Joe, it seems you misinterpreted my remark on roundovers.
Elsinores will exhibit diffraction phenomena because of the dimensions and shape of the enclosure, and placement of the drivers.
The tweeter is waveguided so that should be OK, but in the lower midrange, diffraction will have influence, and not for the better; 9mm roundover or not will not help there.
To fight diffraction, roundovers should be much larger; think 2 inches radius. Check this Magico loudspeaker.
But I get it that this is not doable in an Elsinore.
Elsinores will exhibit diffraction phenomena because of the dimensions and shape of the enclosure, and placement of the drivers.
The tweeter is waveguided so that should be OK, but in the lower midrange, diffraction will have influence, and not for the better; 9mm roundover or not will not help there.
To fight diffraction, roundovers should be much larger; think 2 inches radius. Check this Magico loudspeaker.
But I get it that this is not doable in an Elsinore.
Attachments
I do realise and like that the design is mature and supported - that's why I am interested.
It's study time at the moment.
I am sorting out where to source all the parts and making sure I understand everything so I don't end up waiting on anything when I launch into the build.
Thanks Daan. Actually, I think we are broadly in agreement, as we know, there is two different diffractions effects involved. But there is also a third factor...
1. The tweeter is largely OK because of waveguide.
2. Diffractions effects that you allude to, these could well be improved by much larger 2" roundover. Not questioning that at all.
3. The Elsinore with that large 2" radius would need a possible redesign of the crossover low-pass centered near 500 Hertz. Much larger 2" roundover, my main issue is that it would shift that frequency downwards. and create a trough in the mid-to-low midrange.
I am quite happy as to how things are working right now and that smaller roundovers are not an issue, and that they are purely a visual thing when they are kept reasonable. The larger roundover like 2" inch well improve one thing (as you showed) but actually cause something else, a shift in the crossover frequency. So in my discussions it is this 3rd factor that has occupied me and my comments should be taken in that context.
BTW, I wish you could hear a Mk6 version of the Elsinores. The Mk1 that you heard are not a patch on it. Besides, Thomas had mislabeled a connector and did not tell me about it, only the last night and the last hours, when we moved them into the room Bjørn Aaholm was using (smaller) was the problem noticed and connected correctly. Such is life. So for a few hours they worked as intended into past midnight, and the next morning we packed up. Even then, the current Mk6 blows that out of the water. That was 17 years ago.
But visually, but some roundover makes the Elsinores look nicer, even just like this. I just love the look of these, but I am biased:
[Upcoming aluminimum "NBAC" version, shown without the IsoAcoustics base it sits on.]
Which you suggest adjusting in the crossover, is the inductor value the concern?
Here's an alternative. It would be relatively easy to take a 100mm diameter cardboard tube and split it down the middle, thus reducing the change in baffle support compared to rounding the existing cabinet.
Sure, but what about the looks?
When I first posted the Mk1 design, it was published so that even guys capable of plain square butt joints could build them. Just good old square corners. Now we have access to stuff and the situation has changed. Yet anybody with a Mk1 Elsinore can find a way to upgrade them to Mk6 because the fundamental box design has not changed. So that is something that pleases me no end. Yes, some customisations that are common sense, they can still be done. But let us see if anybody takes up your suggestion? OR 
When I first posted the Mk1 design, it was published so that even guys capable of plain square butt joints could build them. Just good old square corners. Now we have access to stuff and the situation has changed. Yet anybody with a Mk1 Elsinore can find a way to upgrade them to Mk6 because the fundamental box design has not changed. So that is something that pleases me no end. Yes, some customisations that are common sense, they can still be done. But let us see if anybody takes up your suggestion?
OR
Back from a short holiday and came home to see the re-lathed waveguides by Steve in Melbourne having arrived, so hopefully we will have a solution to supplying "NBAC" waveguides based on the current "MFC" waveguide that we have all been using with the D2608 tweeter (and that includes "ULD" version too.
Just to categorise the three different versions of the Elsinore Mk6, they can be described along these lines:
"MFC" is the allrounder that most should build, it has high sensitivity and there is nigh on near no amplifier out there that won't sound good with it and better amplifiers even better. The sound has been well described here on this thread. It is very detailed with no hard edges and it fits into most people's systems, but room placement is always a key. This is the version that 90% of you should be building and if you have any doubts as to which version, then just build this one.
"ULD" is the Elsinore taken to the max, albeit at a massive price increase. It has the edge on the "MFC" is the purity of it sound and also sounding slightly faster. It may also be a tad 'cool' sounding, but also very pure is the gain. The "ULD" is for a few people. Sensitivity is the same, so any amplifier can drive it. It will go louder than any of the other versions and has extraordinary power handling, one that will probably never be needed. It is also suitable for very large rooms.
"NBAC" is a specialist version of the Elsinores. Maybe a poor man's version of the "ULD" could be the answer, but it is not that simple. There is a significant drop in sensitivity and it requires at least twice the power from the amplifier. In fact I will soon be doing a more accurate measurement/assessment and it may well need 3-4 times the overall voltage and current taken into consideration. The impedance around 10KHz falls to 4.5 Ohm and is the lowest of all versions. This speaker sound more like a "ribbon" speaker that I have ever heard any dynamic speaker sound. But that is my personal assessment. I like it and use it with a 30 Watt Triode amplifier. But the "NBAC" is not for everybody.
(No description of the "NRX" version above as it is based on a discontinued driver, but it fits along the "MFC" in performance).
I hope the above is helping in making your choices.
The "NBAC" details with +crossover schematic will be published soon. The driver is exactly the same as for the "MFC" in every way, only the drivers, waveguides and Crossovers are different.
It is not in the public domain but is single-user license and intellectual property rights are retained. Same deal as before.
Just to categorise the three different versions of the Elsinore Mk6, they can be described along these lines:
"MFC" is the allrounder that most should build, it has high sensitivity and there is nigh on near no amplifier out there that won't sound good with it and better amplifiers even better. The sound has been well described here on this thread. It is very detailed with no hard edges and it fits into most people's systems, but room placement is always a key. This is the version that 90% of you should be building and if you have any doubts as to which version, then just build this one.
"ULD" is the Elsinore taken to the max, albeit at a massive price increase. It has the edge on the "MFC" is the purity of it sound and also sounding slightly faster. It may also be a tad 'cool' sounding, but also very pure is the gain. The "ULD" is for a few people. Sensitivity is the same, so any amplifier can drive it. It will go louder than any of the other versions and has extraordinary power handling, one that will probably never be needed. It is also suitable for very large rooms.
"NBAC" is a specialist version of the Elsinores. Maybe a poor man's version of the "ULD" could be the answer, but it is not that simple. There is a significant drop in sensitivity and it requires at least twice the power from the amplifier. In fact I will soon be doing a more accurate measurement/assessment and it may well need 3-4 times the overall voltage and current taken into consideration. The impedance around 10KHz falls to 4.5 Ohm and is the lowest of all versions. This speaker sound more like a "ribbon" speaker that I have ever heard any dynamic speaker sound. But that is my personal assessment. I like it and use it with a 30 Watt Triode amplifier. But the "NBAC" is not for everybody.
(No description of the "NRX" version above as it is based on a discontinued driver, but it fits along the "MFC" in performance).
I hope the above is helping in making your choices.
The "NBAC" details with +crossover schematic will be published soon. The driver is exactly the same as for the "MFC" in every way, only the drivers, waveguides and Crossovers are different.
It is not in the public domain but is single-user license and intellectual property rights are retained. Same deal as before.
Sounding "faster"? That reminds me of popular hifi review practice (snake oil).
In other words: enclosure volume and port dimension are equal for the three versions.
However, the drivers for MFC, ULD and NBAC version have different TSP.
So bass reflex tuning is different which will cause difference in bass (slower..., faster...., better called PRAT), but for which driver the tuning is correct?
However, the drivers for MFC, ULD and NBAC version have different TSP.
So bass reflex tuning is different which will cause difference in bass (slower..., faster...., better called PRAT), but for which driver the tuning is correct?
Hi Daan
Maybe "faster" is an inadequate word. But it was reported back to me by an end user, so I thought fair to report exactly as was said.
A certain sense of lack of compression? Unhindered transient response? Vocabulary often falls short...
A lot to comment on. So I will type quickly as if thinking out aloud.
Here first is the topic is why NBAC ended up having the same box as MFC.
The drivers may be different but the MFC and NBAC is the same driver except the cone material is different. Mainly the difference here is the cone mass that changes the T-S Params. The NBAC alignment is definitely a slightly higher Q than MFC. So you are right, the two alignments is are fractionally different. They are not exactly the same. That was never claimed. Indeed if you go back, I did say that it would be maybe a bit more "fruity" or fuller. That proved to be the case, but less than I thought. I liked the end result in its own right.
Indeed NBAC seems to indicate that perhaps a longer port may be required, but the lower Fs helps here. But there is something else involved. The highly resistive nature of the box, the Q of the box (that sits in parallel) is much lower than normal due to higher level of fill than mere lining commonly used in vented box. If it had not been this feature of the box, I would not have gotten away with it.
This means that you are right, the alignments are not the same, but also they are not as different as you might think because of the parallel box Q. It all comes out in the wash. Should I have used a longer port in the NBAC? I seriously considered it but decided against it, a longer port also introduces other issues, such as peak acceleration in the port goes up with length and hence noise (chuffy). At the very end I still had a choice to lengthen the port, but I decided against it.
The ULD was a different issue here. It caused me the most angst even to choose this driver, and then you have to consider the cost. The driver T-S Params turned out to be well within 10% of what they had published and since their setup was different from mine, it was also a kind of 'calibration' of my own setup compared to theirs.
The issue here is the even lower Qes. I considered tuning it to a higher frequency and reducing box volume. But I did not want to change the overall box design.
Note the years below. I chose only drivers that would work in the box design I had, that the end user could select the driver. The Crossover would of course be a different in every case.
So can you see that I selected carefully a few out of many drivers over many years:
NRX: 2014, MFC: 2018, ULD: 2022. Maybe 2025: MW16.
You can see the spread in years. I choose the driver to suit the box. The vast number of drivers out there are not suitable (I got suggestions all the time and said no). I have eyed the drivers so far used to work in that box. I have another driver spotted that I know would work as well. That is SB Satori MW16P-08. It would slot into this box quite well, I am confident of that.
Back to ULD. Trust me, I spent endless hours before deciding on the ULD because it has the lowest Q of the three drivers. In the end it was a bit risky. But the end result worked and I have a few theories as why the ULD did work despite the fact I had reservations (that low Q). Again the low Q of the box and past experience in working with this kind of alignments in other speaker designs that I have done, and I figured it would work, maybe even change the tuning to the higher 30's - but again it didn't seem to need it. If anybody wants to try shorten the port to say 50-70mm, then we could talk through it here on the forum. I would welcome that. I heard no 1-2dB depression just above 50Hz.
Greater acceleration in a driver can lead to increased perceived loudness.
This I don't think has been discussed. Flat response is equal to flat acceleration across all frequencies. But frequency response measurements don't show that, hence there can be a perceived different frequency response not the same as the classically measured one. Lack of acceleration means the response will be down, increased acceleration leads to a more prominence.
Our normal Frequency Response measurements are based on amplitude, but our ears are based on acceleration.
Is that considered a controversial statement?
As for PRAT, I have some theories about that. I think it comes down to the fact that we are incredibly sensitive to midrange distortion (even beyond -80dB) and less sensitive (not insensitive) to dynamics. Gradually under 500 Hertz we become less sensitive to distortion whilst increasingly sensitive to dynamics. Invert Fletcher Munson and it supports that, not distortion, but dynamics. Below 100 Hertz and distortion can be several percent. That would be dissastrous in the midrange. Purifi has made comments about this in their blogs. But dynamics at these frequencies are paramount. Better dynamics in the 50-100Hz Octave also means it sounds louder. (is this what I heard in the ULD?). At 30Hz we don't hear the threshold until 70dB-SPL which means we hear harmonics come to the ear before the fundamental. This is timing and maybe the basis for a PRAT theory?
Now please Daan, don't try to pick the above to pieces, I was just thinking aloud and typing. Nothing is 100%, right?
Cheers, Joe
EDIT: Above!
Maybe "faster" is an inadequate word. But it was reported back to me by an end user, so I thought fair to report exactly as was said.
A certain sense of lack of compression? Unhindered transient response? Vocabulary often falls short...
A lot to comment on. So I will type quickly as if thinking out aloud.
Here first is the topic is why NBAC ended up having the same box as MFC.
The drivers may be different but the MFC and NBAC is the same driver except the cone material is different. Mainly the difference here is the cone mass that changes the T-S Params. The NBAC alignment is definitely a slightly higher Q than MFC. So you are right, the two alignments is are fractionally different. They are not exactly the same. That was never claimed. Indeed if you go back, I did say that it would be maybe a bit more "fruity" or fuller. That proved to be the case, but less than I thought. I liked the end result in its own right.
Indeed NBAC seems to indicate that perhaps a longer port may be required, but the lower Fs helps here. But there is something else involved. The highly resistive nature of the box, the Q of the box (that sits in parallel) is much lower than normal due to higher level of fill than mere lining commonly used in vented box. If it had not been this feature of the box, I would not have gotten away with it.
This means that you are right, the alignments are not the same, but also they are not as different as you might think because of the parallel box Q. It all comes out in the wash. Should I have used a longer port in the NBAC? I seriously considered it but decided against it, a longer port also introduces other issues, such as peak acceleration in the port goes up with length and hence noise (chuffy). At the very end I still had a choice to lengthen the port, but I decided against it.
The ULD was a different issue here. It caused me the most angst even to choose this driver, and then you have to consider the cost. The driver T-S Params turned out to be well within 10% of what they had published and since their setup was different from mine, it was also a kind of 'calibration' of my own setup compared to theirs.
The issue here is the even lower Qes. I considered tuning it to a higher frequency and reducing box volume. But I did not want to change the overall box design.
Note the years below. I chose only drivers that would work in the box design I had, that the end user could select the driver. The Crossover would of course be a different in every case.
So can you see that I selected carefully a few out of many drivers over many years:
NRX: 2014, MFC: 2018, ULD: 2022. Maybe 2025: MW16.
You can see the spread in years. I choose the driver to suit the box. The vast number of drivers out there are not suitable (I got suggestions all the time and said no). I have eyed the drivers so far used to work in that box. I have another driver spotted that I know would work as well. That is SB Satori MW16P-08. It would slot into this box quite well, I am confident of that.
Back to ULD. Trust me, I spent endless hours before deciding on the ULD because it has the lowest Q of the three drivers. In the end it was a bit risky. But the end result worked and I have a few theories as why the ULD did work despite the fact I had reservations (that low Q). Again the low Q of the box and past experience in working with this kind of alignments in other speaker designs that I have done, and I figured it would work, maybe even change the tuning to the higher 30's - but again it didn't seem to need it. If anybody wants to try shorten the port to say 50-70mm, then we could talk through it here on the forum. I would welcome that. I heard no 1-2dB depression just above 50Hz.
Greater acceleration in a driver can lead to increased perceived loudness.
This I don't think has been discussed. Flat response is equal to flat acceleration across all frequencies. But frequency response measurements don't show that, hence there can be a perceived different frequency response not the same as the classically measured one. Lack of acceleration means the response will be down, increased acceleration leads to a more prominence.
Our normal Frequency Response measurements are based on amplitude, but our ears are based on acceleration.
Is that considered a controversial statement?
As for PRAT, I have some theories about that. I think it comes down to the fact that we are incredibly sensitive to midrange distortion (even beyond -80dB) and less sensitive (not insensitive) to dynamics. Gradually under 500 Hertz we become less sensitive to distortion whilst increasingly sensitive to dynamics. Invert Fletcher Munson and it supports that, not distortion, but dynamics. Below 100 Hertz and distortion can be several percent. That would be dissastrous in the midrange. Purifi has made comments about this in their blogs. But dynamics at these frequencies are paramount. Better dynamics in the 50-100Hz Octave also means it sounds louder. (is this what I heard in the ULD?). At 30Hz we don't hear the threshold until 70dB-SPL which means we hear harmonics come to the ear before the fundamental. This is timing and maybe the basis for a PRAT theory?
Now please Daan, don't try to pick the above to pieces, I was just thinking aloud and typing. Nothing is 100%, right?
Cheers, Joe
EDIT: Above!
Last edited:
Hearing is acceleration based.. Now you are inventing your own universe again Joe.... why?
"Is that considered controversial..." - you kind of seek these situations deliberately and check if you can get away with it... acceleration and the ear has to do with balance, not hearing.
Speakers seem fine - no mystique needed...
//
"Is that considered controversial..." - you kind of seek these situations deliberately and check if you can get away with it... acceleration and the ear has to do with balance, not hearing.
Speakers seem fine - no mystique needed...
//
Yup! But I hear balance.
Oddly enough, you disagree even when we agree. I don't know why, but going back with my dealings with you and your posts about me, you somehow want to personalise everything. I just want to point out to you and other reading these public posts, this is entirely one-way traffic. Do you care to explain why? I am at a loss to explain it. My other Swedish friends don't treat me this way.
Are you here to build Elsinores or critique what you have not heard?
These days,, you are not allowed to think out aloud and somebody will take something in the middle and can't help themselves to nit-pick on it. Note that I predicted in the post it would happen and it did. I am no prophet, but I must be better than just a good guesser. Sigh.
To have a perceived flat frequency response, you must have flat acceleration across the board. I meant no more and no less.
To have a perceived flat frequency response, you must have flat acceleration across the board. I meant no more and no less.
I dont post about you. I post about what you claim/say. It would be very hard to separate these in some way as it is you who say things.
I mean, "Our normal Frequency Response measurements are based on amplitude, but our ears are based on acceleration." is a very bold confident statement for being thinking out loud...
A cone is accelerated to follow the path to produce a 500Hz sinus. The accelerations is greatest leaving top and the bottom as the speed there is zero and reaches the highest speed (and thus, lowest acceleration) between top and bottom - the zero passing, as it where. This is basics. If one would to change the acceleration somewhere between the top and the bottom, there would be the same as introducing distorsion as an increased acceleration would mean that no longer a 500Hz tone is generated but some higher frequency. So what you describe is a speaker with high(er) distortion.
//
- You hear "balance" - please expand... how does this work?
- I have had no intention to build your designs. But your statements about things has made me interested and I have followed some of what you have posted. I'm e.g. still waiting for your group of experts to release the "paper you promised a few years back.
- If you predicted a statement would be challenged, maybe it would have been better to state it as a question?
I mean, "Our normal Frequency Response measurements are based on amplitude, but our ears are based on acceleration." is a very bold confident statement for being thinking out loud...
A cone is accelerated to follow the path to produce a 500Hz sinus. The accelerations is greatest leaving top and the bottom as the speed there is zero and reaches the highest speed (and thus, lowest acceleration) between top and bottom - the zero passing, as it where. This is basics. If one would to change the acceleration somewhere between the top and the bottom, there would be the same as introducing distorsion as an increased acceleration would mean that no longer a 500Hz tone is generated but some higher frequency. So what you describe is a speaker with high(er) distortion.
//
So we are playing this game again and it is not personal? Disagree with me when we mostly agree, and it isn't personal? Of course it is, you are engaged in a form of character assassination and don't even have the honesty to admit it. I am sure you will come back with more baiting, but I won't responding. Have a good day, sir!
We were discussing reflex alignments for different drivers in the same box.
My conclusion is that optimal alignment, taking room acoustics and placement into account, will be driver specific.
The design as it is (box Q, port dimension) can only be a starting point.
Anyway, I did not mean to pull out another can of worms
My conclusion is that optimal alignment, taking room acoustics and placement into account, will be driver specific.
The design as it is (box Q, port dimension) can only be a starting point.
Anyway, I did not mean to pull out another can of worms
I know what you mean and I know where you were coming from. But this is DIY and we had a box design and it was a case of looking for what drivers could be made to work and lots of those that were suggested were not. If you have Mk-1 boxes circa 2006, they can be still adapted to one of the Mk-6 variations. This has happened about a dozen times I know of. The Elsinores are almost unique in this respect? I only know of one other driver right now, that could be used. The SB Satori MW16P-08, and I am not even saying that will ever happen, only that is the only other one that I know right now. There must be a thousand other drivers out there not suitable.
I am working on another design right now, one that is not DIY. That is a very different kettle of fish. Here the box is particularly designed for the driver, not the other way around. That is how it should be. It is a 2-Way and uses a TeXtreme cone (very expensive). No details will be posted, other than maybe a picture.
I am working on another design right now, one that is not DIY. That is a very different kettle of fish. Here the box is particularly designed for the driver, not the other way around. That is how it should be. It is a 2-Way and uses a TeXtreme cone (very expensive). No details will be posted, other than maybe a picture.
