for those experienced with speaker measurements, I have a few questions please:
there is a lot of references to different hrdw/sftw kits dedicated to speaker measurements. those automate a lot of the settings and make the job easy, so no one ever comments on those details. my issue is that I have access (through work) to some hrdw/sftw for general purpose spectral analysis. this is actually professional level equipment and a lot more expensive, but not designed for audio applications only. thus I have to manage the details of setting up data acquisition and processing. I will be testing drivers as mounted in a finished OB and outside.
my first question has to do with choosing excitation i.e. DAC output to feed to speakers. I can do any type: swept sine, random (white) noise within a band, pink noise or impulse. let's assume swept sine for maximum accuracy:
Q1: how do I choose the level that will not damage the tweeter for example when sweeping the slopes of its FR curve?
Q2: my signal generator DAC is capable of +-3.9Vmax and max +-35mA output, at 0.01Ohm output impedance. Would that be enough or should I insert an amp with attenuator?
Q3: any special precautions I should take for testing a pro bass driver (15" Definimax) on an OB near its resonance?
Q4: others often try to keep the same level of excitation across all drivers and measure SPL only. I was thinking of doing a true FRF (SPL/V). that way I can get the phase relative to the excitation and also be able to use different excitation levels on different drivers if needed. these curves then can be scaled to 2.83V input for estimating efficiency? excess phase (or driver alignment) I can correct for by using Jeff B's method in PCD. Does this sound Ok?
well that's enough for a starter. please feel free to share any bits of knowledge here. Please stick with the topic and THANKS a bunch!
there is a lot of references to different hrdw/sftw kits dedicated to speaker measurements. those automate a lot of the settings and make the job easy, so no one ever comments on those details. my issue is that I have access (through work) to some hrdw/sftw for general purpose spectral analysis. this is actually professional level equipment and a lot more expensive, but not designed for audio applications only. thus I have to manage the details of setting up data acquisition and processing. I will be testing drivers as mounted in a finished OB and outside.
my first question has to do with choosing excitation i.e. DAC output to feed to speakers. I can do any type: swept sine, random (white) noise within a band, pink noise or impulse. let's assume swept sine for maximum accuracy:
Q1: how do I choose the level that will not damage the tweeter for example when sweeping the slopes of its FR curve?
Q2: my signal generator DAC is capable of +-3.9Vmax and max +-35mA output, at 0.01Ohm output impedance. Would that be enough or should I insert an amp with attenuator?
Q3: any special precautions I should take for testing a pro bass driver (15" Definimax) on an OB near its resonance?
Q4: others often try to keep the same level of excitation across all drivers and measure SPL only. I was thinking of doing a true FRF (SPL/V). that way I can get the phase relative to the excitation and also be able to use different excitation levels on different drivers if needed. these curves then can be scaled to 2.83V input for estimating efficiency? excess phase (or driver alignment) I can correct for by using Jeff B's method in PCD. Does this sound Ok?
well that's enough for a starter. please feel free to share any bits of knowledge here. Please stick with the topic and THANKS a bunch!
Koja,
A) A sine wave only excites the driver at one frequency at a time, which is not specifically representative of music, which contains a wide band of frequencies.
1) Sine waves have a crest factor of only 3 dB, to be safe one would limit the test to at most 1/2 the level of the older AES standard using band limited pink noise with a 6 dB crest factor. The newer AES standard uses pink noise with 12 dB crest factor.
Given many tweeters have no decent power specs, but few will handle less than 2.83 v, you are generally safe with that level.
2) Should be plenty for near field measurements, though near field measurements are not indicative of far field response.
3) No, other than the points in #1. I have tested all the Eminence woofers at their rated AES power using sine waves (for short duration) before realizing that was double their rating 😱. Tweeter voice coils heat up and burn much more rapidly.
4) Using 2.83v at one meter measures sensitivity, not efficiency. In room response does not fall off at the usual 6 dB per doubling of distance as outdoors, so scaling may not be representative of actual sensitivity. Not familiar with Jeff B.
Art
Instead of building the test equipment from scratch, it is much easier to use a dedicated test software like Arta, plug a microfon into the mic input of your soundcard input and connect soundcard line out to any audio amp that drives your speaker.
This is what I do for a quick comparision where calibration of absolute level is arbitrary.
This is what I do for a quick comparision where calibration of absolute level is arbitrary.
Weltersys: Thanks for the pointers.>
Volt: I know a dedicated system is easier to use, but I already have a much more elaborate (and accurate) system I use daily in my job, so for me it would be a waste of $$s. I only need to research a bit so i can configure it for the job.
Volt: I know a dedicated system is easier to use, but I already have a much more elaborate (and accurate) system I use daily in my job, so for me it would be a waste of $$s. I only need to research a bit so i can configure it for the job.
- Status
- Not open for further replies.