Those looking for the opinions and experiences of other builders of the Modulus-86 may want to keep an eye on this thread on the Siegfried Linkwitz builders forum:
ORION/PLUTO/LX521 Users Group • View topic - LXmini Amplifier
~Tom
ORION/PLUTO/LX521 Users Group • View topic - LXmini Amplifier
~Tom
I was interested in the transition point distortion from biased to Class B. It's usually visible as a crowbar shape or a flattening of the signal at the peaks.
With a sine wave input, I'm not seeing anything other than a pure sine wave output on the oscilloscope, regardless of signal level.
I stand by my previous comment that you will not be able to pick up 0.00018 % THD on an o'scope. 1.8 ppm (parts per million) of the output signal will be distortion products. You won't resolve that on an oscilloscope. This is why Audio Precision can charge a premium for their measurement gear.
~Tom
I stand by my previous comment that you will not be able to pick up 0.00018 % THD on an o'scope. 1.8 ppm (parts per million) of the output signal will be distortion products. You won't resolve that on an oscilloscope. This is why Audio Precision can charge a premium for their measurement gear.
~Tom
Rather obvious when you think about it ... the ones made by your ear/brain as you "digest" the music - the problem is that test gear at the moment, no matter how expensive, doesn't mimic that analysis path.
Nothing's going to found there, I can guarantee it. Systems, rather than components, have introduced distortion as an element of the sound - that's why they nearly always sound like a hifi system, rather than a convincing recreation of a musical event. The real trick is understanding that fact, and learning how to isolate the misbehaviour by using the right source tracks to emphasise, highlight the problems, and then homing in on the causes.
Having high performance components like Tom's reduces the number of issues, but doesn't guarantee optimum sound - your ears will tell you whether you've succeeded or not. The process is to find an issue, and knock it over, solve it - do this over and over again, and when sufficient are resolved the sound you're after will emerge ...
In the thread above they talk about the TPA SYMPATICO amplifier, comparing it to the Modulus in some aspects... Is the Sympatico also a composite amplifier design? What are the major differences in TPA's approach?
Thanks
Do
Thanks
Do
Yes but not everyone has the same ear brain and that is where that poster's idea I had linked earlier had merit. The only way to accurately account for taste and perception is to graph it as a distribution curve. Once those preference numbers are in, the digging for what matters most can begin. It may turn out that the lowest THD+N trumps all other electrical characteristics in achieving the most realistic reproduction and the modulus will become the new standard but there is lots of stuff that isn't measured (or can't be measured well) that could be more important like signal degradation from the number of signal stages (a.k.a. sound thinning or overbuffered sound) and transient response (reaction accuracy). For the record, I can see how these would fall neatly into the categories of timbral accuracy and PRAT so kudos to yldouright if he happens to read this.fas42 said:
I'm not so sure about that ... do you have any problem if you hear a piano playing out of sight distinguishing whether it's a real instrument, or a recording? I suspect most people find it trivial to separate "real" from "recreation" - I focus on minimising that gap, or eliminating it entirely - that's the end "effect" I'm after, that the listener is always "fooled".
Focusing on the typical things that audiophiles worry about, like "thin" or "bright" or PRAT doesn't cut it for me - either the sound is audibly faulty, or it isn't; I listen for that, and that makes the process very easy ... 🙂.
Sympatico Amplifier you can read for yourself. the Sympatico is a composite, using the LM4780 for 100W/8Ohm output. Beyond that, with no published measurements hard to tell how good it really is and this is not the forum for that sort of discussion and better held in the chipsamps forum.
TPA's amp uses the two channels in an LM4780 in a bridge configuration. Not a bad idea in theory, but even with an 8 Ω load, the LM4780 is pushed quite hard as each half of the amplifier 'sees' a 4 Ω load. I would not use the TPA Simpatico for driving 4 Ω loads at any appreciable power levels. You really need a bridge-parallel amp for that.
Other differences: The Modulus-86 will have about 40 dB higher CMRR, hence, reject EMI and noise much better than the TPA amp.
I don't see any measurements on the TPA amp either, so one is left to guess about the performance. The layout of the PCB matters tremendously at the performance level achieved by the Modulus-86. Without test gear, you'll be operating completely open loop for the PCB optimizations.
With a dual LM3886 (aka LM4780), there are two options for a mono block: Bridge and parallel. TPA went for the high-power market with the bridged solution. I'd rather go after the parallel market because, as far as I can tell, more people are driving low-impedance loads (4 Ω with dips below).
~Tom
Last edited:
I think you may be missing the point. Thin is my adjective but there are many more than that one to describe what we hear in playback but he managed to narrow these down to three that can fit them all. Not too shabby IMO. By separating what aspect of the sound is 'wrong' and defining it we have a method to pinpoint and address it more efficiently. The fact that those definitions alone can describe characteristics we can not yet measure is alone worth the price of admission. Tom has shown himself to be an engineer's engineer. He has more than the technical acumen, he has a genuine passion for improvement and a high BS rejection comportment. I suspect he would be genuinely interested in seeing something like that framework adopted and think his product would score rather well in that subjective arena.fas42 said:
Last edited:
My goodness, Frank making some sense. Must be this rather good beer I am drinking.
I had an interesting experience a couple of days ago whilst taking the kids to see the ballet at covent garden. Seats up in the gods, looking down into the orchestra pit. I could read the score with my binoculars. unusual perspective. Expecting there to be sound reinforcement I closed my eyes on a couple of occasions to listen for sound coming from odd sources and failed. The sounds were coming from the instruments.
The presentation was unlike any hifi, but wonderful.
</completely off topic>
Tom: An engineer's engineer with a passion for audio and a high BSRR. I like it... 🙂
The perceived sound quality is important. There's no doubt about that. The issues I have with subjective testing is that they are ... subjective. Many try to conclude too much from corrections and/or sighted tests. Then add the ambiguities in the vocabulary. One person may describe the rise in THD resulting from speaker cone break-up as "harsh" (negative) where as others describe it as "precise" (positive) depending on their personal preferences. Then there are the terms that just flat out don't make sense. I mean... What does it mean when an amplifier has "good tempo" or is "melodic"?
This is why I rely on measurements. At least by measurements I can quantify the performance of the amp in ways that give my customers confidence that my products will sound good when assembled. That said, no product leaves my test bench without a listening test.
~Tom
The perceived sound quality is important. There's no doubt about that. The issues I have with subjective testing is that they are ... subjective. Many try to conclude too much from corrections and/or sighted tests. Then add the ambiguities in the vocabulary. One person may describe the rise in THD resulting from speaker cone break-up as "harsh" (negative) where as others describe it as "precise" (positive) depending on their personal preferences. Then there are the terms that just flat out don't make sense. I mean... What does it mean when an amplifier has "good tempo" or is "melodic"?
This is why I rely on measurements. At least by measurements I can quantify the performance of the amp in ways that give my customers confidence that my products will sound good when assembled. That said, no product leaves my test bench without a listening test.
~Tom
I don't think we're too far apart, Terrence, 🙂 ... but to me using a term like "thin" is far too vague - there are types of music which when reproduced correctly could be described as being "thin", that's just how the composer, player wanted it to sound for an effect - if you're going to use such then you would need to nominate a piece, a section of music that you know definitely does not have that quality and contrast that with the sound the equipment being assessed adds to it, by distorting it, to give it a "thin" quality.
I don't use the concept of a whole piece of music being 'wrong', I zoom in to precisely one point, one element of the sound - say 5 secs worth, at a point 3 minutes in, and say, "There, at that point the system gets a particular, defined aspect of the sound 'wrong' " ! ... I'm nailing the exact flaw that the system has shown up, by provoking the misbehaviour - that gives a very clear handle on what the problem may be, and an easy way of seeing whether I'm getting closer to solving it.
I don't use the concept of a whole piece of music being 'wrong', I zoom in to precisely one point, one element of the sound - say 5 secs worth, at a point 3 minutes in, and say, "There, at that point the system gets a particular, defined aspect of the sound 'wrong' " ! ... I'm nailing the exact flaw that the system has shown up, by provoking the misbehaviour - that gives a very clear handle on what the problem may be, and an easy way of seeing whether I'm getting closer to solving it.
well I should point out no one records looking down at 45 degrees from 120ft up normally, so if a hifi sounded like this you would be very confused!
The great thing about non-amplified music is that, at least for me, it always sounds good - no matter where you listen to it, how far away, above it, below it - the "signature" of the sound always comes through, clean and satisfying. IME that also happens when an audio system is in good shape, you don't have to be in the "right place" to get a buzz from listening to it. Why live sound always seems 'right' is perhaps because we grow up learning to adjust our internal DSP to make sense of the sounds we hear; so the trick is to get playback to follow that same groove.
AJH's recordings aren't quite from 120' up. The microphone is 10-20' above the conductor. On a Periphonic (with height) Playback rig, there is a clear impression of the orchestra below the horizon.
Stravinsky: Pulcinella Suite — Ambisonic Surround Sound. Ambisonics, 5.1, audio recordings
You need VVMic to play the uncompressed *.amb Ambisonic files.
VVMic for TetraMic | vvaudio.com
This can decode to any stereo coincident mike arrangement (and also for various speaker arrangements.) and point this in any direction including up & down. Try it with cardioids at 180 on headphones.
There's more stuff on www.ambisonia.com but you might find AJH's stuff more easily at Ambisonic Info | Aaron Heller
BTW, the Beethoven Sy 4 is one of the best performances on record and the recording probably the finest.
I got curious and did the math. The typical measurement image attached to this thread is about 650 pixels high. If one (generously) assumes an oscope image was captured with the signal going all the way from top to bottom in the image the Mod's 4 ppm distortion works out to be a deviation from ideal of 0.0026 pixels. Good luck seeing that. 😉
There is also the problem that, if one wants to plot the error residual to see the distortion more clearly, an 18 bit HRO is required for the error to be 1 LSB. I am aware of one VHRO (very high resolution oscilloscope) which claims 20 bit resolution as a headline figure but its underlying hardware, like all other HROs I know of, tops out at 16 bits. And, to deliver that resolution, the sampling frequency's well under 1kHz, so it's not really suited to audio. A few ADCs capable of 18+bits at 44.1+kHz do exist but I don't, offhand, know of any whose linearity and DNR/SNR exceed that of the Mod's. For context, look at parts like the AD1555. And be sure to note the price tag.
Hopefully this fills in a bit of perspective around the comments earlier in this thread about composite amplifiers just handing test and measurement gear its own limitations back.
What's your thinking as to how to do that? If you've a blog or such I'd like to check it out.
It's not necessarily fair or just but the pragmatic reality is, if you will, the world tends to be run by the folks who show up. Mr. Olive's downloads are easy to get to whereas few, if any, folks are likely to jump the J. AES paywall.
I've been a beach bum for nearly 2 decades now and have no need to run the world. I only returned to the civilised world in 2005 when the www reached Cooktown at a blistering 28kBaud 🙂
Actually, the paper you linked to of Sean's is probably the best I've seen of his. (today, I only see AES stuff courtesy of friends who are prepared to risk the wrath of the AES mafia 😱) Many of his findings confirm mine.
But his methods would achieve quicker and better statistical significance (especially on his 'untrained' listeners) if he changed a few things.
An important one is letting the listener choose his own music and playback levels .. but there are other points which I've mentioned on this and other threads.
Last edited: