Again, it's strictly for finding a driver based on the XO slope order, so the means must match for best performance and the way your calculating one 'threw' me off, so to speak combined with my needing to do an example why I wouldn't use this pairing's 1st or 2nd iteration pairing(s) (my preference) combined with severely limited time to post........ but coming back to it later, did the equivalent of the way folks use complex XOs to splice seeming poor match-ups by using the 1st iteration of the HiVi and 2nd for FE that yields a plenty close enough one to your 500 Hz preference and use whichever polarity wiring works best overall.
IOW, need to rearrange the formulas to find the ideal VC dia., Fs that yields the desired match to whatever driver(s) you've chosen (or keep doing more iterations to see if one finally works).
For me it was always pretty simple since I always used cheap HF and some times mid, horn drivers or so called full range drivers, leaving only a woofer splice, which is where your formula comes into play (minus the mean calc).
IOW, need to rearrange the formulas to find the ideal VC dia., Fs that yields the desired match to whatever driver(s) you've chosen (or keep doing more iterations to see if one finally works).
For me it was always pretty simple since I always used cheap HF and some times mid, horn drivers or so called full range drivers, leaving only a woofer splice, which is where your formula comes into play (minus the mean calc).
Last edited:
Okay, so my understanding of what you're saying is the means must match, this being the individually calculated means such as the ones below:
I assume they don't need to match correctly, but be quite close?
But then you said
So I take it this mean is the crossover you'd use if you wanted to make the drivers work together
Sorry for the million questions, I'm just trying to make sure I understand correctly. I think the issue is me misunderstanding the terms you use and not knowing exactly the context you are using them for such as iterations.
I've been trying to find some information on this elsewhere but can't seem to, so sorry for all the questions.
I still don't quite understand why crossing the woofer lower (such as 300hz 2nd order, ignoring other factors) wouldn't solve the mismatch issue
I assume they don't need to match correctly, but be quite close?
But then you said
So I take it this mean is the crossover you'd use if you wanted to make the drivers work together
Sorry for the million questions, I'm just trying to make sure I understand correctly. I think the issue is me misunderstanding the terms you use and not knowing exactly the context you are using them for such as iterations.
I've been trying to find some information on this elsewhere but can't seem to, so sorry for all the questions.
I still don't quite understand why crossing the woofer lower (such as 300hz 2nd order, ignoring other factors) wouldn't solve the mismatch issue
Last edited:
I do quite like the look of them the more I look at them. They seem to have a smooth and a very nice price (also the 6" version).
Once I understand GM's approach to driver matching, then I'll decide what to do as I'd like to do everything to make the drivers match as best as possible
Ideally, yes, but can be off some depending on XO point due to our hearing acuity varies with frequency.
Yes, if using this driver combo.
Understood.
TTBOMK you won't per se; I had to derive it from studying many of the pioneer's technical papers, patents, get hints from Altec Engineers, a 'crusty' old EE that had apprenticed at W.E. during cinema (pro) sound's early development and last, but not least, was how the mid 1930s W.E./Altec 755 three way 'full range' driver was designed and TTBOMK yet to be superseded WRT blending a 3 way (mechanical XOs!), matching polar response WRT holding a time/phase coherent pattern over a long distance.
Regardless, the folks that understand all about XO design theory should de facto know it, but apparently choose to concentrate on driver choice by other means and 'stitch' them together even if they otherwise match up like one's a Bull, another a Lamb and topped off with a Bird and/or Bat for top end 'air'. 😉 When they get it right, they're spectacular performers, but way too time consuming, costly for this 'jack o' trades, master of none'. Most IME though were 'meh' performing overall. I was raised on live music, Cinema Palace sound systems, so anything less, is well, less, i.e. the Memorex ad: 'Is it live or is it Memorex?' and Cerwin Vegas' mantra 'Loud is Beautiful if it's Clean'
Well, for one we want as much overlap as practical for 1st order and also we want the big driver to handle the large WLs that otherwise can modulate the smaller driver's output (harmonic distortion, among others) at higher SPL, especially if the source has a high dynamic headroom, but fail to see how it solves the polar response mismatch except in the nearfield (< ~1m/4ft).
Adding considerable room treatment will help, but the acoustic difference is great enough that you'll likely need a special XO rather than basically textbook w/reversed polarity when matched 'good enough', i.e. measure in room and design in one of the programs I never even imagined/dreamed of in my DIY lifetime.
Last edited:
GM, thanks for that detailed response. It's helped to clarify this for me and makes me remember I definitely need to study driver behaviour more.
Where do you stand on matching cone diameter, as well as voice coil diameter? I guess a perfect match would be two identical fullrange drivers that have a 1st order crossover or one just rolled off at a few hundred hz. But then we are running into the size vs dispersion issue. I am not sure how much this is an issue for me as I have only have experience with two small fullrange drivers - the 7ms and Pluvia 7.2HD.
Maybe something like two Markaudio CHN-110 working together would work as a sweet spot (they seem to hold a good position for size, performance and cost). They don't seem to publish their VC diameter, but I assume this would also marry up to a larger woofer better if required. Using the values 20mm and 25mm for VC diameter, the mean would be 490hz or 439hz.
Out of curiosity, are you able to share any images of your current system or any speakers you have made in the past? I am curious where you stand in regards to the driver diameters, baffle diamters, etc
Maybe the 3FE22 is my bat. I just wanted the best possible dispersion from a wide band driver. I'll have to think about what I'll do with them. I think the high FS makes the mean matching method fail a lot, and to make something match with this doesn't make sense as then the woofer would be so small with a high FS.
Where do you stand on matching cone diameter, as well as voice coil diameter? I guess a perfect match would be two identical fullrange drivers that have a 1st order crossover or one just rolled off at a few hundred hz. But then we are running into the size vs dispersion issue. I am not sure how much this is an issue for me as I have only have experience with two small fullrange drivers - the 7ms and Pluvia 7.2HD.
Maybe something like two Markaudio CHN-110 working together would work as a sweet spot (they seem to hold a good position for size, performance and cost). They don't seem to publish their VC diameter, but I assume this would also marry up to a larger woofer better if required. Using the values 20mm and 25mm for VC diameter, the mean would be 490hz or 439hz.
Out of curiosity, are you able to share any images of your current system or any speakers you have made in the past? I am curious where you stand in regards to the driver diameters, baffle diamters, etc
You're welcome!
Hmm, still misinterpreting this design routine in that we haven't addressed driver size, which is a function of desired BW, just seeing how well (or not) the ones you've chosen match up.
Multi-way cone dome for coworkers, relatives, neighbors, none and just a few proofs of various concepts I sold or gave away when I'd learned all I could from them; as I said, mostly used cheap horns except for woofer since folks usually could only afford small used tube amps/receivers, so needed relatively high efficiency.
Was never any need for pics of my horn systems to the point where Christmas, etc., pics were framed to keep as much of them hidden as practical for insurance purposes because even though mostly DIY, the cabs were very costly material wise (very large horn cabs made with 3/4" no void marine grade plywood, high quality, warp free lumber) and component cost is well into the thousands nowadays.
What I have left is just my BP6 corner bass horns sans the horns for stereo with some stock Altec 500 Hz horns perched on top and even then need to have new diaphragms installed, but my hands are too arthritic to do it. In the attached I'm about 5'8" circa 2000. The yellow paint was just some left over to seal the wood; as originally installed, all that was visible as a subwoofer was its matte black backside stuck in the corners behind full room height grills (horn system from dim memory was ~ 7 ft tall IIRC), which are currently tuned to 14 Hz. This is the current iteration of the woofers.
As for performance..........
Regardless, more to come re theoretical ideal driver sizes as time permits. Re baffle sizing, there's probably weeks (months?) of reading already posted by many of us here and other DIY forums over the decades to research, so time to do a bit of 'due diligence'.
Hmm, still misinterpreting this design routine in that we haven't addressed driver size, which is a function of desired BW, just seeing how well (or not) the ones you've chosen match up.
Multi-way cone dome for coworkers, relatives, neighbors, none and just a few proofs of various concepts I sold or gave away when I'd learned all I could from them; as I said, mostly used cheap horns except for woofer since folks usually could only afford small used tube amps/receivers, so needed relatively high efficiency.
Was never any need for pics of my horn systems to the point where Christmas, etc., pics were framed to keep as much of them hidden as practical for insurance purposes because even though mostly DIY, the cabs were very costly material wise (very large horn cabs made with 3/4" no void marine grade plywood, high quality, warp free lumber) and component cost is well into the thousands nowadays.
What I have left is just my BP6 corner bass horns sans the horns for stereo with some stock Altec 500 Hz horns perched on top and even then need to have new diaphragms installed, but my hands are too arthritic to do it. In the attached I'm about 5'8" circa 2000. The yellow paint was just some left over to seal the wood; as originally installed, all that was visible as a subwoofer was its matte black backside stuck in the corners behind full room height grills (horn system from dim memory was ~ 7 ft tall IIRC), which are currently tuned to 14 Hz. This is the current iteration of the woofers.
As for performance..........
Regardless, more to come re theoretical ideal driver sizes as time permits. Re baffle sizing, there's probably weeks (months?) of reading already posted by many of us here and other DIY forums over the decades to research, so time to do a bit of 'due diligence'.
Attachments
Interesting. I've never heard horn speakers (not counting those little horn tweeters).
That's one very big speaker, thanks for sharing. I'd love some big woofers like that one day. Getting a bit off topic now, but it's good to see where other people have been taken on their journey.
I've read all sorts with people preferring both big and small. But I'll do some proper reading this weekend if I get a chance.
I'm going to try to go away and come back to this thread once I've done more research and have a better proposition for a speaker.
Thanks for all your help so far.
Re choosing speakers; sticking with one of the pioneer's most famous driver designs, they were primarily interested in the AM (250-2500 Hz) and later FM (300-3000 Hz) radio BWs at high speech intelligibility over long distances, (Public Address apps), so for the W.E. 755A.........
VC dia = 34400/pi/2500 = 4.38 cm/1.72" rounded to 1.75"
Fs = 250/2^2 = 62 Hz
mean = (250*2500)^0.5 = 790.57 Hz
diaphragm dia = 34400/pi/790.57 = 13.85 cm/5.45" = existing 8" frame apparently due to (originally) using leather for the surround and later was made as a single molded diaphragm cum dustcap, surround, mechanical XOs with 2" VC = ~2155 Hz/734 Hz mean, 14.92 cm/5.87" to fit existing frame.
If these AN Super 8 specs are reasonably accurate (they didn't use to be), then a good HE mids driver choice for both BW's or to at least use for comparison.
VC dia = 34400/pi/2500 = 4.38 cm/1.72" rounded to 1.75"
Fs = 250/2^2 = 62 Hz
mean = (250*2500)^0.5 = 790.57 Hz
diaphragm dia = 34400/pi/790.57 = 13.85 cm/5.45" = existing 8" frame apparently due to (originally) using leather for the surround and later was made as a single molded diaphragm cum dustcap, surround, mechanical XOs with 2" VC = ~2155 Hz/734 Hz mean, 14.92 cm/5.87" to fit existing frame.
If these AN Super 8 specs are reasonably accurate (they didn't use to be), then a good HE mids driver choice for both BW's or to at least use for comparison.
Interesting. Thanks for providing that example. AN seem very modest - "Audio Nirvanas are the best sounding speakers in the world--at any price"
For my driver choice I'm now considering just two Mark Audio drivers such as CHN110 + Alpair 11ms or 2x CHN 110 as I believe they have the same VC diameter (MA doesn't publish VC Diameter, but one site seems to have all the VC diameters listed, not sure how accurate it is. I emailed MA but they won't tell me unfortunately).
Acording to toutlehautparleur, all the "11" drivers have 30mm VCs and also the 12s. I know they will have similar cone profiles too, so similar polar response. I wouldn't mind a 12 such as CHR120 or the Alpair 12pw for the bigger SD.
I guess my question to you is, in your opinion, how different can the means actually be. Ideally I want the Midtweet to have the smallest diameter possible for improved dispersion up top, but the VCs get smaller once you step down from the 11 ranges. Here is some calculations using the VC Diameters I got from toutlehautparleur (so I can't say if they are 100% accurate):
Alpair 12pw - VC diameter 30mm, FS 38.4
34400/PI/VCD = 3649.9
2^4 = 228.1215517
Mean = 374.4
CHR 120 - VC diameter 30mm, FS 34.2
Basically the same as above. Mean = 353.3 instead
CHN110 - VC diameter 30mm, FS 44
Mean = 400.7
11ms - VC diameter 30mm, FS 43.7
Mean = 399.3
The smaller drivers I'd consider are the Pluvia 7.2HD and the CHP90. The CHP90 has unknown VC diameter (I'm assuming around 25mm for my calculations) and I am confident the Pluvia 7.2HD is either 19mm or 20mm.
Pluvia 7.2HD - VC diameter 20mm, FS 72.5
34400/PI/VCD = 5474.91724
2^4 = 342.1823275
Mean = 630
CHP90 - VC Diameter 25mm, FS 48.5
34400/PI/VCD = 5474.91724
2^4 = 342.1823275
Mean = 515.3
I am particularly interested in your opinion on mixing the CHP90 with the Alpair 12pw. Due to my lack of understanding, I've been going round in circles for days now so it would be good to hear what you think. I'm thinking of a 2nd order XO around 300-400hz (depends how much the parts for the XO will cost).
Thanks
Would the mean not be (Fs*2500)^0.5 = (62*2500)^0.5 = 393.7 - I believe this is how we calculated it earlier and that's how I've been calculating it this whole time.
For my driver choice I'm now considering just two Mark Audio drivers such as CHN110 + Alpair 11ms or 2x CHN 110 as I believe they have the same VC diameter (MA doesn't publish VC Diameter, but one site seems to have all the VC diameters listed, not sure how accurate it is. I emailed MA but they won't tell me unfortunately).
Acording to toutlehautparleur, all the "11" drivers have 30mm VCs and also the 12s. I know they will have similar cone profiles too, so similar polar response. I wouldn't mind a 12 such as CHR120 or the Alpair 12pw for the bigger SD.
I guess my question to you is, in your opinion, how different can the means actually be. Ideally I want the Midtweet to have the smallest diameter possible for improved dispersion up top, but the VCs get smaller once you step down from the 11 ranges. Here is some calculations using the VC Diameters I got from toutlehautparleur (so I can't say if they are 100% accurate):
Alpair 12pw - VC diameter 30mm, FS 38.4
34400/PI/VCD = 3649.9
2^4 = 228.1215517
Mean = 374.4
CHR 120 - VC diameter 30mm, FS 34.2
Basically the same as above. Mean = 353.3 instead
CHN110 - VC diameter 30mm, FS 44
Mean = 400.7
11ms - VC diameter 30mm, FS 43.7
Mean = 399.3
The smaller drivers I'd consider are the Pluvia 7.2HD and the CHP90. The CHP90 has unknown VC diameter (I'm assuming around 25mm for my calculations) and I am confident the Pluvia 7.2HD is either 19mm or 20mm.
Pluvia 7.2HD - VC diameter 20mm, FS 72.5
34400/PI/VCD = 5474.91724
2^4 = 342.1823275
Mean = 630
CHP90 - VC Diameter 25mm, FS 48.5
34400/PI/VCD = 5474.91724
2^4 = 342.1823275
Mean = 515.3
I am particularly interested in your opinion on mixing the CHP90 with the Alpair 12pw. Due to my lack of understanding, I've been going round in circles for days now so it would be good to hear what you think. I'm thinking of a 2nd order XO around 300-400hz (depends how much the parts for the XO will cost).
Thanks
That example was for choosing a driver size based on a specific BW such as a mid bass, mid, etc., in a multiway Vs the other that's for finding theoretically optimal XO points for a given driver.
Again, folks splice together all manner of driver types, sizes and use measurements to find optimal XO points and/or make as elaborate a dividing network as required whereas the pioneer's way was to polar match for highest speech intelligibility over HxWxD area and worked backwards to find/design the requisite sound system specs and in their case made/used whatever/existing was required; so WRT 'how close', all you can do is what folks do using measurements, which in essence is finding the 'mean of the means'
.
No clue ATM, will have to research them as time permits.
Again, folks splice together all manner of driver types, sizes and use measurements to find optimal XO points and/or make as elaborate a dividing network as required whereas the pioneer's way was to polar match for highest speech intelligibility over HxWxD area and worked backwards to find/design the requisite sound system specs and in their case made/used whatever/existing was required; so WRT 'how close', all you can do is what folks do using measurements, which in essence is finding the 'mean of the means'
.No clue ATM, will have to research them as time permits.
Alpair 12pw - VC diameter 30mm, FS 38.4
34400/PI/VCD = 3649.9
2^4 = 228.1215517
Mean = 374.4
CHP90 - VC Diameter 25mm, FS 48.5
34400/PI/VCD = 5474.91724
2^4 = 342.1823275
Mean = 515.3
I am particularly interested in your opinion on mixing the CHP90 with the Alpair 12pw. Due to my lack of understanding, I've been going round in circles for days now so it would be good to hear what you think. I'm thinking of a 2nd order XO around 300-400hz (depends how much the parts for the XO will cost).
WRT 'how close', all you can do is what folks do using measurements, which in essence is finding the 'mean of the means', so (374.4*515.3)^0.5 = ~439.2 Hz/1st order.
For 2nd order 12pw: 38.4*2^2 = 153.6 Hz
(153.6*3649.9)^0.5 = 748.75 Hz mean
CHP90: 48.5*2^2 = 194 Hz
(194*5474.91724)^0.5 = 1030.6 Hz
(748.75*1030.6)^0.5 = 447.1 Hz/2nd order
While the few I did using RS drivers performed very well to me and the folks I did them for, we had precious few to compare to and super expensive (McIntosh and I forget the brand/model of the huge bunch of cubes stacked in a curve); hopefully someone that's done a multiway design using measurements, polar response plots Vs just calculating them, how close would they be and whether or not that at least on paper might be superior in some ways and/or just overall......
PS: please double check the math as I did this under a higher than normal duress.
PS: please double check the math as I did this under a higher than normal duress.
Thanks GM. The maths checks out, based on your calculations.
I have the 12pw sitting here on the floor so I think I'll just finally use them. For some reason I keep trying to think of projects using other stuff. I'll build a ported box for them and then try active XOs for now to the Pluvia 7.2HD and the 3FE22 since I have all this stuff and it won't cost much to experiment. Just building the boxes for the 12pw and the 3FE22. The final goal would be moving to passive
I'll report back once I have an update 🙂