Another Sensitive OB Woofer Design
For your enjoyment, here is another open baffle woofer I recently threw together. I say 'another', because this design evolved out my use of a popular OB woofer, MJK's Alpha 15a in an H-frame. See here for the background on MJK's design: http://www.diyaudio.com/forums/showt...ight=alpha+15a
My issues with the Alpha 15 were pretty simple: using a 15" driver in a 17" by 17" by 17" enclosure, its too big to pass as 'elegant'. Also, although it rolls off at about 30Hz, it's nonlinear distortion (NLD) below 45-50Hz is relatively high. Related, some people have concerns about the drivers Xmax being reached easily at lower frequencies. This may be a realistic worry considering how high the NLD is below 45Hz. IMO, NLD is overrated - in the bass region, psychoacoustic masking limits our dynamic range to about 40dB, up to at least the 5th harmonic (see Floyd Toole's latest book, and references therein). Also, just as important, any actual musical instrument (as opposed to synthesized tones), has harmonics that are barely lower in level than the fundamental - and subsequently are much louder than the NLD harmonics produced by the driver itself. If in doubt, put an instrument in front of a fft. There are two reasons I'm concerned with NLD here - first, it shows that the driver is reaching its volume displacement limit, and hence SPL limit. Second, IMD creates a 'dynamic noise floor', which may or may not be an issue, depending on the recording. I'm also just a sucker for 'cleanliness', so lowering NLD is nice if it is easy.
I might add that I have never found issue with how the Alphas sound, even at loud volumes.
So here are a few pictures of the design - nothing revolutionary, just simple and useful, and I hadn't seen something like this around here yet.
Above, you can see MJK's H-frame next to my recent 'pet'. I don't know if it deserves a name, so lets just call it GWx4 for now - the drivers are four Goldwood GW-210 10" drivers, available from PE for $18 each. The Alpha driver is $60, while the Goldwoods are $72 (per side). So we are talking $12 more (but there's more to that). The drivers are high Qts types, so need little EQ in their lower end.
The footprint of the GWx4 is about 13" by 11", and stands about 23" tall. I more or less fixed how big I wanted the enclosure to be, and then worked with it. A larger enclosure would make the whole thing more sensitive, but one of the primary goals of this design was to not be 'big'. As you can see, the enclosure is somewhere between an H-frame, and an OB with wings. Using one set of drivers behind a front set is something I've only seen in an Emerald Physics design, and of course only works in the bass range. Of course, there is no free lunch - above 400Hz, one sees interference patterns, depending on where the mic is. But 400Hz is high enough.
So lets talk goals:
1) Smaller size - particularly footprint
2) Greater SPL output
3) Bandwidth of 45Hz to 200Hz
4) High Sensitivity
5) Low Cost
to be continued after lunch...
I don't get the complete picture from your pictures. Are both volumes behind the driver baffles open to the outside? Since you've already got a central cavity, why don't you go full circle - ending with a W-frame? If the pictures tell the full truth, I see a vast mismatch between the impedance on either side of the cone. Wouldn't that be a severe source of NLD?
Interesting cuibono. It looks like a 'compound woofer' as described by SL. According to him, it won't play any louder than 2 drivers in an H-frame the same size but the cavity resonance should be higher. Seems like it works if it plays well up to 400.
1) Small size: 13" by 11" versus 17" by 16" footprint. I adjusted the final dimensions to be able to use 12" wide pine board, and a height of 23" is a good base for the mid/tweeter.
2) SPL output: the 4 10" drivers have 60% greater surface area (Sd) than the Alpha, which means less linear movement (and NLD) for a given SPL. The volume product (Sd times Xmax) is 483cm cubed for the GWx4, and 325cm cubed for the Alpha. So greater total dBSPL, by about 2dB, with less distortion. Also having the drivers facing different directions reduces some orders of NLD harmonics, a good thing. Power handling is also increased, which is important - less VC heating, less NLD, less chance of melting something. Another possibility for more output is using the GW-212, a 12" driver with probably the same motor as the GW-210, but that would increase the footprint of the speaker more than I'd like. Later, I'll show the drivers NLD measurements.
3) Bandwidth of 45-200Hz: I set the lower limit to avoid having to use larger or more expensive drivers. 45Hz is also good place to have subs take over, which is something I'll add sooner or later. In my current OB design, I found the woofer was much more capable of handling the 100-200Hz region compared to the mid driver - which was dealing with needing a lot of EQ (aka, low sensitivity due to OB roll off), and a lot of room interference. The woofer had neither of these problems, so I've been crossing them at about 180Hz. I'm still not sure if this skews the soundstage downwards, but seems okay for now. Anyway, the multiple drivers don't develop interference patterns till above 400Hz, which is plenty of bandwidth. In the future, I'll use a different setup where the woofers will be crossed at about 115Hz, so everyone will be happy. Adding chokes to a couple of the woofer is an option to prevent midrange cancellations, particularly with mid's crossed high.
Something to note is that SL's Orions use a pair of 10" drivers, but they are way more expensive, have way more Xmax, and are used down to 20Hz. Cutting off OB woofers at 45Hz makes things a lot easier to do, including dealing with room modes.
4) High Sensitivity - 'high sensitivity' designs have a couple of benefits - the most obvious is less thermal heating through the voice coil. Another is the purported improvement in musical 'transients'. This is still a grey area to me, particularly as it applies to either the bass, mids or treble, and the inability to subjectively quantify music's 'transient' character. But I'll give it a go, considering the reputation horns have as being 'dynamic'.
Particularly for OB use, woofers are usually less sensitive than mids or tweeters. The main way to get around this is by using multiple drivers, but size matters too. As it turns out, the GWx4 and MJK H frame end up being just about the same sensitivity (within .5dB, depending how you measure it), although there is a major caveat to this - the GW woofers are 8ohm, and I'm using them in parallel - but I'm not trying to drive a 2ohm load - I'm using 2 amps, each driving 4ohm loads. This makes a bit of a difference to the cost and actual sensitivity. I decided to do this because amps are cheap these days, and GC or Class D amps would be a good match to these. 50W per channel should give plenty of head room.
So, remembering I'm using one amp with the Alphas, and two with the GWx4, they are both approximately 87.5dBSPL/2.83V/1m sensitive at 50Hz, measured 33" above the floor, outdoors. (Moving closer to the floor greatly inflates the SPL; I recently made the mistake of using that number in another thread...) They are about doing 90.5dBSPL at 100Hz, under the same conditions (and with no LP filtering). I was actually hoping the design would be closer to mid-90's, as it looked that way simulated, but didn't work out that way in real life.
So far, that has been the only disappointment in this design. But that's okay, 90dB at 100Hz on OB is better than a lot designs are doing.
5) Low Cost: the GWx4 drivers are $12 more than the Alpha, but then add about $50 for another channel of amplification. For me, the improvement in size, output and distortion character are worth it. Amps are cheap. Particularly relative to other high-output OB designs, I think this design qualifies as a budget/low cost/high value design. Actually, I haven't listened to the design yet, and I don't expect it to sound particularly different to the Alpha 15a - but I do know it is more aesthetic and more capable, which makes me happy.
ACTUAL DATA COMING UP NEXT
I've actually done several days of measuring, but I'm fairly tired of looking at all the data, so I'm only going to post the most concise.
First is an overlay of the FR of each woofer design, measured under the same conditions: 2.83V at the drivers terminals (but remember there are two amps driving the GWx4), mic at a distance of 2m, .84m off the ground (which would approximate the listeners ears), all done outdoors with long MLS signals.
As you can see, between 45 and 100Hz, the are doing just about the same thing. Remember, at a distance of 2m and outdoors, there are some environmental effects on the data - so lets call these close approximations. 1m data had more similar results.
Green is the MJK H frame with the Alpha 15a, Brown is the GWx4. Note the nasty FR above 400Hz... Also, the GWx4 has less of a peak at 200Hz - the GWx4 baffles are less than 4" deep, as opposed to 7.5" for the Alpha 15.
Here is the nonlinear distortion data, done using sweeps. These two graphs were measured the same way, but differently than above. The driving signal was 6dB higher, at 5.66Vrms at the terminals, and measured at the opening of the H-frames indoors. I would have rather measured at a greater distance, outdoors, but there is way too much noise and environmental effects. Indoors, I barely had enough noise free dynamic range.
5.66V represents about 93.5dB at 50Hz and 96.5dB at 100Hz at 1m, so decently loud, but not insane.
The first graph is the Alpha 15a, the second is the GWx4. What you can see, in both of them, is they really loose control of things between 40 and 50Hz. The GWx4 maintains control lower. The other thing I see is that at 50Hz, the GWx4 tall order NLD is 20dB lower than the Alpha. This makes me happy. At -75dB, 4th and 5th order distortion is 'clean' and I'd postulate inaudible, and indicates that the woofers have plenty of headroom till distortion becomes an issue. I need to redo the simulations, but when considering both sets of speakers, I would guess the GWx4 should be able to hit 108dBSPL in the mid bass without problems - which is a lot of output, and comparable to some of the more high end designs. Just not below 50Hz unfiltered. Again, due to masking, instruments natural harmonic content and recording lack of dynamic range, I wouldn't worry about NLD till its around -45dB compared to the fundamental.
Oh - my computer makes these annoying glitches during the sweeps that cause spikes in the graphs below - so ignore the tall sharp spikes.
Hi all, thanks for the questions.
It looks very interesting! And very good performance.
Have you ever considered the T-bass circuit?
That is a very interesting design. Goldwood offers quite a few drivers with Qts values that are ideal for OB. The prices are also very attractive. You are definitely thinking outside of the box.
If I understood you correctly, you measured higher NLD using the Alpha 15A compared to the Goldwood drivers but at these low frequencies you did not think it was audible. Is that correct? How loud are you playing these driver so that they are approaching Xmax?
|All times are GMT. The time now is 03:08 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio