Hi,
I am currently interested in the most precise as possible location of the drivers emission. I carried out some tests, mainly with tweeters at a distance varying from 34.4 cm and 10 cm. Results using sine and square wave seem to be quite consistent.
However there is a part of uncertainty, from 2 to 5 mm, for the exact location of the cell of the measuring microphone : what is its distance from the mike front ?
Can anybody provide some information about it for the Berhinger ECM8000 and the mike provided in the Clio Pocket kit ?
I am currently interested in the most precise as possible location of the drivers emission. I carried out some tests, mainly with tweeters at a distance varying from 34.4 cm and 10 cm. Results using sine and square wave seem to be quite consistent.
However there is a part of uncertainty, from 2 to 5 mm, for the exact location of the cell of the measuring microphone : what is its distance from the mike front ?
Can anybody provide some information about it for the Berhinger ECM8000 and the mike provided in the Clio Pocket kit ?
Spl is normally quoted at 1Metre from the cone/diaphragm.
Eg. 102dB for 1Watt at 1Metre , (Eminence Gamma) range.
Eg. 102dB for 1Watt at 1Metre , (Eminence Gamma) range.
Let me give more details.
One of my experience was to drive a tweeter with a sine at 1720 Hz in a region where its impedance is resistive. I fed one channel of a dual trace scope with the output of the amp and the other channel with the amplified output of the microphone.
With some precautions to avoid the influence of anything close to the apparatus, I slightly moved the mike forward and backward until the traces superposed.
The distance between the emitting area of the driver and the mike cell then should correspond to a full wavelength, 20 cm.
If I knew the position of the mike cell with precision, I could locate the location of the emitting area. I've done many tries and the procedure seems reliable.
If space and time allow, I'll try longer distances.
I am not measuring the standard sensitivity of a driver.
If I have to do, I use 2.83 Vrms, like everybody else, because loudspeakers are driven by voltage (or sometimes by current), never by power.
One of my experience was to drive a tweeter with a sine at 1720 Hz in a region where its impedance is resistive. I fed one channel of a dual trace scope with the output of the amp and the other channel with the amplified output of the microphone.
With some precautions to avoid the influence of anything close to the apparatus, I slightly moved the mike forward and backward until the traces superposed.
The distance between the emitting area of the driver and the mike cell then should correspond to a full wavelength, 20 cm.
If I knew the position of the mike cell with precision, I could locate the location of the emitting area. I've done many tries and the procedure seems reliable.
If space and time allow, I'll try longer distances.
I am not measuring the standard sensitivity of a driver.
If I have to do, I use 2.83 Vrms, like everybody else, because loudspeakers are driven by voltage (or sometimes by current), never by power.
Last edited:
You mean exactly where the diaphragm is. This would probably require some deconstruction. The spacer from the front of the actual capsule to the diaphragm inside is small, (I lost my picture of a deconstructed typical capsule). There is usually a small port under a felt disk that covers the front then there is a chamber between the diaphragm and front of the capsule that has a critical dimension. IIRC you can look in and make a decent estimate of the distance without destroying the capsule.
EDIT - Someone on the Yahoo micbuilders list might be able to give you the brand of capsule they use in the ECM8000, IIRC it is a common cheap one.
Last edited:
Thanks for your reply.
I've contacted Clio and asked them about the diaphragm location of the mike included in their Pocket system.
A kind reply said it's 1.2 mm from the front.
I then put this mike alongside an ECM8000, submitted them to periodic signals coming from a dome tweeter at a distance about 50 cm.
I look at their output signals on a dual trace scope.
After some adjustments, moving back and forth the Clio mike, I found that I can count on a distance of 4 mm for the ECM8000.
The uncertainty is about 1 mm, I can live with that in my measuring process.
Regards.
Do not assume that the diaphragm's physical location has no delay from the electrical drive. There is usually some significant acoustic delay as the drive propagates from the current in the voice coil to moving the diaphragm. And it changes with frequency. This is what limits applying feedback around drivers (or noise cancelling headphones, which is a similar problem).
Should you try to make your measurement process rely on relative comparison rather than absolute? You are not building a distance measuring apparatus - right?
Do all measurements from one fixed location and the diaphragm position is out i the equation?
//
Do all measurements from one fixed location and the diaphragm position is out i the equation?
//
I am conscious of that. The precise distance I am looking for may help to evaluate the delay with frequence you refer to.
Note that Hiquphon provides such a distance for its tweeters.
Once known, the distance of the emitting area from a reference point (my preference goes to the rear of the fixing plate) of various drivers (medium and tweeter) should make easier to mount them together with the correct alignment and delay.
I think the idea is worth some investigation.
Last edited:
Don't forget that different mic patterns can effect the phase between electrical output and stimulus. IIRC the usual measurement mic is a small omni.
All mics for the usual measurements done by diyers look omnidirectionnal and have quite similar shapes. When using sine signals, the results are very sensible to anything at a distance less than 1 m reflecting sound , first of them being the body of the experimenter himself.
My next tests aim to verify if various tools, methods and programs I have give coherent results for the phase response.
Mics manufacturer ISEM I contacted evaluates the distance diaphragm-front of his EMX-7150 ( emx-7150 ) as being less than 2 mm.
Hi Forr,
May be it's a bad idea but I think you would have more information by using a step instead of a sine for your measurements.
The differences between the signal front shape (amplifier and mike) would give you information on the "delay" during the electrical / acoustical transduction.
All the best !
May be it's a bad idea but I think you would have more information by using a step instead of a sine for your measurements.
The differences between the signal front shape (amplifier and mike) would give you information on the "delay" during the electrical / acoustical transduction.
All the best !
Hi Hervé,
I try both. My idea using sine is to be sure to obtain 0° phase shift between the drivers in position on their baffle at the crossover frequency as required, for example, by the 4th order Riley-Linkwitz.
As reminded above by 1audio, there can be unsuspected delays in the transductance process.
That was the reason why knowing the exact position of the mic diagram can help.
It is quite fun to do such little experiences because it teaches a lot to the experimenter.
I hope to find some correlation between the sine and step signals emitted by the drivers.
Last edited:
In post #6, using peridodic signals shown on a dual trace scope and a Clio Pocket mic as a reference, I've found that the distance between the front and the diaphragm of the Behringer ECM8000 mic (connected through a Stage Line MPA-102 preamp to the input of the Pcket) should be 4 mm.
As suggested in various posts, I made new tests with the Pocket system using log-chip signals converted to impulses.
This time, the distance was repeatedly found (three times, the sound source being a tweeter at 343.5 mm and 500 mm) to be 4.5 mm.
I consider the results with impulses more reliable than those using sine signals (because of my lack of an anechoic chamber).
But with some precautions, sine signals can do a quite good job.
This experience allowed me to familiarize with Clio Pocket system, once accustomed to it, it can make to earn a lot of time.
Having now answers to my question, it's time for me to measure drivers and loudspeakers.
As suggested in various posts, I made new tests with the Pocket system using log-chip signals converted to impulses.
This time, the distance was repeatedly found (three times, the sound source being a tweeter at 343.5 mm and 500 mm) to be 4.5 mm.
I consider the results with impulses more reliable than those using sine signals (because of my lack of an anechoic chamber).
But with some precautions, sine signals can do a quite good job.
This experience allowed me to familiarize with Clio Pocket system, once accustomed to it, it can make to earn a lot of time.
Having now answers to my question, it's time for me to measure drivers and loudspeakers.
Last edited:
I forgot to mention I went through this exercise 30yr. ago using a 10mm omni and an HP (it was still HP) network analyzer programmed to do the chirp. The instrument allowed you to delay the local oscillator which gives the exact delay, ideally the phase can be made flat (perfect impulse) but reality is usually different. BTW the speakers ended up very disappointing.
As you mentioned, the results from a sine sweep can be converted to an impulse response or a step response. Once you have output as a function of input over the full bandwidth (and have time/phase info included) you have all the responses via simple mathematical conversion (it's just Fourier theory). You can measure with a narrow pulse, but its hard to get enough power in the test signal that way to get far above ambient noise, so other methods (MLS, swept-sine, log-swept sine) are usually used for more practical results and with no more effort or time these days.
With a sine log-sweep, you can also measure or reject harmonic distortions and balance the signal to noise relative to typical ambient noise spectra. The sound during the test tends to attract curious neighbors, though.
With a sine log-sweep, you can also measure or reject harmonic distortions and balance the signal to noise relative to typical ambient noise spectra. The sound during the test tends to attract curious neighbors, though.
I did find a cheap ribbon tweeter that "stuck" at high level in a spurious resonance the results were then garbage. Accounting for different gain/phase with level is where a lot of differences that we don't measure lie.
Well, yeah. I was assuming sufficiently linear operation there. If you are looking at frequency response where it is changing as a function of level, then all bets are off (and it would be hard to say what such a measurement would be useful for other than to say 'try not to use it that loud'). Sometimes you have to shape the levels of a swept-sine test to keep LF drive down to things like sensitive tweeters which wouldn't normally see that (and then deconvolve with the same shaped drive signal).
- Status
- Not open for further replies.