Maybe this is about the word 'voicing'. I believe that your Class D amps sound similar enough (not identical) to your OTL amps and that is not by accident
Review: Atma-Sphere Class D Amplifiers
- Thread starter tima
- Start date
Maybe this is about the word 'voicing'. I believe that your Class D amps sound similar enough (not identical) to your OTL amps and that is not by accident
...double-huh. I did order a pair of the monos today, BTW. Helpful review Tima.
Which part of…
…is not clear?
However I do suspect that they do listen for both, but are focused on primarily on the microphonics. (Maybe subconsciously on the noise.)
Maybe we need @Atmasphere to expound upon his statement.
Are you fellows suggesting that I am reading more into it than his words are saying?
...in my case, the statement was unclear. Perhaps you meant: this is what I am hearing? I couldn't quite suss out your intent.
At this point is it that if Ralph’s words seem unclear to you, and my words seem unclear to you, then I am sensing that this is a YOU problem…
Before that last post, I would have said that microphonics are different than noise (like a hissing noise).
There is no intent to “suss out”, other than it seems to be clear to me.
The specifically wording of my question was… “which part of <the quote> is not clear?”
The only possible thing to “suss out” is whether II am reading more into things than I should be?
Sigh ... This is my last comment on this topic. You can have the last word.
post to Ralph:
What is the referent of "this" ? I have no idea what you are "seeing". I do not understand why you wrote about seeing something. Ergo I responded ...
post to Ralph:
What is the referent of "this" ? I have no idea what you are "seeing". I do not understand why you wrote about seeing something. Ergo I responded ...
I forgot I had this. It is a video made during the 500hr break-in period of the Class D amps. I was playing CDs on repeat to do the break-in. I do not remember how many hours I had on them when I made this video but I'll guess 400 or more. The bass was the last sonics to evolve.
From a live concert in Tokyo. John Williams conducting music from Raiders of the Lost Ark.
From a live concert in Tokyo. John Williams conducting music from Raiders of the Lost Ark.
I don’t need to have the “last word”, I am just trying to figure out if we are on different pages.
”This” == this (next part - directly below):
We didn't set out to make them sound like our OTLs. What we did instead was to make sure the distortion signature was benign like you expect of any good amplifier.
Correct. A noisy tube generates hiss and may not be microphonic. A microphonic tube generates distortion and signals of its own through mechanical means but might otherwise be low noise. If the microphonic tube is in a low vibration environment you might never know that microphonics is a problem. But most of our preamps will be used in a room in which the speakers also reside, so we have to vet tubes for microphonics independently of noise.
Tubes should be tested for microphony with a soft rubber mallet. George Kaye Small Signal Tube Tester is great for such test.
Nor, presumably did you set out to make them sound different from your OTLs. Perhaps " voicing" as an active verb may not be accurate. Nonetheless in both cases we have the Ralph Karsten version of doing that. It would be unusual if they sounded different presuming consistent application of your theories/principles.
Edit: I understand voicing as choosing the components (transformers, resistors, etc.) and the character of the circuitry to achieve a particular sonic result.
Last edited:
I assumed that noise was random.
And microphonics had a causal mechanism that it depends on, and hence follow the input like IMD, HD, etc.
So I am glad I was on the same page, and it looks like we are all getting there.
If I hear tube hiss coming out of my speakers or if I hear distortion from a microphonic tube, that to me is a random fluctuation disturbing the signal -- it is tube noise regardless how its cause is labeled.
I'll take you at your word that the 3.3 release is quieter than the 3.0 release. If noisy is the antonym of quiet, I'll stick to my statement that Atma preamps need quiet tubes.
Only for you though.
- For the majority of engineering types, noise is stochastic. (Ie.e totally random)
- And distortion has a transfer function applied against the signal… (or… it depends on the signal to manifest itself.)
The difference between electronic distortion and noise
The concepts of distortion and noise in signal processing overlap to a limited extent, but they are distinct phenomena. Distortion in a signal is the
www.testandmeasurementtips.com
In the case of microphonics, the transfer function may be delayed by the speakers exciting the DUT, and may have some frequency dependence based upon whatever thing it sits on… It can be a lot more complicated than some simple non-linear device in a circuit that is instantaneous.
Basically mircophonics are totally NOT random.
We didn't do that.
You can 'voice' a zero feedback circuit that way (if you use cheap/junk parts you might have to spend a lot of time 'voicing' something, but if you use good parts and have a good circuit, no voicing required). So zero feedback circuits really should have the best parts you can find.
If you have a lot of feedback its a different game since high feedback allows the circuit to reject that which is not the signal, including influences from parts. But that is not an argument for using cheap parts, especially if you want longevity.
FWIW we've never 'voiced' our tube amps- they lack compensation of any kind. What we did do in both the tube amps and the class Ds is pay attention to the math and make sure everything is working as the math predicts.
We had a functional amp pretty early on but there were artifacts from switching which can be quite common in a prototype. Once we had settled on the final circuit it was then a matter of shutting down all the stray inductances since the GaNFETs can switch at some crazy speeds. Once that was done the distortion of the amp was the remaining signature as it is with any amp. In our circuit, there are two sources of non-linearity: the encoding scheme and the 'deadtime' used to prevent the outputs from overheating. The former is responsible for about 3-5% of the distortion and the latter the rest; both producing lower ordered harmonics.
We did not audition any of the parts for their sonic merit. There literally was no need since all the parameters of the circuit are knowable through the math. But we did use good quality parts throughout since we want it to hold up.
We found the Russian 12AT7s to be low noise. But they have a harsh microphonic 'ping' to them we found intolerable. I know of one tube tester that allows you to audition the tube under test so you can vet it for both parameters of noise and microphonics. On our tube tester, the strength of the tube shows easily and that tells us how noisy it will be as there is a 100% correlation in our circuits. But the tube tester can't tell us if the tube is microphonic. You have to play it to know that.
The Mk3.3 revision level affected the MP-1 more than any of our other products. It has an additional 12AT7 at the input of the phono, paralleled with two others to allow for lower noise. So its less critical of the tubes than its predecessors. We also went to a solid state constant current source instead of a tube CCS. That increased the performance of the CCS circuits by about 2 orders of magnitude and allowed for a CCS circuit on every stage of gain in the preamp, something we'd not done before. A CCS circuit is vital if you want the most performance (gain, noise, distortion and bandwidth) out of a differential circuit. So the preamp is lower noise on that account too. These changes also increased gain of the phono by about 6dB. We also went to a shunt volume control system for the Gain Trim controls, so they have full range control of the volume and are other wise dead silent. The servo circuit was improved too so we have nearly two orders of magnitude lower DC Offset at the output of the preamp- its measures in fractions of a milliVolt now. It settles faster and can control a wider range of tubes in the line stage than before.
Any high performance tube product needs the tubes vetted if you want the most out of that product. A tube phono section that can work with any LOMC cartridge made going direct in (no SUT) is going to need low noise tubes. In our case that really only applies to the input tubes since they make 95% or more of the noise in the phono section; but we really are only testing the tubes to see how strong they are. So its much easier to vet them than it used to be.
Of course you did. You chose the parts and you chose the circuit according to what you count as good.
Similar threads
 | Steve Williams Site Founder | Site Owner | Administrator |  | Ron Resnick Site Owner | Administrator |  | Julian (The Fixer) Website Build | Marketing Managersing |