As a brief introduction, I find myself circling the rabbit hole of DIY speaker building. I'm on the precipice, and I am trying to get an understanding of appropriate expectations.
On one hand I am aware that mass produced speakers typically use fairly inexpensive drivers and components, but they manage to end up with a final package that can be quite good and inexpensive. On the other hand, a single driver with a really flat response curve may cost as much or more than a pair of these mass produced speakers.
This leads me to wonder if the driver with a "jagged" response curve is likely to behave better in an ideal enclosure? Or is the data sheet pretty accurate to what you are going to get before adding crossover/ dsp? (I don't mean jagged in the break up region; I mean cases where the driver's best range is jaggy.)
My current system is active studio monitors. They aren't super expensive, but they perform pretty well with a pretty flat in room response. But I'm not sure how easy or hard it will be to recreate that type of response. It seems to me the options are to fix every "jaggy" with a crossover or DSP (I expect to use DSP), or to get an expensive driver that doesn't need much correction. Or is it the third option where the driver's real world response is better than the data sheet would indicate?
I hope that is enough to start a conversation. I know speaker building is more complicated than meets the eye, so I guess this is my way of asking for some real world experience that can help me get beyond the basics of speaker building. If you know what you want to end up with, how do you get there without buying all the drivers and testing them?
On one hand I am aware that mass produced speakers typically use fairly inexpensive drivers and components, but they manage to end up with a final package that can be quite good and inexpensive. On the other hand, a single driver with a really flat response curve may cost as much or more than a pair of these mass produced speakers.
This leads me to wonder if the driver with a "jagged" response curve is likely to behave better in an ideal enclosure? Or is the data sheet pretty accurate to what you are going to get before adding crossover/ dsp? (I don't mean jagged in the break up region; I mean cases where the driver's best range is jaggy.)
My current system is active studio monitors. They aren't super expensive, but they perform pretty well with a pretty flat in room response. But I'm not sure how easy or hard it will be to recreate that type of response. It seems to me the options are to fix every "jaggy" with a crossover or DSP (I expect to use DSP), or to get an expensive driver that doesn't need much correction. Or is it the third option where the driver's real world response is better than the data sheet would indicate?
I hope that is enough to start a conversation. I know speaker building is more complicated than meets the eye, so I guess this is my way of asking for some real world experience that can help me get beyond the basics of speaker building. If you know what you want to end up with, how do you get there without buying all the drivers and testing them?
I'd start out with the best naturally sounding driver, then use a little EQ to make it even better. I've had some luck picking drivers based on other's impressions as something to be weighed - along with published FRs. If it were me, I wouldnt try to tame a lion with EQ or by putting it in a nice enclosure.
Depends on how it was measured, though nowadays it's probably on an IEC flat baffle unless otherwise stated, so irrelevant for the most part, especially since T/S theory peters out at the driver's upper mass corner (Fhm), so above this point the driver's impedance, inductance sets the basic response curve.
A point source driver is considered to be pistonic up to the VC's frequency (Fvc), hence its 'rising on axis' response, then transitions to its radiating TL, flex modes and finally its dust cap breakup modes.
Baffle/box + any local boundaries of course sets its fundamental polar response.
Fhm = 2*Fs/Qts'
Qts': 2*Fs/Fhm
[Qts']: [Qts] + any added series resistance [Rs]: http://www.mh-audio.nl/Calculators/newqts.html
[Rs] = 0.5 ohm minimum for wiring, so may be higher if a super small gauge is used as a series resistor and/or there's other series resistance.
Fvc = SoS/pi/VC diameter
Irrelevant, huh? That isn't what I was expecting! But it does reconcile the seemingly conflicting pieces of information I was wrestling with.
And if I'm not mistaken, the second part is something like saying that the anechoic frequency response, on an infinite baffle for any driver will be "flat" until the cone or dust cap begins to distort, resonate, or fail in some other way? (I guess "flat" isn't accurate, because driver area and displacement will detemine how the low frequencies will drop off.)
So if I'm looking at a frequency response curve, it's mostly showing me the highest frequency where the driver begins to "fail" in some way. But the low frequency response can be calculated based on driver size and xmax...so a graph doesn't show anything you can't easily calculate.
If I am applying this to driver selection, if I am seeking a speaker with a flat and neutral response, either I should be looking for a "reasonably priced" driver that has a cone that "fails" in a subtle or gradual way (i.e. paper cone), or go to an incredibly expensive driver that is designed to not "fail" at all in it's intended frequency range.
I'm still chewing on this. I've I'm starting top catch on, let me know!
Last edited:
To my way of thinking unless it's what one is planning to use. I assume this is still relevant, so how much you figure it will represent a typical box?
Yeah, 'flat' isn't the way to view it. Attached is an old Altec tech bulletin that should clarify how point source drivers radiate, give a bit of insight WRT your other Qs.
Yeah, 'flat' isn't the way to view it. Attached is an old Altec tech bulletin that should clarify how point source drivers radiate, give a bit of insight WRT your other Qs.
Thanks! I am going to hang on to this one! It clears up a lot of questions that I've had, and it dispells myths I've come across. The point about cone shape or even domes not having an effect on directivity is quite interesting. I've seen people talk about the ATC dome mid driver, but I wasn't sure what to make of the claims that domes have wider dispersion than cones.
Here is a more practical question. I have been looking at the Morel 6" drivers for a mid range, and they seem like a good example since they have several versions that are mostly the same, except for different cone material.
If the frequency response curve is mostly irrelevant, is there anything I can learn from the data that will indicate differences between the different versions? The biggest difference is that frequency response of the poly cone is up to 4.5khz, and the carbon/ rohacell cone is 6khz. But that seems irrelevant, since I would plan to cross them down closer to 2khz (~200hz to ~2khz).
Based on the input in this thread so far, I would assume there should be no difference between the two cone materials until the forces on the cone cause one of them to "fail" in some way. So I would expect differences to show up at high SPL and at the upper and lower edges of its frequency response. But if I'm only using the middle of its range, it seems even high SPL wouldn't differentiate the two. (I don't mean to discount the differences in motor structure between the different models. But I expect the cone material is a much bigger difference.)