Dear Sirs,
I would like to try to build a 3 way speaker able to reach down 50 Hz flat.
I have some preliminary doubts anyway:
1) is it better to choose a cone mid or a dom mid?
2) any suggestion of a really great woofer? it should be reproduce the 50 Hz very strongly (read at least 95 dB/2 meters).
Thank you very much indeed.
Kind regards,
beppe61
I would like to try to build a 3 way speaker able to reach down 50 Hz flat.
I have some preliminary doubts anyway:
1) is it better to choose a cone mid or a dom mid?
2) any suggestion of a really great woofer? it should be reproduce the 50 Hz very strongly (read at least 95 dB/2 meters).
Thank you very much indeed.
Kind regards,
beppe61
Sounds like you may want PA drivers.
What's the budget?
How big do you want your boxes?
Dome mids are not as effeicient as a good PA mid.
What sort of tweeters do you like?
There are so many things to consider that you have to give us some more info before we can really help. Tell us about what you are trying to achieve.
What's the budget?
How big do you want your boxes?
Dome mids are not as effeicient as a good PA mid.
What sort of tweeters do you like?
There are so many things to consider that you have to give us some more info before we can really help. Tell us about what you are trying to achieve.
Perhaps a 2-way with sub? This is still a 3-way but allows you to use a wide range cone mid to cover the majority of the spectrum and crossover to a low-reaching woofer at a lower point than you could with a dome mid.
I really don't make a big distinction between a 3-way and a 2-way with sub - it's just semantics - you're using three drivers so it's a 3-way. You can have one sub unit per speaker, even built in to the main cabinets so it's all one speaker (though many like to isolate the bass from the rest).
For the efficency you are looking for the PA idea may still be the way to go, but I can't make any specific recommendations.
I really don't make a big distinction between a 3-way and a 2-way with sub - it's just semantics - you're using three drivers so it's a 3-way. You can have one sub unit per speaker, even built in to the main cabinets so it's all one speaker (though many like to isolate the bass from the rest).
For the efficency you are looking for the PA idea may still be the way to go, but I can't make any specific recommendations.
Dear Mr. Weldon,
please excuse my belated reply but I took time to clear my mind.
I strongly think this discussion will help me very much.
First of all I would like to say that I have thought about a 3 way system beacuse I do not think that is possible to reproduce accurately the range 50-20.000 with just 2 drivers.
Am I wrong ?
>
My general requirements are:
1) let's say frequency response 50-20.000 Hz +-3dB
2) good efficiency and easy load for the amp
3) good soundstage reproduction in the depth dimension
But for instance I do not know if it is better to use a cone or a dome for the midrange.
So the first step would be to select suitable drivers for the project.
Thank you very much again.
Kind regards,
beppe61
please excuse my belated reply but I took time to clear my mind.
I strongly think this discussion will help me very much.
First of all I would like to say that I have thought about a 3 way system beacuse I do not think that is possible to reproduce accurately the range 50-20.000 with just 2 drivers.
Am I wrong ?
>
Cal Weldon said:
My general requirements are:
1) let's say frequency response 50-20.000 Hz +-3dB
2) good efficiency and easy load for the amp
3) good soundstage reproduction in the depth dimension
But for instance I do not know if it is better to use a cone or a dome for the midrange.
So the first step would be to select suitable drivers for the project.
Thank you very much again.
Kind regards,
beppe61
Dear Sir,
thanks for the reply. I am still trying to choose the way to go.
>
I have a question: to reproduce accurately the range 50-20.000 Hz do you think that at least 3 ways are needed?
Once I read that to reproduce the full range from 20-20.000 a minimum of 4 drivers are needed (like in the JBL 250Ti for instance).
Is it correct?
Thank you very much again.
Kind regards,
beppe61
thanks for the reply. I am still trying to choose the way to go.
>
sdclc126 said:
I have a question: to reproduce accurately the range 50-20.000 Hz do you think that at least 3 ways are needed?
Once I read that to reproduce the full range from 20-20.000 a minimum of 4 drivers are needed (like in the JBL 250Ti for instance).
Is it correct?
Thank you very much again.
Kind regards,
beppe61
Ok, if you need a narrow speaker then perhaps consider a thin floorstander 2 way with a separate woofer. Maybe a 6.5 woof and a dome tweeter. Do you have a spot to place a woofer box?
If you only need 90 dB then you don't need PA drivers (public address or pro drivers) as they are known for efficiency and toughness, not necessarily SQ.
I would only recommend a dome mid if you have heard it and like it. It would be better in a three way without the separate woofer.
Under approx. 80 hz is non-directional.
I agree that if you have a separate woofer you should have a plate amp to power it. As CLC said, you can have the woofers and amp in the same cabinet as the rest. I have two sets like that, one with a dome mid. The picture is a 4 way and you can see the front and back. Maybe this is the way for you also. You can't make the exact speaker as I have used old and discontinued parts for these, but you can still do it with new if the idea appeals to you. It is really two cabinets in one. It's approximately 10" (25 cm) wide by 12" deep (30cm) by 44" high (112cm)
It can be done for $800USD
I also noticed I have repeated a lot of what CLC said, oops.
Cal
If you only need 90 dB then you don't need PA drivers (public address or pro drivers) as they are known for efficiency and toughness, not necessarily SQ.
I would only recommend a dome mid if you have heard it and like it. It would be better in a three way without the separate woofer.
Under approx. 80 hz is non-directional.
I agree that if you have a separate woofer you should have a plate amp to power it. As CLC said, you can have the woofers and amp in the same cabinet as the rest. I have two sets like that, one with a dome mid. The picture is a 4 way and you can see the front and back. Maybe this is the way for you also. You can't make the exact speaker as I have used old and discontinued parts for these, but you can still do it with new if the idea appeals to you. It is really two cabinets in one. It's approximately 10" (25 cm) wide by 12" deep (30cm) by 44" high (112cm)
It can be done for $800USD
I also noticed I have repeated a lot of what CLC said, oops.
Cal
Attachments
More questions. Is it primarily for music or home theater? If music, what kind? I agree with the above posters that it does not sound as if you need special drivers for what you're after.
I would suggest that you look at kits; that helps get around the crossover design issue. In the US, Madisound speakers would be a good website to visit, and you could get a feel for cost and availability. Do ou have the tools and ability to build your own cabinets? They are also available as kits. And for a strong 50 Hz, I would go a little larger, to a good 8 inch; and actually if it is used for music I would suggest setting 40 Hz as the lower limit.
It can be done with either a three way or a two way; I prefer three ways, with the mid crossing over at no more than 150 Hz, and the tweeter at 3kHz or so, which puts a real burden on the midrange, but makes woofer and tweeter choices a bit easier. And since most of the signal is in the midrange, it is a good place to put more money relatively. But there are MANY different approaches that work.
I would suggest that you look at kits; that helps get around the crossover design issue. In the US, Madisound speakers would be a good website to visit, and you could get a feel for cost and availability. Do ou have the tools and ability to build your own cabinets? They are also available as kits. And for a strong 50 Hz, I would go a little larger, to a good 8 inch; and actually if it is used for music I would suggest setting 40 Hz as the lower limit.
It can be done with either a three way or a two way; I prefer three ways, with the mid crossing over at no more than 150 Hz, and the tweeter at 3kHz or so, which puts a real burden on the midrange, but makes woofer and tweeter choices a bit easier. And since most of the signal is in the midrange, it is a good place to put more money relatively. But there are MANY different approaches that work.
I should have phrased the question better, home theater/music need not be just one or the other, and universal is certainly possible.
Driver selection information This site takes a little patience to navigate, but shows many popular drivers, their costs, and some notes about performance. It is getting to be a bit out of date, but still very useful.
Madisound offers a very inexpensive crossover design service if you buy the drivers from them. A friend has used it, and it seemed worthwhile.
"The best" drivers are probably not what you really want. They would exceed your budget, and should be used with excellent electronics, cables, and crossover components. Otherwise the money would be wasted, and worse, if the associated equipment has unpleasant sounding flaws,the flaws will be reproduced clearly as well as the music. A balanced approach is often best.
I just went and looked at the Madisound site, and I see that they've changed the kit area of the online catalog; you can see more and better information about the kits if you look at the manufacturers' sections; Eton, ScanSpeak, etc. Time-tested, reliable manufacturers are: Scanspeak, Vifa, Peerless, Eton, Seas and LPG. All are European, and you might do well to investigate European sources, due to $USD/Euro fluctuations.
As a general rule, avoid drivers with peaks, especially sharp, high peaks in the frequency response. A dip in the frequency response is much less of a problem than a peak. Polypropylene is a good choice for a cone material, but paper, and nomex can be very good also. Kevlar is tricky, and the little scanspeak coated Kevlar mid and the eton honeycomb mids are very good, but uncoated Kevlar often has fierce high frequency peaks. I've had good luck with the Scanspeak carbon fiber/paper coned woofers, but they can have a nasty peak at the high end that the crossover has to be particularly designed to suppress. Tastes vary a great deal however, and you will find on this and other DIY forums many opinions, and a lot of experience.
If you are interested in speaker design as a continuing hobby, I suggest Vance Dickason's Loudspeaker Cookbook as an excellent start. However, designing your own really requires some measurement equipment and design software; cost and capability vary widely.
It is a hobby that combines art with technology, that is ever changing, and that rewards patience and experience.
Driver selection information This site takes a little patience to navigate, but shows many popular drivers, their costs, and some notes about performance. It is getting to be a bit out of date, but still very useful.
Madisound offers a very inexpensive crossover design service if you buy the drivers from them. A friend has used it, and it seemed worthwhile.
"The best" drivers are probably not what you really want. They would exceed your budget, and should be used with excellent electronics, cables, and crossover components. Otherwise the money would be wasted, and worse, if the associated equipment has unpleasant sounding flaws,the flaws will be reproduced clearly as well as the music. A balanced approach is often best.
I just went and looked at the Madisound site, and I see that they've changed the kit area of the online catalog; you can see more and better information about the kits if you look at the manufacturers' sections; Eton, ScanSpeak, etc. Time-tested, reliable manufacturers are: Scanspeak, Vifa, Peerless, Eton, Seas and LPG. All are European, and you might do well to investigate European sources, due to $USD/Euro fluctuations.
As a general rule, avoid drivers with peaks, especially sharp, high peaks in the frequency response. A dip in the frequency response is much less of a problem than a peak. Polypropylene is a good choice for a cone material, but paper, and nomex can be very good also. Kevlar is tricky, and the little scanspeak coated Kevlar mid and the eton honeycomb mids are very good, but uncoated Kevlar often has fierce high frequency peaks. I've had good luck with the Scanspeak carbon fiber/paper coned woofers, but they can have a nasty peak at the high end that the crossover has to be particularly designed to suppress. Tastes vary a great deal however, and you will find on this and other DIY forums many opinions, and a lot of experience.
If you are interested in speaker design as a continuing hobby, I suggest Vance Dickason's Loudspeaker Cookbook as an excellent start. However, designing your own really requires some measurement equipment and design software; cost and capability vary widely.
It is a hobby that combines art with technology, that is ever changing, and that rewards patience and experience.
Dear Sir,
thank you very much again for your extremely valuable support.
One last question: when you finish a prototype you judge it on the basis of a listening tests or just measurements ?
And in the case of a set of measurements, what would be the measurements more important to you ?
Thank you very much again.
Kind regards,
beppe
I'm not really up to date on current driver offerings. In addition, the major Danish companies have been purchased and consolidated. And the way I select drivers is not for all; I buy one each, mount, listen, measure, and choose that way.
I use both measurements and listening. It is very difficult to hear enough about a problem to be able to try and fix it correctly. One can hear the problem, generally identify the issue and region, ("sounds like a peak in the upper midrange" for example) but then I turn to measuring equipment to see exactly where the problem is, and what it's magnitude is. Then you can design it out accurately. Measurement is also very useful for basic verification. Reversed driver polarity can sound like some other problem for example, but shows up immediately on measurement.
In general, speakers today are a lot less "flavored" than the JBL's and AR's were. AR-3 (a's) were smooth, rolled off at the high end and had poor transient capability. The JBL's (especially the mid and tweet horns) were peaky, the high end sounded tipped up, and they had excellent transient capability. I don't know of any mainstream drivers that are that far off center today. And I didn't think your point was trivial.
Yes, 5" mids do get a bit directional at 3 kHz, but at that point the tweeter is contributing half the signal, which helps a lot. I tend to prefer first order crossovers at the crossover point, steepening to 2nd order about an octave out, and steeper yet 3 octaves out. This is not always possible, just a design guide. Note that lower order crossovers demand good out-of-band performance by the drivers, making the choice more difficult yet. However, the overlap helps make the transition from one driver to the next more "seamless". Or, drop the crossover frequency to 2.5 kHz, but as the crossover frequency drops the burden on the tweeter increases very rapidly; and stressing the tweeter is not a good thing. On the other end, many woofers do poorly at higher frequencies, so I just try to keep them out of the human voice range. You may notice that there are very few commercial designs using mid range domes; I think that their limitations at low frequencies may be responsible. Might work out nicely in a two way if you are willing to sacrifice some top end, and use a bass/mid driver. Most bass/mid drivers sacrifice some low end to get the extension, which is why I prefer the three way with a low bass/mid xover point.
Directionality is not necessarily a bad thing. In a bright room, reducing the energy "sprayed" at high frequencies can be a good thing. But more generally, good design is a matter of balancing compromises. I would sacrifice some evenness in dispersion if I gained on axis smoothness in return. This is one area where the listening is very important; your taste can guide in selecting the trade offs in a way that measuring cannot. A lot of money will reduce the severity of the compromises, but you can never have it all... So balancing the compromises is a large part of the art.
The drivers and crossover components (not all agree with the need for components to be broken in) need to be broken in before the listening tests (and to a lesser degree measurements) are reliable. Forty hours is my rule of thumb, played at a little above normal listening levels with wide bandwidth program material. This can be not fun.
Leave aside some money for improved crossover components. You can find a LOT of discussion about capacitors, wire, resistors, and connectors. The more transparent the drivers, the greater the importance.
Edit: The most important measurement for me is the waterfall. It shows resonances which are often very detrimental. Next would be good ol' frequency response, and impulse response perhaps next. And I measure and listen at each stage. Usually woofer, mounted in the cabinet is first. I use the Sheffield Drum Record, (any material with a low, DEAD, kickdrum will do) to listen for any signs of bass ringing. I prefer a tight bass without any overhang. Then bass/midrange crossover and transition. This is fairly easy. Then the bass/mid/tweet; this is much harder. The mid/tweet crossover design and tweaking usually takes me more time than all the rest put together. On and off axis frequency response are what I use here. Internal box reflections, diffraction issues, mounting plane offsets, are all things that show up at this time.
The problems can be overwhelming if you build the full proto and then start debugging, at least for me. So step by step works best for me.
I use both measurements and listening. It is very difficult to hear enough about a problem to be able to try and fix it correctly. One can hear the problem, generally identify the issue and region, ("sounds like a peak in the upper midrange" for example) but then I turn to measuring equipment to see exactly where the problem is, and what it's magnitude is. Then you can design it out accurately. Measurement is also very useful for basic verification. Reversed driver polarity can sound like some other problem for example, but shows up immediately on measurement.
In general, speakers today are a lot less "flavored" than the JBL's and AR's were. AR-3 (a's) were smooth, rolled off at the high end and had poor transient capability. The JBL's (especially the mid and tweet horns) were peaky, the high end sounded tipped up, and they had excellent transient capability. I don't know of any mainstream drivers that are that far off center today. And I didn't think your point was trivial.
Yes, 5" mids do get a bit directional at 3 kHz, but at that point the tweeter is contributing half the signal, which helps a lot. I tend to prefer first order crossovers at the crossover point, steepening to 2nd order about an octave out, and steeper yet 3 octaves out. This is not always possible, just a design guide. Note that lower order crossovers demand good out-of-band performance by the drivers, making the choice more difficult yet. However, the overlap helps make the transition from one driver to the next more "seamless". Or, drop the crossover frequency to 2.5 kHz, but as the crossover frequency drops the burden on the tweeter increases very rapidly; and stressing the tweeter is not a good thing. On the other end, many woofers do poorly at higher frequencies, so I just try to keep them out of the human voice range. You may notice that there are very few commercial designs using mid range domes; I think that their limitations at low frequencies may be responsible. Might work out nicely in a two way if you are willing to sacrifice some top end, and use a bass/mid driver. Most bass/mid drivers sacrifice some low end to get the extension, which is why I prefer the three way with a low bass/mid xover point.
Directionality is not necessarily a bad thing. In a bright room, reducing the energy "sprayed" at high frequencies can be a good thing. But more generally, good design is a matter of balancing compromises. I would sacrifice some evenness in dispersion if I gained on axis smoothness in return. This is one area where the listening is very important; your taste can guide in selecting the trade offs in a way that measuring cannot. A lot of money will reduce the severity of the compromises, but you can never have it all... So balancing the compromises is a large part of the art.
The drivers and crossover components (not all agree with the need for components to be broken in) need to be broken in before the listening tests (and to a lesser degree measurements) are reliable. Forty hours is my rule of thumb, played at a little above normal listening levels with wide bandwidth program material. This can be not fun.
Leave aside some money for improved crossover components. You can find a LOT of discussion about capacitors, wire, resistors, and connectors. The more transparent the drivers, the greater the importance.
Edit: The most important measurement for me is the waterfall. It shows resonances which are often very detrimental. Next would be good ol' frequency response, and impulse response perhaps next. And I measure and listen at each stage. Usually woofer, mounted in the cabinet is first. I use the Sheffield Drum Record, (any material with a low, DEAD, kickdrum will do) to listen for any signs of bass ringing. I prefer a tight bass without any overhang. Then bass/midrange crossover and transition. This is fairly easy. Then the bass/mid/tweet; this is much harder. The mid/tweet crossover design and tweaking usually takes me more time than all the rest put together. On and off axis frequency response are what I use here. Internal box reflections, diffraction issues, mounting plane offsets, are all things that show up at this time.
The problems can be overwhelming if you build the full proto and then start debugging, at least for me. So step by step works best for me.
I don't know if this is the right thread to throw this querry, but I tried DIYing a 3-way speaker. I used a 4 inch full range driver and crossed it with the tweeter at 4khz. Then used an 8incher for low end. I used 3rd order butterworth for the tweeter,thus, for the tweet, its a cap, then inductor to ground, then another cap. For the mids, I kinda tried using the same 3rd order crossover as if it were a 2-way design with the tweeter, it has an inductor, a cap to ground, then another inductor, then an inductance corrector. For the woofer, I just used 2nd order butterworth, with a seried inductor then a cap to ground. My problem is this, all the drivers used are 8ohm driver, all the inductors have very low resistance, these are diyed and measured with speakerworkshop, but when i attach all of the drivers to the x-over, the impendance of the system drops to 4ohms. When I tried to debug it, I removed the woofer and the x-over network attached to it, still it was 4 ohms, so I tried removing the mid driver only from the system, it got back to 8 or so ohms. I reattached the woofer, but not the mid driver, it still remained at 8ohms. Could it be because of the mid xover I used, since it was not a full bandbass network? Thanks for taking time on my querry.
It's hard to say anything without a lot more information; not least component values, and the frequencies at which the impedance was measured.
But if you are driving the speaker with a conventional transistor amplifier with reasonable current capability, the dip in impedance probably does not make a difference.
But if you are driving the speaker with a conventional transistor amplifier with reasonable current capability, the dip in impedance probably does not make a difference.
Hi Do,
the mid and treble as you have described appear to be a full range two way design. I would expect the impedance to remain the same as the two drivers.
You have added a low pass crossover and bass only unit which will be running in parallel with the two way. This will reduce the low end impedance to the average of the two systems. You have effectively made a 2.5way. Many manufacturers do this to increase the bass sensitivity to better match the treble unit. in the old days the treble was padded down to reduce it's sensitivity.
Your description of measuring the joined up and separated impedances do not quite make sense but a diagram and component values may explain all.
the mid and treble as you have described appear to be a full range two way design. I would expect the impedance to remain the same as the two drivers.
You have added a low pass crossover and bass only unit which will be running in parallel with the two way. This will reduce the low end impedance to the average of the two systems. You have effectively made a 2.5way. Many manufacturers do this to increase the bass sensitivity to better match the treble unit. in the old days the treble was padded down to reduce it's sensitivity.
Your description of measuring the joined up and separated impedances do not quite make sense but a diagram and component values may explain all.
Hi,
I agree wholeheartily with Curmudgeon.
Select the best mid driver you can sensibly afford and restrict it's output to what it can handle really well.
Then add in the bass and treble drivers to fill the extremes.
And again I agree, a well designed 3way will give a quality advantage over an over extended 2way design. The problem as Cur has said is designing and tweaking a relatively complex 3way.
Could a digital crossover help here in determining the frequencies and Qs of all the equalisation needed to optimise and then extract the info and dupilcate it in an analogue crossover (active or passive?)
I agree wholeheartily with Curmudgeon.
Select the best mid driver you can sensibly afford and restrict it's output to what it can handle really well.
Then add in the bass and treble drivers to fill the extremes.
And again I agree, a well designed 3way will give a quality advantage over an over extended 2way design. The problem as Cur has said is designing and tweaking a relatively complex 3way.
Could a digital crossover help here in determining the frequencies and Qs of all the equalisation needed to optimise and then extract the info and dupilcate it in an analogue crossover (active or passive?)
How about doing a MTMWW with (2) 10" Dayton Reference Woofers (RS270S8), (2) 5" Dayton Reference Woofers (RS125S8) and one Seas 27TBFCG tweeter?
The multi-drivers will help increase SPL a little. Not sure what the sensitivity of the 5" drivers are off the top of my head. With the dual 10", below 50 Hz should not be a problem. I would think you would be able to blend the 5" with the Tweeter ok. If I recall, the 27TBFCG has a lower Fs.
I have attached a concept picture of someone's design that I liked. I think it was posted on this forum a while back.
The multi-drivers will help increase SPL a little. Not sure what the sensitivity of the 5" drivers are off the top of my head. With the dual 10", below 50 Hz should not be a problem. I would think you would be able to blend the 5" with the Tweeter ok. If I recall, the 27TBFCG has a lower Fs.
I have attached a concept picture of someone's design that I liked. I think it was posted on this forum a while back.
Attachments
Curmudgeon said:
Dear Sir,
you have given me a complete lecture on loudspeaker design.
I thank you sincerely for your extremely kind, valuable and precious support.
I will study all this material very deeply before starting to build anything.
It is indeed a complex topic.
Thank you so much again.
Kind regards,
beppe
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.