'ats a nice driver. Perhaps more importantly for those of us without Michael's unusually low SPL requirements it has enough xmax to produce SPLs comparable to the Orion when placed in an H baffle of depth comparable to the Orion. The effective path difference in Linkwitz's SPL spreadsheet is not well defined in the case of a nude driver, but from what I can tell it's probably on the order of 25mm. That suggests a 25dB decrease in SPL when reducing from the Orion's ~380mm H baffle. Regaining the lost 25dB requires a 15 fold increase in volume displacement, which implies either multiple line arrays of smaller nude drivers or taking StikErik's four 21" Beymas per side and running them nude.
Unless one likes to listen loud, a more practical solution would be to lower the system's maximum SPL by 20dB from the Orion's 109dB. This would require "only" a single line array of eight nude 12" woofers per side. Assuming you negotiated decent volume discount on 285 GMFs that'd be something like 3000 euro just for the deep bass drivers. A line array of 830669s is a more cost effective option but still around 750 euro (I'm converting US pricing a the current exchange rate here). I think I've just rediscovered the virtues of baffle coloration.
Perhaps the most interesting cost effective option I came across is a pair of Selenium 15WP5s per side. Run nude these come in 23dB below the Orion when excursion limited. Should be enough headroom to provide reasonable distortion at what I'd consider to be typical listening levels. They're unfortunately too large to work in my listening room but they could be an interesting starting point if someone else is so inclined.Only for point source drivers; if you look at Linkwitz's website you'll see the effective path difference is defined as the H baffle's depth and is independent of the size of the drivers. In the case of a nude speaker the obvious starting point for finding an equivalent H baffle depth is the difference between the frame diameter and the driver's effective diameter as given by its Sd. For the drivers I've looked at that's fallen between 25 and 27mm.
The question is how well the H baffle model holds up. I'm looking at Linkwitz's power limited SPL spreadsheet at the moment and getting odd looking results above a couple hundred Hertz. Off the top of my head I'd guess an effective path difference closer to a sixth of the diameter which, for a 12 inch driver, is within a factor of two of an effective diameter based calculation. The Fequal model for SPL calculation may be a larger source of error and I'm not, at this point, convinced an effective path difference is necessarily the best approach.
This is starting to turn into a thread hijack; I'll spin up another thread on nude driver SPL. Need to measure the one type of woofer I have handy nude and think about the theory some, but that's going to take a while.
The question is how well the H baffle model holds up. I'm looking at Linkwitz's power limited SPL spreadsheet at the moment and getting odd looking results above a couple hundred Hertz. Off the top of my head I'd guess an effective path difference closer to a sixth of the diameter which, for a 12 inch driver, is within a factor of two of an effective diameter based calculation. The Fequal model for SPL calculation may be a larger source of error and I'm not, at this point, convinced an effective path difference is necessarily the best approach.
This is starting to turn into a thread hijack; I'll spin up another thread on nude driver SPL. Need to measure the one type of woofer I have handy nude and think about the theory some, but that's going to take a while.
Last edited:
I don't think that's right. D is the average of the distance from all points on the back of the cone to all points around the frame and then forward. It's not just from any point at the edge to the nearest point on the frame.
That sounds about right. The sound is making its way around obstructions on the back side and that would increase the path length. I had to model the Neo3 in Edge with a bigger frame as well to get it to match measurements posted here, most likely because of the thickness of the frame.
Last edited:
Wow guys – great discussion – keep it going....
###########
regarding Rms – you may certainly be right in analysis of the root of high Qms – on the other hand Rms is nothing closer to what I'm after with respect to low-SPL distortion – simply because its not purely represented in TSP - we would have to measure the way I did for a quick and dirty try – meaning – detecting the relationship of 2nd and higher order and have a look at what SPL this is more or less equal.
This possibly might provide a more objective guideline to what I stumbled across
###########
regarding "Dampening pads"
*if * they do what the name suggests – these are quite different in outcome to what a swing does.
Given they do not only provide a mere spring function, the "dampening" part is actually not welcome here. This may be of some surprise as it is not immediately intuitively at all.
We have to go back to the roots of what the aim was for my suggestion of a swing. Basically the "complete impulse compensation by local mass " works best when there is *no* leakage path for any energy.
What this means is best described in the "outer space" example I like to use. If our beloved speaker would be operated in outer space then - while the membrane is going in and out - the rest of the speaker would move *exactly* in the opposite direction by the ratio of masses involved.
I probably should underline that the *speaker as a whole* would perfectly stay in place, though.
We clearly see that there is no energy leakage whatsoever involved – meaning "perfect impulse compensation by local mass".
Going back to the dampening issue - there are two different types of dampening to be kept apart.
Both have in common that they provide an "energy leackage path" - but they differ considerably on *how* the energy gets lost and thus the net movement of the diaphragm also gets affected differently.
#########
LOL – I may have fallen too much in love with my dipole horn
Anytime you have a wish for more bass-quantity just increase the inductor and put a L-pad in front of the tweeter....
Its just like turning the bass knob of vintage receiver..
##########
Graaf, your pix did not open for me...
Michael
###########
regarding Rms – you may certainly be right in analysis of the root of high Qms – on the other hand Rms is nothing closer to what I'm after with respect to low-SPL distortion – simply because its not purely represented in TSP - we would have to measure the way I did for a quick and dirty try – meaning – detecting the relationship of 2nd and higher order and have a look at what SPL this is more or less equal.
This possibly might provide a more objective guideline to what I stumbled across
###########
regarding "Dampening pads"
*if * they do what the name suggests – these are quite different in outcome to what a swing does.
Given they do not only provide a mere spring function, the "dampening" part is actually not welcome here. This may be of some surprise as it is not immediately intuitively at all.
We have to go back to the roots of what the aim was for my suggestion of a swing. Basically the "complete impulse compensation by local mass " works best when there is *no* leakage path for any energy.
What this means is best described in the "outer space" example I like to use. If our beloved speaker would be operated in outer space then - while the membrane is going in and out - the rest of the speaker would move *exactly* in the opposite direction by the ratio of masses involved.
I probably should underline that the *speaker as a whole* would perfectly stay in place, though.
We clearly see that there is no energy leakage whatsoever involved – meaning "perfect impulse compensation by local mass".
Going back to the dampening issue - there are two different types of dampening to be kept apart.
Both have in common that they provide an "energy leackage path" - but they differ considerably on *how* the energy gets lost and thus the net movement of the diaphragm also gets affected differently.
#########
Based on the attached I'll file you under "allergic to bass" as 0.5dB boost at 150Hz isn't much.
LOL – I may have fallen too much in love with my dipole horn
Anytime you have a wish for more bass-quantity just increase the inductor and put a L-pad in front of the tweeter....
Its just like turning the bass knob of vintage receiver..
##########
Graaf, your pix did not open for me...
Michael
It just seemed counter-intuitive to me that high Mms should be good for a free-swinging driver.
Oh, OK, I won't split the thread then.
Agree it's the average of the integrals over all points and paths. Also agree backside obstructions will add a few percent to the path lengths with the most influence on the integral's value. And I'd really like for the effective path length to work out to be as along as possible since more SPL means less baffle and cheaper drivers.
However, I'm missing something. If we're following Linkwitz's formulation then the dipole peak occurs when D is half a wavelength. I forget if StikErik's using W12s, W14s, W15s, or W16s---all are labeled as 5" drivers at least some of the time---but the basket diameters are in the range of 120 to 146mm. I recall it being W15s, which are 146mm, and if I simulate 146mm in Edge I get a dipole peak around the expected 1750Hz. However, (1.1 to 1.15) * 146mm is not in good agreement with a dipole peak at 1550Hz, where D would be 0.5 * 343 / 1550 = 110mm. There seems to be a factor of 1.5 missing in here somewhere, or maybe pi/2, though that would be a little large.
The horn is cool. We understand these things happen.LOL – I may have fallen too much in love with my dipole horn
I think Stig Erik has used so many drivers he's getting them confused. Let's take Keyser's dipole with the Visaton 170 which has a 180mm frame. 343/.18 would predict a 1.9K peak but it actually measures closer to 1.5K, about 20% lower. His speaker isn't quite naked with a small rear baffle and the tweeter attached so Stig Erik's rule of thumb to add 10-15% is probably pretty close. Measurements will tell you exactly what you have but that's close enough for estimating how many drivers you need.
Let's take Keyser's dipole with the Visaton 170 which has a 180mm frame. 343/.18 would predict a 1.9K peak but it actually measures closer to 1.5K, about 20% lower. His speaker isn't quite naked with a small rear baffle and the tweeter attached so Stig Erik's rule of thumb to add 10-15% is probably pretty close. Measurements will tell you exactly what you have but that's close enough for estimating how many drivers you need.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Thanks! My guess from the datasheet would be the driver's SPL peak at 1550Hz is masking the actual dipole peak. If you've still got the data handy would you mind posting the curves? I'd like to take a closer look and I suspect catapult would be interested as well.
I suspect the multiplier shrinks a bit as driver size increases---backside structures tend to get smaller relative to the overall diameter----but if 12% is about right then switching from the Orion's H baffled 10 inchers to nude 12 inchers maintains the SPL, give or take minor differences in excursion. That would indicate nude drivers are thoroughly viable on a less than StigErik scale. I haven't closed on a 12" driver yet, but so far I keep coming back to the 830669 as the best combination of availabilty, affordability, and performance. I can buy four of them for just about what it'd cost me to get one of the Beymas Michael's using.
I'm currently thinking of building a four way. Nominally a Neo3 W, Neo8 (non-PDR), ~8" woofer, and an 830669 per side. Crossovers around 2kHz, 750Hz, and 350Hz. However, I'm having difficulty finding an ~8" I like, so I'm debating various options such as a three way that extends the 830669 to 750Hz, a four way with two 830669s with one going to 750Hz and the other acting as a filler in the low bass, and so on. I can run up to five way, so possibilities like Neo3/Neo8/7"/10"/830669 are open as well. Suggestions? I'd prefer to have a couple of complete designs in mind before ordering 830669s. Partly to have a better sense of what the 830669s would need to measure to, partly to mitigate the risk of some better design coming up and leaving me with 830669s I turn out not to need.
This sounds like a concept, fun to do.
Keep in mind though that the 830669 is a very good driver as long as combined with usual tweeters but does not mate that well with planar's, IMO.
I'm interested what you or others findings are in this respect.
Open baffle and even more so nude speaker reveal such details very clearly...
Its not only been because of the low-SPL distortion / Qms aspect I was looking for an alternative - also I simply don't like the sound of rubber surrounds. Pretty hard to find anything else than PA speakers if you have such preferences
Whats the problem with 8"? If you cross active, you should be able to find a 8" easily as FR wiggles should not bother you that much ?
Michael
Last edited:
According to basic dipole theory, dipole distance equals the distance from the centre of the driver to its edge (flat baffle), for a circle it's simply the radius. However, as has already been pointed out, in practice there is not one single value for D, yet there are an infinite number of D's - to estimate the effect you'd have to calculate a quite difficult integral. Luckily Tolvan Edge is a nice little program that does all the dirty work for you! Home of the Edge
It seems quite accurate, as the results from the simulations are similar to my measurements.
Stig Erik's 10% to 15% seems to hold up very well when I simulate unbaffled drivers in Edge, for small as well as larger drivers. Therefore I don't think the size of the motor system has much to do with it. I think the size and shape of the backside structure mainly cause reflections and some kind of acoustic low-pass filter of the rear radiation.
BTW, I like your idea about using the Neo8 for the midrange, maybe even without a tweeter. That would solve many of the problems we are facing. I've been thinking along the same lines. It should be possible to use an efficient unbaffled 10" driver from about 100 hz to 1 khz. That would leave you with a very compact 'satellite' 3-way main dipole. Below 100 hz you could use the multiple (small) subs approach. Just some thoughts. High fidelity and high WAF.
Yep. But I've not found any 10 or 12 inch pro drivers with sufficient excursion to handle the subwoofer region. Fs on the lower cost 15 inch pro drivers I've looked at is too high to yield good dipole bass extension. So, not having rubber allergies
, I figure I'll start with the 830669 and see how that works. I'll be paying considerably more attention to Rms for the midwoofer, though.I do have digital crossover and equalization but I find one gets better results if the driver does its job well rather being patched up after the fact. There's of course some guesswork about how much smoothing's been applied to the datasheet SPLs but I tend to cringe looking at pro 8" drivers. "Wiggles" are 5dB; the 8W4P's not a bad example. With hi-fi 8" drivers many of them have significant diffraction around 600-700Hz. Maybe I'm being too fussy, but for the most part I'm guessing the pro drivers are targeted either at lower fidelity reproduction or at a frequency range narrower than 100Hz to 1kHz.
I was originally planning to use the 830884 for its nicely flat SPL but it's been discontinued. That's probably OK as it's not particularly low Rms and hence might not have been a good choice anyway.
Yep, I'm quite familiar with it.
It's the obvious solution for consistent directivity, especially the backwave; every small woofer I've looked at (and that's a lot!) has major problems to the rear (except possibly a heavily modified ScanSpeak Discovery D7608). Vertically the Neo8 gets a bit directional for tweeter use; my forum trawls show near unanimous agreement it's preferable to cross to a Neo3. BG has a longer throw version---the Neo8 S---that's supposedly going to be available this month along with the Neo10. I'm not finding application for the Neo10 in a dipole as it's too big to cross well to the Neo3 but the Neo8 S is of interest. From what I've read the regular Neo8 is good down to 600-700Hz in point source box applications and may become more efficient in dipole. So I'd hope to be able to cross it somewhat below 1kHz to get better integration between it and the "mid" woofer. Makes woofer selection easier, too, especially if a cross below under 700Hz is possible. I've been holding off on getting a pair of Neo8s to measure as I want to see specs on the Neo8 S first.
Unfortunately I seem to be allergic to remote subs, probably due to being fussy about linear phase.
Fortunately, I've no WAF issues to contend with in hanging 12 inch subs under the rest of the dipole. 100Hz is probably pushing a pro driver of typical excursion hard and the high Qms ones I'm finding don't have enough xmax to get that low.A decently flat SPL, low Rms 10 inch driver from ~100Hz to 1kHz would be nice. Unfortunately these two attributes appear to be mutually exclusive, though there are a few reasonable, lower cost options with Qms and Rms that's at least not worse than the 830669---notably the Eminence Delta 10A, though I'm suspicious of the smoothing. However, I'm not finding good options with a Qms much above 7 or Rms under 1.5. Suggestions anyone? The Legend 1058 could be interesting; Qms of 17.8 but no Rms or linear xmax spec.
Have a link handy? I'm not finding how to calculate it on a search.
Last edited:
Stig Erik - I'm missing your portrait !?
Michael
Haha. That was actually "Walter" (John Goodman) from the movie "The big Lebowski"....
Here are the formulas for calculating TSPs:
HAUPT
For the bass section I would ask for "free air" woofers in a car hifi forum. I'm sure there are similar drivers in the US as the Ciare I have shown above. The ones for sealed or ported enclosures have high Qms only because their diaphragms weight a ton.
HAUPT
For the bass section I would ask for "free air" woofers in a car hifi forum. I'm sure there are similar drivers in the US as the Ciare I have shown above. The ones for sealed or ported enclosures have high Qms only because their diaphragms weight a ton.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.