I'm not sure this is true. Once you send a signal thru a triode (6SN7 for example), our spectrum analyzer showed a fast fall off of harmonic distortion products relative to any of the solid state amps we looked at, and the 2nd harmonic (arguably the only desirable one) was dominant. If this triode is outside of any negative feedback loop, its distortion signature should still be there at the speaker, and theoretically dominate over the smaller but ugly sounding distortion of most solid state gear. I tend to think of this distortion spectrum shape as being the main magic of tubes. Maybe I'm fooling myself. There's of coarse also the soft clipping, the usually bigger dynamic range, and a few other things that I've heard about but not verified. Putting a tube in the path could "warm" things up a bit by adding some 2nd harmonic (same note an octave higher), below the level of consciousness, but present enough to affect the feel of the sound.
Not to get way off original subject, but here's more of my findings on tube distortion vs. transistor distortion, if anyone is interested:
http://www.spiritone.com/~rob_369/audio/Distortion%20Spectrum%20Shape.htm
http://www.spiritone.com/~rob_369/audio/Distortion%20Spectrum%20Shape.htm
It's just the gain structure, not "tube pre amp" per se.
We had to crank up the gain which then caused clipping and all (or was is to reduce..... I can't remember).
If the preamp was pre-DSPxover, why would you need to crank up the gain of the DSP crossover?
Or are you referring to cranking up the gain (volume) of the tube preamp?
Regardless, that portion of your test description doesn't make much sense. If your tube preamp(s) had less voltage gain than your solid-state preamp(s) then it's clearly an issue that's not related tube/transistor related and the description is confusing.
Cheers,
Dave.
Or are you referring to cranking up the gain (volume) of the tube preamp?
Regardless, that portion of your test description doesn't make much sense. If your tube preamp(s) had less voltage gain than your solid-state preamp(s) then it's clearly an issue that's not related tube/transistor related and the description is confusing.
Cheers,
Dave.
Well in that first test the guy (Keith) used various ultra-expensive speakers, tube amps, tube preamp. Cary amps, Acapella Violon, etc. So we let him "do his things"... there was no way I can pay him if something goes wrong !! 
So yes, both three of us could not understand why it did not work and distorted so much so we discarded the test. But at least I was not $80,000 poorer lol.
The second test we used a DIY unity system which we were more comfortable to play around with and encountered no issues.
So yes, both three of us could not understand why it did not work and distorted so much so we discarded the test. But at least I was not $80,000 poorer lol.
The second test we used a DIY unity system which we were more comfortable to play around with and encountered no issues.
I certainly understand and value your opinion. But SL chooses not to post here, does that mean I can not? He can comment on his Orion group. I don't go there. It is his turf and I respect that. But this is DIY Audio, and as I have said several time, it's an open board and if the moderators feel my comments are out of place or out of line they will certainly let me know. On the other hand, I am discussing the LX521, and in doing so, I have looked at technical issues and in some cases made comparisons. Why not? It should not be hard to look at what I post objectively because my posts generally deal with issues which can be objectively quantified; excursion, group delay, power distribution. When I have expressed a subjective opinion, I place it in the context of what I experience in my system with only a suggestion that it might be beneficial in SL's. SL is probably more likely to read them in a post about his speaker than mine. And if he believes the comments have merit, he is free to experiment with them in his system.
Just for a moment take comments about my speaker out of it. Does than make my discussion of the order of the coupler with regard to GD and excursion or acoustic overlap any more or less relevant?
I really don't think it is thread high jacking when the issues I raise directly relate to the design of the speaker in the thread tittle. And I think if the same issues were bought up in a thread about my speaker then I would still be accused of trying to denigrate SL's design in favor of mine. In that sense, I will always be in a no win situation because there will always be this undertone that I am in competition with SL, jealous of SL, or just that what ever SL does or say is always going to be the only correct/best answer/approach.
And should anyone remain offended by my making comparison to my system let me suggest you turn on you TV and I'm sure you will not that Comcast advertises on Dish and Direct TV, and ATT Uverse and vise versa. And they all compare one another's service to the other in their commercials and they all point out what is better about theirs......
And once again, the thread seems to be headed OT. Certainly my comments have not been OT.
Anyway, It's been an interesting and enlightening discussion. By partaking in it, I've discovered things about my own design that I had not previously considered and been able to improve, though acousticaly I'm not convinced they resut in any difference. And those discoveries are out there in the public domain for anyone to try. Isn't that what DIY Audio is suppose to be about?
The main thing I'd worry about with the Linkwitz LX521 is the issue of rotating beaming with the one pole crossover. Not that it's necessarily a big problem, but at any crossover point you have two drivers putting out the same thing from different physical locations relative to your ear, which opens up the probability of comb filter effects. As you sweep a sinewave through the crossover region, the two acoustic signals will shift in phase, which is time, causing the maximum amplitude addition to move vertically (with vertically aligned drivers) depending on how their phases add at any given angle and frequency. I hope that makes sense. With a one pole xover, you've got a fairly wide band of frequencies that will be affected by this. Maybe 2-3 octaves, often close to the frequencies where our ear is most sensitive (800 - 6kHZ).
Below 800HZ, image location is all about timing comparisons, left to right. If in that range of frequencies we have two acoustic signals with different phase timings, it could be argued that you've blurred the imaging cue information, much like inter-aural crosstalk blurs and effectively scrambles much of the imaging cue info that our brains try to interpret in order to place the images in the soundfield.
Apparently David Griesinger (top engineer at Lexicon) has written a paper recently where he argues in favor of no crossover points between about 200HZ and 6kHZ, but whoever put it up here has it in a filetype that I've never heard of. I'd be interested to hear what he says about this.
Below 800HZ, image location is all about timing comparisons, left to right. If in that range of frequencies we have two acoustic signals with different phase timings, it could be argued that you've blurred the imaging cue information, much like inter-aural crosstalk blurs and effectively scrambles much of the imaging cue info that our brains try to interpret in order to place the images in the soundfield.
Apparently David Griesinger (top engineer at Lexicon) has written a paper recently where he argues in favor of no crossover points between about 200HZ and 6kHZ, but whoever put it up here has it in a filetype that I've never heard of. I'd be interested to hear what he says about this.
There is a danger of misleading using a photo like this when we don't know what the studio tech does before actually mixing something. In other words, this is a photo oportunity, not a setup for actually mxing something.
Last edited:
Why would that be? I'm not convinced. It seems to me that you could insert the tube distortion spectrum before or after the A_D_A process and still have it at your ears.
A correction on my above mention of David Griesingers recent paper on phase issues:
He states that imaging clarity is significantly improved if there is consistent phase from 630HZ - 4kHZ, if I understood his paper correctly. (I got a software upgrade so I could view the PP presentation).
I believe he WAS refering to the practice of dropping the sound back onto a reel to reel to "tapify" the sound. We used to do it all the time in tape cal class to show the students what tape can do to alter the sound. I would not say it is common practice, but there are those who do it. Of course now that there are tape emulation plug-ins that are really pretty convincing, this practice will die out for all but the most nostalgic of folks.
Greg
PPTX = Microsoft Power Point 2007 with XML format
Interesting paper -- SL owns a Lexicon MC-8 (though he does not use it currently) -- there is a DIYAudio thread that covers this paper HERE.
The file format *.pptx is Microsoft Power Point 2007 in "XML" format (the 'x' in *.pptx)
On JohnK's commentary on the LX521 -- I used to be (many years ago) in the "annoyed" camp, but now I welcome his commentary -- there is nothing 'wrong' with stating you would design it differently and why, and I think that all of us that build SL's systems understand that he offers a set of designs that compete with SL's offerings, and can put these comments into that context. As I like to experiment, it will be interesting to see if making a second order crossover for the LX521 lower-mid to upper-mid makes any big changes -- as I am on the eve of building this loudspeaker (as soon as all the parts are available) I will compare them to my ORION's and then tweak them, but only after listening to them as SL intended for many months -- but all ideas are (or at least should be) welcome....
Charles
[SL's former Brazilian Orion builder]
Interesting paper -- SL owns a Lexicon MC-8 (though he does not use it currently) -- there is a DIYAudio thread that covers this paper HERE.
The file format *.pptx is Microsoft Power Point 2007 in "XML" format (the 'x' in *.pptx)
On JohnK's commentary on the LX521 -- I used to be (many years ago) in the "annoyed" camp, but now I welcome his commentary -- there is nothing 'wrong' with stating you would design it differently and why, and I think that all of us that build SL's systems understand that he offers a set of designs that compete with SL's offerings, and can put these comments into that context. As I like to experiment, it will be interesting to see if making a second order crossover for the LX521 lower-mid to upper-mid makes any big changes -- as I am on the eve of building this loudspeaker (as soon as all the parts are available) I will compare them to my ORION's and then tweak them, but only after listening to them as SL intended for many months -- but all ideas are (or at least should be) welcome....
Charles
[SL's former Brazilian Orion builder]
I appreciate you comments Charles. Might I add that in regard to Bob's comments about the 1st order x-o I think it is important to recognize that from what is discussed on SL's site, the LX521 is not using a 1st order crossover but rather 1st order HP and LP filters with staggered poles so that the crossover is actually at the -6dB point. Then with one driver connected with inverted polarity and with the AC offsets, the crossover behaves like a LR type crossover, summing in phase, at least over a portion of the overlap region. This results in fairly symmetrical vertical lobing at the crossover point. With my system I find that vertical polar does depart from symmetric as the frequency rises above the crossover point because the phase does not remain aligned. With 2nd order its gets a little better and, as would be expected, with 4th order even better, due to the decreased region of overlap as the order gets higher. So all in all, with different order couplers there are a number of issues: excursion, group delay, overlap, and vertical polar response. It is really difficult to say which ultimately will play the most important roll and what trade offs each presents.
One other point I would add specifically for SL, should he be reading this, is that since SEAS is making custom drivers for him, he might ask that they make a lower Qes, Qts version of the upper midrange with increases sensitivity. I noticed that the SEAS upper mid he is using is quoted at 85dB/2.83v. I also noticed that on his site where he shows the preliminary coupling of the upper and lower mids the lower mid has a resistor in series for attenuation. The SS Discovery upper mid I am using is rated at about 90dB/2.83V and with the baffle design blends perfectly with the lower mid without any need for attenuation. Obviously I do not know what the situation is in SL's final design, but the observation that the dipole peak shifts up in frequency for the small upper mid driver/baffle also indicates that it begins the dipole roll off sooner and is effectively less sensitive than the lower mid in the dipole region. This suggests that the upper mid should be of higher nominall sensitivity than the lower mid to start with. In my case this is why I choose the 4 ohm version of the 10F to be mated with he 8 ohm version of the 22W.
Is this called a solen split? I think I read about this once.
It always made sense to me to do it this way.
Since (according to Madisound) neither of the drivers (SEAS MU10RB-SL, H1658-04, Curv cone and SEAS U22REX/P-SL, H1659-08, Curv cone) is currently available ("going into production mid December and will hopefully be in stock before the end of the year") and there may even be some variation from the not-fully-described performance of the prototypes demonstrated at BurningAmp we're all just going to have to wait until samples from the actual production run are in hand "to know what the situation is" or seriously discuss details of their performance. I do look forward to seeing your assessment of them once you have them in hand and have had a chance to test them on their intended baffle (or some functional equivalent).
I'm particularly interested to see whether UE comes up with a crossover topology that gives better overall integration of the two into a "wide-midrange" than that accomplished by the simple passive crossover used in the LX521 design. I suspect that many of us would be happy to mod the design to digital-active-4-way (LX521-D ? ? ?) should that further improve the already excellent sound.