Martin, crossover frequency....
I read through the paper describing the system and, it does echo many of the same things I've found in the "Spirits"---but one thing is very glaring in relation to the Jordan JX92S drivers, that the reader should be aware of.
The Jordans are absolutely ruthless to anything put in series or parallel to them Even using a passive line level cross over, reveals the quality (or lack of ) of the capacitors/resistors being used. A "J" spec (+/- 10%), 104 cap (.01 uF), results in a roll off in the bass of 3dB per octave starting at 200Hz (-6db by 50 Hz), but at the loss of the Jordan "sparkle". The JX92S drivers lose there ability to emotionally involve the listener, to the point of preferring no cross-over at all on the Jordans.
What has been your experience with the Jordans in this regard? Using an active crossover/biamp system could very well be the answer, but depends somewhat on the quality of the crossover. I see that you use a 4th order LR @100Hz for the bass, and a 2nd order LR @ 350 Hz for the Jordans.
The subs we have used 12" M&Ks in 3.25 cu.ft. sealed enclosures (originally to be aperiodic), each with it's own 250W plate amp, and low pass crossover. I am surprised your best results were obtained with a crossover frequency as high as 350Hz for the Jordans, and 100 Hz for the Goldwoods.
Looking forward to reading more about your experiences, and thank you for another very well explained and documented project.
stew
I read through the paper describing the system and, it does echo many of the same things I've found in the "Spirits"---but one thing is very glaring in relation to the Jordan JX92S drivers, that the reader should be aware of.
The Jordans are absolutely ruthless to anything put in series or parallel to them Even using a passive line level cross over, reveals the quality (or lack of ) of the capacitors/resistors being used. A "J" spec (+/- 10%), 104 cap (.01 uF), results in a roll off in the bass of 3dB per octave starting at 200Hz (-6db by 50 Hz), but at the loss of the Jordan "sparkle". The JX92S drivers lose there ability to emotionally involve the listener, to the point of preferring no cross-over at all on the Jordans.
What has been your experience with the Jordans in this regard? Using an active crossover/biamp system could very well be the answer, but depends somewhat on the quality of the crossover. I see that you use a 4th order LR @100Hz for the bass, and a 2nd order LR @ 350 Hz for the Jordans.
The subs we have used 12" M&Ks in 3.25 cu.ft. sealed enclosures (originally to be aperiodic), each with it's own 250W plate amp, and low pass crossover. I am surprised your best results were obtained with a crossover frequency as high as 350Hz for the Jordans, and 100 Hz for the Goldwoods.
Looking forward to reading more about your experiences, and thank you for another very well explained and documented project.
stew
Re: Martin, crossover frequency....
I think it was 2nd order LR @ 100 Hz for the bass, and 4th order LR @ 350 Hz for the Jordans.
By the way, what I don't like about higher order crossovers (active in this case, 4th order) is that signal must go through more capacitors and, worse, opamps. If you use J-fet buffers instead of opamps, this could be remedied to an extent.
Vix
Nanook said:
I think it was 2nd order LR @ 100 Hz for the bass, and 4th order LR @ 350 Hz for the Jordans.
By the way, what I don't like about higher order crossovers (active in this case, 4th order) is that signal must go through more capacitors and, worse, opamps. If you use J-fet buffers instead of opamps, this could be remedied to an extent.
Vix
Re: Martin, crossover frequency....
Hi,
Both c/o functions are 2nd order L/R electrical originally, 100Hz for the low pass, 250Hz for the high pass, moved to 4th order L/R 350Hz later.
The acoustic functions are different, they are not L/R or 2nd / 4th order and the effective acoustic crossover point is ~ 300Hz.
You should not be surprised, but you are correct that such settings would not work with a sealed box bass end open baffle mid / treble.
I would be surprised that L/R's are optimum, but they do allow standard active c/o's to be used. The passive version will not be L/R I assume.
/sreten.
Nanook said:
Hi,
Both c/o functions are 2nd order L/R electrical originally, 100Hz for the low pass, 250Hz for the high pass, moved to 4th order L/R 350Hz later.
The acoustic functions are different, they are not L/R or 2nd / 4th order and the effective acoustic crossover point is ~ 300Hz.
You should not be surprised, but you are correct that such settings would not work with a sealed box bass end open baffle mid / treble.
I would be surprised that L/R's are optimum, but they do allow standard active c/o's to be used. The passive version will not be L/R I assume.
/sreten.Re: Re: Martin, crossover frequency....
I am not sure I am buying into all of this discussion of op-amps or passive circuitry in series or parallel automatically significantly degrading the sound of a system. If each is designed carefully it should do its job and any degradation to the signal would be a minimal effect compared to everything else going on. My position should be no surprise since I am a strong proponent of BSC filters and the improvements they provide which the audio purists immediately condemn. If you get to the point where this is the dominant problem in your system and room you have achieved an unbelievable level of performance. Based on the drivers, designs, and room I am working with I am no where near that level and my safe guess is that nobody else is either. I think this type of discussion is a convenient distraction from bigger issues in speaker design. It may be the thing that is easy and convenient to blame for other problems we have not identified or understood. If focused on exclusively it is pure snake oil. That is my opinion.
Vix said:
I am not sure I am buying into all of this discussion of op-amps or passive circuitry in series or parallel automatically significantly degrading the sound of a system. If each is designed carefully it should do its job and any degradation to the signal would be a minimal effect compared to everything else going on. My position should be no surprise since I am a strong proponent of BSC filters and the improvements they provide which the audio purists immediately condemn. If you get to the point where this is the dominant problem in your system and room you have achieved an unbelievable level of performance. Based on the drivers, designs, and room I am working with I am no where near that level and my safe guess is that nobody else is either. I think this type of discussion is a convenient distraction from bigger issues in speaker design. It may be the thing that is easy and convenient to blame for other problems we have not identified or understood. If focused on exclusively it is pure snake oil. That is my opinion.
Quite. I was all against BSC filters on my Jordan system until I tried one which was properly adjusted to the enclosure. It made a significant difference for the better. Earl Geddes is reported to get exemplary results using a properly matched speaker - room design and only modest electronics.
But that's OT - thanks Martin for a well researched article and a splendid read.
But that's OT - thanks Martin for a well researched article and a splendid read.
MJK said:
I hesitate to give you actual *advice*, Martin. I don't design speakers. This may all be wrong, but at least it's food for thought.
The driver becomes directional at around 1kHz (driver diameter = 1/3 wavelength), so we're only concerned below that. Is the magnet structure and cover acoustically transparent below that? I don't know how to figure that out. But if it distorts the polar response by blocking the on-axis rear wave, then the monopole field will have too much energy elsewhere and move the cancellation plane.
You could simply measure the off axis response all the way around the speaker and see.
This might help a bit: http://www.linkwitzlab.com/models.htm#D
But it also made me think of another issue. The Jordan is offset from the edges so the dipole distance varies from perhaps 18" to the furthest corner and 7" (minimum to one edge). Fequal ranges from 125Hz to 325Hz (0.17v/D) and the 6db peak before the comb filtering starts (0.5v/D) ranges from 372Hz to 957Hz. Whatever the actual frequency, which seems to lie between the crossover and beaming frequencies, there will be comb filtering above it.
What to do about it if a problem? How about living with the design as a quasi-dipole with a little more ambient sound than theoretically possible. If I'm correct about the large magnet distortion the dipole field, choose a driver with a smaller motor. Maybe the resonant peak caused by the cavity is a good thing, offsetting a null above Fequal.
FYI, I've had a pair of Orions for nearly four years and expect to keep them until I'm old enough for hearing aids. Even with all the nasty opamps, nothing I've heard in the past 40 years gives me more enjoyment and, I believe, more accurate reproduction.
- Eric
I really have to wonder if we are thinking and talking ourselves into a problem that does not exist .... because we are so smart. Compared to room interactions and baffle edge interactions and everything else going on I guess I am struggling to see how this would stick up as the biggest issue, maybe even worse then the op-amp problems. Every driver has a frame and magnet structure so it should be to some degree a common accurance. Look at the magnet and frame structure on the back of my Lowther PM2A which seemed to work fine.
I am not trying to shoot these ideas down, I am just trying to play devils advocate. So many times in audio we rationalize rocket science explanations for something and it is just not that significant. We totally focus on minutia and neglect the simple things we can do to improve a system's performance. My experience leads me to question things that sound complex or seem to elaborate, most of the time the simple answer is correct.
I am not trying to shoot these ideas down, I am just trying to play devils advocate. So many times in audio we rationalize rocket science explanations for something and it is just not that significant. We totally focus on minutia and neglect the simple things we can do to improve a system's performance. My experience leads me to question things that sound complex or seem to elaborate, most of the time the simple answer is correct.
If you're worried about the rear wave obstruction there are two reasonably simple solutions to improve the situation. The first is to chamfer/roundover/bevel the area between the mounting screws on the back of the baffle. The second way is to roundover the driver cutout on the front of the baffle and then mount the driver from the rear. This is the more aesthetically pleasing route and is very easy to do if you have a router and a roundover bit.
Xover questions for MJK, mods: please move this post if in an inappropriate
I had written my notes on the paper correctly but (almost always) as per usual, mistyped the comments (I am ADHD and dyslexic so typing most anything is a chore). Yes Jordans:350Xz, LR 4th order,---subs :100Hz, 2nd order LR.
Our plate amps are from the M&K VX1250 , at 250 watts each driving a M&K 12" sub (from the same system). Unfortunately integrating them with the Jordans is becoming a bit of a handful. The potential is there, we need to basically provide a high pass to the Jordans, mainly to gain a little more efficienct use of the amplifiers available and limit the excursion (and improve the midrange). Just because a driver has 9mm Xmax doesn't mean you should use it, and we're wanting the Jordans to be the best possible in this situation. We've aimed at a 12 dB passive line level crossover, to that end. as Fs is stated at 45 Hz, (and the impedance plot of our matched drivers clearly indicates this), so we are 4X's Fs+ above where our slope begins. Due to the input impedance of the amp that is being used for the Jordans of 22kΩ, it had been calculated that only a .01uF cap in series with the signal to the "Jordan amp" should provide us with that.
Regarding the quality of components upstream of the Jordans and individual capacitors and resistors, I used a "104 J" cap, which is +/- 10% and of good quality. We simply need to know if this can be done if we go to Mudorfs, etc, or do we go to an active electronic crossover.? Using these caps killed the life from the Jordans.
If so, Mr. King , can any suggest something that could be DIY or a known good crossover that does not impart any character to the signal (I understand the Behringer products are quite high performing for the money, but may need a modification to sound their best), that ir reasonably priced, or do we (me and my audio partner) simply go to something like the Harrison Fmods, and hope for the best?
thanks
stew
I had written my notes on the paper correctly but (almost always) as per usual, mistyped the comments (I am ADHD and dyslexic so typing most anything is a chore). Yes Jordans:350Xz, LR 4th order,---subs :100Hz, 2nd order LR.
Our plate amps are from the M&K VX1250 , at 250 watts each driving a M&K 12" sub (from the same system). Unfortunately integrating them with the Jordans is becoming a bit of a handful. The potential is there, we need to basically provide a high pass to the Jordans, mainly to gain a little more efficienct use of the amplifiers available and limit the excursion (and improve the midrange). Just because a driver has 9mm Xmax doesn't mean you should use it, and we're wanting the Jordans to be the best possible in this situation. We've aimed at a 12 dB passive line level crossover, to that end. as Fs is stated at 45 Hz, (and the impedance plot of our matched drivers clearly indicates this), so we are 4X's Fs+ above where our slope begins. Due to the input impedance of the amp that is being used for the Jordans of 22kΩ, it had been calculated that only a .01uF cap in series with the signal to the "Jordan amp" should provide us with that.
Regarding the quality of components upstream of the Jordans and individual capacitors and resistors, I used a "104 J" cap, which is +/- 10% and of good quality. We simply need to know if this can be done if we go to Mudorfs, etc, or do we go to an active electronic crossover.? Using these caps killed the life from the Jordans.
If so, Mr. King , can any suggest something that could be DIY or a known good crossover that does not impart any character to the signal (I understand the Behringer products are quite high performing for the money, but may need a modification to sound their best), that ir reasonably priced, or do we (me and my audio partner) simply go to something like the Harrison Fmods, and hope for the best?
thanks
stew
MJK said:
Agreed. But all you need to do is measure all around the speaker off-axis to see if it's an issue of significance or not. Maybe dipole behavior isn't a major prerequisite for this design anyways.
Another option is a larger driver. Here's a shot of the back of the driver used in the Orion. But if one of your criteria is to use a full ranger, then your options are limited.
- Eric
Attachments
I think the phrase is 'that's life. Live with it.' Nothing's ever perfect. Good enough for the job is good enough for the job. Perfect is usually a PITA, and rarely worth the extra effort you put in.
One of my favourite engineering stories is from Stanley Hooker. When they were designing the Olympus turbojet (think TSR2 & Concorde) somebody kept insisting on total perfection, which to their mind was a 1,000 mile full-throttle range. Did it actually require a 1,000 mile full-throttle range for any of its applications? Nope. 800 miles was perfectly sufficient. Stanley's comment: 'why are you insisting on 1,000 miles? You do know that for the last 100 of those, it's going to cost about £1million per mile in development costs?' Turned out that was an understatement, but they had to keep on plugging away, pouring money down the drain for no practical reason.
I know what you mean -champhering the rear of the baffle might be useful for the midrange as that motor is big & does block things a bit. The devil, to an extent, is in the detail. But beyond that, I wouldn't get overly worked up. Life's too short.
One of my favourite engineering stories is from Stanley Hooker. When they were designing the Olympus turbojet (think TSR2 & Concorde) somebody kept insisting on total perfection, which to their mind was a 1,000 mile full-throttle range. Did it actually require a 1,000 mile full-throttle range for any of its applications? Nope. 800 miles was perfectly sufficient. Stanley's comment: 'why are you insisting on 1,000 miles? You do know that for the last 100 of those, it's going to cost about £1million per mile in development costs?' Turned out that was an understatement, but they had to keep on plugging away, pouring money down the drain for no practical reason.
I know what you mean -champhering the rear of the baffle might be useful for the midrange as that motor is big & does block things a bit. The devil, to an extent, is in the detail. But beyond that, I wouldn't get overly worked up. Life's too short.
I am using a dBx Driverack Pro crossover with XLR cables used for all interconnects. I believe that this unit is similar in function and probably a direct competitor for the Behringer DXC2496. No idea which is better. If you wanted to spend more money a Marchand crossover might be an option. Not sure if a Marchand is better but it definitely is not as flexible as the digital crossovers.
I don't know anything about DIY active crossovers, line level crossovers, or DIY electronics. I am only really interested in speaker design and acoustics.
as a slight refinement to the technique shown in the image attached to post 55 above, when I chamfered the back of the cut out hole for this type of driver, I would do it by hand with a Stanley "Sureform" (rasp) , the one with the curved cutting surface, and cut away only the areas between the bolts - leaving lots of meat for the bolts/screws, but cutting away most of the rest to almost a knife edge.
I can't remember having done a listening comparison, though. This was 35 years ago!
I can't remember having done a listening comparison, though. This was 35 years ago!
Eric Weitzman said:
The opening is probably a little better then what is depicted in the picture, the angle I took the picture makes it look smaller.
But in any case, what I ended up with is a 3/4 inch long annular TL behind the driver. The tuning frequency for a 3/4 inch long TL is about 4500 Hz. At that frequency the acoustic impedance at the open end of the TL, based on the available open area, is probably purely resistive. So sound is not reflected back to the driver and is efficiently transmitted into the room. Below the tuning frequency air just moves as a slug in and out of the cavity adding a parasitic mass load to the back of the cone.
That is my simplistic thinking about what happens in this region. Maybe I am right or maybe I am full of crap. The sound is all that matters.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.