XXHighEnd

(http://www.stordiau.nl/orelo/IMG_0830a.JPG)
Here you see a glimps of the amplification part of the Orelo MKII.

But why an all-active speaker ?

For some this will not come as a surprise, but this is all about noise. Noise and how it can be eliminated from our "chain";

As we may know, many topics deal with noise as such and how actually devistating it is to the music signal. However, no matter how many times - and where - I ask, nobody is able to eliminate noise to such degree that it should be harmless.
Harmless : not more than the source implies. And this explained more :

The output noise of the Phasure NOS1 D/A converter is at something like -141dB. When this is "gained" with for example 20dB, then the output noise from the power amp should not exceed -121dB. So it means that whatever noise comes from the DAC, this is not worstened by the power amp and better it won't get (unless the noise from the DAC is brought down somehow).

Sadly, everybody (I ever asked) perceives noise from his speakers and usually this is not a 118dB sensitive speaker (like the top horns from Orelo MKII). This means that so much devistating noise must be there (because audible) that this all destroys the good intentions of everything. And for example, might we present that fine speaker then nobody should take the risk that it doesn't perform as how it does over here. And well, "it" is the chain and its apperantly always implied noise.

To remember : What we present anyway is no usage of any preamp or other analogue attenuator; this implies noise itself plus unlinearities which "deadens" the sound. However, this now obviously implies that the power amplifiers are always at full gain.

And so the sheer reason for including the amplification in the speaker is to guarantee that no harmful noise will be your share. And *if* there's noise after all, it can only be from relatively easy to find sources. So just think of it :
You have a NOS1 (so that assumed) with a pair of interlinks into the speaker, and that is really all. This is noiseless over here and so there is no reason it should be noisy at your place. But if it is it should be easy to find.
And notice that in no single circumstance so far, from the distance, "we" could eliminate audible noise completely from "inherently" noisy systems. This in other words, no matter the experience and knowledge we have ourselves on the matter, it appears undoable - and despite all the guidance provided - to make systems noise free at "your" place. And oh, we will get somewhere (improve) but what always remains is the uncertainty about the specific devices in use and how noise free they are themselves, or how they may imply noise (groundloops) in combination with other connected devices.
Not so with the Orelo MKII and NOS1 D/A converter combination because this is all under our control.

---

Assumed that our (human) dynamic range is something like 72dB, roughly put any audible noise perceived should be play at -72dB. Edit : What I mean here is what we can perceive from loudspeakers while windows stay in. This needs more outlay to fully understand (and needs incorporation of the speaker sensitivity), but something like that. Now that assumed, we can see that system noise which can be at -121dB now will have added 49dB of noise. Now I don't know whether I must show a signal with 49dB of noise riding on that, but maybe you should not see it because it looks monstrous.

Though I should prove it with plots, the output noise of the amplification for the woofer section is at -119dB. This needs more "reasoning" (actually math) for myself, because first off this includes "DSP noise", secondly it also includes "preamp noise" from the mid-section (that's how it works) and lastly it incorporates the DSP attennuation which is needed because of too much headroom in the bass section (two bridged amplifiers feed that). So generally think of the mid(-high) section doing what it needs to do (at 118dB of sensitivity) and the woofer section is brought in (measured) balance with that. Anyway resulting output noise is at -119dB assumed I did the analyser calibration (math) right which is also not an easy job because it depends on the frequency (DSP implied). Possibly it's best to say that the -119dB counts for 19Hz.

What I did not measure yet is the output noise of the mid and high section, but theoretically this should be at something like -125dB.
Notice that the gain has been set to 15-16.

People knowing the NOS1/XXHighEnd will know that (at least theoretically) the best SQ will be achieved when the attenuation is the least while enough headroom is present for the softer (for level) albums. So, the speaker (in standard setup) is set to this lower gain just because not more is needed. Think like playing with a digital attenuation of -12dBFS as the loudest (that I experience myself) and then there's still 12dB of headroom for emergency parties.
Less gain also implies less inherent (amplifier) noise, but that's another matter (and helps again).

Might someone want to know ...
The amplification has been built around what we tend to call "GainClone chips" which is just a chip for amplification. But nothing has ever been "GainClone" as such from the BD-Design amplifiers and the specs are way better than datasheets will tell you because of the way it is applied (dig deep in those datasheets and you'll have the hunches). For this speaker a complete new design has been made (on new PCBs) already because all had to fit somewhere, but with again better specs.
Of course this all runs cool at something like 60W + 30W + 30W per speaker when playing, and idle I am yet to measure (but should be less than 20W all together).

---

Summarized

Any noise, no matter how low and no matter how inaudible in itself, is detrimental to good SQ. So what has been done is creating a self-contained system (active speaker plus D/A converter) which is under our complete control and which exhibits the lowest noise levels probably unseen elsewhere. Everything from head to tail is designed/created/soldered in our own house and with all the means available to check all what is involved for not only this low noise but also for good specs. This includes the amplification which for THD measures equal or better than the D/A converter (this is crucial) and which sounds the very best or otherwise it would not be in there in the first place.

When now noise etc. is audible in your home, it can only be external factors and its source *will* be found with the help of us, just because it is so much more easy now.

With the recent development of the NOS1a (upgrade) another source of less under our control noise has been eliminated (noise from the PC in the digital domain, that inflending jitter) and things again got (way) better;
At this moment of writing we are heading for no less than live sound through loudspeaker reproduction.

Lastly and talking about "upgrades", do not forget that the Orelo MKII is fully ready for near-future all-DSP with external amplification and where the DSP itself is not detrimental to the sound (no A/D conversion involved).

(http://www.stordiau.nl/orelo/Inputs01.png)
So have our (with the DAC integrated) DSP module (or your own), drag in a 6 channel amplification, plug in, set the input selector appropriately and play.
(notice that the Orelo MKII will always be shipped with its internal amplification regardless).

Peter

(http://www.stordiau.nl/orelo/IMG_1050a.jpg)

(http://www.stordiau.nl/orelo/IMG_0010e.jpg)special genuine Ferrari Rosso Scuderia
(PPG 229172 / Fer 323)

Impressive and versatile as allways.


Ok four amplifiers for a three way system?
That might be one driving the entire system, one after the fm correction with digital eq for the bass, one after the fm eq an bp for the mid and one with fm an hp for the high frequencies?

Another one of the riddles?!  :)

Regards, Coen

Why would anyone want to use external amplification if the internal amps are so quiet?

Peter mentions bridged amps for the woofers...maybe that accounts for the fourth amplifier?

Yep, that's my understanding. I currently have a pair of BD-Design bridged amps which I believe are similar, if not identical, to those in the Orelo speakers' bass section. They are extremely good full-range amps in themselves.

Mani.

Hey Matt,

This is partly a psychological thing; people exist with 20K+ power amps and the sheer fact that they cost that amount always makes them "better". And maybe they are - who knows. So although the speaker is active it can be used in passive mode just the same. Well, that "the same" is not entirely true because the filters in there now will behave differently. Say not optimally. So it is not advised.

What I didn't tell explicitly (but I did tell it) is that when the full non-detrimental DSP it utilized, this can only work with external amplifiers. So now the internal filtering is replaced by the DSP module and *now* all is OK. So :
[DSP -> DAC (3ch)] -> Power Amp (3ch) -> Speaker drivers Bass/Mid/High (this for each speaker).

Summarized the input possibilities :

1. Line out from DAC (or preamp) into Speaker;
2. High out from Power Amp (1ch) into Speaker for mid-high (Internal DSP and Bass amplification remains active);
3. High out from Power Amp (3ch) "directly" into Speaker Drivers (filtering must have been done in adfvance of this).

This may look a bit complicated but it only really was when we were thinking of what ever could be possible and how to switch all which took days and recaps and "impossible" and "maye after all" and so on. The really only more or less valuable thing which could not be done is having your own power amp for mid-high and let the filter work the same. This is because internally two amps do this job with filter parts in between and physically this just can not be switched. Or maybe it can but at some stage things become too difficult.

I hope it is a bit clear ? If not, ask away.
Peter

Hi Coen,


So, clear by now - the "first two" amps are bridged indeed. Still there is some truth in what you suggested and although there isn't a real reason to think like you diid, you went the complicated direction which ... we did too.
I don't feel like elaborating further but I do want to grant you the not-so-much wrong thinking.

Regards,
Peter

Thanks Peter,

I got the picture now, though the non-detrimental dsp is somewhat cryptic, because it must include AD conversion to be able to be fed from the analog out of the NOS1a...

In the end the final non detrimental solution will be the three way NOS1x I guess! ;)

regards, Coen

Nah Coen, now it should be too obvious ! But OK :

PC (or anything) -> DSP -> D/A.

Clear now ?
This of course is similar to DSP-in-software but hardware-DSP is infinitely faster and this time again onder "my" convenient control (includes Arc Prediction for example). But also, no 6x (or 8x) 768 from the PC is possible, while the DSP is sheerly unlimited regarding this (read : only the D/A needs to be able to digest it). Thus for example :

Source = 44.1 -> PC -> 44.1 -> DSP -> 705600 -> D/A.
And that 8x.

And oh, there's an ADC neeeded but that is merely to show what all has been friggeled with the XOvers etc. (see the result of it all).

Peter

Clear,


I missed the 'when' in the first sentence.

A three (or six) channel Dac aha.. hmm, I see a new project in the cooking.

regards, Coen