Tweet
I have been helping my father build his very first set of speakers which are the Sunflower Redux from Paul Carmody. More specifically, I have been trying to provide technical expertise and guidance for him as he constructs it half way across the country where he lives. I have never built a set of speakers before either, but I am a mechanical engineer by profession and have some car audio experience in my past, so I like to think I can at least learn what is needed to get the job done right if I know where to start. It has been an an interesting and fun experience, and he is getting close to the point where it can hooked up to actually play something through it!
I have read several resources regarding port tuning, that in essence, said calculations will only get you in the ballpark and true port length should be tweaked during the last few steps of construction. If I understand correctly there are several methods to determine what a port is tuned to, but I am unsure about which methods will work best (or at all) and am looking for more information in general. He does not have any audio analyzers or precision acoustic measurement tools at his disposal. I'd guess wouldn't want to justify a huge cost in acquiring them either since I don't think he is looking into speaker building as a long term hobby and may only build one other set depending on the outcome of this project. The methods and ideas I am aware of are listed below, but please correct me if I am wrong or have missed something. I am open to all good input on this one so please set me straight.
FYI if it helps to answer my question, the box is ~43L and uses Dayton RS225-8 woofers. Paul says the tuning frequency should be around Fb=23 Hz which would be a 2" diameter x 8" long round vent.
1.) Measure system frequency response using a calibrated microphone
I infer that this is the best way to do it, and if I understand correctly is simply measuring the frequency response output with a signal generator input. I didn't look too much more into this one after I saw the price of OmniMic. I guess this post is really asking about alternatives to this method.
2.) Measure impedance of the woofer
I believe this one basically amounts to incrementing a frequency input and measuring the impedance(resistance?) across the driver at the same time. I was thinking that we could use an iPod with test tones and play something from around 15Hz - 50Hz one at a time in ~5Hz increments or so. By running the multimeter leads through the port tube we could then measure impedance and plot it very roughly vs. frequency until we mapped out the double-peak bass reflex impedance chart. IIRC, the actual Fb is the frequency at the lowest impedance point in the valley between the two impedance peaks. I would expect by starting with a port tube cut an inch or two longer than the calculated length, that the interpolated Fb from the measurements would be lower than the 23 Hz target. So then in theory, this process could be repeated and the port could be trimmed in 1/2" increments until the impedance valley measured for a given frequency sweep came down to the 23 Hz region. Of course once the general Fb was found after the first round of measurements, the frequency tones could be varied in 1Hz increments for a smaller region so as to just save time and get more accurate results.
I know it would be pretty tedious, but I am wondering if this method is even valid. One other concern I have is that I'd guess his multimeter will probably only measure DC resistance and likely does not have a way to measure true impedance. Given that impedance has AC components, I am unsure if this method would work using a DC multimeter. Surely the two peaks around the Fb would not be accurate, but wouldn't the valley (the part I think we care about) still be apparent enough to determine the corresponding frequency with some degree of certainty? After all, I think that at this resonant Fb where impedance is at a low, that phase angle is also 0 degrees which would then imply that there are no imaginary components to the impedance and it is completely equal to DCR in that region. Is this correct and would this method work using a DC multimeter?
3.) Visually observe woofer excursion
If not mistaken, when a woofer in a tuned enclosure is playing at the tuning frequency, that excursion is nearly zero and the port is doing most of the heavy lifting--most observable sound should be coming from there instead. So then the same test tones used in the method #2 above could be played in ~5 Hz increments and eventually 1Hz or 2Hz increments to identify the Fb by looking for when the woofer cone moves the least amount. As I think about it, I believe this may actually be using the same underlying principles as method #2 except instead of measuring impedance we'd just be visually observing the effects. I'm not sure if that's actually true or not, but is this method good enough to get within 5 Hz of the target Fb or so?
I have read several resources regarding port tuning, that in essence, said calculations will only get you in the ballpark and true port length should be tweaked during the last few steps of construction. If I understand correctly there are several methods to determine what a port is tuned to, but I am unsure about which methods will work best (or at all) and am looking for more information in general. He does not have any audio analyzers or precision acoustic measurement tools at his disposal. I'd guess wouldn't want to justify a huge cost in acquiring them either since I don't think he is looking into speaker building as a long term hobby and may only build one other set depending on the outcome of this project. The methods and ideas I am aware of are listed below, but please correct me if I am wrong or have missed something. I am open to all good input on this one so please set me straight.
FYI if it helps to answer my question, the box is ~43L and uses Dayton RS225-8 woofers. Paul says the tuning frequency should be around Fb=23 Hz which would be a 2" diameter x 8" long round vent.
1.) Measure system frequency response using a calibrated microphone
I infer that this is the best way to do it, and if I understand correctly is simply measuring the frequency response output with a signal generator input. I didn't look too much more into this one after I saw the price of OmniMic. I guess this post is really asking about alternatives to this method.
2.) Measure impedance of the woofer
I believe this one basically amounts to incrementing a frequency input and measuring the impedance(resistance?) across the driver at the same time. I was thinking that we could use an iPod with test tones and play something from around 15Hz - 50Hz one at a time in ~5Hz increments or so. By running the multimeter leads through the port tube we could then measure impedance and plot it very roughly vs. frequency until we mapped out the double-peak bass reflex impedance chart. IIRC, the actual Fb is the frequency at the lowest impedance point in the valley between the two impedance peaks. I would expect by starting with a port tube cut an inch or two longer than the calculated length, that the interpolated Fb from the measurements would be lower than the 23 Hz target. So then in theory, this process could be repeated and the port could be trimmed in 1/2" increments until the impedance valley measured for a given frequency sweep came down to the 23 Hz region. Of course once the general Fb was found after the first round of measurements, the frequency tones could be varied in 1Hz increments for a smaller region so as to just save time and get more accurate results.
I know it would be pretty tedious, but I am wondering if this method is even valid. One other concern I have is that I'd guess his multimeter will probably only measure DC resistance and likely does not have a way to measure true impedance. Given that impedance has AC components, I am unsure if this method would work using a DC multimeter. Surely the two peaks around the Fb would not be accurate, but wouldn't the valley (the part I think we care about) still be apparent enough to determine the corresponding frequency with some degree of certainty? After all, I think that at this resonant Fb where impedance is at a low, that phase angle is also 0 degrees which would then imply that there are no imaginary components to the impedance and it is completely equal to DCR in that region. Is this correct and would this method work using a DC multimeter?
3.) Visually observe woofer excursion
If not mistaken, when a woofer in a tuned enclosure is playing at the tuning frequency, that excursion is nearly zero and the port is doing most of the heavy lifting--most observable sound should be coming from there instead. So then the same test tones used in the method #2 above could be played in ~5 Hz increments and eventually 1Hz or 2Hz increments to identify the Fb by looking for when the woofer cone moves the least amount. As I think about it, I believe this may actually be using the same underlying principles as method #2 except instead of measuring impedance we'd just be visually observing the effects. I'm not sure if that's actually true or not, but is this method good enough to get within 5 Hz of the target Fb or so?
Comment