In Reply to: Re: Use which high Effiency Speaker with which company Tube Amplifier by Using Digital X-Over to Bi or Tri Amp Speaker posted by sanjaysahani on January 15, 2001 at 06:30:32:
Hello again Sanjaysahani!You wrote:
> > Hello please suggest chain of components,
> > Please also suggest suitable best of best world
> > class companies making the following below
> > mentioned products :I think you will find as many recommendations as there are people in this forum, and probably better than half of them will have merit. So there's several ways you can do this thing.
My main focus is on the technologies involved, and on your choices between them.
While I understand the tube/transistor thing - I'm a big fan of bipolar amps. We don't have to worry about mechanically modulating the elements - so we never feedback or "ring" the grid. And really, the main difference is the issue of what happens in the two types of amplifiers is what happens at zero crossing and at clipping.
The tube amp clips "gracefully" and doesn't turn the sine wave into a square like the semiconductor does. Its gain is non-linear - especially at the extreme "edges" - which makes it "feel" as though it clips less. Some express this as the harmonics thing - the simple harmonics are expessed with high order harmonics less represented. They say it as transistor amps having more high order harmonics.
But that's a reductionistic view. There's an easier way to view it - at a holistic level - that gives a clearer picture. It's the simple fact of non-linear gain, and particularly at the extremes. The transistor amp will provide full gain to clipping or zero crossing and will then cleanly stop - zero gain. The tube amp will have begun to reduce gain near these limits, and so will "round off the edges" of the clipped signal.
So while we can express these in terms of harmonics - that's because square waves are comprised of certain (multiple) harmonics of the fundamental. Mathematically, we can express this fact. But a better way to view it is to look at the signal, and see what's happening at clipping.
In my book, I'd much rather have this compression effect applied by design than as a by-product of the technology chosen. Just make sure that the amp is powerful enough to never enter clipping, and do not run the preamps with a signal that approaches it.
Another method that can be employed is to use a signal compression device to begin compression near the limits. The "standard" preamp signal is 0.775 volts, and one could install a compressor that was effective only at signals of 0.700 volts and above - and then made certain that compression grew more as voltage neared 0.775 - never allowing a signal to exist greater than this.
This would be very attractive for making "tube sound" and still having transistor power.
Still, the differences between Class A and Class AB amplifiers are still there; The differences of bandwidth and linearity are still there; And the differences of harmonics below cutoff are still there too.
But for my money, I'd buy a large bipolar amplifier such as a Crown or Hafler.
Of course - consider the source - I'd build or buy a set of Pi Speakers, and I'd use the best motors I could afford. That would probably be a 2227 and a 2426. The type of enclosure would depend on my listening room. If it were very, very large - I'd run the folded horn. If it were mdeium sized and had good corners, I'd run the corner horns. And if it were very small, I'd run the bass reflex.
If you want digital signal modifiers - I can't help you here. Someone else may have recommendations, but digital designs - by their nature - sort of "level the playing field." There isn't as much difference in quality between the top end gear and the mid-priced gear because the circuits involved don't lend themselves to improvement (or to reduction of quality) by the selection of components. It's all pretty much the same, and the better units offer more features without considerably better performance.
One can certainly cut corners on D/A design, but not by much. That sort of removes the possibility of getting a really bad design. Probably the worse "offense" a designer can make is to fail to take into consideration the upper limits of the system, and to ensure that what is produced grows increasingly more "rounded." Failure to do this can cause the top octaves to be "squared off" or present sampling "spikes" - but even here, we have some of the same sorts of "pseudo effects" that the tube/transistor debate considers.
For example, if we have an 18Khz sine wave that is presented to a A/D converter and then is reproduced poorly by a D/A converter as a sort of "stepped square wave" - it can be shown to have created lots of harmonics. But even the second harmonic is 36Khz, so the loudspeaker can't play it and if it could - only our dog could hear. Perhaps the psychoacoustics guys will have something to say, but I expect that we'd be talking about the "number of angels on the head of a pin."
So as far as your digital signal processors go - choose a good brand name of your choice, and pick the features you want. Quality will be similar among them all.
If you go digital on the front end, you'll be able to get all the "little do dads" without any further effort. But if you go analog, you will have to pick and choose.
Pick a parametric equalizer with as many "tank sections" as you can afford. Having said that - You'll probably only need to use one frequency and maybe two to EQ out the problems in any room. I've had situations where an auditorium had more than this, but not a home theater environment. And even in the live production situations where there are more than a couple of anomolies - usually at least one of them is a "wide band" effect that is better to EQ with a high pass or low pass filter (Bass or Treble control).
So I don't suspect you'll need more than three sections on a parametric EQ.
Active crossovers are always nice, since they do the driver splitting at before the amp instead of after. Particularly important for the tube guys, in my opinion, where we want to squeeze all the bandwidth we can out of our systems. If I were to run tubes, I'd probably run the midrange and treble with my tube amp and the lows would go out bipolar, making sure that I limited the bass bandwidth to keep my harmonics limited to the tubes.
My preference for bipolar does not mean that I discount the tube designs. I understand why they hear a difference, what it is that they hear, and how to go about achieving it. So in order to accomplish this, if it is my goal, I'd make sure that its compression effects were realized by using the bass amps for only the bottom octaves and make sure they never clipped. They'd be gain-matched, of course, but well oversized and underdriven. And the bandwidth would be limited too, so that harmonics could not be developed. Run an inductor in series with its output in addition to the active crossover on the input.
On the subject of active crossovers and digital devices (or all kinds), it is interesting to note that while I maintain that a harmonic of the (error prone) high frequencies is a non-issue (because it can't be heard) - I also suggest that the tube amp will better "reshape" an affected signal. Lack of bandwidth and a natural compression tendency will round-off the edges.
As funny as it sounds to me - mating digital with vaccuum tubes - the effects will be sort of complementary.
Then again, maybe getting those 35Khz super tweeters and a vaccuum tube amp to drive them will add "subtlety." Run a digital sound processor to ensure that phase angles are restored and DC couple it to the grid resistor. Put a monster bipolar amplifier run in that weird "class X" - whatever it was for the peaks. It had some class B looking "clip section" that augmented an AB style push pull. Yeah, that's the ticket.
Just having fun guys...
Wayne
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- Recomendations - Wayne_Parham 01/15/0110:46:17 01/15/01 (0)