I find it strange but it seems as these speakers age the Fs actually goes up - I would think it would 'loosen' with age, which would result in a lower Fs - but it hasn't/doesn't!
My simulations on WinISD don't show a gap, though I have definitely considered adding a subsonic filter to the crossover, though I'm not sure how that will affect everything else just yet.
If nothing else I can switch in the rumble filter on my preamp.
The thing is the mechanical excursion maximum on the 2226H is something like 40mm, which is pretty good I think.Vented bass reflex I suppose.
You can check the max SPL with a given driver excursion (displacement) at different frequencies yourself. Just open the link and fill in the values.
The story with the relations between displaced air per cycle and maximum attainable SPL with that displacement is a sad story.
Because most people including me before a year or so, would think that way:
- well, we got a 95 db@1 w sensivity;
- then it would be 98@2w, 101@4w, 104@8w, 107@16w, 110@32w, 113@64w, 117@128w and 120@256w...
- then they look at the power rating of the bass driver and find the maximum SPL that the driver is capable of.
The ugly truth though is that a 15 inch driver with ~10mm p-p excursion is going to reproduce no more than 105 db SPL @ 30hz in a bass reflex no matter how much power you feed to it... it probably won't be 10-12 w as suggested by the 1w SPL, because the suspension is not linear in it's elasticity, but 30 or 40 w at low frequencies will consume all available excursion of the woofer.
Recently I was able to feed continuously (10sec) steady 120W at 20 hz sine wave into a 12 inch 92 dbWm 100w woofer with p-p excursion of ~8mm... in a vented box... it was not moving allot more than at 15W... And it wasn't producing more SPL... - luckily it's suspension has progressive stiffness...
So having in mind the laws of nature and the 105-107 db SPL limitation, you could design your system towards the most LF extension and the most flat SPL graph. And it will again be a conventional one, that's how conventional systems are designed - conventionally.
But a system with just ported bass box and two other drivers with a textbook crossover is less than conventional.
Conventional is flat on axis response, good +/- 15 degree off axis response, good LF extension, all drivers as much time aligned as possible, avoided baffle step anomalies. recently me and a friend of mine built a 3 way where the midrange is 85 dbWm and the bass is 91-92 dbWm and those measure flat on axis...
front baffle is 380x1000 mm with baffle step +3 db point around 150 hz and + 6 db from 250-300 hz on with crossover point around 600 hz BSC was avoided. - that is as conventional as it can be and it is purely design driven achievement. it is another story that we used a basic design and were aided by two acoustical engineers
T101 I'm not arguing any of that. It sounds like you have a sweet little speaker system there - I'm jealous!
Yes, I'm well aware of the limitations of what I propose. You've certainly described the 'classic' limitation.
What's killing me is the ''classic' conundrum of trying to extend the bass as much as possible with a single (relatively) high efficiency driver. This is problematic no matter what the size.
I figure the real way to fix it would be to chop the lower cutoff at around 100 Hz using the 2226H as just a mid-bass and stick a plate amp powered subwoofer under it. That would fix it good!
But, for the 'Classic 3-way' that would also be cheating, and as I have explained, the purpose of this exercise is more minimalist.
But thanks for the explanation! Too true, all of it.
Yes, I'm well aware of the limitations of what I propose. You've certainly described the 'classic' limitation.
What's killing me is the ''classic' conundrum of trying to extend the bass as much as possible with a single (relatively) high efficiency driver. This is problematic no matter what the size.
I figure the real way to fix it would be to chop the lower cutoff at around 100 Hz using the 2226H as just a mid-bass and stick a plate amp powered subwoofer under it. That would fix it good!
But, for the 'Classic 3-way' that would also be cheating, and as I have explained, the purpose of this exercise is more minimalist.
But thanks for the explanation! Too true, all of it.
T101 I'm not arguing any of that. It sounds like you have a sweet little speaker system there - I'm jealous!
Yes, I'm well aware of the limitations of what I propose. You've certainly described the 'classic' limitation.
What's killing me is the ''classic' conundrum of trying to extend the bass as much as possible with a single (relatively) high efficiency driver. This is problematic no matter what the size.
I figure the real way to fix it would be to chop the lower cutoff at around 100 Hz using the 2226H as just a mid-bass and stick a plate amp powered subwoofer under it. That would fix it good!
But, for the 'Classic 3-way' that would also be cheating, and as I have explained, the purpose of this exercise is more minimalist.
But thanks for the explanation! Too true, all of it.
I'm not arguing too, I'm just contributing here
have in mind that my original language is not English.There are ways for extending the LF on the price of loss of efficiency, and you have lots to spare
Don't sacrifice the beautiful 2226 for a sub with plate amp. The forum will help you if given the right guidelines.
My apologies T101 - I didn't mean to imply that we were arguing, only that I didn't disagree. Your english is pretty good! and I certainly do appreciate your contribution.
Exactly! I know I could sacrifice some efficiency for a little more bass extension - but then I might as well sell my 2226Hs and just get some regular low efficiency woofers, quite honestly.
Probably, I will compromise somewhere. The question is where.
Exactly! I know I could sacrifice some efficiency for a little more bass extension - but then I might as well sell my 2226Hs and just get some regular low efficiency woofers, quite honestly.
Probably, I will compromise somewhere. The question is where.
Multiple subs are a common thing these days, used in a way that works with the room modes. This is a room augmentation/correction system and so I consider it a part of the room, not a part of the main speaker system. For the most part they can be left independently when changing out different main speakers.
You do not cross them over. If you used 3 subs, for example, you'd have 5 woofers producing bass including the JBLs. This is the way I do it and I run my 15s in a plain closed box with modest fs.
You do not cross them over. If you used 3 subs, for example, you'd have 5 woofers producing bass including the JBLs. This is the way I do it and I run my 15s in a plain closed box with modest fs.
Well, you could start with simulations. I have two ugly 8 inch bass drivers with Fs 29 hz that can do ~17 hz@-10 db, 20 hz@ -6 db and 23 hz @ -3db in reasonably sized box.
If you play a little with bass box 6 or another simulation software, you could find the Hz's below Fs. The JBL all mighty is a very capable one and Fs is only one parameter.
The way to go is to accumulate knowledge, but it is not obtained by simply building an enclosure (you already know that), but with struggling with the presented by nature and hardware problems.
Double tuned box, band pass box and etc are all conventional too. Shaping the electrical output to the drivers is not uncommon too though I am prejudiced in favor of such devices and my opinion might be tilted.
See if a transmission line won't give you more extension towards the LF.
If you play a little with bass box 6 or another simulation software, you could find the Hz's below Fs. The JBL all mighty is a very capable one and Fs is only one parameter.
The way to go is to accumulate knowledge, but it is not obtained by simply building an enclosure (you already know that), but with struggling with the presented by nature and hardware problems.
Double tuned box, band pass box and etc are all conventional too. Shaping the electrical output to the drivers is not uncommon too though I am prejudiced in favor of such devices and my opinion might be tilted.
See if a transmission line won't give you more extension towards the LF.
I suppose external subs can be a future alternative, should I want them. I remember seeing a web page where a guy built a pair of subs to match a pair of La Scalas he had. While I don't think my situation is that bad, it's certainly not quite ideal either.
In any case, I see active woofers in my future...
I have neither the space, the time nor the know-how to develop a TL or some other fancy box. No, I'm afraid for the 'Classic 3 way' I'm just going to have to stick to a basic vented box and manipulate the signal for the best low extension I can.
The 2226H does have some features that are less than ideal. Though a large speaker, it wasn't designed to produce frequencies lower than 30 Hz, and going lower than that definitely presents some difficulties. Another difficulty, and the one that flies full in the face of the current fad of making speakers as small as possible, I suspect causesing the 2226H to be used far less than it might for home use, is that it requires a very large cabinet.
I want to persevere with the 2226H however. It's clearly a speaker that doesn't fit everyone – but it is a speaker I want very much to build with and listen to.
At this point it's looking like the baffle width will match the lower crossover frequency, about 71 cm. The cabinet will stand about 122 cm. This will make for a rather shallow cabinet, less than 35 cm overall, meaning it can be set close to the wall, hopefully to further mitigate the weak bass.
These dimensions should be big enough to encompasss the drivers, vents, bracing, crossover, and internal mid enclosure, and still leave the necessary 200 litres volume.
Obviously, however, the critical part of this will be the crossover. I am thinking some type of subsonic filter to protect the driver from over-excursion, and then a high-attenuating filter that kicks in around 75 Hz that inserts ~3db attenuation. This will have the effect of boosting the lowest pass-band. Perhaps this will give me another 5 Hz?
I have no idea how to do this as yet, but it seems like where we are headed. What do you all think?
In any case, I see active woofers in my future...
I have neither the space, the time nor the know-how to develop a TL or some other fancy box. No, I'm afraid for the 'Classic 3 way' I'm just going to have to stick to a basic vented box and manipulate the signal for the best low extension I can.
The 2226H does have some features that are less than ideal. Though a large speaker, it wasn't designed to produce frequencies lower than 30 Hz, and going lower than that definitely presents some difficulties. Another difficulty, and the one that flies full in the face of the current fad of making speakers as small as possible, I suspect causesing the 2226H to be used far less than it might for home use, is that it requires a very large cabinet.
I want to persevere with the 2226H however. It's clearly a speaker that doesn't fit everyone – but it is a speaker I want very much to build with and listen to.
At this point it's looking like the baffle width will match the lower crossover frequency, about 71 cm. The cabinet will stand about 122 cm. This will make for a rather shallow cabinet, less than 35 cm overall, meaning it can be set close to the wall, hopefully to further mitigate the weak bass.
These dimensions should be big enough to encompasss the drivers, vents, bracing, crossover, and internal mid enclosure, and still leave the necessary 200 litres volume.
Obviously, however, the critical part of this will be the crossover. I am thinking some type of subsonic filter to protect the driver from over-excursion, and then a high-attenuating filter that kicks in around 75 Hz that inserts ~3db attenuation. This will have the effect of boosting the lowest pass-band. Perhaps this will give me another 5 Hz?
I have no idea how to do this as yet, but it seems like where we are headed. What do you all think?
200 l tuned at 35 hz gives pretty good response with the given driver.
The group delay is ~ 22 ms, but you'll live with it, the slopes of the GD curve are not very steep, so it would not be very audible. Maybe some definition will be lost under 40 hz.
If you reduce the frequencies above 70 hz with 3db, your - 6 db point (F3) will become around 28 hz which is pretty good.
But before that you would have to deal with the baffle step, which is considerable wit a baffle sized 71x122 cm.
Without BSC you have a 8 db peak at 300 hz.
If you add the suggested by the Edge BSC, you will be left with a little over 2 db peak at 300 hz.
So a second order LP crossover at 300-350 hz and HP at 400-450 hz becomes an obvious solution. The final response should be flat with those measures taken.
Then the - 3-4 db correction can be added if you wish. A bigger version of this would do it : http://www.diyaudio.com/forums/full-range/180491-l-here-l-there-l-everywhere-2.html it can be applied to the bass and midrange section together and it can be adjustable.
You can do it without a transformer, with just an adjustable rheostat in the place of the bypass resistor, but then the demand for power rating is about double and the price of the rheostat is about triple if you even find one that powerful.
If you go with the adjustable correction circuit (which is an over sized BSC and thus it's relative), you could have an adjustable L-pad for the compression driver. Then you'll have adjustable midrange and adjustable treble.
All this will come to you at the price of two big inductances per speaker cabinet and two rheostats.
The third photo is the transformer with a powerful rheostat added to it, the same if wired with all 3 leads is a powerful continuously adjustable L-pad and comes at a price of about 30$ here.
A very strong suggestion is that you put the bass driver at ear level! The whole perception of low frequencies changes massively in the good sense of the word.
Best regards!
The group delay is ~ 22 ms, but you'll live with it, the slopes of the GD curve are not very steep, so it would not be very audible. Maybe some definition will be lost under 40 hz.
If you reduce the frequencies above 70 hz with 3db, your - 6 db point (F3) will become around 28 hz which is pretty good.
But before that you would have to deal with the baffle step, which is considerable wit a baffle sized 71x122 cm.
Without BSC you have a 8 db peak at 300 hz.
If you add the suggested by the Edge BSC, you will be left with a little over 2 db peak at 300 hz.
So a second order LP crossover at 300-350 hz and HP at 400-450 hz becomes an obvious solution. The final response should be flat with those measures taken.
Then the - 3-4 db correction can be added if you wish. A bigger version of this would do it : http://www.diyaudio.com/forums/full-range/180491-l-here-l-there-l-everywhere-2.html it can be applied to the bass and midrange section together and it can be adjustable.
You can do it without a transformer, with just an adjustable rheostat in the place of the bypass resistor, but then the demand for power rating is about double and the price of the rheostat is about triple if you even find one that powerful.
If you go with the adjustable correction circuit (which is an over sized BSC and thus it's relative), you could have an adjustable L-pad for the compression driver. Then you'll have adjustable midrange and adjustable treble.
All this will come to you at the price of two big inductances per speaker cabinet and two rheostats.
The third photo is the transformer with a powerful rheostat added to it, the same if wired with all 3 leads is a powerful continuously adjustable L-pad and comes at a price of about 30$ here.
A very strong suggestion is that you put the bass driver at ear level! The whole perception of low frequencies changes massively in the good sense of the word.
Best regards!
Attachments
Last edited:
The Edge by Tolvan Data is the one for predicting the baffle step. It is recommended by Martin King. Also on his site "quarter wave" you'll find allot of information on the topic.
The other is Bass Box 6 pro for the boxes and LTspice for circuits.
But there is another option - an MJK mathcad sheet for calculating ported boxes and another one for Mass Loaded Transmissionn Lines MLTL. An MLTL is constructed exactly as bass reflex, with only difference the proportions and driver and port placement. This again is from MJK's quarter wave site.
And you can't go wrong with an MLTL, because it is easily convertible to bass reflex.
The other is Bass Box 6 pro for the boxes and LTspice for circuits.
But there is another option - an MJK mathcad sheet for calculating ported boxes and another one for Mass Loaded Transmissionn Lines MLTL. An MLTL is constructed exactly as bass reflex, with only difference the proportions and driver and port placement. This again is from MJK's quarter wave site.
And you can't go wrong with an MLTL, because it is easily convertible to bass reflex.
Last edited:
Well hello!
It has been quite a while since my last post here, and though progress has been slow, it has not been altogether lacking.
I have started construction on the cabinet and done some test finishes. I will add pictures of this later.
One might well wonder why I have been quiet on this for so long. I noticed this myselfp) and I know that when I find myself procrasting on something like this, it is because something is not quite right. I have had to re-evaluate some of my starting decisions on the project.
The primary issue, much discussed earlier in the thread, is the lack of LF extension, especially for such a big speaker. The truth is that the 35 Hz I was looking at just was not making me happy. It is true that it is 'enough' to reproduce most music, but I also know that it shouldn't be too hard to do better. Having come to the realization, I made the decision to search out a better woofer.
My original choice, the JBL 2226H, is a venerable and well respected driver. The problem, I realized, is that it just didn't fit my application very well. I really wanted to get below 30 Hz. Now, some will doubtless argue that 35 to 30 Hz isn't much of a difference, but I felt it really shouldn't be that hard, and if I am putting this much effort and resources into something, I should do what I could.
So a couple of weeks ago I set out, WinISD in hand, slowly slogging through the various venders catalogs plugging in likely candidates to see how they would work in the cabinet I had already envisioned.
The results were rather surprising to me. For one thing, a lot of 15" speakers exhibited the same f3 as the 2226H. Of course that wasn't the only thing I was looking at: Distortion, xmax, bandwidth, fb, SPL, and power handling were all considerations, but the sheer general uniformity of bass cutoff was somewhat surprising.
It seems many woofers are made with close to the same general lower limit. That is not always apparent looking at the specifications, but when modeled that is what shows. Exactly why this is so is probably somewhat complex. Is it a general compromise between frequency coverage and mechanical capability? Or is it a convention that has more to do with the prevalence of the separation of monitor boxes and subwoofers? Dunno, but I did notice that the speakers that went really low tended to be far less efficient and far narrower of bandwidth.
Others reading this may well be saying 'Duh!', but this newb found it all very interesting, so I thought I'd write it up here as part of my 'voyage of discovery' that this project is turning out to be.
The driver I eventually settled on did not look too promising to start. Fs @ 42Hz, f3 listed as 35Hz. Speaker was unrated at partsexpress. Still, it didn't look too bad, and by then I was getting desperate, so I plugged it into the simulator, and I found an astounding thing: It worked.
Vas was stated as 115.5L, but when put into a 200L vented cabinet the lower end just kept extending. Tuned to 29Hz the curve gently dropped off to a 29Hz f3, and the Xmax curve was within limits to 300 watts. In fact, box volume could be extended all the way up to 300L, and properly tuned the f3 was good down to 25Hz! 10 cu ft is just too big, even for me, but I thought WOW, why haven't I ever heard of this driver?
http://www.parts-express.com/pdf/290-578s.pdf
Additionally, it features a copper shorting ring, and even better xmax than the 2226H. The cost isn't too bad either, less than half the going price for a new JBL, and even better from another vender I found online.
More to come soon.
It has been quite a while since my last post here, and though progress has been slow, it has not been altogether lacking.
I have started construction on the cabinet and done some test finishes. I will add pictures of this later.
One might well wonder why I have been quiet on this for so long. I noticed this myself
p) and I know that when I find myself procrasting on something like this, it is because something is not quite right. I have had to re-evaluate some of my starting decisions on the project.The primary issue, much discussed earlier in the thread, is the lack of LF extension, especially for such a big speaker. The truth is that the 35 Hz I was looking at just was not making me happy. It is true that it is 'enough' to reproduce most music, but I also know that it shouldn't be too hard to do better. Having come to the realization, I made the decision to search out a better woofer.
My original choice, the JBL 2226H, is a venerable and well respected driver. The problem, I realized, is that it just didn't fit my application very well. I really wanted to get below 30 Hz. Now, some will doubtless argue that 35 to 30 Hz isn't much of a difference, but I felt it really shouldn't be that hard, and if I am putting this much effort and resources into something, I should do what I could.
So a couple of weeks ago I set out, WinISD in hand, slowly slogging through the various venders catalogs plugging in likely candidates to see how they would work in the cabinet I had already envisioned.
The results were rather surprising to me. For one thing, a lot of 15" speakers exhibited the same f3 as the 2226H. Of course that wasn't the only thing I was looking at: Distortion, xmax, bandwidth, fb, SPL, and power handling were all considerations, but the sheer general uniformity of bass cutoff was somewhat surprising.
It seems many woofers are made with close to the same general lower limit. That is not always apparent looking at the specifications, but when modeled that is what shows. Exactly why this is so is probably somewhat complex. Is it a general compromise between frequency coverage and mechanical capability? Or is it a convention that has more to do with the prevalence of the separation of monitor boxes and subwoofers? Dunno, but I did notice that the speakers that went really low tended to be far less efficient and far narrower of bandwidth.
Others reading this may well be saying 'Duh!', but this newb found it all very interesting, so I thought I'd write it up here as part of my 'voyage of discovery' that this project is turning out to be.
The driver I eventually settled on did not look too promising to start. Fs @ 42Hz, f3 listed as 35Hz. Speaker was unrated at partsexpress. Still, it didn't look too bad, and by then I was getting desperate, so I plugged it into the simulator, and I found an astounding thing: It worked.
Vas was stated as 115.5L, but when put into a 200L vented cabinet the lower end just kept extending. Tuned to 29Hz the curve gently dropped off to a 29Hz f3, and the Xmax curve was within limits to 300 watts. In fact, box volume could be extended all the way up to 300L, and properly tuned the f3 was good down to 25Hz! 10 cu ft is just too big, even for me, but I thought WOW, why haven't I ever heard of this driver?
http://www.parts-express.com/pdf/290-578s.pdf
Additionally, it features a copper shorting ring, and even better xmax than the 2226H. The cost isn't too bad either, less than half the going price for a new JBL, and even better from another vender I found online.
More to come soon.
Last edited:
Why do you ask that? It's only 65 cm wide, 132 cm tall, and 40 cm deep.
As you can imagine, this has been a bit of an issue, but she'll come around eventually - I have faith!
A thread discussing the tweeter horns can be seen here:
http://www.diyaudio.com/forums/multi-way/209952-horn-impedance.html
http://www.diyaudio.com/forums/multi-way/209952-horn-impedance.html
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.