Hi guys,
I'm ready to design the enclosure of my 2-way active monitor. Now I have to decide the size... I know the Q matter has been widely discussed, but please let me start from the following assumptions:
1)sealed enclosure
2)enclosure volume is no design constraint
3)bass roll-off can be digitally EQd through linear phase FIR filters
4)max SPL is no object
Which Q would you choose in order to get the best overall sound quality?
I'm ready to design the enclosure of my 2-way active monitor. Now I have to decide the size... I know the Q matter has been widely discussed, but please let me start from the following assumptions:
1)sealed enclosure
2)enclosure volume is no design constraint
3)bass roll-off can be digitally EQd through linear phase FIR filters
4)max SPL is no object
Which Q would you choose in order to get the best overall sound quality?
Last edited:
Do you want Butterworth, or Linkwitz-Riley, or some other?
Do you want 1pole, 2pole, 3pole, 4pole, or a mixture?
Do you want 1pole, 2pole, 3pole, 4pole, or a mixture?
The "flexible" approach is to use the largest box possible and then EQ the response. This assumes that you have no limits on power (e.g. as much amplifier power as you need) and that the thermal effects of additional power are negligible (not always the case, see "power compression"). This approach gives you:
flexibility to change the low end response at will
low(est) distortion
A more "optimum" approach would be to decide in advance what cutoff frequency you want to have. Then design the box for that. If you cannot achieve that F3 with the driver while staying under Q=0.71 then you will need to EQ with an LT or the like to extend the response lower.
Also, keep in mind that the Q value is a continuum and that there is no magic in specific values like 0.707, 0.577, 0.5, etc. These simple correspond to certain properties like "maximally flat", "no overshoot", critically damped", etc.
Furthermore, many do not realize that the driver's TS parameters will CHANGE (sometimes significantly) with the position of the voice coil, e.g. when the cone moves in and out. So DO NOT obsess over the Q, because as soon as you cause the cone to move your driver's own Qts is changing. Bonus points for getting Klippel data that shows small AND large signal values for the driver parameters and then designing the system so that it has good performance under both regimes.
Thanks Charlie for your clear explanation! So probably I will build something between 0.5 and 0.55, for a total volume of 15-20 liters. Anyway we can summerize that, given my assumtpions and no amp-power constraints, the bigger the better, right? (Or at least up to critacally damped)
I resolve this by measuring "TS" params at both low drive levels as intended and then also at high drive levels to see what extent of "mistuning" happens at higher powers. That provides knowledge of the effect before cutting wood.
High power will not give "true" TS parameters (the TS derivation assumes linearity and absence of 2nd order effects that creep up at high power) but it is still a quite good prediction.
Follow-up question: I've seen often cited the sound of low Qtc (0.4-0.5) bass to sound "anemic" (whatever that means in this context) vs 0.707.
Assuming a situation like the OP describes where DSP is fully available to achieve a flat response, would the dry/anemic sound qualities of a low Qtc design persist?
Anemic means low bass SPL, thin bass. You raise it electronically
and it's no longer anemic, the Qtc raises accordingly. Q is a FR
curve.
and it's no longer anemic, the Qtc raises accordingly. Q is a FR
curve.
There are a couple of things against very low Q (meaning 0.5 or less) alignments. The "anemic" quality comes primarily from a frequency response that is slowly decreasing (rolling off with a shallow initial slope) as frequency decreases. This means that the region around resonance might be -6dB to-10dB down. You hear this as "anemic" or "thin" sound but it's just a consequence of the frequency response. Correspondingly the efficiency around resonance is pretty low. Again, just a consequence of the frequency response IIRC. Finally, the transient response is "overdamped" which translates to a "slow" response to input and a gentle, no-overshoot rise. This is not as much of a "good thing" as it has been purported to be IMHO, and as many things the middle ground is best. You don't want a very high Q (Q>1.0), which generates ringing in the time domain response, but a very low Q is also not great. So the "middle ground" of 0.5-1.0 is best.
perfect! But what you can't do with dsp is changing the mechanical effect of the air inside the cabinet acting like a spring. That's the crucial point!
ok, I get it!
Not your fault, but here is exactly where I can't get the grasp of the problem..
I can't understand why a "softer air-spring" (eg 0.5 softer than 0.7), if pulled from the rest position (driver's membrane goes out), causes the driver to react slower. Sholdn't be the opposite? Anyway the reason must be that I'm thinking wrongly about what really happens mechanically...
Hi,
What bass roll off you build into the cabinets
is pretty much irrelevant given the above.
The choice of raw box Q will depend on the drivers,
in ways rarely discussed on DIY sites. Basically its
question of getting the right box size for a driver
designed to cancel box nonlinearity at higher SPL.
TBH the 4 statements above are very naive.
rgds, sreten.
Hi,
You want an answer to a question you can't ask properly.
Like I said given your conditions, box Q hardly matters
without further rather involved driver information, and
more practical conditions than airy fairy waffle.
There is no simple answer.
rgds, sreten.
By measuring the impedance when driven by an amp into a high power resistor to use as a divider. By calibrating it open cct it allows complete and accurate magnitude and phase
I used this technique just tonight to weed out a dodgy D26NC55: impedance changed with drive level. [shakes head, Vifa, Vifa, Vifa....]
As I mentioned, its not true "TS params" but I think you get that from your parenthesis
0.7, even as high as 1. I've done lower, all they do is roll off the bass too prematurely, no other real benefit.
Two caveats
1. if the box is really small and the woofer has good power handling, and you have a big amp, I would consider a lower Q just to allow more stuffing to the box to help clean up the mids
2. If you plan to over-eq it and get super low bass with poor power handling, the lower Eq will give some benefits, but that would be a non-ideal system
Go to SL's site and download his spreadsheet for the Linkwitz transform and play around with it. It'll give you a good idea of the trade offs
With the driver I'm working with (Beyma 10G40), in a sealed box, I get:
FWIW the amp can deliver 150W cleanly, and UniBox shows at 150W and 6.3 liter it would get to 105dB SPL at 70Hz. Should I go with 1) and have the amp work hard and call it a day?
- If Vb= 6.3 liter, then Qtc=0.71 and Fb=122Hz
- If Vb=20 liter, then Qtc=0.49 and Fb=78Hz
FWIW the amp can deliver 150W cleanly, and UniBox shows at 150W and 6.3 liter it would get to 105dB SPL at 70Hz. Should I go with 1) and have the amp work hard and call it a day?
I suggest attacking the problem differently. The box choice totally depends on the sub's high pass crossover, as you need to match to that.
1. Is it second order? If so, its probably Butterworth of Linkwitz Riley, so tune the box Q to 0.6 to 0.7. You don't care about Fb, its F3 (if Butterworth) or F6 (if Linkwitz Riley) that matter
2. Is it 4th order? Then you probably need a high pass xover to this box/driver. Then you need to design the box/driver and xover together to match the sub low pass
There's no cookie cutter answer but I recommend to quit thinking in terms of transient response and just get the right frequency blend with the sub
If this stuff is all too complicated, then pick a Q that gives an F6 no lower than 70 Hz and consider that. What is the box size/Q with that?
Neither work.
The Butterworth rolls off @ 122Hz but you want/need roll off at 70Hz.
The LR (Q=0.5 for a cascaded Butterworth) rolls off at 78Hz but will be down by many dB at this frequency.
You need EQ which will overload your speaker, or you need a bigger speaker.
Try to find a driver and box that with Butterworth loading rolls off at your preferred frequency.
Alternatively use this smaller driver in it's smaler box with the 122Hz Butterworth and add on a bass only speaker for the octave below that, i.e. 61Hz to 122Hz Use LR crossover to integrate them. An 8" in a medium size box will give you reasonable bass volume for 60 to 120Hz.
Thanks for the ellaborate answer!
The sub is in a sealed box and is servo driven, FWIW. The HPF is 12 dB/octave and can be switched to 24 dB/octave. I'm inclined to go with 24 dB/octave because the xo at 350 (or 500) and 2500 Hz will be high order, probably 96 dB/octave. They will be linear phase, again FWIW.
In the examples above:
- If Vb= 6.3 liter, then Qtc=0.71 and Fb=122Hz, and F3=121Hz
- If Vb=20 liter, then Qtc=0.49 and Fb=78Hz, and F3=119Hz
- If Vb=10 liter, then Qtc=0.60 and Fb=101Hz, and F3=119Hz
How do you recommend I move forward with the design?
Thank you!
- Status
- Not open for further replies.