Hi Bill
The best article on PWT was by Roberto Magalotti currently at B&C in AES about 15 years ago. I believe that he did his MS thesis on them. I worked with this tube when I was working with him at B&C.
But his design was well before good gating measurements were available and I now believe that what you suggest is the way to go.
A couple of points:
1) the mic should not be too close to the aperature connection because of evanescent wave effects. Further is better, but that means a longer tube.
2) The use of two microphone can actually surpress the open end reflection, but this is probably more trouble than its worth.
3) while at B&C I developed a technique to measuring how flat the wavefront is at the aperature by measuring at different distances along the PWT. From this one can calculate the curvature of the wavefront which is very important to designing a good waveguide.
4) it is also know that non-axisymmetric performance in a Compression Driver is a big issue and by rotaing the tube on its mount one can calculate not only the axisymmetric modes of the device, but the first few non-axisymmetric ones as well.
I was reading the driver reviews in Voice coil and thought that what is clearly missing in our world are good compression driver measurements. I thought of offering to do them for voice coil which would involve making a PWT. Compression drivers measured on horns are dominated by the horn and there simply are no good sources of data like this that is of the driver itself. This is sorely needed.
So I would like to volunteer to help out this project in any way that I can as I believe that this kind of data is crucial to moving the needle of waveguide designs. We have no idea what the aperature waveshape looks like, except for some cursory data done at B&C. Based on some other conversations that I had (Don Keele for example) drivers can have substantial nonaxial issues. There is simply no data available anywhere on this kind of problem.
The best article on PWT was by Roberto Magalotti currently at B&C in AES about 15 years ago. I believe that he did his MS thesis on them. I worked with this tube when I was working with him at B&C.
But his design was well before good gating measurements were available and I now believe that what you suggest is the way to go.
A couple of points:
1) the mic should not be too close to the aperature connection because of evanescent wave effects. Further is better, but that means a longer tube.
2) The use of two microphone can actually surpress the open end reflection, but this is probably more trouble than its worth.
3) while at B&C I developed a technique to measuring how flat the wavefront is at the aperature by measuring at different distances along the PWT. From this one can calculate the curvature of the wavefront which is very important to designing a good waveguide.
4) it is also know that non-axisymmetric performance in a Compression Driver is a big issue and by rotaing the tube on its mount one can calculate not only the axisymmetric modes of the device, but the first few non-axisymmetric ones as well.
I was reading the driver reviews in Voice coil and thought that what is clearly missing in our world are good compression driver measurements. I thought of offering to do them for voice coil which would involve making a PWT. Compression drivers measured on horns are dominated by the horn and there simply are no good sources of data like this that is of the driver itself. This is sorely needed.
So I would like to volunteer to help out this project in any way that I can as I believe that this kind of data is crucial to moving the needle of waveguide designs. We have no idea what the aperature waveshape looks like, except for some cursory data done at B&C. Based on some other conversations that I had (Don Keele for example) drivers can have substantial nonaxial issues. There is simply no data available anywhere on this kind of problem.
Earl,
Excellent.
A great idea. Glad I ran across this thread today.
Did the curvature calculation depend on arrival times or phase measurement with mics plane to the surface of the tube, carefully spaced, or a combination of a surface mic and a mic in the center of the tube along the way? Trying how to envision determining the shape of a bubble sliding down the length of the tube... and would there not be adhesion or surface tension effects at the boundary of the tube surface and the wave??
_-_-bear
Excellent.
A great idea. Glad I ran across this thread today.
Did the curvature calculation depend on arrival times or phase measurement with mics plane to the surface of the tube, carefully spaced, or a combination of a surface mic and a mic in the center of the tube along the way? Trying how to envision determining the shape of a bubble sliding down the length of the tube... and would there not be adhesion or surface tension effects at the boundary of the tube surface and the wave??
_-_-bear
The "bubble" does not "slide" down the tube, the wavefront for all modes except the plane wave one will bounce off of the walls of the tube. By measuring at points along the tube - at the tubes surface - one can calculate (after some "simple" algebra
) the contribution of every non-planar mode. Of course if the wavefront is perfectly planar then every microphone signal will report the same thing except for a phase delay. This is not likely to be the case, especially at HF.To the extent that there is nonplanar motion, any single point measurement will be in error because of these same effects. So sorting it all out is necessary for any high accuracy measurement.
This is, in my mind, the biggest problem. Since we know that the wavefront is not flat - the walls always have some angle to them and hence some wavefront curvature - then there is always a slope discontinuity when a straight walled tube is used. There are two ways to deal with this. 1) inverse taper the "throat" of the PWT to properly align it with the driver (this would require a differnt "coupler" for every driver exit slope. Not impossible just not very attractive.) 2) mathematically account for this effect and remove it (I would be inclined to do this, except that I have not yet done it and it may be more problematic than it appears at first.)
Of course plane wave tubes are always good below the point where the first cross mode can occur - that's why they are called "plane wave" tubes. But most of the interesting and differentiating effects will occur just about this transition point and certainly above. And it is often not realized that non-axisymmetric modes have lower cuit-in frequiencies than do the axially symmetric ones. So the upper limit of a PWT is much lower than often assumed.
The complexity of these issues is why we don't see this technique used all that much. However, these days, with power PCs at our disposal all of these complications CAN BE surmounted. Its just a lot of work.
Thanks for weighing in, Earl.
Here's an idea: In the interest of advancing the cause of PWT measurements and getting more of them out there, how about writing a How-To wiki here on DIYAudio that will teach audio hobbyists how to do a rudimentary PWT simply and cheaply? Assume a low common denominator of equipment--say a Behringer test mic, Holmimpulse and hardware-store pipe. Outline a set of standardized protocols to obtain useful results that could be compared with the results of others using the same standardized setup. What kind of testing would be achievable with such a setup?
I think that the issue is simplicity. No one here can afford to make a PWT like is shown above, I can't. So if it isn't simple then no one will do it.
The problem that I am faced with is simply time.
As I outlined above, below say 5 kHz for a 1" driver almost anything works. There are only plane waves and the mic doesn't matter so much, neither does the mounting, etc. But above about 5 kHz and all the problems that I write about above start to come into play.
So what may be the most reasonable approach would be to just start with a system, get widespead usage and then refine it as required.
I just want to be clear that the simple approach is not going to be very good above about 5 kHz.
The problem that I am faced with is simply time.
As I outlined above, below say 5 kHz for a 1" driver almost anything works. There are only plane waves and the mic doesn't matter so much, neither does the mounting, etc. But above about 5 kHz and all the problems that I write about above start to come into play.
So what may be the most reasonable approach would be to just start with a system, get widespead usage and then refine it as required.
I just want to be clear that the simple approach is not going to be very good above about 5 kHz.
e o
Tell us a bit about that one you show. Are the fittings on left end of the tube custom made? Did you design it? What are you testing with it?
Are there 'gotchas' when using or building the device?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.