Preliminary HIVI B3S design

[Zaph]

Original posted on the Parts Express forum where the B3S was being discussed.

http://home.new.rr.com/zaph/audio/te...3S-xover-1.gif

Here's the design. Note that this is preliminary and untested. This would be my first choice for a filter for this driver, as it gives a little more positive control over the response. The schematic, final response curve, impedence and transfer function are shown. The BSC can be adjusted with the R, and in some case the whole BSC circuit can be probably be removed - like when the system is up tight against a wall or sitting on top of a TV for center channel usage. Generally, away from walls is going to sound better, and putting these on little "surround stands" will offer the best sound. The .08 inductor will probably have to be unwound from a .10 Ask me for help on that if you need it.

http://home.new.rr.com/zaph/audio/te...3S-xover-2.gif

Here's another option similar to what I think Paul W was describing. The impedence dip isn't bad at all, but this sort of notch gives up a little control in the in the response shaping. Lowering the notch Q to widen the effective range and not cause a peak at 4Khz results in a peak that's not properly flattened. It would work, have a smoother impedence phase swing and probably sound good, but it's not my first choice.

The general rule for notches is that those within the bandwidth should be a parallel RLC in series with the driver + and those outside the bandwidth should be a series RLC shunt across the driver terminals. It's just a guidline really, but is true most of the time. The more you break this rule, the less control the notch has.

http://home.new.rr.com/zaph/audio/te...-enclosure.gif

Here's the box design. It's about .12 cu ft [3.3 l] and has a slim profile. I have designed the baffle to be removeable and fastened with socket head cap screws. there are .75" x 2" x 4" gussets to hold the T-nuts. One side of this gusset is cut at 45 degrees to minimize internal edge reflections near the cone. The gussets and internal width of 4" leave a 4" square to fit the crossover through with a removed baffle.

The baffle driver opening MUST have the inside routered at a 45 deg angle to promote airflow.

http://home.new.rr.com/zaph/audio/te...x-nofilter.gif

Here's the predicted driver response in this box with no filter. Note that it only needs about 3 or 4 dB of BSC. This is a simulated baffle added on the the infinite baffle measurement. The 1/2" roundover on the baffle is also part of the simulation. These are usually surprisingly accurate though, and I wouldn't expect the final real in-box measurement to differ more than .5 dB.

http://home.new.rr.com/zaph/audio/te...-voicecoil.jpg

And just for the heck of it, here's what the voice coil and back of the cone look like. It's really a long voice coil and has more than twice the output of the old TB W3-871. The former is attached to the concave cone with a strengthening/alignment aluminum piece. The back side of the cone outside edge is coated with a damping material. Overall, pretty good construction for a cheap driver although the motor is really nothing special.

This is all preliminary. I'm in the process of building this myself, and there is a chance I may change something when I'm done and I post the final project.

The whole goal for this project is to address all possible usages for small speakers:
* Normal 2 channel stereo - on stands
* HT usage - TV top mounting
* Architectural audio - wall mounted
* Near field - desk mounted
I intend to test and provide options for each type of usage. I'll also provide upgrade options for a higher output and lower distortion system using the Aura NS3-193. This makes it a rather ambitous project, but the ever popular TB W3-871 system on my site is due for an upgrade. No promises on when I get it done. I'm easily distracted.

[HeatMiser]

Great news John. Quality designs for small/inexpensive drivers are hard to come by, and all your recent work in this area is certainly appreciated. I look forward to seeing what all you come up with.

[Izrun]

I for one am extremely excited about the NS3-193 version you will be making. I've been waiting to take the plunge on my first project. I needed small, and I was planning on building the W3-871 design, but didn't start because I wasn't sure it was right for me. The NS3-193 seems to be the exact project I'm looking for. Keep up the good work, and I can't wait to get started!

[/end butt kissing]

[cjd]

Hmm. Would you happen to have numbers for Bl and Sd that are any good? Trying to make a very vaguely educated guess suggests that Bl may be around 3 but that's a bit dependant on Sd... at any rate, it's shaping up to look surprising in a TQWP, but that is a bit dependant on Bl and Sd for me to get it right. In fact, if it *does* work out, the potential for a rather elegant (and perhaps insane) design is quite high.

Thanks for all your work, whether you have these numbers or not!

Can't wait to hear whatever you bring this year.

C

[coolkhoa]

How about rear-mounted design for the B3S? (With routing of the outside, of course.) IMO it looks cleaner than with the frame exposed but perhaps has some adverse effects?

[Zaph]

Originally posted by cjd
Hmm. Would you happen to have numbers for Bl and Sd that are any good?

The BL is about 3.5, and the Sd is about 30.2 cm^2. If you want to play around with bass alignments, most of what you need can be found here. Notice how high the Qts is, which I suspect is why this driver only needs 3 or 4 dB of BSC - the midbass peaking partially makes up for the low end 2pi-4pi rolloff. The scary thing about putting these in any kind of enclosure other than small sealed is the power handling.

Originally posted by coolkhoa
How about rear-mounted design for the B3S? (With routing of the outside, of course.) IMO it looks cleaner than with the frame exposed but perhaps has some adverse effects?

There's a round flange version of this driver which I'm pretty confident is the same thing. It would need to be countersunk though. Rear mounted could work ok, but I don't think I'd do it with a baffle more than 1/4" thick even if it's rounded over. The hole edge tends to affect the response curve a little. As far as looks, it looks pretty classy with 4 socket head cap screws holding the driver in place.

[cjd]

Thanks for those specs! I'll have to run 'em through Mathcad tomorrow. If I remember, power handling peaks out at about 85dB at 1M in the TQWP I was messing with, but I was pondering something that could give good bass and be interesting enough (perhaps) to pass more SAF tests than usual. Not for HT, but perhaps for music. Given that I usually listen with nominal levels in the 50's and peaks in the mid 70's, that still would work at 3M for me with a little room left over. And, response to 40Hz. Out of a single 3" driver. Your numbers are different enough that I don't believe a bit of that, and I'll go re-run the model to see how bad it screwed things up.

IF it still works out, expect to see some at DIY Chicago 2005. If my wife doesn't kill me first.

C

[coolkhoa]

Originally posted by Zaph

There's a round flange version of this driver which I'm pretty confident is the same thing. It would need to be countersunk though. Rear mounted could work ok, but I don't think I'd do it with a baffle more than 1/4" thick even if it's rounded over. The hole edge tends to affect the response curve a little. As far as looks, it looks pretty classy with 4 socket head cap screws holding the driver in place.

Thanks for your input. I do know of the round-flange version of the B3S (the B3N) but I'd rather not go through the trouble of recessing. I guess surface mounting isn't so bad.

[SQ Kid]

so is this the successor to the throne once held by the TBs? they look pretty good, just wondering if a tweeter would be necessary for these things to be well rounded (at least in the upper frequencies)? i'm looking into doing a new speaker setup for my HT and these things would work great with a sub. thinking maybe single drivers (plus tweet?) for center and rears, while doing the fronts with 4 each plus maybe the tweeter. i have to keep it 8 ohm (reciever not 4 ohm capable, and i know this can be discussed over in regards to higher freq.)

[AAXenoStrange]

Yea I'm interested in these things -- or more possibly the Aura one that will follow this.... When they are all made and optimised we will know which is the winner after all...

Funny thing is that both the Aura and the B3S model well in a ported design with a high pass filter and a linkwitz transform... REALLY well actualy

Flat to 60hz and the hivi hits its excursion limit 3dB lower thn its thermal limit - -and the aura hits them at the same time...

Aura in a 1.5 cuft box tuned to 60 hz, 4th order LR highpass @ 55-60 hz depending on how much of a peak you want near tuning, linkwitz transform of:

f(0) = 97.15 Hz
Q(0) = 0.83
f(p) = 50 Hz
Q(p) = 0.500

C2 = 0.047 µF
R1 = 16.84 kOhms
R2 = 8.32 kOhms
R3 = 63.57 kOhms
C1 = 0.201 µF
C3 = 0.0534 µF

B3S also looks good in:

1.5 cuft tuned to 55hz, 4th order LR hp @ 55-60, linkwitz transform:

f(0) = 115 Hz
Q(0) = 1.22
f(p) = 60 Hz
Q(p) = 0.707

C2 = 0.047 µF
R1 = 10.41 kOhms
R2 = 3.31 kOhms
R3 = 38.26 kOhms
C1 = 0.376 µF
C3 = 0.1023 µF

Someone else should give this a looksie over before anyone goes and actualy does anything with it but yea... It looks good to me...

BTW Ive never seen anyone use a LT with a ported design along with a highpass to cut the peak.... Utilizing this leads to less cone excursion arround the tuning frequency as well as less gain needed due to slightly increased efficency. Is this a first? Fullrange speaker in a small box from 60hz to ~16-20khz!!!



This same idea can be used in a small ported or PR box as well Say a small subwoofer tuned to 16-18hz getting the group delay peak below 20hz and less audable. The problem with a small box tuned so low is simply the port needs to be very long. A high mass PR would work though... Try modeling in winisd any sealed box subwoofer you want, add a port tuned to 16-18hz to the same box volume and compare. Now compare it to a sealed box responce of about .1qtc lower (ie if it was .707 compare, a .607 to the ported...) look simmilar dont they? Also note cone excursion plots... Add a linkwitz transform and you wont bottom the driver so quickly now will you? Past the tuning things go off the chart however but a subsonic filter will take care of that.