Separate ways

[bvbellomo]

Often people would expect that a lighter woofer cone might mean "faster" behavior more faithful as regard linear response (SPL versus frequency) but the T01 has some rather high Q cone modes, which will result in distortion products resonant amplification- and its very visible in the ripples in the impedance curve.

Sorry, I'd pass on that driver. Especially for the money, I think you'd be better off with even a Dayton RSS315FH-4.

Very visible to you, but I still can't see it. Do you mean just the T01, or the T00, or both? On the T00, I see a small ripple at 65Hz on the voice coil graph, but not scanspeak's site, is that it? ScanSpeak's graph look clean to 700Hz for both drivers. The RS315 shows a small ripple at 37Hz too.

I agree I won't know for sure until I measure the driver myself. The c173-6-191e's ended up being great drivers but not even close to the spec that was published at the time I ordered them.

[bvbellomo]

Hobby HiFi measured it, too.

The Peerless/Vifa NE315W looks pretty good by comparison. The 8-ohm version would be more suitable for a sealed cabinet, which is convenient because that's the one that PE carries.

I would definitely buy the driver Voice Coil tested with over the one Hobby HiFi tested. How does this happen? Was one made in Denmark and the other outsourced to China? Or just different testing methods?

The Vifa looks good except for the huge H3 spike at 100Hz, which rules it out for me.

Other than size, any reason to recommend the RSS315FH-4 over the larger Dayton Reference subwoofers?

[bvbellomo]

JonMarsh, I know you mentioned the Wavecor drivers being a poor value proposition, but the SW312WA03 is another driver that looks good to me, and cheaper than the ScanSpeak if shipping is reasonable.

I know we don't agree on value proposition. Considering the total cost of this hobby, a few hundred more for drivers isn't that much, and factoring my time (and what I'd be paid working on a paying project) it isn't much at all. Lots of people making half my salary are driving more expensive cars, value is a function of what it is you value.

[nopomo]

Have you looked at Zaph's write-up for his SB12.3? He measures the SB34NRX75-6 (the small magnet version) and shows some box models. The 3rd harmonic is at -50 dB (~0.3%) or lower below 250 Hz. The C173 has higher HD3 than that from 400 to 1200 Hz, going by Accuton's own measurements. Neither driver has shorting rings in its motor.

I don't see what that Wavecor offers that would make it a better choice than anything else mentioned thus far. It's optimized for sealed enclosures but modeling shows that it won't go any deeper than the SB in a 100-ish liter sealed enclosure. The SB is more efficient, has a bigger voice coil, and its cone is better-damped.

[bvbellomo]

Have you looked at Zaph's write-up for his SB12.3? He measures the SB34NRX75-6 (the small magnet version) and shows some box models. The 3rd harmonic is at -50 dB (~0.3%) or lower below 250 Hz. The C173 has higher HD3 than that from 400 to 1200 Hz, going by Accuton's own measurements. Neither driver has shorting rings in its motor.

I don't see what that Wavecor offers that would make it a better choice than anything else mentioned thus far. It's optimized for sealed enclosures but modeling shows that it won't go any deeper than the SB in a 100-ish liter sealed enclosure. The SB is more efficient, has a bigger voice coil, and its cone is better-damped.

I have seen it, and it is an interesting build.

The 3rd harmonic is indeed -50db down, but not from 0db, the reference level is much lower at the frequencies this driver would be used for. And this is 2 drivers playing only 90dB. So distortion isn't bad, but nowhere in the same class as the Accuton's or ScanSpeak.

Any idea why Zaph says these should be no closer than 4 feet to a wall? Seems a strange requirement for a non-ported speaker. Of course you will have room issues, especially in a smaller room, but that is all speakers in all rooms.

As I said before, this will almost certainly be a large sealed design with the goal of best 40-200Hz performance, efficeny isn't super important. If I keep a seperate sub, I question the value below 40Hz.

I am not sure how the Accuton's keep distortion so low without shorting of some kind, possibly a bigger issue with higher XMax drivers, but this is a driver I already have and am familiar with.

It is really hard to compare woofers I don't have, but:

Dayton:
Pros: Reputed and advertised to have good distortion. Great value for the price. Aluminum cone.
Cons: Very low mechanical Q. Some people say this doesn't make a difference, but IMO drivers with higher mechanical Q sound better.

WaveCor:
Pros: Very high mechanical Q, aluminum cone. Ardent design with the same tweeter and similar mid uses a similar sub with great results.

ScanSpeak:
Pros: Near perfect BL and class leading distortion as measured by Voice Coil. Plenty of positive subjective reviews on the Internet.
Cons: Hobby HiFi measured it as pretty bad, Jon doesn't like the impedence graph, price is very high, paper cone.

SB12:
Cons: Paper cone, good but not great distortion measured by Zaph

[nopomo]

I have seen it, and it is an interesting build.

The 3rd harmonic is indeed -50db down, but not from 0db, the reference level is much lower at the frequencies this driver would be used for. And this is 2 drivers playing only 90dB. So distortion isn't bad, but nowhere in the same class as the Accuton's or ScanSpeak.

You must be looking at the system graph. He measured a single speaker at 90 dB/1 M in a 2 cu. ft. enclosure.

The fundamental is at the 0dB level at about 60 Hz, where the 3rd order is at -60 dB. HD3 stays at that level as the fundamental goes down to -5 dB around 200 Hz, and HD3 is still -55 dB.

While we're discussing test conditions: either Vance Dickason made a typo or he actually tested that Scan 32W at 1.4 V.

"For distortion measurements, I set the voltage level with the driver mounted in an enclosure with a 14” × 30” baffle and increased it until it produced an 1-m SPL of 94 dB (1.4 V)"

If he did test at 1.4 V that might explain why the results are so exemplary compared to, say, the 28W Revelator (which was tested at 6 V).

I imagine that these drivers all have such low HD that it isn't a useful way of differentiating them. I've heard it said that mechanical Q is only relevant if the speaker is operating down near fs (which none of these are doing), which would make sense, but I don't really have an opinion about that.

[bvbellomo]

For distortion measurements, I set the voltage level with the driver mounted in an enclosure with a 14” × 30” baffle and increased it until it produced an 1-m SPL of 94 dB (1.4 V), which is my SPL standard for home audio drivers. I made the distortion measurement with the microphone placed near-field (10 cm) and the woofer mounted in the enclosure.

The graph draws the fundamental at about 108db. He measured at 1/10th the distance as the efficiency measurement, so volume should be 23db higher if we are radiating into a hemisphere, putting the measurement at 85db at 1.4V/meter. We can expect 89db at 2.8V/meter. I don't see a typo.

Zaph's graph is indeed down 55db at 50Hz, although lower frequencies are worse, we don't get them on the 32W graph, and the 32W is still 10db better at 50Hz.

While the voltage of 28W is indeed more, he is targeting the same output, as he should. Imagine a terribly inefficient driver with distortion is 70db down at 10 volts but only putting out 20db of fundamental. It would be misleading to say that driver measured well.

Do you have any vested interest pushing the SB34NRX75-6?

[cochinada]

I imagine that these drivers all have such low HD that it isn't a useful way of differentiating them. I've heard it said that mechanical Q is only relevant if the speaker is operating down near fs (which none of these are doing), which would make sense, but I don't really have an opinion about that.

...Do you have any vested interest pushing the SB34NRX75-6?

Hi, I think nobody here is trying to push any driver. They are just giving their opinion. Isn't that the purpose of your and everybody's else thread after all? In my opinion at a certain point, distortion is irrelevant. Namely because you can't tell the difference and also because it is much less noticeable in lower frequencies. That being said, I don't agree at all with your statement "but nowhere in the same class as the Accuton's or ScanSpeak.". I think Accuton for instance is highly over rated although they do have excellent drivers. You may not know but SB Accoustics origins came from Scan Speak. Anyway this is just my opinion and I have no intention to confront you or anybody for that matters.

[bvbellomo]

Hi, I think nobody here is trying to push any driver. They are just giving their opinion. Isn't that the purpose of your and everybody's else thread after all?

This wasn't a driver I mentioned and he didn't mention personal experience using it. Of course he is welcome to an opinion based only on a spec sheet, but he was pushing pretty hard if that was all he went by. I don't mind people pushing specific drivers for any reason (I appreciate it), but interests should be disclosed.

In my opinion at a certain point, distortion is irrelevant. Namely because you can't tell the difference and also because it is much less noticeable in lower frequencies.

I would agree past a certain point distortion is irrelevant. I won't get into whether it is more important for high or low frequencies, that is a complex topic that doesn't affect my choice of woofer. Voice Coil uses a standard of 20% distortion being acceptable, if that is HD3, it isn't acceptable to me. IMO people only accept distortion for subwoofers because so many subwoofers perform poorly.

That being said, I don't agree at all with your statement "but nowhere in the same class as the Accuton's or ScanSpeak.". I think Accuton for instance is highly over rated although they do have excellent drivers. You may not know but SB Accoustics origins came from Scan Speak. Anyway this is just my opinion and I have no intention to confront you or anybody for that matters.

My statement was in regard to my C173-6-096E's in specific and not Accuton in general. I do know where SB Accoustics came from but I have no idea where Wavecor came from, and their driver is top of my list now. I am not trying to be confrontational, especially about brands. As far as I care about brands, I respect all the brands discussed in this thread.

[nopomo]

This wasn't a driver I mentioned and he didn't mention personal experience using it. Of course he is welcome to an opinion based only on a spec sheet, but he was pushing pretty hard if that was all he went by. I don't mind people pushing specific drivers for any reason (I appreciate it), but interests should be disclosed.

The only 12" driver that's been mentioned and that I've used is the Dayton RSS315HF-4, as a sealed sub. The SB34 is noteworthy because not many 12" drivers have parameters optimized for a large sealed box. Wavecor has a new 10", the SW275BD0-, that seems to be intended for such use, but Solen doesn't stock it. The HiVi D10.8 (which is really an 11"er) looks interesting, too—and it has a shallow mounting depth of ~4 inches. I'm not selling a pair of SB34s, and I didn't mean to come off as confrontational.

My statement was in regard to my C173-6-096E's in specific and not Accuton in general. I do know where SB Accoustics came from but I have no idea where Wavecor came from, and their driver is top of my list now. I am not trying to be confrontational, especially about brands. As far as I care about brands, I respect all the brands discussed in this thread.

Some former Vifa/DST employees are behind Wavecor.

[bvbellomo]

The SB34 is noteworthy because not many 12" drivers have parameters optimized for a large sealed box

I'd am curious what you consider "optimized for a large sealed box".

Do you mean higher Qts? This would optimize for a larger sealed box vs a smaller sealed box. All things equal, I'd actually prefer the largest possible driver in the smallest possible box, simply because IMO larger cones sound better (everything else being equal, assuming they aren't playing so high they beam) and while I want a build that takes advantage of being allowed to build a large box, there are practical limits. But this is pretty far down my list of criteria. The Wavecor's are very high Q drivers and probably the top of my list now. If the Qts is at factory parameters, a critically damped sealed box would be too big.

People use Efficiency Bandwidth Product (Fs/Qes) to decide between ported and sealed. I am not sure the value of this, but Wavecor SW312WA03 are 28/0.50=56
vs SB's 19/.43 = 44. So both look better sealed from that angle, although the SB's a bit more so.

Wavecor vendors are horrible about keeping stock online, but I am not sure how to build a box for a Qtc > .5 which is underdamped in free air.

[bvbellomo]

I just sent an e-mail to Meniscus to preorder a pair of Wavecor SW312WA03. I was torn between this and the Scan Speak 32W/4878T for a long time. Either fits nicely with my design and I don't have enough information just from online to know which is best. I am not opposed to sending these back or selling them 2nd hand and using the SS32W4878T or something else if they disappoint. But I have to start somewhere, and I am not getting any new information without a physical driver in hand.

Regarding crossover design, how do I know if I need Zobels? Will lack of Zobel's just show up as too much high frequency from the woofer? Or is this like notching a tweeter - easy to hear but hard to measure?

[bvbellomo]

Big box arrived yesterday. They misspelled "LOUD" but no other issues:


These things seem so much bigger and heavier in person:


Meniscus was nice enough to supply DATS printouts with each woofer:
Woofer 1 f(s) 29.61Hz Q(es) 0.5483 Q(ms) 8.777 Q(ts) 0.5161 R(e) 3.460 Hz Le(1k) = 1.539 mH Le(10k) = 0.8371 mH
Woofer 2 f(s) 29.61Hz Q(es) 0.5331 Q(ms) 8.433 Q(ts) 0.5014 R(e) 3.431 Hz Le(1k) = 1.539 mH Le(10k) = 0.8377 mH
Factory f(s) 28Hz Q(es) .5 Q(ms) Q(ts) .48 11.5 Le(1k) 1.55

So I'd prefer the factory, but these aren't far off, and more importantly, they match well. I will verify these with my own DATS, get other parameters, and then see how they measure broken in. Might start cutting front baffles, since I know what height will be whatever I need to put the tweeter where I want it, and the the width will be as narrow as fits the woofer nicely, but I want measured parameters before I determine final enclosure volume.

Any advice on Qtc? Especially if the Qts doesn't drop much as they break in? I prefer a lower Qtc, critically damped if it makes sense. But with the Qts this high, a really small drop in Qtc can mean a really big increase in enclosure size.

As a Browns fan, I really needed something good yesterday, and I think I got it - although certainly not from Dorsey way overreaching for his players.

[Carl V]

I am sure you are aware that LOURE is=Heavy
Browns Fan...oh my. an Texas/Oklahoma QB draft pick.

QTS is a matter of Preference .5 -.7 etc.,
it's not always obvious as the trout in the milk

have fun

[bvbellomo]

I am sure you are aware that LOURE is=Heavy
Browns Fan...oh my. an Texas/Oklahoma QB draft pick.

QTS is a matter of Preference .5 -.7 etc.,
it's not always obvious as the trout in the milk

have fun

My preference is for a lower QTS. But the math makes volume size ridiculous. With a Qts of 0.48 (I don't know where I'll be after break-in)
.707 is 36 liters
.577 is 85.2 liters
.52 is 165 liters
.51 is 190 liters
.500 is 236.1 liters

Will I hear a difference between .5 and .577? Is it enough to have boxes 3 times as big?

[Bear]

My preference is for a lower QTS. But the math makes volume size ridiculous. With a Qts of 0.48 (I don't know where I'll be after break-in)
.707 is 36 liters
.577 is 85.2 liters
.52 is 165 liters
.51 is 190 liters
.500 is 236.1 liters

Will I hear a difference between .5 and .577? Is it enough to have boxes 3 times as big?

That's a definite maybe. When you add room effects and placement effects, you can get lost in what's happening at <40Hz pretty easily. If you have UniBox, you can look at the impulse response as the cabinet increases in volume. Also look at the shape of the driver rolloff towards sub-aural thresholds. If you play around with it a bit, you will see what happens physically when people talk about "sloppy", "boomy", or "one note" bass. This is why people like "fast bass", but this is still fundamentally the manifestation of Hoffman's Iron Law.

[bvbellomo]

Almost no change running a 30Hz wave over 24 hours. Qts actually went up a tiny bit (0.5276), so these won't get critically damped.

I did play with Unibox, but it doesn't seem to give me anything beyond what Speakerboxlite does. I can see the response curve, but I can't tell whether I will hear a difference. This is closed box - badly designed vented boxes show obvious problems, and very small closed boxes also show obvious problems. But .707 or .5 or anything in between looks almost the same on the graph.

[bvbellomo]

With a 42"x12" inside area (since we are going by inside dimensions) of my front panel, this gives me:

A critically damped (0.5 Qtc) enclosure would be 1466.5 liters, or 14 feet and 9 inches deep. Even without a wife yelling at me, I don't see that happening.

A 0.577 Qtc enclosure would be 17.5 inches deep. This is a little more than I planned, but certainly doable.

A 0.707 Qtc enclosure would be 6 inches deep, nearly impossible to build and likely to fall over. Of course I could have part of the box not contribute to the volume, but that completely defeats the purpose.

A Qtc of 1 would be a ridiculously slim 2 inches deep.


So having such a high Qts pretty much forces me into a Qtc between 0.7 and 0.57 for a rectangular speaker of the height and width I'd want. I've thought of tapering the box - wider at the bottom and tilted in some, but am not sure I want to do it. It will be a subtle taper if anything, and my woodworking skills aren't the best.

[bvbellomo]

Looking at the Ardent design, does anyone know why they slope back 10 degrees? I thought the whole point of the LR3 was to phase shift at crossover frequencies to time-align the waves. You do either this or slope the baffle, but by doing both, wouldn't you end up with the same problem, just with the axis tilted up instead of down? I was considering a 5 degree slope to center the drivers vertically with LR2's, but this may not be enough. Does anyone know a reliable way to determine how far back a driver is set? I can whip out a tape measure and say it is roughly 3 inches on the Wavecor, but you'd think the manufacturer would include this on the spec sheet - instead we get measurements for just about every other other distance. The Wavecor's don't really matter, 200Hz is 17 inches for a quarter wave, but the Accuton's do matter.

The Ardents use 3/4 inch Baltic Birch bracing. Would MDF baffles work just as well or is plywood better? If we use plywood, wouldn't thinner be better? No matter how loud I play, I can't see breaking a 3/4 inch MDF brace or a 1/2" plywood brace.

[nopomo]

Looking at the Ardent design, does anyone know why they slope back 10 degrees? I thought the whole point of the LR3 was to phase shift at crossover frequencies to time-align the waves. You do either this or slope the baffle, but by doing both, wouldn't you end up with the same problem, just with the axis tilted up instead of down?

Post #12

Does anyone know a reliable way to determine how far back a driver is set? I can whip out a tape measure and say it is roughly 3 inches on the Wavecor, but you'd think the manufacturer would include this on the spec sheet - instead we get measurements for just about every other other distance. The Wavecor's don't really matter, 200Hz is 17 inches for a quarter wave, but the Accuton's do matter.

You would have to take a frequency response measurement with phase and an accurate distance measurement from some reference on the driver, probably to .001" or better. Then extract minimum phase from the response. Next you would need to determine the time delay that exists in the excess phase between the measured phase and the minimum phase and convert that to distance, and finally add or subtract that distance from your reference distance. This might get you pretty close, but probably still not on the nose, as you will find that the acoustic center may be different at different frequencies for some drivers.

As a rule of thumb, you can estimate that the AC is located just ahead of the point at which the voice coil former attaches to the diaphragm. SBA is the only manufacturer I know of to publish the relative AC offsets of its drivers.

[bvbellomo]

I saw post 12. "The usable vertical window starts just below the midrange driver axis as is". Does "as is" mean tilted or straight? And what is "usable"?

With no tilt or crossover induced phase shift, we'd have a drop in the vertical axis at the crossover frequency between tweeter and mid. Tilting back is not a great fix, as we get a rise in the vertical axis at the non-crossover frequencies. This is a really minor problem, but it seems the Ardent (or Avalon Isis) is on a level where it is trying to fix this. The LR3 is an elegant solution, and should fix the vertical axis at crossover without a rise outside the crossover frequency.

I hate to be critical of Jon's design, but it looks like he copied the aesthetics of the Isis and their concept of fixing vertical alignment in the crossover, without actually fixing the vertical axis (in the Isis the tweeter is forward of the mid). As I said, this is a tiny problem, and the Ardent is still a great speaker, but I wonder if it wouldn't be just as great in a simple rectangle with an LR2.

For my design, I am going to stick with an LR2. I am tempted to slope my baffles backwards, not only because of this issue, but also aesthetics and balance. My design will be much taller, and with kids (I may have some someday, but my brother's will visit), I want to minimize the chance of knocking it over and inhaling beryllium.

[Bear]

Regarding crossover design, how do I know if I need Zobels? Will lack of Zobel's just show up as too much high frequency from the woofer? Or is this like notching a tweeter - easy to hear but hard to measure?

I was playing around with a design and wasn't having luck with hammering a woofers cone breakup using a traditional notch filter. I also could get great phase alignment right at the crossover point, but the "slope" of the woofer phase wasn't lining up with the tweeter phase just above the XO frequency. So, I sucked it up that I wasn't going to find an easy-out parts-wise, and added a Zobel/conjugate notch (basically, adding an additional parallel leg after the main crossover components). By adjusting the values for some of those components, I was able to tackle the cone break-up and also better align the phasing of the two drivers.

Net: you'll know it when dialing in the more foundational elements of your crossover aren't giving you the results that you want.

[Bear]

I saw post 12. "The usable vertical window starts just below the midrange driver axis as is". Does "as is" mean tilted or straight? And what is "usable"?

With no tilt or crossover induced phase shift, we'd have a drop in the vertical axis at the crossover frequency between tweeter and mid. Tilting back is not a great fix, as we get a rise in the vertical axis at the non-crossover frequencies. This is a really minor problem, but it seems the Ardent (or Avalon Isis) is on a level where it is trying to fix this. The LR3 is an elegant solution, and should fix the vertical axis at crossover without a rise outside the crossover frequency.

I hate to be critical of Jon's design, but it looks like he copied the aesthetics of the Isis and their concept of fixing vertical alignment in the crossover, without actually fixing the vertical axis (in the Isis the tweeter is forward of the mid). As I said, this is a tiny problem, and the Ardent is still a great speaker, but I wonder if it wouldn't be just as great in a simple rectangle with an LR2.

For my design, I am going to stick with an LR2. I am tempted to slope my baffles backwards, not only because of this issue, but also aesthetics and balance. My design will be much taller, and with kids (I may have some someday, but my brother's will visit), I want to minimize the chance of knocking it over and inhaling beryllium.

You, uh, may want to read some of the back issues (i.e., older threads) here a bit more completely before talking about people copying other people. You may also want to start with Jon's writings on Duelund phase perfect crossover topologies. After that, things became compromises that spanned about seven to nine years of incremental work and multiple lifestages, including drivers that weren't originally available to drivers that are no longer available.

People slope baffles for LR2 crossover targets because it's the easiest way to align the acoustic centers of the drivers (yes, there can be a positive aesthetic, as well, but there is a functional purpose as well). The alternate approach that is "easy" is a ladder delay network, but then you are potentially adding a lot of parts into the mix. Other options include using waveguides on tweeters (good fits between a given tweeter and a given waveguide is more miss than hit) and using only planar transducers. Accuton's line of "Cell" drivers was founded on the principle that all of the Cell drivers would have the same acoustic center, regardless of use or purpose.

[bvbellomo]

I was playing around with a design and wasn't having luck with hammering a woofers cone breakup using a traditional notch filter. I also could get great phase alignment right at the crossover point, but the "slope" of the woofer phase wasn't lining up with the tweeter phase just above the XO frequency. So, I sucked it up that I wasn't going to find an easy-out parts-wise, and added a Zobel/conjugate notch (basically, adding an additional parallel leg after the main crossover components). By adjusting the values for some of those components, I was able to tackle the cone break-up and also better align the phasing of the two drivers.

Net: you'll know it when dialing in the more foundational elements of your crossover aren't giving you the results that you want.

I never considered Zobels (or lack of) would affect phase as well as response, but I suppose I need to consider that as well. I suppose the rising impedance would behave like a first order crossover with 90 degree shift? This will probably be an exercise in trying different components and measuring what works, but I wouldn't have thought to measure phase.

[bvbellomo]

You, uh, may want to read some of the back issues (i.e., older threads) here a bit more completely before talking about people copying other people. You may also want to start with Jon's writings on Duelund phase perfect crossover topologies. After that, things became compromises that spanned about seven to nine years of incremental work and multiple lifestages, including drivers that weren't originally available to drivers that are no longer available.

People slope baffles for LR2 crossover targets because it's the easiest way to align the acoustic centers of the drivers (yes, there can be a positive aesthetic, as well, but there is a functional purpose as well). The alternate approach that is "easy" is a ladder delay network, but then you are potentially adding a lot of parts into the mix. Other options include using waveguides on tweeters (good fits between a given tweeter and a given waveguide is more miss than hit) and using only planar transducers. Accuton's line of "Cell" drivers was founded on the principle that all of the Cell drivers would have the same acoustic center, regardless of use or purpose.

I've read everything Jon's written, at least on htguide, and the relevant stuff multiple times. I am well aware he was working on this project (or what became this project) years before I even considered building a speaker. This is why "I hate to be critical of Jon's design". But ultimately, this is engineering, and we should be critical of Jon's design (or any other engineer's design).

"copying" might not be the most elegant choice of words, but Jon makes no secret that the Ardents didn't come from a vacuum. He was inspired with the Isis Avalon, one of the best speakers ever made, and scaled the design down a much smaller size, made something friendly to his amps, and built something for about 5% the cost. He intentionally picked the mid to be similar to the Avalon's, and intentionally used a similar crossover. He could have made the Ardents look like anything, but I don't think it is by coincidence they look like the Avalon.

Where the design doesn't make sense, at least to me, is the 10 degree sloped baffle. Jon himself wrote "you could do these in a straight tower box with the same baffle width, it will not behave quite the same off axis horizontally, but on axis will be pretty much identical". So it is not unreasonable to say that it doesn't make a huge difference.

Jon may have made this decision based on factors irrelevant to my design. He wrote "they won't sound as good standing up close to them, as with the cabinet tilt back", which assumes the listener is taller than the Wavecor. This is a fair assumption, but completely irrelevant with a 60 inch tall speaker and a listener sitting further back. He may have intentionally wanted an axis off vertical.

[Steve Manning]

Looking at the Ardent design, does anyone know why they slope back 10 degrees? I thought the whole point of the LR3 was to phase shift at crossover frequencies to time-align the waves. You do either this or slope the baffle, but by doing both, wouldn't you end up with the same problem, just with the axis tilted up instead of down? I was considering a 5 degree slope to center the drivers vertically with LR2's, but this may not be enough. Does anyone know a reliable way to determine how far back a driver is set? I can whip out a tape measure and say it is roughly 3 inches on the Wavecor, but you'd think the manufacturer would include this on the spec sheet - instead we get measurements for just about every other other distance. The Wavecor's don't really matter, 200Hz is 17 inches for a quarter wave, but the Accuton's do matter.

The Ardents use 3/4 inch Baltic Birch bracing. Would MDF baffles work just as well or is plywood better? If we use plywood, wouldn't thinner be better? No matter how loud I play, I can't see breaking a 3/4 inch MDF brace or a 1/2" plywood brace.

I can't speak to the why's of the 10 deg slant, though as far as how far back the drivers sit, if you use any type of CAD software that would be fairly easy to determine. Also dust off the old math skills and that's also fairly straight forward to figure out.

As for the differences between mdf and Baltic Birch bracing. BB is stronger and thinner is not going to be better with either material choice if your trying to stop cabinet flexing. I imagine if you wanted to go with mdf they would sound just fine. Just for fun you could build a set of each and let us known if they sound any different. :W

[bvbellomo]

I can't speak to the why's of the 10 deg slant, though as far as how far back the drivers sit, if you use any type of CAD software that would be fairly easy to determine. Also dust off the old math skills and that's also fairly straight forward to figure out.

My difficulty is not with the math, which is straightforward. The manufacture doesn't give me the dimensions, and short of cutting an $800 woofer in half, the best I can do is guesstimate the distance from the acoustic center to the baffle plain.

As for the differences between mdf and Baltic Birch bracing. BB is stronger and thinner is not going to be better with either material choice if your trying to stop cabinet flexing. I imagine if you wanted to go with mdf they would sound just fine. Just for fun you could build a set of each and let us known if they sound any different. :W

A box flexes because there is varying pressure on the inside. But a brace should never flex because pressure is equal except where it is attached - in theory material shouldn't matter unless the brace breaks. But Jon specified BB for his braces, and I wondered if he had a reason.

[Steve Manning]

I hate to tell ya, but the materials do matter, otherwise you could just use balsa wood. Thing is, this is DIY, do your own thing and experiment, let us know how it turns out and that way we all get to learn.

[bvbellomo]

I hate to tell ya, but the materials do matter, otherwise you could just use balsa wood. Thing is, this is DIY, do your own thing and experiment, let us know how it turns out and that way we all get to learn.

Yes, but I save a lot of time and money whenever I find someone who already made the same mistake before I do. I don't understand why balsa wood wouldn't work, it isn't particularly strong so it would have to be thick to avoid breaking, and it isn't particularly cheap, so I wouldn't try it. Does it resonate?

[Steve Manning]

What learning from others mistakes ..... smart man.

Not sure I've seen any speakers that have used balsa .... good for model planes though.

[Bear]

But ultimately, this is engineering, and we should be critical of Jon's design (or any other engineer's design).

You can always ask for your money back...

Where the design doesn't make sense, at least to me, is the 10 degree sloped baffle. Jon himself wrote "you could do these in a straight tower box with the same baffle width, it will not behave quite the same off axis horizontally, but on axis will be pretty much identical". So it is not unreasonable to say that it doesn't make a huge difference.

Jon may have made this decision based on factors irrelevant to my design. He wrote "they won't sound as good standing up close to them, as with the cabinet tilt back", which assumes the listener is taller than the Wavecor. This is a fair assumption, but completely irrelevant with a 60 inch tall speaker and a listener sitting further back. He may have intentionally wanted an axis off vertical.

I'm not quite following where you got the 60" dimension. The Ardent baffle is 40", with a 10* slope. That gives you a starting height of 39.39", plus a base that is 2" - 4" or so, depending upon builder. Because the drivers are not coaxial with a constant acoustic center, the waveforms emanating from each will interact. This causes combing and lobing, among others, though lobing can be caused by other factors (e.g., too big of a cone and too small of a wavelength). If you're coming up with your own design, then you probably already know these things, so I'm not 100% sure what the question is, from an engineering standpoint (e.g., how much does the slop account for differences in acoustic centers? What does the vertical polar response look like, and how does it change vis-a-vis the listener, etc.).

When you tilt a speaker forward, the vertical polar response goes with it. Whether that puts the listener's ears into a null or not for some set of frequencies might be answerable by the quality of the software and the willingness of the designer. Jon indicated that the speaker can be "straight", so I would take that as indicating that there would be no noticeable nulls based upon some assumptions about where the target listener is. How that relates to your design is truly unclear to me.

My difficulty is not with the math, which is straightforward. The manufacture doesn't give me the dimensions, and short of cutting an $800 woofer in half, the best I can do is guesstimate the distance from the acoustic center to the baffle plain.

This is where measurement comes into play, since even chopping up a driver isn't going to give you a very good answer. A good design software that does FEA (e.g., ANSYS) will give you a pretty precise answer, but that begs a few additional questions (e.g., how long have you worked for Scan-Speak, and why not ask someone down the hall?). Your baffle will also cause diffraction, and while there are tools that can estimate these effects, measurement helps confirm theory. Remember, these are analog devices produced by low-cost manufacturers. You should expect unit-to-unit variation.

Best of luck!

[bvbellomo]

I'm not quite following where you got the 60" dimension. The Ardent baffle is 40", with a 10* slope. That gives you a starting height of 39.39", plus a base that is 2" - 4" or so, depending upon builder. Because the drivers are not coaxial with a constant acoustic center, the waveforms emanating from each will interact. This causes combing and lobing, among others, though lobing can be caused by other factors (e.g., too big of a cone and too small of a wavelength). If you're coming up with your own design, then you probably already know these things, so I'm not 100% sure what the question is, from an engineering standpoint (e.g., how much does the slop account for differences in acoustic centers? What does the vertical polar response look like, and how does it change vis-a-vis the listener, etc.).

I am working on my own design, which will be 60 inches or so. My point is that standing next to a 36 inch speaker, the listener (unless super short) is not at ear level with the tweeter, and this could be Jon's reason for the 10 degree tilt, which would not apply to my own design.

This is where measurement comes into play, since even chopping up a driver isn't going to give you a very good answer. A good design software that does FEA (e.g., ANSYS) will give you a pretty precise answer, but that begs a few additional questions (e.g., how long have you worked for Scan-Speak, and why not ask someone down the hall?). Your baffle will also cause diffraction, and while there are tools that can estimate these effects, measurement helps confirm theory. Remember, these are analog devices produced by low-cost manufacturers. You should expect unit-to-unit variation.
Best of luck!

This isn't so much a tools issue as it is a measurement issue. If I have something to start from (either manufacture supplied or measured) either math or tools can tell me what the offset is, but I don't have anything to start with. "Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?".

[Carl V]

What learning from others mistakes ..... smart man.

Not sure I've seen any speakers that have used balsa .... good for model planes though.

Richard Vandersteen has demonstrated that High Quality Balsa Core with Carbon Fiber makes for
a strong, rigid & non resonate material. It can be equally strong etc., with traditional marine Polyester
Bagged laminate for off shore race boats. It's like wise been used in Aviation. Richard is an experienced
Pilot of both Planes & Boats.

[bvbellomo]

To summarize for those who don't want to read this whole thread, I am building 2 large 3-way floor standers using SW312WA03 woofers, C173-6-096E mids and D3004/664000 tweeters by adding the Wavecor's to a ported 2-way using the mids as woofers.

I haven't done much speaker related this summer, as my divorce is going badly. My ex is going to make this as long, painful and expensive as she can, the existing speakers and materials are shared property, and she is trying to get the court to value my speakers at twice what I paid for parts. So I am hesitant to take them apart and reuse the drivers.

For now, I am planning on keeping the same drivers, but not building a finished cabinet yet. I'd like to build temporary boxes for the Wavecors, and set the existing speakers on top. This also lets me work modify the crossover easily, as my new design is sealed. Once I have finished crossovers and a finished divorce, I will build really nice cabinets.

My existing cabinet is ported and smaller than I'd ideally like for my mids. I can seal it easily enough, but I'd like a critically damped sealed box for the finished Accutons and the Q probably around 0.8, I'd need to measure. Assuming I cross between the wavecore's and accuton's well above the F3 of my current cabinet, does this even matter? I know I will have to fine tune my crossover for the finished cabinets, but can't I get close now?

[Bear]

My existing cabinet is ported and smaller than I'd ideally like for my mids. I can seal it easily enough, but I'd like a critically damped sealed box for the finished Accutons and the Q probably around 0.8, I'd need to measure. Assuming I cross between the wavecore's and accuton's well above the F3 of my current cabinet, does this even matter? I know I will have to fine tune my crossover for the finished cabinets, but can't I get close now?

Qtc for critically damped is 0.5. I don't know enough to reconcile the Qtc of 0.8 vs. critically damped, without resorting to a Linkwitz circuit. And those are generally used to make smaller boxes seem larger. What I do know is that the Q will affect the resonance peak, and associated impedance peak. If you cross well above that peak, then you should focus more on the rising impedance as frequency increases (this is where a Zobel comes into play). If you are using an active crossover, then the latter issue isn't relevant.

I think most designers use a sealed midrange because the chamber is smaller and the overall construction/execution is simpler (cheaper). If you have pre-existing TM speakers that you're happy with, then yeah, I'd just focus on the bass bins and be done for a bit. People routinely report spending _years_ tweaking crossover designs, so getting something "good enough" is often just the start.

[bvbellomo]

Qtc for critically damped is 0.5. I don't know enough to reconcile the Qtc of 0.8 vs. critically damped, without resorting to a Linkwitz circuit. And those are generally used to make smaller boxes seem larger. What I do know is that the Q will affect the resonance peak, and associated impedance peak. If you cross well above that peak, then you should focus more on the rising impedance as frequency increases (this is where a Zobel comes into play). If you are using an active crossover, then the latter issue isn't relevant.

I think most designers use a sealed midrange because the chamber is smaller and the overall construction/execution is simpler (cheaper). If you have pre-existing TM speakers that you're happy with, then yeah, I'd just focus on the bass bins and be done for a bit. People routinely report spending _years_ tweaking crossover designs, so getting something "good enough" is often just the start.

Thanks - not looking to do active or use a Linkwitz circuit, just hoping that it won't matter if I cross more than an octave above resonance peak.

This isn't that I am happy with the existing design (which IMO would benefit with Zobels on the Accutons), as the ex-wife fighting in court and claiming an outrageous value for my cabinetry work. If I can't get these fairly valued in court, I will either try to give them to her at the value she suggested, or sell them - either of those requires having the speakers whole, and not in pieces.