Protecting an Empty Barrel
My research library is huge, as it has been built up over many years; and I paid for each and every ASA & AES article in it, other than those provided to me free by their authors. The intent of copyright laws is to prevent the unauthorized publication of these and other texts in large quantities for the purposes of sale and profit. As these same works are available at most college libraries, where they may be freely read and copied for just the cost of running a copy machine, the redaction here of several old articles, of only a few pages each, seems to me, to be somewhat pedantic. When copyright laws are enforced for the purpose of knowledge control, we all lose, author and reader as well. In the universe of ”publish or perish”, the motivation to author a technical work does not come from the revenues to be had by the sale of copies of it, particularly when the market is as esoteric as that of acoustics. In any event, in the future I will share such information in other, friendlier, venues where redaction does not occur for trivial reasons.
WHG
My research library is huge, as it has been built up over many years; and I paid for each and every ASA & AES article in it, other than those provided to me free by their authors. The intent of copyright laws is to prevent the unauthorized publication of these and other texts in large quantities for the purposes of sale and profit. As these same works are available at most college libraries, where they may be freely read and copied for just the cost of running a copy machine, the redaction here of several old articles, of only a few pages each, seems to me, to be somewhat pedantic. When copyright laws are enforced for the purpose of knowledge control, we all lose, author and reader as well. In the universe of ”publish or perish”, the motivation to author a technical work does not come from the revenues to be had by the sale of copies of it, particularly when the market is as esoteric as that of acoustics. In any event, in the future I will share such information in other, friendlier, venues where redaction does not occur for trivial reasons.
WHG
Last edited:
Wonder if you would care to expand on how plane wave tubes are used for microphone calibration? Quite frankly, that entire topic has typically made my brain seize up. Every once in a while I have read something on it that I think I understood, but then a short while later I realize that no, I didn't quite get it. 
_-_-bear
_-_-bear
Testing microphones in a plane-wave tube
> expand on how plane wave tubes are used for microphone calibration
Sure. The frequency response curves that appear on microphone mfgr. web sites are typically said to be made with plane waves. That is, the assumption is that the source is far away and that the wave has expanded far enough to be (nearly) plane. Historically, those measurements are made in anechoic chambers. More recently frequency response measurements may be made using quasi-anechoic techniques. But that typically only works down to about 300 Hz or so. I created a technique to extend that range but it's difficult to do.
An alternative to an anechoic chamber is to use a plane wave tube. A loudspeaker on one end, an absorbing termination on the other end, and the wave propagates down the tube as a (mostly) plane wave. With a sufficiently long termination the bandwidth of such a tube can extend down to very low frequencies. Gerhardt Bore writes about the plane-wave tube technique in section 5 here:
www.neumann.com/download.php?download=lect0044.PDF
He describes a plane-wave tube that is 44 cm in diameter and 14 m long (!).
It seems to me that this problem is like everything else. It's easy to build a plane-wave tube but difficult to build a good one.
> expand on how plane wave tubes are used for microphone calibration
Sure. The frequency response curves that appear on microphone mfgr. web sites are typically said to be made with plane waves. That is, the assumption is that the source is far away and that the wave has expanded far enough to be (nearly) plane. Historically, those measurements are made in anechoic chambers. More recently frequency response measurements may be made using quasi-anechoic techniques. But that typically only works down to about 300 Hz or so. I created a technique to extend that range but it's difficult to do.
An alternative to an anechoic chamber is to use a plane wave tube. A loudspeaker on one end, an absorbing termination on the other end, and the wave propagates down the tube as a (mostly) plane wave. With a sufficiently long termination the bandwidth of such a tube can extend down to very low frequencies. Gerhardt Bore writes about the plane-wave tube technique in section 5 here:
www.neumann.com/download.php?download=lect0044.PDF
He describes a plane-wave tube that is 44 cm in diameter and 14 m long (!).
It seems to me that this problem is like everything else. It's easy to build a plane-wave tube but difficult to build a good one.
You realize that microphone response within a plane wave tube would have a very different high end? In a free field normal diffraction effects gives the microphone directivity and flatter response on axis. The pressure response (response when the mic is driven within a small volume) rolls off prematurely.
See if you can find the old B&K application notes on their instrumentation microphones. They cover this very well.
David
See if you can find the old B&K application notes on their instrumentation microphones. They cover this very well.
David
Maybe someone can help me out, but I am not understanding where the basis for "calibration", where the reference comes from. Everything is non-flat, so how is "flat" determined if you do not start with some sort of transducer that is "flat"??
In other words, day one, before anyone figured anything out, how was a flat reference determined if there is nothing to compare against? There is probably a straightforward method/explanation, but I don't know what it is...
_-_-bear
In other words, day one, before anyone figured anything out, how was a flat reference determined if there is nothing to compare against? There is probably a straightforward method/explanation, but I don't know what it is...
_-_-bear
Last edited:
Hi Bear,
Measurement microphone calibration - Wikipedia, the free encyclopedia
Good question. See the section on Reciprocity Calibration.
David
Measurement microphone calibration - Wikipedia, the free encyclopedia
Good question. See the section on Reciprocity Calibration.
David
> microphone response within a plane wave tube
> would have a very different high end
If the plane-wave tube produces a plane wave, then the response of the microphone would be the same as in free space, that is, the same as in normal use. But I'm not planning on using the tube for high-frequency measurement. I'll use quasi-anechoic techniques for that. And my tube will be large enough in diameter that its high-frequency cutoff will be below 20 kHz in any case.
Low-frequency response is a problem for cardioid and figure-eight microphones. Typical ones roll off at 150 Hz or so.
The actual calibration comes from using the so-called substitution method. I'll place an instrumentation microphone in the measurement position first, and measure the system response. Then the device under test is placed in the same position and the response is measured again. Finally, the reference response is subtracted (in dB) from the response for the DUT, and that gives the 'calibrated' measurement. All of this is predicated on the sound in the tube being an actual plane wave, and not having any reflections from the terminated end.
> would have a very different high end
If the plane-wave tube produces a plane wave, then the response of the microphone would be the same as in free space, that is, the same as in normal use. But I'm not planning on using the tube for high-frequency measurement. I'll use quasi-anechoic techniques for that. And my tube will be large enough in diameter that its high-frequency cutoff will be below 20 kHz in any case.
Low-frequency response is a problem for cardioid and figure-eight microphones. Typical ones roll off at 150 Hz or so.
The actual calibration comes from using the so-called substitution method. I'll place an instrumentation microphone in the measurement position first, and measure the system response. Then the device under test is placed in the same position and the response is measured again. Finally, the reference response is subtracted (in dB) from the response for the DUT, and that gives the 'calibrated' measurement. All of this is predicated on the sound in the tube being an actual plane wave, and not having any reflections from the terminated end.
Redaction Not Required
AES put it on the web.
http://www.aes.org/tmpFiles/aessc/20120622/aes-1id-1991-r2003-i.pdf
AES put it on the web.
http://www.aes.org/tmpFiles/aessc/20120622/aes-1id-1991-r2003-i.pdf
ok, thank you.
I'm currently trying to design a setup for finding the response of a self-built probetube microphone which I plan to use for (pressure fluctuation) measurements in a gas burner exhaust. A good way might be connecting a regular pressure microphone and the probe microphone (which uses the same mic) through the wall of a plane wave tube at equal distance from the speaker. That way, since there's a plane wave in the tube, I could be sure the same sound pressure at the same phase arrives at both the reference mic and the probe tip simultaneously.
What I'm thinking of is something like the nonreflecting closed tube from the Boré paper posted earlier. I need to go down to about 100 Hz, so I might make a tube with a 12" woofer, 30 cm ID, put my mics at 1 m away from the speaker and start the absorbing wedge right after the mics. That should be at least 1 m long, maybe more. Cone shape would be ideal, I guess.
I'd appreciate your thoughts on that.
I'm currently trying to design a setup for finding the response of a self-built probetube microphone which I plan to use for (pressure fluctuation) measurements in a gas burner exhaust. A good way might be connecting a regular pressure microphone and the probe microphone (which uses the same mic) through the wall of a plane wave tube at equal distance from the speaker. That way, since there's a plane wave in the tube, I could be sure the same sound pressure at the same phase arrives at both the reference mic and the probe tip simultaneously.
What I'm thinking of is something like the nonreflecting closed tube from the Boré paper posted earlier. I need to go down to about 100 Hz, so I might make a tube with a 12" woofer, 30 cm ID, put my mics at 1 m away from the speaker and start the absorbing wedge right after the mics. That should be at least 1 m long, maybe more. Cone shape would be ideal, I guess.
I'd appreciate your thoughts on that.
The Mic is a 1/4" GRAS pressure mic.
Yes, I'm looking for the transfer function of the added tube.
The length of the tube is at least 30 cm and it includes a 90° curve.
There's two seperate things I want to measure:
1. a ~50 Pa pressure pulse of about 10 ms (so corresponding frequency is 100 Hz)
2. fluctuations of about 80 Pa at about 100 and 250 Hz.
Klaus
Yes, I'm looking for the transfer function of the added tube.
The length of the tube is at least 30 cm and it includes a 90° curve.
There's two seperate things I want to measure:
1. a ~50 Pa pressure pulse of about 10 ms (so corresponding frequency is 100 Hz)
2. fluctuations of about 80 Pa at about 100 and 250 Hz.
Klaus
Any ideas? Is it a feasible idea or completely weird?
One thing I forgot in my last post: the probe tube is 2mm internal diameter.
John Eargle's Loudspeaker Handbook has this formula for sound pressure in a plane wave tube: L_p = 94 + 20 log(sqrt(W_A*rho_0*c/S_T))
where W_A is the acoustic power in W and S_T the cross-sectional area of the tube. That should still be valid even for a larger tube. In that case a 10" woofer and thus a smaller tube might be enough.
One thing I forgot in my last post: the probe tube is 2mm internal diameter.
John Eargle's Loudspeaker Handbook has this formula for sound pressure in a plane wave tube: L_p = 94 + 20 log(sqrt(W_A*rho_0*c/S_T))
where W_A is the acoustic power in W and S_T the cross-sectional area of the tube. That should still be valid even for a larger tube. In that case a 10" woofer and thus a smaller tube might be enough.
Sorry I missed your reply;
I don't think you need anything exquisite to accomplish your goal. I would simply mount a small speaker face down on a piece of something rigid with a pluggable port for the mic and a port for the tube at some same distance from center. Test with the tube inserted and plug the port or cap the tube when comparing to insure the compression chamber volume is unchanged.
One caution, with a sealed compression chamber you can very easily overdrive the mic and possibly damage it. Carefully creep up on the drive level, when you reach the voltage limit of the mic pre the frequency response curve will artificially flatten out and the magnitude will cease to increase. You must however find and know this level and be sure to stay below it to insure that your real test data is valid.
Barry.
I don't think you need anything exquisite to accomplish your goal. I would simply mount a small speaker face down on a piece of something rigid with a pluggable port for the mic and a port for the tube at some same distance from center. Test with the tube inserted and plug the port or cap the tube when comparing to insure the compression chamber volume is unchanged.
One caution, with a sealed compression chamber you can very easily overdrive the mic and possibly damage it. Carefully creep up on the drive level, when you reach the voltage limit of the mic pre the frequency response curve will artificially flatten out and the magnitude will cease to increase. You must however find and know this level and be sure to stay below it to insure that your real test data is valid.
Barry.
Thanks Barry,
I'm not entirely convinced. Is it possible to exactly recreate those amplitudes with that setup? I just saw a thesis where they used something like this for pressure fluctuations of 2000 Pa. Not sure if the system would be fast enough. Also, symmetry is critical, so that the two pressure ports see exactly the same signal in terms of phase and amplitude. The size of a 12" ID (steel) tube wouldn't be a problem...
I'm not entirely convinced. Is it possible to exactly recreate those amplitudes with that setup? I just saw a thesis where they used something like this for pressure fluctuations of 2000 Pa. Not sure if the system would be fast enough. Also, symmetry is critical, so that the two pressure ports see exactly the same signal in terms of phase and amplitude. The size of a 12" ID (steel) tube wouldn't be a problem...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.