Looks like the same drivers and flush mounted. Here's a quote from the builder:Another diy purifi floorstander has emerged:
http://www.troelsgravesen.dk/Purifi-6661.htm
Workshop total mockup to hear if all of this had a chance. It certainly did!
I thought I'd over the years heard most - from el cheapos to diamond drivers. It turned out I hadn't.
The bass from these two 6" drivers is phenomenal, deep and dynamic.
The midrange: I probably haven't heard better. Utterly transparent, dynamic, yet non-aggressive.
The Be dome - well, I knew that one - but silky smooth. OK, I used an Amber-Z in my mockup!
Last edited:
If you look down the page you'll see some photos of the boxes prior to speaker mounting - they’re flush mounted/recessed…Looks like the same drivers and surface mounted.
I really hope we can surface mount.. going to be very difficult to get my mark Vs altered for flush mounting..
Saw that...edited my post. Good to hear his response to the Purifi's.If you look down the page you'll see some photos of the boxes prior to speaker mounting - they’re flush mounted/recessed…
I really hope we can surface mount.. going to be very difficult to get my mark Vs altered for flush mounting..
Seems like there's another one of his builds where he did not flush mount.If you look down the page you'll see some photos of the boxes prior to speaker mounting - they’re flush mounted/recessed…
I really hope we can surface mount.. going to be very difficult to get my mark Vs altered for flush mounting..
Looks good the driver combo with that tweeter -superb literally the same drivers I was planning except DEQX active rather than hypex and no farty port sealed bass. Still no rush who knows what other drivers Purifi are releasing soonAnother diy purifi floorstander has emerged:
http://www.troelsgravesen.dk/Purifi-6661.htm
Are you flush mounting or surface mounting the Purifi's?
Since I am getting the speakers made by a professional, they will be recessed by 4mm and the diameter 178mm which is 2mm more than the driver's 176mm diameter. This will leave 1mm all around. The Waveguide too is done the same way. It is 148mm diameter and the cutout 150mm, also leaving 1mm all around.
But you can mount them straight into a 145-146mm hole, no problem. The old Peerless HDS drivers were likewise designed to be used both ways.
Burn-In:
I have burned in the driver for 10 hours measured the Thiele-Small Parameters and they have hardly budged at all. It behaves as it has already been run in.
I am going to give it another 10 hours, that will be 20 hours in total, and test again. I am going to do that will all eight drivers before they are used. I will do similar to the two tweeters HDS/D2608, but of course, the setup is a bit different, here a series capacitor is used. When I fit the drivers to the new boxes and do the calibrated measurements, the drivers will not be in 'cold' condition.
For those interested, I will be running a commentary as we go along - nothing too complex, just that you guys get a general idea of progress.
As to the bass alignment, measuring both the Purifi driver and the current SB17MFC driver and looking at the difference, I can tell pretty much what needs to be done in a practical sense. It basically comes down to the Fs, Qts and Vas of the driver:
The Purifi driver has a lower Qts than SB, a lower Vas too. The Fs is near the same, only slightly higher. The reason that box will need to be tuned at a bit higher frequency is the lower Qts. and to some extent the slightly higher Fs.
I will post the T-S Parameters for both drivers shortly.
A quick note of the midrange version PTT6.5M08, it appears the only difference is the shorter voice coil which results in much higher sensitivity and sure enough they say it is 91dB. It also results in higher Fs because, the voice coil, will have a lower mass. This also boosts sensitivity (and efficiency).
The cone is the same, the suspension too. Same re the magnet system.
The cone is the same, the suspension too. Same re the magnet system.
OK, the 30 hours of burning in the Purifi driver has yielded an interesting result. Usually, my system measures the Fs just slightly high, and compared to Purifi's spec, this is very close to theirs.
This is Purifi's Key Specs and the important ones we are looking at here in Red:
Next is my T-S Parameters test using ClioFW01 and the same Key Specs highlighted in Red:
Very close indeed. About the best correlation, I have seen, by a fair margin.
The parameters hardly changed at all over 30 hours of burning in. This seems to indicate the suspension is remarkably stable. The key to that is not the surround, but rather the spider part of the suspension. It has a look that I have not seen before, an intricate weave and if you shine a light behind it, it is partly transparent as I can see one of the cone breathing holes behind it. So to me, this strikes me as a very high-tech suspension and not just the 'funky' surround, as Lars Risbo calls it. I will now burn in all the other drivers for 20 hours each and then test them and compare them to the above. If I find any anomalies, I will report them.
Other than that there is nothing I can do until I get the boxes, hopefully, Friday 7th January and back from the painter by Friday 14th.
Now back to my musings about the bass alignment. We need to stay with the current volume of the box, which is about 70-75 Litres. The fill makes it behave more like 100-105 Litres. Troels uses two of the exact same driver in 37 Litres box volume. So this fits in well as the Elsinore physical volume is double that again and needs to be that since we are using double the drivers (4). But he lines the box whereas I use quite a bit of fill with some lining. My biggest surprise is that he has chosen to tune the box at slightly under 30 Hertz and yet my initial inclination is to tuning higher than the current 33 Hertz and go up to 37-38 Hertz. But what Troels has done has got me thinking. Troels method has increased the output below 33 Hertz at the expense of decreasing above 33 Hertz. In fact, it looks 2nd order above and gradually goes into 4th order below that. Maybe that is the way to go, he is certainly impressed by the bass, indicating that it is superior to what he has achieved before. So I played around with those numbers and based on the assumption that the QL of his box is around average 7 and that the Elsinores are like 2-3 because of the fill. This changed the picture somewhat. In the real world, the reduction in the midbass behaves in a certain way that I have to account for.
I am now inclined not to change the box tuning we currently use. If anything, increase the port length to 130mm and get near 30 Hertz tuning, this might be something worth trying. When doing this we get better below 33 Hertz extension than we have now (not by a big margin) and at the same expense that Troels is getting.
When the drivers are finally mounted in the box, then we shall get into the measurements. There is also a potential option available in the crossover, one that may work out or maybe not. This is to try compensate more for the baffle step by 1-2dB, but the downside is that this will lower overall system sensitivity, something I won't give up easily, but if it works, I might do it. It will still be well above average sensitivity.
Not much more I can do now, but wait for the boxes.
Cheers, Joe
This is Purifi's Key Specs and the important ones we are looking at here in Red:
Next is my T-S Parameters test using ClioFW01 and the same Key Specs highlighted in Red:
Very close indeed. About the best correlation, I have seen, by a fair margin.
The parameters hardly changed at all over 30 hours of burning in. This seems to indicate the suspension is remarkably stable. The key to that is not the surround, but rather the spider part of the suspension. It has a look that I have not seen before, an intricate weave and if you shine a light behind it, it is partly transparent as I can see one of the cone breathing holes behind it. So to me, this strikes me as a very high-tech suspension and not just the 'funky' surround, as Lars Risbo calls it. I will now burn in all the other drivers for 20 hours each and then test them and compare them to the above. If I find any anomalies, I will report them.
Other than that there is nothing I can do until I get the boxes, hopefully, Friday 7th January and back from the painter by Friday 14th.
Now back to my musings about the bass alignment. We need to stay with the current volume of the box, which is about 70-75 Litres. The fill makes it behave more like 100-105 Litres. Troels uses two of the exact same driver in 37 Litres box volume. So this fits in well as the Elsinore physical volume is double that again and needs to be that since we are using double the drivers (4). But he lines the box whereas I use quite a bit of fill with some lining. My biggest surprise is that he has chosen to tune the box at slightly under 30 Hertz and yet my initial inclination is to tuning higher than the current 33 Hertz and go up to 37-38 Hertz. But what Troels has done has got me thinking. Troels method has increased the output below 33 Hertz at the expense of decreasing above 33 Hertz. In fact, it looks 2nd order above and gradually goes into 4th order below that. Maybe that is the way to go, he is certainly impressed by the bass, indicating that it is superior to what he has achieved before. So I played around with those numbers and based on the assumption that the QL of his box is around average 7 and that the Elsinores are like 2-3 because of the fill. This changed the picture somewhat. In the real world, the reduction in the midbass behaves in a certain way that I have to account for.
I am now inclined not to change the box tuning we currently use. If anything, increase the port length to 130mm and get near 30 Hertz tuning, this might be something worth trying. When doing this we get better below 33 Hertz extension than we have now (not by a big margin) and at the same expense that Troels is getting.
When the drivers are finally mounted in the box, then we shall get into the measurements. There is also a potential option available in the crossover, one that may work out or maybe not. This is to try compensate more for the baffle step by 1-2dB, but the downside is that this will lower overall system sensitivity, something I won't give up easily, but if it works, I might do it. It will still be well above average sensitivity.
Not much more I can do now, but wait for the boxes.
Cheers, Joe
Attachments
Troels biamps his speaker, with the bottom section running on a 250W hypex plate amp. He also offers a passive option, but even then it is still meant to be biamped, although it probably doesn’t have to be provided the amp can handle the varying impedances and has enough power. The stated minimum requirement is 50W which is much higher than the current Elsinores. Sadly the crossover frequencies are not included in his write-up.
Anyway, maybe this biamped approach is required to get away with the lower box tuning?
Anyway, maybe this biamped approach is required to get away with the lower box tuning?
The amplifiers he uses are voltage sources, so the alignment should hardly move at all. The DC series resistance of the coil in the passive version might change it slightly, more than a Hypex amplifier would. But it is already well damped. Re crossover frequencies and values, you will need to buy the kit to find out. At least with the Elsinores and Hamlets, I put everything out there. Even the Waveguides can be made because I put out the drawings, but they are tricky to do.
Correct. It's good that he's generous with the amplifier settings because the Hypex Fusion amp appears to be capable of changing the speaker damping via its DSP.you will need to buy the kit to find out
Correct. It's good that he's generous with the amplifier settings because the Hypex Fusion amp appears to be capable of changing the speaker damping via its DSP.
I tried to find out about that. But the Hypex FA-251 is simply a voltage source, then how can you change damping by changing voltage because what else can the Hypex do (other than invoking some kind of voodoo, which we can all agree is not the case). I tried to find out as much as I could about the Hypex. It comes down to what the software controls and we can see that they have put that here:
The HFD v4 software consists of the following areas:
Also, I don't think Troels is doing that anyway as shown by below:
That indicates that the Hypex amp is set to being a low-pass that matches the passive low-pass. It has to do that as it needs to match the same high-pass of the midrange "M" driver. It is the same for both versions. So the filter responses must be the same.
Another thing that points to this:
Curious, in this nearfield measurement of the bass, he does not state if this particular measurement was done with the passive or Hypex low-pass. But the message here is that they should measure the same, or so close as to not matter.
But the thing that is getting my interest is that dip in the high 20Hz range. Because this is the driver itself and the port tuning and no sign that it is being interfered with. We know that the port is filling in that frequency.
When I first saw this I said hello and this is tuning way below the Fs of the driver. This is OK provided that the Qts is not has low as we have in the "W" driver. And then his comments about the quality of the bass, which is not easy just to ignore. BTW, I have been to Troels place and seen his setup in the basement of his house, so I recognise his room etc.
I digress, the fact is that this is an indication that Troels has tuned this very low. Looking at his port, being about 70mm in diameter and somewhere above 200mm long, with a 37L box, this confirms to me what I see above as the tuning frequency.
The problem with the "W" driver used here is that this decreases the output above the 30-40Hz and up to a bit above 100Hz. So studying the above measurement something occurred to be: See how the roll-off starts around 200Hz? Pull down the relative response of the midrange "W" driver (using R6 and R7) by 1-2dB, and that issue of an overdamped system, can be filled in. Of course, the system sensitivity will take a bit of a hit, but that is sometimes the kind of solution you need to use. He still gets a healthy 90dB.
Above: We can see the extension we get tuning the "W" driver to a lower Fb and that the penalty of lower output above 35Hz to 150Hz and how this can be dealt with by lowering system sensitivity, bring up that response and also get lower LF extension. A classic trade-off.
In the Elsinores, to do something similar, the port length needs to be increased from 100mm to 150mm. I use a much larger diameter port than Troels does. I did some calculations and found that with 50 Watt input, the peak Mach with 100mm long port is around 0.18 (this is very loud). Increasing the length of the port and with the new tuning, this had increased moderately to Mach 0.20 - and this would still be a much better result compared to Troel's example.
All this has to be figured out in the second half of January if our timetable. The crossover values may end up rather different and not just a minor tweak. We could also end up with a lower sensitivity of 90dB, in fact similar to Troels, but he is rated at 4 Ohm and the Elsinores are 8 Ohm nominally.
So these are the things that are on my mind as we speak.
PS: This demonstrates what I have in mind. Flatten out that red line and you have now compensated for those frequencies whilst getting more LF extention. This may well be something equal that Troels is doing, consciously or not. The hit is sensitivity:
Last edited:
You answered this question yourself when you wrote earlier about modifying a conjugate circuit against a finite source impedance to vary damping. Didn't you?then how can you change damping by changing voltage because what else can the Hypex do (other than invoking some kind of voodoo, which we can all agree is not the case).
Think of the possibilities. What's the main strength of this driver, compared to others of the same size (apart from linearity)?
Must be some kind of confusion, I never actually said that, and if it was interpreted as such, then not intended.
An amplifier (voltage source) cannot increase damping, it can only decrease damping. This simple statement stumbles many. Despite the fact that there is a lot of folklore about that, the physics is actually quite clear. You can decrease damping, not increase it. That was a lesson I got from none other than Richard H. Small when he was at Sydney University, this was back in 1975. The Hypex can't increase damping (it would literally have to produce a negative impedance to do that). Since the o/p Z can only be a positive number, it can only make damping less (some may say worse, I just say less).
Re conjugates, what they do is to equalise the current of the amplifier, the aim to make the amplifier produce the same current versus frequency, where this will create a current phase angle of zero (looks resistive). Over that bandwidth, the output impedance will have no effect on the damping. I can demonstrate the Elsinores to anybody who comes here and hook up my transconductance amplifier with and o/p impedance of 270 Ohm. Damping stays locked in, the tonal balance is locked in.
It doesn't increase damping, rather it maintains it in the real world. It is a way of not losing damping.
It also locks in the crossovers, no matter what amplifier you use. Many speakers the crossover goes all over the place with different amplifiers (tubes of course), but not so the Elsinores.
Does it also improve sound quality? IMO, yes!
Curiously, if the amplifier is a current source and you adjust the LCR correctly, you can turn a 2nd order Butterworth into a Bessel (more damped) alignment. When using a voltage source it will stay Butterworth, but using a current source, the damping will increase because it becomes a Bessel. The reason is that we can control the alignment by limiting current versus frequency, something you cannot do with a voltage source, because then you cannot put a limiter on the current (a voltage has zero control over current). Before the trolls have a go at me for saying that, may I point out that this is not theory, it has been done, and also the current can be mathematically mapped and measured. It totally hangs together.
As for the Hypex, I could email Troels and ask him. Pretty sure I know the answer already.
An amplifier (voltage source) cannot increase damping, it can only decrease damping. This simple statement stumbles many. Despite the fact that there is a lot of folklore about that, the physics is actually quite clear. You can decrease damping, not increase it. That was a lesson I got from none other than Richard H. Small when he was at Sydney University, this was back in 1975. The Hypex can't increase damping (it would literally have to produce a negative impedance to do that). Since the o/p Z can only be a positive number, it can only make damping less (some may say worse, I just say less).
Re conjugates, what they do is to equalise the current of the amplifier, the aim to make the amplifier produce the same current versus frequency, where this will create a current phase angle of zero (looks resistive). Over that bandwidth, the output impedance will have no effect on the damping. I can demonstrate the Elsinores to anybody who comes here and hook up my transconductance amplifier with and o/p impedance of 270 Ohm. Damping stays locked in, the tonal balance is locked in.
It doesn't increase damping, rather it maintains it in the real world. It is a way of not losing damping.
It also locks in the crossovers, no matter what amplifier you use. Many speakers the crossover goes all over the place with different amplifiers (tubes of course), but not so the Elsinores.
Does it also improve sound quality? IMO, yes!
Curiously, if the amplifier is a current source and you adjust the LCR correctly, you can turn a 2nd order Butterworth into a Bessel (more damped) alignment. When using a voltage source it will stay Butterworth, but using a current source, the damping will increase because it becomes a Bessel. The reason is that we can control the alignment by limiting current versus frequency, something you cannot do with a voltage source, because then you cannot put a limiter on the current (a voltage has zero control over current). Before the trolls have a go at me for saying that, may I point out that this is not theory, it has been done, and also the current can be mathematically mapped and measured. It totally hangs together.
As for the Hypex, I could email Troels and ask him. Pretty sure I know the answer already.
Not to get tied down to the amplifier itself, there are other sources of damping. As you say...
Anyway, it's your choice 😉
And since equalising current equates to equalising voltage (linearity aside), you can see the possibilities are endless.what they do is to equalise the current of the amplifier, the aim to make the amplifier produce the same current versus frequency,
Anyway, it's your choice 😉
And since equalising current equates to equalising voltage (linearity aside), you can see the possibilities are endless.
No, equalising current does not equate to equalising voltage. The physics doesn't support that. Sound only comes from the driver due to current, not voltages. There are a number of ways that the voltage of the amplifier (assuming a voltage source) can be perfect and the sound less so bnecause the current has deviated. I have said it many times, voltage tells us what we ought to hear, but the current tells us exactly what we hear. You can make an amplifier distort on the current side and show distortion there that is not there when analysing the voltage. Lars Risbo from Purifi has pointed to "impedance modulations" and I made the statement that it is the "current of the amplifier" that is being modulated by the driver, by the very imperfection of the driver itself. He agreed. Even though he was discussing "impedance modulations" in relation to excursions, it applies to any imperfection of the driver.
The driver effectively causes a corrupted and polluted environment for itself. This is because a voltage source can only control the voltage, But to do that it must relinquish any control over the current. Many find this hard to accept, but it is what is happening. This indeed is one area that a current source has an advantage. Purifi themselves recognises this, that in the critical midrange and a bit lower, typically you see a reduction of distortion up to 15dB and even more. They know this. By making the driver more linear, you don't let the amplifier in and mess things up further. Get the driver right, get the driver to control the current, then you get the benefits of a current source with a voltage source.
I hope this is understandable. I know that Purifi must be thinking along similar lines.
So with voltage sources, we cannot leave it to the amplifier to control the current, because it does not have any control there, at all. Hence we must get the driver to control the current. And what are we listening to? Current!
So Allen, I hope you can see this. Yes I know that this is not spelled out it textbooks. You asked what was the main thing about Purifi drivers? It is about "Force-Factor Modulations" which is not yet being measured by others. They have expressed a desire to have FFM as a future spec along with the current recognised specs. I believe we will be reading about it in the textbooks of the future.
AES Convention - Force Factor Modulation in Electro Dynamic Loudspeakers
Please Allen, I am going to demonstrate for you a crude way of showing a driver's FFM imperfection in percentage terms, and anybody with a 1KHz LCR Meter can do this. Stay tuned. It is very interesting and convincing - and yet oh so simple to do.
- Home
- Loudspeakers
- Multi-Way
- The "Elsinore Project" Thread