Hi there! Though I have been reading here for some time, this is my first thread, and will be my first real speaker project. I look forward to your thoughts and suggestions as things progress.
As the title suggests I am looking to build some classic big box 3 way speakers. I like a fairly big sound (though it may not be so big to some of you) and while I would probably be very happy with a pair of Klipsch Horns or La Scalas, I just can't afford that - so a good alternative for me is to build my own.
The design criteria for these speakers I have decided are thus:
* 3 way design, utilizing a woofer, mid, and tweeter.
* Passive crossover with minimal insertion loss
* Fairly high efficiency (92~95 db)
* Fairly high power handling (~200 watts)
* Fairly high max SPL (~120 db)
* Fairly flat frequency response (+-3db)
* As close to 20-20KHz as I can realistically get, holding within other parameters
* Fairly low distortion
* Attractive cabinet (WAF!)
As the 'realistically' and all the 'fairly's' suggest I have no intention of being fanatical about any of it, but I do want to give it a good effort and build some speakers I can enjoy listening to and looking at for a long time.
Of course I have been considering this for a long time and have already done a lot of comparisons and so have chosen and acquired the drivers.
For the woofers I have acquired a set of JBL 2226H. These will require a fairly large cabinet, but as I kind of had in mind something rather imposing anyway, I'm quite happy with that.
For the tweeters I bought a pair of the venerable Selenium D220Ti, mated to Selenium HC23-25 horns. These seem to have a very good reputation, very clear, not harsh, with low distortion.
For the mids I have chosen a pair of B&C 6MD38. These should fill in the gap nicely between the 2226H's and the D220Ti's, giving the latter some breathing room for higher power handling.
I'll continue to post as things progress.
As the title suggests I am looking to build some classic big box 3 way speakers. I like a fairly big sound (though it may not be so big to some of you) and while I would probably be very happy with a pair of Klipsch Horns or La Scalas, I just can't afford that - so a good alternative for me is to build my own.
The design criteria for these speakers I have decided are thus:
* 3 way design, utilizing a woofer, mid, and tweeter.
* Passive crossover with minimal insertion loss
* Fairly high efficiency (92~95 db)
* Fairly high power handling (~200 watts)
* Fairly high max SPL (~120 db)
* Fairly flat frequency response (+-3db)
* As close to 20-20KHz as I can realistically get, holding within other parameters
* Fairly low distortion
* Attractive cabinet (WAF!)
As the 'realistically' and all the 'fairly's' suggest I have no intention of being fanatical about any of it, but I do want to give it a good effort and build some speakers I can enjoy listening to and looking at for a long time.
Of course I have been considering this for a long time and have already done a lot of comparisons and so have chosen and acquired the drivers.
For the woofers I have acquired a set of JBL 2226H. These will require a fairly large cabinet, but as I kind of had in mind something rather imposing anyway, I'm quite happy with that.
For the tweeters I bought a pair of the venerable Selenium D220Ti, mated to Selenium HC23-25 horns. These seem to have a very good reputation, very clear, not harsh, with low distortion.
For the mids I have chosen a pair of B&C 6MD38. These should fill in the gap nicely between the 2226H's and the D220Ti's, giving the latter some breathing room for higher power handling.
I'll continue to post as things progress.
Hi,
3 way speakers are notoriously difficult to design properly.
The "classic" method is to select 3 drivers and sling them together.
It usually doesn't work very well because too much is ignored.
You need a high quality methodology to get a high quality result.
The links below contain a wealth of information.
Troels has quite a lot of info on JBL drivers.
Zaph has measurements of your midrange.
rgds, sreten.
undefinition (see FAQs)
Zaph|Audio
FRD Consortium tools guide
RJB Audio Projects
Speaker Design Works
HTGuide Forum - A Guide to HTguide.com Completed Speaker Designs.
Humble Homemade Hifi
DIY Loudspeaker Projects Troels Gravesen
Quarter Wavelength Loudspeaker Design
The Frugal-Horns Site -- High Performance, Low Cost DIY Horn Designs
Linkwitz Lab - Loudspeaker Design
Music and Design
3 way speakers are notoriously difficult to design properly.
The "classic" method is to select 3 drivers and sling them together.
It usually doesn't work very well because too much is ignored.
You need a high quality methodology to get a high quality result.
The links below contain a wealth of information.
Troels has quite a lot of info on JBL drivers.
Zaph has measurements of your midrange.
rgds, sreten.
undefinition (see FAQs)
Zaph|Audio
FRD Consortium tools guide
RJB Audio Projects
Speaker Design Works
HTGuide Forum - A Guide to HTguide.com Completed Speaker Designs.
Humble Homemade Hifi
DIY Loudspeaker Projects Troels Gravesen
Quarter Wavelength Loudspeaker Design
The Frugal-Horns Site -- High Performance, Low Cost DIY Horn Designs
Linkwitz Lab - Loudspeaker Design
Music and Design
Thanks guys. I appreciate it.
Lowpoke I'll be sure to have plenty of pictures up as I go. I'm sure to have some posts about the drivers pretty quick - I just bought a WT3 and will be posting the measurements I get. The cabinet I only have a vague idea of because there are just too many variables at this point.
Thanks for the links Sreten! Some of those I've seen before but most I haven't and a couple are exactly what I've been thinking I need! As many questions as I have about the cabinet, I'm pretty much dreading trying to figure out the crossover.
But anywho, when in doubt, onwards and upwards!, I always say.
Lowpoke I'll be sure to have plenty of pictures up as I go. I'm sure to have some posts about the drivers pretty quick - I just bought a WT3 and will be posting the measurements I get. The cabinet I only have a vague idea of because there are just too many variables at this point.
Thanks for the links Sreten! Some of those I've seen before but most I haven't and a couple are exactly what I've been thinking I need! As many questions as I have about the cabinet, I'm pretty much dreading trying to figure out the crossover.
But anywho, when in doubt, onwards and upwards!, I always say.
btw, I just recently finished this low budget 4way
different approach
but I rarely play loud
and if I do, its really not very loud
funny thing
when my sister comes visiting she always says; why do you play so loud
and thats when its not loud
well, what is realistic SPL
at a disco ?
or an intimate concert ?
is it a blasting trumpet ?
a jazz singer ?
rock concert ?
Attachments
if your max SPL goal really is 120db as you suggets, maybe better go active
hard to avoid increased distortion at loud SPL
even if very very small, its still audible
less clean you might say
even if driver do not distort, the ultimate exstreme precision and control suffers
you might get disco sound rather than smooth hifi
hard to avoid increased distortion at loud SPL
even if very very small, its still audible
less clean you might say
even if driver do not distort, the ultimate exstreme precision and control suffers
you might get disco sound rather than smooth hifi
yeah, I know
mid moves too much when loud
but those big caps are bloddy expencive
but if I wanted to play loud, yes it should have a series cap, and a paralel inductor
and it will probably screw up the present filter
but yeah, I have some bipolars I have to try
and a zobel on woofer
but I'm sure it will cause a lot of trouble
I will try it when I have some free time
but right now I just enjoy the music
but to be honest, I'm done with audiophile listening
instead the music is pounding many many hours every day
Thanks for the suggestion Tinitus.
I may well build an active system in the future, but for this project my real interest is a 'Classic 3 way' (yes, the ubiquitous 'monkey coffins' of the 60's, 70's, and 80's) so will undoubtedly stay passive at this time (besides, I really can't afford the extra amps at this time). Yes, that means I will be stuck with the limitations of the form - but I also mean to learn a lot about those limitations.
Speaking of which - what do you know (or anyone else?) of the mechanism of the high SPL distortions that an active crossover resolves? Do these distortions originate in the passive crossover then? What in the crossover causes them?
I may well build an active system in the future, but for this project my real interest is a 'Classic 3 way' (yes, the ubiquitous 'monkey coffins' of the 60's, 70's, and 80's) so will undoubtedly stay passive at this time (besides, I really can't afford the extra amps at this time). Yes, that means I will be stuck with the limitations of the form - but I also mean to learn a lot about those limitations.
Speaking of which - what do you know (or anyone else?) of the mechanism of the high SPL distortions that an active crossover resolves? Do these distortions originate in the passive crossover then? What in the crossover causes them?
If an amp distorts a little, some of the distortion components will play through the tweeter etc. and if you put the tweeter on a separate amp, the bass amp can distort a lot while the tweeter plays clear, and distortion isn't as obvious or annoying in the bass.
Passive crossovers do not cause problems when done properly, just as active systems won't (when done properly).
Re: distortion in active systems. I have a first edition of Martin Colloms classic "High Performance Loudspeakers" from 1978 and on the subject of active crossovers the first point he makes is a reduction in the intermodulation distortion in the power amps due to the reduced bandwidth they are asked to handle.
This may explain the common reaction that active systems sound cleaner or "louder" than their passive counterparts. (Btw I put louder in quotation marks! I am just reporting comments......don't want to debate it...ha ha.)
The subjective improvement often surpasses expectation.
This may explain the common reaction that active systems sound cleaner or "louder" than their passive counterparts. (Btw I put louder in quotation marks! I am just reporting comments......don't want to debate it...ha ha.)
The subjective improvement often surpasses expectation.
you plan to use a compression driver
its well known that you can lower its xo point if you don't play too loud
or use shallow filter slopes
if you want to play loud you have to use either higher xo point, or steeper filter slopes
or both
notice that power rating of a CD is sometimes listed with either 12db, 24db, or both
its well known that you can lower its xo point if you don't play too loud
or use shallow filter slopes
if you want to play loud you have to use either higher xo point, or steeper filter slopes
or both
notice that power rating of a CD is sometimes listed with either 12db, 24db, or both
Hi,
The point of an active system is that in the case of a 3 way the peak
transient output far exceeds the total rms output, by about 10dB,
this is why they sound cleaner, and no other particular reason.
FWIW the c/o makes or breaks a speaker, and done well, using at least
proper simulation tools (in the previous links) gives you a flying chance.
Do not, under any circumstances, think you can get away with a cookbook
c/o or any of the oversimplified approaches to c/o design expounded.
rgds, sreten.
The point of an active system is that in the case of a 3 way the peak
transient output far exceeds the total rms output, by about 10dB,
this is why they sound cleaner, and no other particular reason.
FWIW the c/o makes or breaks a speaker, and done well, using at least
proper simulation tools (in the previous links) gives you a flying chance.
Do not, under any circumstances, think you can get away with a cookbook
c/o or any of the oversimplified approaches to c/o design expounded.
rgds, sreten.
Last edited:
Ah. Yes, that all makes sense.
In consideration of something like that one of the reasons I want such high efficiency and high power handling isn't (just) to blow out my neighbors. The amp I'm currently using is an old Adcom 5500, which will put out 200 watts @ 8 ohms, but while I probably will listen to it on occasion at higher levels, for the most part I hope all this over-building will result in taxing the system as little as possible at 'normal' listening levels.
Then again it may make very little difference, if any - but again again I am forced to consider the limitation of the format, and try to plan accordingly, as best I can.
I have given a lot of thought to the crossover, even though I haven't actually committed to a design yet because of what you both Tinitus and Streten mention. I have collected some software packages and have some ideas, but we shall have to wait until the right time to discuss in depth, methinks.
Lowpoke asked for a picture, so...
The one 2226H is still in the stand I threw together to 'burn it in'. I'll have to tell about that in another post.
In consideration of something like that one of the reasons I want such high efficiency and high power handling isn't (just) to blow out my neighbors. The amp I'm currently using is an old Adcom 5500, which will put out 200 watts @ 8 ohms, but while I probably will listen to it on occasion at higher levels, for the most part I hope all this over-building will result in taxing the system as little as possible at 'normal' listening levels.
Then again it may make very little difference, if any - but again again I am forced to consider the limitation of the format, and try to plan accordingly, as best I can.
I have given a lot of thought to the crossover, even though I haven't actually committed to a design yet because of what you both Tinitus and Streten mention. I have collected some software packages and have some ideas, but we shall have to wait until the right time to discuss in depth, methinks.
Lowpoke asked for a picture, so...
The one 2226H is still in the stand I threw together to 'burn it in'. I'll have to tell about that in another post.
This effect (not sounding taxed) will be helped greatly by your choice of woofer and compression tweeter regardless of your amp power. The open backed cabinet might work against though. FWIW even though you hinted you don't need to go there, I doubt you'd get 120 with a single driver on an open baffle when fully EQed.
dont let it scare you
with those drivers it shouldn't be too difficult to make it play ok
spectacular, no ofcourse not in the first try
toss something simple together and give it a listen
I suggest to do that first
then do the simulations with more advanced filters
wouldn't it be nice to have them playing a bit in the background, while fiddling with sims
I can promisse you, even if not perfect, you will enjoy it
and when you have listened to simpler filters, you may better hear clearly what goes wrong when more componets are added
you will need more components, and thats the tricky part
a real big grand master of crossovers once 'admitted' to me personally that for every component you can remove without compromising the function, the sound will improve
that said, he was known to use bloddy many components, and more than usual, and thus a realtive term, still holds true
I suggest to have fun, and learn
such takes time
with those drivers it shouldn't be too difficult to make it play ok
spectacular, no ofcourse not in the first try
toss something simple together and give it a listen
I suggest to do that first
then do the simulations with more advanced filters
wouldn't it be nice to have them playing a bit in the background, while fiddling with sims
I can promisse you, even if not perfect, you will enjoy it
and when you have listened to simpler filters, you may better hear clearly what goes wrong when more componets are added
you will need more components, and thats the tricky part
a real big grand master of crossovers once 'admitted' to me personally that for every component you can remove without compromising the function, the sound will improve
that said, he was known to use bloddy many components, and more than usual, and thus a realtive term, still holds true
I suggest to have fun, and learn
such takes time
I don't think this implies that you must have few components, just that if components can be combined to achieve an identical function, then they probably should... and remove them if they aren't doing anything significantly positive. Eg, I can't see any harm in adding an impedance resonance peak filter (damped flat) if it is required .
@Lafeeric, not to derail anything but if you happen to get the itch again a ways down the track, it seems that you have almost all the parts needed for an econowave.
No no Allen, that simple baffle/stand isn't anything permanent. My 'Classic 3 way Hi-Fi' will definitely be a vented enclosure.
As I mentioned above, my budget is somewhat restricted. 'So why have you spent nearly a thousand dollars on drivers?' you may ask. The truth is I didn't. The D220Ti's and the 6MD38's are new, but the 2226H's I got off fleabay.
Yes, I well know the dangers of buying anything off ebay, and especially the dangers of buying used drivers. I did my best though, paying attention to descriptions, photos, feedback numbers and price, and I think I did all right.
I did, however, get two drivers of rather diverse condition. This is one reason I have invested in the WT3. One 2226H looks quite used, and the other looks brand new. Of course that was also a factor in choosing the 2226H - they have a reputation for being nearly indestructible.
Measuring them using the WT3 added mass method I got the following numbers:
(My goodness - this board sure isn't table friendly!)
2226H 1
SPL = 97
R(e) = 4.868
F(s) = 55.18
Q(ts) = 0.435
Q(es) = 0.504
Q(ms) = 3.183
L(e) = 2.33
M(ms) = 93.53
V(as) = 3.439
2226H 2
SPL = 96.41
R(e) = 4.744
F(s) = 46.43
Q(ts) = 0.406
Q(es) = 0.485
Q(ms) = 2.464
L(e) = 2.171
M(ms) = 93.78
V(as) = 4.844
JBL published 2226H specs
SPL = 97
R(e) = 5
F(s) = 40
Q(ts) = 0.31
Q(es) = 0.33
Q(ms) = 5
L(e) = 1.75
M(ms) = 98
V(as) = 6.2
After looking at my results and comparing them to the published specs I deduced that '2226H 1' could possibly use some 'breaking in', so I built the baffle/stand you see above, set them to about 1cm excursion at 60 Hz and let them run about 6 hours. The next day I tested it again:
2226H 1, 2nd test
SPL = 96.77
R(e) = 4.913
F(s) = 51.81
Q(ts) = 0.419
Q(es) = 0.483
Q(ms) = 3.172
L(e) = 2.353
M(ms) = 96.84
V(as) = 3.768
This looked somewhat better to me, so I ran it again, then tested again the next day:
2226H 1, 3rd test
SPL = 96.3
R(e) = 4.932
F(s) = 47.78
Q(ts) = 0.4
Q(es) = 0.451
Q(ms) = 3.511
L(e) = 2.389
M(ms) = 106.6
V(as) = 4.025
I feel that this is now close enough for me to load some numbers into Win ISD Pro and design and build my first actual test box.
As I mentioned above, my budget is somewhat restricted. 'So why have you spent nearly a thousand dollars on drivers?' you may ask. The truth is I didn't. The D220Ti's and the 6MD38's are new, but the 2226H's I got off fleabay.
Yes, I well know the dangers of buying anything off ebay, and especially the dangers of buying used drivers. I did my best though, paying attention to descriptions, photos, feedback numbers and price, and I think I did all right.
I did, however, get two drivers of rather diverse condition. This is one reason I have invested in the WT3. One 2226H looks quite used, and the other looks brand new. Of course that was also a factor in choosing the 2226H - they have a reputation for being nearly indestructible.
Measuring them using the WT3 added mass method I got the following numbers:
(My goodness - this board sure isn't table friendly!)
2226H 1
SPL = 97
R(e) = 4.868
F(s) = 55.18
Q(ts) = 0.435
Q(es) = 0.504
Q(ms) = 3.183
L(e) = 2.33
M(ms) = 93.53
V(as) = 3.439
2226H 2
SPL = 96.41
R(e) = 4.744
F(s) = 46.43
Q(ts) = 0.406
Q(es) = 0.485
Q(ms) = 2.464
L(e) = 2.171
M(ms) = 93.78
V(as) = 4.844
JBL published 2226H specs
SPL = 97
R(e) = 5
F(s) = 40
Q(ts) = 0.31
Q(es) = 0.33
Q(ms) = 5
L(e) = 1.75
M(ms) = 98
V(as) = 6.2
After looking at my results and comparing them to the published specs I deduced that '2226H 1' could possibly use some 'breaking in', so I built the baffle/stand you see above, set them to about 1cm excursion at 60 Hz and let them run about 6 hours. The next day I tested it again:
2226H 1, 2nd test
SPL = 96.77
R(e) = 4.913
F(s) = 51.81
Q(ts) = 0.419
Q(es) = 0.483
Q(ms) = 3.172
L(e) = 2.353
M(ms) = 96.84
V(as) = 3.768
This looked somewhat better to me, so I ran it again, then tested again the next day:
2226H 1, 3rd test
SPL = 96.3
R(e) = 4.932
F(s) = 47.78
Q(ts) = 0.4
Q(es) = 0.451
Q(ms) = 3.511
L(e) = 2.389
M(ms) = 106.6
V(as) = 4.025
I feel that this is now close enough for me to load some numbers into Win ISD Pro and design and build my first actual test box.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.