Hello forum! I have been bitten by the bug for DIY audio (wife: "mid life crisis?" me: "be glad it's speakers!") and I'm hoping to collect some criticism/warnings/sage advice on my design concept before committing so much cash to the fires of invention.
Yes, I'm sure I should start with a 2 way kit or something. And I did look at a bunch of them. But A, what fun is that? and B, woodworking, yuck! and C, ambition - foolish, foolish ambition. So I riffed through the Cookbook and Murphy's Intro book and found Linkwitz Labs and said hey...how about some tubes? And mocked up a notional crossover in the 'Active XO Designer' tool that can be had via miniDSP's website, and it looks like it's time to 'designer up' and order some hardware if I intend to proceed.
So here's my concept as developed so far -- please point out anything that seems like an especially bad idea!
Purpose: L/R mains 50% music, 50% home theater. Matching center channel to follow if successful (currently commercial soundbar, ASB-2).
Room: 13' x 17' slightly irregular rectangle, listening position at 10-16'. 8' ceilings.
Drivers/ranges:
SEAS 27TBCD/GB-DXT ... 1" AL/MG w/ waveguide ... 20k-2.5k Hz
Tang Band W3-1364SA ... 3" Paper/Bamboo mid .... 2.5k-500 Hz
Dayton Audio RS180P ... 7" Paper mid/woofer ...... 500 Hz - TBD Hz
Sub(s), e.g. Rhythmik F12G ... 12" subwoofer ...... TBD Hz - 20 Hz
Using spec sheet FRD data (with phase extraction per a Jeff Bagby tool), the mock up active XO has LR2 filters close to the indicated breakpoints, a few dB level matching offsets, and a little parametric EQ to knock down a couple of peaks in the TB 3" and flatten the waveguide tweeter out. (Seems a shame to 'waste' the waveguide boost? But how else are you going to use the thing? A separate super-tweeter to fill out 10k+ seems pretty egregious.) Seems to promise 90dB sensitivity. Interesting (?) feature, all three T-M-W drivers are contributing significantly at ~1200 Hz. Good for integration, or bad for off-axis, other...? Got me. Not sweating the low end details without measurements in my planned enclosure, but hoping the 7" can play down to 80 Hz, with Linkwitz transform if needed. I think off-axis/directivity issues are okay given the breakpoints and size of the drivers? But that kind of leads to....
Enclosures
Stand the 7" upright in a sealed 6" Schedule 80 PVC pipe with feet, and interior treatment TBD but thinking of Dynamat, wall foam, wool, Acoustastuff. 36" tube height makes the volume about 0.5 cu ft -- about double what the driver wants to see, so pack the damping stuff in.
Science fair project detour: try inserting a cone out of poly sheeting (layered with sorbothane or dynamat, etc.?) to provide a waveguide taper from the 6" behind-driver diameter down to 3" diameter or so at 3 or 4" above the sealed end. Optional 45 degree PVC elbow bend or soft rubber tube to redirect at end of cone taper? If that is not at all clear to visualize, I can add a sketch. The idea is to not let it be so easy to reflect a wave off the sealed end and right back up the waveguide to the driver. (Yes, I've seen the Nautilus pictures!) Note, I used the 45 degree FR response in the XO mock up since that was greatest angle provided, but it should all be omni to 500 Hz and even higher, for a 7", nicht wahr?
The 3" would live in a separate sealed PVC line mounted over the 7" vertical pipe, a la Linkwitz Pluto or LXmini. With two 45 deg elbows to make the driver point at the listener but the tube's main length would parallel the 7" tube. Putting a bend in the pipe hurts my soul but 18" or so of straight pipe floating horizontal above the vertical woofer is an ergonomic non-starter.
The tweeter would be free-standing mounted on top of the 3" pipe, so also pointing at the listener...seems to include a sealed rear housing. This puts the tweeter at about 42"...seated ear height. Separation between the center of upfiring 7" woofer and forward firing 3" should only be a few inches, and similar or less for mid to tweeter.
And the sub just sits wherever the sub needs to sit. So this is 'infinite transmission line' aka big, heavily treated sealed enclosure; but without Linkwitz's dipole/cardioid preoccupation. Is that okay, ha? I could experiment with that with the TB 3" I guess, an open back plus diffuser like LXmini. But I think all sealed would be if anything more orthodox?
Things I pretty much expect to mess up, could really use some input on:
Amplifier pairing? I'm in for a lot of channels and would be looking to something like an Outlaw Audio or ATI product, or some mid-level HT receiver like from Denon. This is a bit embarrassing to ask, but how do I ensure I don't send 100W into the 3" TB if it's attached to a 100W per channel amp output? Some way to scale the 'full scale' audio signal to less than full amp output? Please don't say 'resistors', ha. (Although I guess 'potentiometer' is fine. But on the amp? Need a 'pre-amp'?)
Otherwise, driver protection? When I was looking at planar tweeters I read it's good to have a cap in line for transient protection; do dome tweeters need something like this? Some way to simulate playing with crossover and look at power dissipation, excursion? I've spotted some software tools, free and otherwise (VituixCAD, WinSpeakers) but not entered that arena yet other than the ACD spreadsheets. Before I smell smoke, how do I know if I need a steeper crossover?
In the crossover frequencies I have tried to stay well clear of driver rolloff and non-rigid modes so I think the stated design hits close to LR2 in acoustic reality not just DSP land. But for how important this aspect is, my understanding is too limited. Particular acid tests? (I've heard of reverse polarity null, try it and listen, and by implication, check +/- dB and slope at XO frequency....what's the acid test?) Can post graph of crafted (pseudo-)FR if helpful.
If there's anyone still awake after all that....thanks for any insights!
-Patrick R, Alexandria VA, USA
Yes, I'm sure I should start with a 2 way kit or something. And I did look at a bunch of them. But A, what fun is that? and B, woodworking, yuck! and C, ambition - foolish, foolish ambition. So I riffed through the Cookbook and Murphy's Intro book and found Linkwitz Labs and said hey...how about some tubes? And mocked up a notional crossover in the 'Active XO Designer' tool that can be had via miniDSP's website, and it looks like it's time to 'designer up' and order some hardware if I intend to proceed.
So here's my concept as developed so far -- please point out anything that seems like an especially bad idea!
Purpose: L/R mains 50% music, 50% home theater. Matching center channel to follow if successful (currently commercial soundbar, ASB-2).
Room: 13' x 17' slightly irregular rectangle, listening position at 10-16'. 8' ceilings.
Drivers/ranges:
SEAS 27TBCD/GB-DXT ... 1" AL/MG w/ waveguide ... 20k-2.5k Hz
Tang Band W3-1364SA ... 3" Paper/Bamboo mid .... 2.5k-500 Hz
Dayton Audio RS180P ... 7" Paper mid/woofer ...... 500 Hz - TBD Hz
Sub(s), e.g. Rhythmik F12G ... 12" subwoofer ...... TBD Hz - 20 Hz
Using spec sheet FRD data (with phase extraction per a Jeff Bagby tool), the mock up active XO has LR2 filters close to the indicated breakpoints, a few dB level matching offsets, and a little parametric EQ to knock down a couple of peaks in the TB 3" and flatten the waveguide tweeter out. (Seems a shame to 'waste' the waveguide boost? But how else are you going to use the thing? A separate super-tweeter to fill out 10k+ seems pretty egregious.) Seems to promise 90dB sensitivity. Interesting (?) feature, all three T-M-W drivers are contributing significantly at ~1200 Hz. Good for integration, or bad for off-axis, other...? Got me. Not sweating the low end details without measurements in my planned enclosure, but hoping the 7" can play down to 80 Hz, with Linkwitz transform if needed. I think off-axis/directivity issues are okay given the breakpoints and size of the drivers? But that kind of leads to....
Enclosures
Stand the 7" upright in a sealed 6" Schedule 80 PVC pipe with feet, and interior treatment TBD but thinking of Dynamat, wall foam, wool, Acoustastuff. 36" tube height makes the volume about 0.5 cu ft -- about double what the driver wants to see, so pack the damping stuff in.
Science fair project detour: try inserting a cone out of poly sheeting (layered with sorbothane or dynamat, etc.?) to provide a waveguide taper from the 6" behind-driver diameter down to 3" diameter or so at 3 or 4" above the sealed end. Optional 45 degree PVC elbow bend or soft rubber tube to redirect at end of cone taper? If that is not at all clear to visualize, I can add a sketch. The idea is to not let it be so easy to reflect a wave off the sealed end and right back up the waveguide to the driver. (Yes, I've seen the Nautilus pictures!) Note, I used the 45 degree FR response in the XO mock up since that was greatest angle provided, but it should all be omni to 500 Hz and even higher, for a 7", nicht wahr?
The 3" would live in a separate sealed PVC line mounted over the 7" vertical pipe, a la Linkwitz Pluto or LXmini. With two 45 deg elbows to make the driver point at the listener but the tube's main length would parallel the 7" tube. Putting a bend in the pipe hurts my soul but 18" or so of straight pipe floating horizontal above the vertical woofer is an ergonomic non-starter.
The tweeter would be free-standing mounted on top of the 3" pipe, so also pointing at the listener...seems to include a sealed rear housing. This puts the tweeter at about 42"...seated ear height. Separation between the center of upfiring 7" woofer and forward firing 3" should only be a few inches, and similar or less for mid to tweeter.
And the sub just sits wherever the sub needs to sit. So this is 'infinite transmission line' aka big, heavily treated sealed enclosure; but without Linkwitz's dipole/cardioid preoccupation. Is that okay, ha? I could experiment with that with the TB 3" I guess, an open back plus diffuser like LXmini. But I think all sealed would be if anything more orthodox?
Things I pretty much expect to mess up, could really use some input on:
Amplifier pairing? I'm in for a lot of channels and would be looking to something like an Outlaw Audio or ATI product, or some mid-level HT receiver like from Denon. This is a bit embarrassing to ask, but how do I ensure I don't send 100W into the 3" TB if it's attached to a 100W per channel amp output? Some way to scale the 'full scale' audio signal to less than full amp output? Please don't say 'resistors', ha. (Although I guess 'potentiometer' is fine. But on the amp? Need a 'pre-amp'?)
Otherwise, driver protection? When I was looking at planar tweeters I read it's good to have a cap in line for transient protection; do dome tweeters need something like this? Some way to simulate playing with crossover and look at power dissipation, excursion? I've spotted some software tools, free and otherwise (VituixCAD, WinSpeakers) but not entered that arena yet other than the ACD spreadsheets. Before I smell smoke, how do I know if I need a steeper crossover?
In the crossover frequencies I have tried to stay well clear of driver rolloff and non-rigid modes so I think the stated design hits close to LR2 in acoustic reality not just DSP land. But for how important this aspect is, my understanding is too limited. Particular acid tests? (I've heard of reverse polarity null, try it and listen, and by implication, check +/- dB and slope at XO frequency....what's the acid test?) Can post graph of crafted (pseudo-)FR if helpful.
If there's anyone still awake after all that....thanks for any insights!
-Patrick R, Alexandria VA, USA
Where do we start?
First you'll need some basics. How does your 25W light bulb not blow out when clearly you can put a 150W bulb in the same apparatus and it(the 150W) draws six times the power?
For another, what happens when you put a driver in a tiny volume and port it into a larger one?
Start with those, others will chime in, and we'll help you build a foundation for translating all the diverse talking points you have collected into a unified model of how this stuff works. You have read a bunch, but we do not all use the same words in the same way, and you need a sifter in your head to sort it all out. This takes time, and by you asking I don't think you are the guy in Ashley Brilliant's joke- "I need to learn patience- where can I take a crash course?"
So let's take it a few questions at a time, and remember, they didn't build the Saturn V first.
So, good-easy-fast, pick one. (but have fun with it, they each provide a different sort of satisfaction)
David
First you'll need some basics. How does your 25W light bulb not blow out when clearly you can put a 150W bulb in the same apparatus and it(the 150W) draws six times the power?
For another, what happens when you put a driver in a tiny volume and port it into a larger one?
Start with those, others will chime in, and we'll help you build a foundation for translating all the diverse talking points you have collected into a unified model of how this stuff works. You have read a bunch, but we do not all use the same words in the same way, and you need a sifter in your head to sort it all out. This takes time, and by you asking I don't think you are the guy in Ashley Brilliant's joke- "I need to learn patience- where can I take a crash course?"
So let's take it a few questions at a time, and remember, they didn't build the Saturn V first.
So, good-easy-fast, pick one. (but have fun with it, they each provide a different sort of satisfaction)
David
Last edited:
Ok thanks, David. I'm a EE by training btw, although that was some time ago and I've never had an audio amplifier in front of me. If I've understood correctly speaker drivers tend to come in standard 4 and 8 ohm nominal flavors, so it seems they aren't power limiting by current limiting with increased resistance as I would guess filament lightbulbs do. There must be something upstream in the system architecture that is doing that output regulation? Now ordinary op amps, IIRC, you set the gain by connecting different value resistors, etc., to the amp pins. I would prefer to keep the amps outside the driver enclosures and most preferably use a finished product instead of DIY or integrating a stand alone card; so in that case, how is the gain-setting functionality exposed to the end user? If it's a pot knob on the housing or somesuch (per channel?) as headphone amps seem to have, I'm vulnerable to having my 3 year old (or, in due course, 13 year old) come by and twist the knob to a new, hardware and/or person damaging level. So it could be a reason to build fixed-gain amp boards into the enclosure after all. Or, am I misconceiving something basic?
I don't quite follow your question about 'port' a driver; are we talking about a tuned ported enclosure (not what I thought I was thinking of doing) or just putting a driver in a larger-than-critically damped enclosure? If I've understood, ordinarily you'd get a higher Q resonance at a slightly different frequency; but if the volume is filled up enough with damping material, the apparent volume the driver 'sees' is reduced again and damping is restored, and you're absorbing the backwave better than the smaller, more lightly stuffed box would do. Or, not so?
Thanks again, looking forward to learning as I go.
I don't quite follow your question about 'port' a driver; are we talking about a tuned ported enclosure (not what I thought I was thinking of doing) or just putting a driver in a larger-than-critically damped enclosure? If I've understood, ordinarily you'd get a higher Q resonance at a slightly different frequency; but if the volume is filled up enough with damping material, the apparent volume the driver 'sees' is reduced again and damping is restored, and you're absorbing the backwave better than the smaller, more lightly stuffed box would do. Or, not so?
Thanks again, looking forward to learning as I go.
You're working too hard. You have a voltage controlled by an impedance. If you don't push up the voltage, the motor cannot pass enough power to hurt itself. Ribbons(and a few others) sometimes have a very low impedance at low frequencies, and so cannot block a DC impulse should it occur. Hence the blocking capacitor.
For kid protection, how about a hidden(or software) -20db switch ? So they can turn it up to 11.
As to my second question, I was referring to your cone in the tube idea, a way to look at it. If the hole is small it will behave as a small volume ported to a larger. If the hole is large relative to the cross-section of the enclosure it will behave like a holey brace- the wave will diffract around the hole's edges, with a little loss each time a wave passes.
For kid protection, how about a hidden(or software) -20db switch ? So they can turn it up to 11.
As to my second question, I was referring to your cone in the tube idea, a way to look at it. If the hole is small it will behave as a small volume ported to a larger. If the hole is large relative to the cross-section of the enclosure it will behave like a holey brace- the wave will diffract around the hole's edges, with a little loss each time a wave passes.
Ok, I think I've been dense. About the amp, it sounds like a typical audio amp is just a dumb fixed gain (up to its ability limit), and entire responsibility for protecting the driver lies on the music player volume control. So see if I've got this right: If I use a 50W into 8 ohms amp channel to drive a 5W-rated, 8 ohm driver, I can, but then I only dare adjust my music source volume knob up to 1 out of 10, or I'll fry the driver. But that would produce the same SPL as using a 5W into 8 ohms mini-amp, and setting the source volume at 10/10. That's totally simple, and I guess you're right, I was trying too hard. I've -never- done that before, ha. But thanks for helping me through it, now I can set things up appropriately.
On the 'port' aspect...ok, got it now. I was thinking of the cone as a tapered transmission line, but sealed at the end is different of course. The narrow opening version might set up a resonance between the chambers (two masses connected by a spring?) The holey brace scenario doesn't sound that bad, just maybe not all that helpful. From what I've seen of LXmini builds, it looks like they just use Acoustastuff and nothing else, so hopefully that could suffice for for me as well.
Thanks for taking the time!
On the 'port' aspect...ok, got it now. I was thinking of the cone as a tapered transmission line, but sealed at the end is different of course. The narrow opening version might set up a resonance between the chambers (two masses connected by a spring?) The holey brace scenario doesn't sound that bad, just maybe not all that helpful. From what I've seen of LXmini builds, it looks like they just use Acoustastuff and nothing else, so hopefully that could suffice for for me as well.
Thanks for taking the time!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.