Another speaker thread

[chris661]

It'd be very interesting to see some measurements of that setup.

I would be amazed if you could get much meaning from measurements in a room at those sorts of frequencies.

I'm not sure I understand, Nick. You'll definitely be measuring the speaker + room combination, but in my experience that's perfectly acceptable. Where do the measurements become meaningless?

It would indeed Chris.

I was musing yesterday on a single driver per channel. MiniDSP is quite cheap, and class D amps from eBay aren't expensive. Sounds like the 100W jobs I already have aren't up to the job though.

I suspect I probably need to buy a couple of 18" drivers and start playing with them in different arrangements - OB, H frame, sealed, vented, no baffle - to see what works for me, and what doesn't. But a driver that suits one type probably won't suit another, and looking on Blue Aran there are 143 18" drivers to choose from, all with slightly or largely differing parameters. People's thoughts on this would be appreciated.

100w is certainly enough to get the cones moving. Remember, there's 6dB going from 100w to 400w, and 10dB from 100w to 1000w.
I'm not convinced that 18" drivers are the be-all and end-all size for subwoofers. If you want a dipole bass system, though, more really is more.

The closest at an initial look is the Faital Pro 18XL1800 which has an Fs of 29 rather than 19.5Hz, and a Qts of 0.4 rather than 0.53. And they cost around the same as the Daytons at £360 each. I'd guess the Daytons are better suited, but I'm only guessing.

I happen to have a pair of that particular Faital unit. Excellent drivers, and capable of some truly terrifying excursion, but there's quite a lot of wind noise (from the motor vents) associated with that. Put in boxes (to keep the motor noise away), they'd be excellent subwoofers. I probably wouldn't use them for OB, though.

So many variables here. How good is the Behringer DSP compared to say MiniDSP or MiniDSP HD? How good are the Behringer amps compared to say various eBay class D "cheapies"? How important are the Daytons in the overall mix? No easy way of knowing.

Which is I guess why I'm thinking I probably just need to buy some and have a play.

The Behringer DSP is probably about as good as that in the MiniDSP. The amp section of the NX6000s will be more capable than most of the amps on eBay.
Plus, warranty, nice case, etc.

If your prepared to wait until the end of lockdown I can lend you a few drivers to play with..

My experience hearing a few pa type amps and dsp is that can sound impressive, and is good enough really for most.

The best bass I've heard through my speakers has been with the mofo powering them.. about 2 watts or so.
I have one track that made my 12" drivers rattle, but the 18s loved it. Much cleaner than transformers at the very bottom.

I think our Chris demo'd his set up nicely at our last meet. Gave us a taste of what big pa amps do well.

What sorta puzzles me regarding bass and impressiveness..

...we had Netflix on a couple of nights ago, there was a nightclub scene, the bass was banging.. like a nightclub does.. through my tv speakers, but my tv speakers do not go low?

is it say multiple ( say around 40 to 60 hz) signals through speaker that can't differentiate, can make one impressive sound?

Answers on a postcard...

I think I brought a few PA amps, Steve. One was a big Crown amp, which was class AB/H, and there was also a rack of Powersoft T-series amps. I think the Crown got plugged into Colin's floorstanding speakers. The Powersofts got used for a few different things, but mostly (IIRC) the tiny Fostex monitors + sub.

With regards to the TV speakers, you'll find that something that peaks around 70Hz (easily attainable with a 3-4" driver) sounds like it has bass, even if it doesn't go particularly low.


Phew, long post.

Cheers, all.

Chris

[Ali Tait]

just my thoughts were that things like soundstaging and imaging etc are not really a factor at these frequencies, or at least much less of one.

Probably not, but tone and harmonic accuracy (phase) certainly are IMHO. That was one of the biggest things I found I liked when I first started using OB's

No argument there Nick.

[Scottmoose]

Interesting discussion chaps.

I was looking at the Dayton UM15 on the Parts-express website earlier and noticed they sell a sealed sub of 3 cubic feet volume which they reckon is Qtc 0.707 and F3 35Hz...
This struck me as a rather small box (88L) so I tried modelling it and got a result over twice the volume. Are they being economical with the truth or is there something clever going on?

Likely derived from the nominal pulled from BassBox6 Pro, with an assumption of heavy internal damping, and possibly relatively high box losses. Personally I have significant reservations about BB6; as a piece of relatively basic T/S derived LF alignment modelling software it has / had a handful of decent features (notwithstanding a really clunky interface and graphs you can't rescale), but I generally find its accuracy / baseline assumptions questionable; it frequently puts out volumes significantly short of what you would expect, or indeed what most other software does, and you have to start going though its menus &c. to assess what default settings have been applied, then adjust them to something more representative (where possible). Not so good. And even then, it seems to deviate, which makes me suspicious about the details of some of the underlying coefficients.

[Scottmoose]

No need to be sorry, it’s all good discussion. Don’t disagree on it’s importance, just my thoughts were that things like soundstaging and imaging etc are not really a factor at these frequencies, or at least much less of one.

Less, but there are a lot of different factors at work. Our hearing acuity to all but large amplitude variation drops to c. 'lousy' below about 200Hz, give or take -good job too or pretty much everything with output below that would be unlistenable in the majority of spaces. And on paper, below 70Hz, the nominal wavelengths are sufficiently large that it's more or less mono. In practice it doesn't necessarily work that way though; you have different pressurisation and boundary loadings that do still affect things, including image perception; this is one reason, albeit not the primary one, that Harman and others have long-advocated multiple subs (preferably 4 or more) to even out loading and room modes.

You're also in the business of crossover rolloff rates. Even with DSP, it's not a brick wall: the woofers are still producing audible output above the crossover frequency -how much and how high depending on the low-pass slope you use. So assuming 200Hz as an example, the woofers will still be producing harmonic modes of lower bass notes even before you start to roll them off, and will continue to affect these to some varying extent (as noted depending on exactly how you've implemented them) above that until output is attenuated by at least -36dB.

[simon]

Thanks Scott that's helpful. You can't cheat physics then .

[simon]

If your prepared to wait until the end of lockdown I can lend you a few drivers to play with..

I missed this, that would be great Steve. No rush, it's not like I can get any wood at the moment
<insert Kenny Williams image here>

[Scottmoose]

Thanks Scott that's helpful. You can't cheat physics then .

I've yet to achieve that myself. Alas.

For whatever it's worth, assuming a conservative 1/2 ohm series R for typical wire loop, connection losses, moderate levels of damping (Qa of about 20 and relatively small amounts of box leakage, then Butterworth (Qtc = 0.707) is about 240 litres based on the published spec. I can hit a reasonable alignment in about 80 litres with heavy fibre-fill and bleeding off some pressure through resistive venting, but that's getting creative rather than the sealed box referenced, which depending on damping I'd expect based on the published driver data to have a system Qtc in the low 0.80 regions.

[simon]

Thanks again Scott, useful info. Reassuring you got a larger volume like the ajsealed too, though yours is another c60L. I think when I looked at the sim I was up around Qtc 0.8 too IIRC.

I've still a lot to learn about the T/S parameters and how they affect each other and speaker design but it's been an interesting few weeks and I'm starting to understand the trade offs and compromises.

[simon]

Any recommendations for good books for speaker design, akin to Morgan Jones for valve amps?

[Scottmoose]

The default recommendation is Vance Dickason: The Loudspeaker Design Cookbook. Just ignore the sections on transmission lines and on crossovers. The former is outdated, the latter is electrical only so like on-line 'crossover calculators' (or any other crossover calculator for that matter) is basically worthless as it assumes the drive units are all coincident, and have a perfectly flat frequency and impedance response from 0Hz to infinity. There were also some incorrect statements about the filters themselves in previous editions -I haven't checked the most recent to see if they're still present; hopefully not, Vance knows better than that. Otherwise, it's a solid offering.

If you want a shortcut that will give you a basic grounding then two things to keep in mind when it comes to enclosure design:

1/ The three key parameters are: Fs, Vas and effective Qts (Qts = any series R in the circuit, which has the effect of artificially raising it). These completely dominate box size & tuning characteristics aka the low frequency alignment

2/ If you hold two of those three parameters static, and increase the value of the other, the box size will increase (and as you'd expect, if you decreased its value, the box size will also decrease). Of the three, Qts has the greatest impact in that small changes have what many would think to be a disproportionately large effect on box size.

Bonus point:
3/ T/S parameters are not fixed, but vary with drive level, prevailing climatic conditions & a bunch of other factors, not least manufacturer production tolerances and how economical with the actualite they were for marketing purposes (m'lud). The actual T/S parameters are derived from electrical filter theory (a lot of what are lumped in with the term are physical properties) and they are only valid in the rising response bandwidth from Fs up to the mass corner frequency of the drive unit: take the latter as 2Fs/Qts.

[simon]

The default recommendation is Vance Dickason: The Loudspeaker Design Cookbook.

Snip

The actual T/S parameters are derived from electrical filter theory (a lot of what are lumped in with the term are physical properties) and they are only valid in the rising response bandwidth from Fs up to the mass corner frequency of the drive unit: take the latter as 2Fs/Qts.

Thanks again Scott, some really helpful pointers there. Book bought!

The last bit is a nugget - that's actually quite a low range.

[simon]

OBs are somewhat different from sealed and vented boxes. Is it reasonable to say the most important thing is baffle size followed by Fs and Qts?

I can't see why Vas would be significant for OBs?

[Scottmoose]

Narry a problem.

OK, well strictly speaking the absolute bandwidth limits covered by T/S parameters are from the mass-corner frequency Fhm I mentioned above, down to the lower corner frequency from suspension compliance Flc, which is defined as (Qe*Fs)/2. The latter is basically theoretical, and defines the lower acoustical loading limit of the driver; the only way you can technically achieve it is via a full sized horn, so it's not often used. In essence, the mass corner is the critical one from the perspective of defining the upper point where T/S parameters cease to be relevant.

WRT baffles, Vas is no longer a major factor in itself although it can have some influence for H-frames & other variations of heavily folded types, with Fs & effective Qt being the dominant factors as far as alignment / modelling is concerned.

[simon]

Great, thanks Scott

[simon]

I took a some more measurements last weekend to try to understand a bit more about what I'm listening to in the hope it'll better inform what I'm trying to do.

The first thing I did was investigate the dip in the sealed cabinet above 100Hz - measurements are with the speakers toed in, straight down the room, and toed out. No difference. So maybe it's to do with the speakers themselves rather than the room?

Measurements of the sealed cabinet have always been with the "subwoofer" amp i.e. a 180Hz low pass filter so I measured it with a full range amp too, just for interest to see what the 285GMF does.

Looks like it might just get to the magic 400Hz to cross over with the 208, though 300Hz might be more realistic?