Welcome! Need support, you got it. Or share your ideas and experiences.
Return to Planar Speaker Asylum
173.239.212.43
In Reply to: RE: ET LFT-8b in Harry Weisfelds system posted by BDP24 on June 20, 2017 at 03:33:34
He may have been going for phase linearity/minimal crossover components. Had to believe he isn't aware of the issue.
Follow Ups:
Yeah, there's no doubt that was his objective. However, with this type of large/hybrid speaker it's not really achievable anyway. I prefer to have crossover regions sum correctly with the selected filter slopes regardless of whether it achieves a linear-phase result or not.
The woofer issue is something different. Ideally, the Peerless woofer would be electronically equalized at the low end without the weight modification and thus somewhat alleviating the cone resonance issue. However, that requires a different (non-conventional) approach to the whole scheme......which is still something that many audiophiles reject.
Cheers,
Dave.
I didn't know he'd weighted the woofer. That does seem a problematic approach. Magnepan gets away with it because they can segment the woofer and use the tuned resonant sections to shape the response to achieve something reasonably close to 6 dB/octave dipole equalization.
Even then, I think that electronic equalization is preferable. But, as you say, audiophiles aren't wild about electronic equalization and it does stress amplifiers and woofers and add to cost. Still, I assume you've corrected it in your own system? Arguably, woofers should always have electronic EQ anyway, it just isn't possible to get smooth response in a practical room otherwise.
When my system gets to the point at which I can worry about some things I'll use the Mini DSP to equalize the woofers and then I'm thinking about using secondary woofers to take out the dips from backwave cancellation.
there is a problem of latency and group delay if you eq beyond 1st order. Thigpen is opposed in principle to having detached or delayed time behavior in any of the drivers as that eliminates transient fidelity of the whole speaker and phase fidelity at the crossover region. He thinks it is as important and possibly more so than freq response, noise or resonances. For those that appreciate spatial performance most these are the things that must be preserved, while everything else can be compromised.
In my own setup I started with the amazing spatial performance and textural detail of the Neo8 lines and tried to splice in the LF and HF drivers so as to preserve that performance. 1st order symmetrical in an equidistant arc was ideal for freq response vs. spatial performance with xo at 250, 5000Hz but there were "issues" of a general thickness at the lower mids and some loss of instrumental texture. Raising the upper xo to 11-12kHz solved them nearly completely but left a serios head in a vise problem that I just could not live with. So I tried the Apogee ET solution and dropped the mid's LP filter entirely and got all of the single driver performance back and no real head in a vise problem. So I am officially in that camp.
On the woofer xo end I tried to go as low as 100 Hz symmetrical but the lower mids became MIA so despite terrific spatial performance, the woofer had to dominate below 200hz since the line array was just too weak there. But at 1st order above 150hz the bass panels muddy the low mids and create a sense of general pressure. Tolerable but not ideal but preserves the single driver spatial performance.
When trying out wall loading schemes the bass can' be perfectly timed with the mids anyway so I went to steep LP filters and time aligned the deep bass pass band and mids by offset placement. Lost just a bit of spatial performance but removed the mud and the general sense of pressure. That is another usable compromise like the 1st order one just with a different balance of strengths and weaknesses..
I know what you mean about the mud, having fixed the XO problem with the left woofer HPF filter and heard notably less of it.
This is one reason I'm eager to try FIR filters with the MiniDSP, phase linearity and a sharp cutoff, assuming I can do it without too much latency. Pre-ringing shouldn't be a problem once the sound recombines though differences in speaker placement and dispersion could result in imperfect summation.
It seems to me that if you want to run 6 dB/octave filters between the Tympani woofers and midrange you have to do some tinkering with the woofers themselves. Specifically, you'd probably want to direct the higher frequencies to a narrow section of the woofer (or just one section depending on how the dimensions work out -- which is already partly true due to the acoustic EQ). That would fix power response and minimize moving mass though it wouldn't favorably alter the mass density since you'd just be driving a smaller section of the woofer, meaning it wouldn't improve air damping. Another issue is that there's still a LF resonant section down at the bottom of the midbass panel so you could impair LF performance.
"there is a problem of latency and group delay if you eq beyond 1st order. Thigpen is opposed in principle to having detached or delayed time behavior in any of the drivers as that eliminates transient fidelity of the whole speaker and phase fidelity at the crossover region. He thinks it is as important and possibly more so than freq response, noise or resonances. For those that appreciate spatial performance most these are the things that must be preserved, while everything else can be compromised."
Conjecture....regarding his principles??
FYI, the fixes I implemented do not alter this aspect of the design anyways.
Dave.
Things he spoke about a couple of decades ago.
Okay. Well anyways, the fixes that need applied don't change that aspect of the design.
Dave.
Davey, would you mind posting your parts list for the woofer fix again? For those who didn't see it when Davey posted it a while back, you will see it is not a x/o, but more of a notch filter, I believe it would be called.
No, it's not a notch filter.
It's simply an RC series network wired across the woofer terminals. A 35uF cap and 1.5 ohm resistor.
Dave.
Beautiful, thanks Dave. Any candidates for your LFT-8b replacements?---Eric.
I've heard from a couple of people offline, but there isn't that many (relatively speaking) of these systems out there, and most owners probably don't frequent this insane asylum. :)
Dave.
I meant speakers to take the place of the LFT-8b's in your system. Maggies? Quads? Dynamics of some sort? I have a pair of LFT-4's I could let go I suppose, though I'm looking forward to comparing them to the LFT-8b's.
I don't know your age Davey, but are you old enough to remember the ESS TranStatic I speakers? An interesting design from the very early 70's (before they came out with the Heil AMT speakers)---a KEF B139 woofer in a transmission-line enclosure, a KEF B110 Bextrene 5" midrange, and a trio of RTR ESL tweeters. I have a pair of them as well. Too many speakers!
Oh, I have multiple speakers available to use. :)A friend has a pair of ET LFT-6's that need repair. I might take on that project next.
I do remember the ESS TranStatic speakers, but never listened to those. I spent some time with Fulton J's and E's back then and also a few of the main RtR models that optioned the electrostatic tweeter schemes....ESR-15 and ESR-6.
Dave.
Edits: 06/23/17
He definitely is a linear phase guy and his goal was a transient perfect speaker. Hence his emphasis on square wave performance, which he puts in the manual 100 up to 5khz as the 10khz is visibly rounded.
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: