Post B6axnNtLtuqqTqDWoi by funkylab@mastodon.social
(DIR) More posts by funkylab@mastodon.social
(DIR) Post #B6au7rYiuOoOT04sG8 by niconiconi@mk.absturztau.be
0 likes, 0 repeats
Hi-Fi audio be like: For a $1000 amplifier targeting a community that hates interference more than evil spirits, one would expect that the amp uses RF design rules at 20 kHz, so it works at its rated performance absolutely anywhere even with injected RF: controlled PSRR to 1 GHz with multi-stage ferrites, controlled sub-ohm PDN impedance to 100 MHz, full shielding, perhaps 100 dB analog-digital isolation. But no, we still see some DACs with noise floors that are objectively affected by USB cables because of source impedance variations. What's the solution, learn what a VNA is in your next designs? No, just buy more power conditioners and gold-plated outlets!
(DIR) Post #B6auNFWa5Vwr8cPdMu by gsuberland@chaos.social
0 likes, 0 repeats
@niconiconi 100% this. you want good audio electronics engineering, you hire folks who worked on RF T&M gear.
(DIR) Post #B6awFXQ9rirmEarJM8 by lasse@social.tchncs.de
0 likes, 0 repeats
@gsuberland @niconiconi but then the high-end gear just gets even more expensive 😬 🙈
(DIR) Post #B6awFXb9Cremmg05U8 by gsuberland@chaos.social
0 likes, 0 repeats
@lasse @niconiconi yeah but at least it's justified!
(DIR) Post #B6awFXl4bxb3HSe0xM by niconiconi@mk.absturztau.be
0 likes, 0 repeats
@gsuberland@chaos.social @lasse@social.tchncs.de 14 years ago, an audio engineer called "NwAvGuy" built the most boring headphone amplifier possible: a non-inverting opamp gain stage, followed by a unity-gain follower stage, powered by +/- 12 V from rectified AC with 7812/7912, plus a comparator for current limiting. The only difference was analog design rules were actually followed, and an audio analyzer was actually used. The specsheet performance was on par with $1000 amps for $30. Later he did the same with a DAC, with is nothing more than just a vendor reference design with some tweaks and part changes.No RF shielding or filtering though.So my hypothesis is that if you want great PC audio, you should just purchase a reference design PCB, put it onto a carrier board with filtering, shield and isolation at all ports, and you're done (it does get complicated if you want to drive passive loudspeakers).
(DIR) Post #B6awkU3UQdyCm00nwm by funkylab@mastodon.social
0 likes, 0 repeats
@niconiconi @gsuberland @lasse yeah, PA is more complicated, because things actually start getting into the physical product design domain when you need to cool transistors.Also, note, that a lot of modern audio amplifier reference designs' layouts really are pretty bad. As in, obvious "this connection must not be there, have you actually tried using this and measured its output or not wondered why it shuts down after a few seconds?" level bad. Not to mention that seemingly ~all major…
(DIR) Post #B6awv67TlG8JiuwHWC by niconiconi@mk.absturztau.be
0 likes, 0 repeats
@funkylab@mastodon.social @gsuberland@chaos.social @lasse@social.tchncs.de If some vendor reference designs are that bad, it makes you wonder if the IC vendor's datasheet performance just came from thin air. :woozy_baa:
(DIR) Post #B6awysd7rCUD2pXrKi by funkylab@mastodon.social
0 likes, 1 repeats
@niconiconi @gsuberland @lasse analog audio IC vendors (ST, TI, Maxim less so, …) went through a "we need to update our datasheets to match modern corporate identity, let's outsource that to whoever cares so little to correctly copy over figures that they are the cheapest" phase, which took, in case of TI, half a decade to repair – but now the datasheets are as good as they were before.
(DIR) Post #B6axFnRiNtejtDH3gW by funkylab@mastodon.social
0 likes, 0 repeats
@niconiconi @gsuberland @lasse usually, that's not the case. Nobody dedicates a wafer run to something that the large-volume buyers get, test, and say, "OK, your choice; we'll just subtract that impression from the margin of the next product we buy from you".It's just that the analog design team makes the IC, probably just wire-bonded to board or even just directly probes it, verifies it works according to all the design specs (and they're usually good about these kinds of things), and then…
(DIR) Post #B6axnNZr4PFtVN5xQW by funkylab@mastodon.social
0 likes, 0 repeats
@niconiconi @gsuberland @lasse … comes the part where "investing a lot of expensive internal engineering hours into an eval board that the interesting customers don't need, as they get an FAE to hand-hold them through the design process" isn't very interesting. So, to still get to the middle- and low-volume markets, you just get a contractor to design a board from the preliminary data sheet. Just my guess; but it's pretty common to see schematic layouts and design practices between ref designs…
(DIR) Post #B6axnNtLtuqqTqDWoi by funkylab@mastodon.social
0 likes, 0 repeats
@niconiconi @gsuberland @lasse … differ enough to make you suspect that they might not be coming out of a coherent technical organization.
(DIR) Post #B6aym2rchFBkvbKkmO by lasse@social.tchncs.de
0 likes, 0 repeats
@niconiconi @gsuberland I like boring stuff, but listening over headphones not so much - contemporary compressed music (as in 'Loudness Wars') just gets tiring way too fast.
(DIR) Post #B6b4QGbRXAKCtQlRDc by jernej__s@infosec.exchange
0 likes, 0 repeats
@niconiconi Don't forget £21k power cords.
(DIR) Post #B6b4QGpGhlNraJETlg by niconiconi@mk.absturztau.be
0 likes, 0 repeats
@jernej__s@infosec.exchange or Japanese Hi-Fi fans who install personal step-down transformers to the grid. “Electricity is like blood. If it is tainted, the whole body will get sick,” says Mr. Morita. “No matter how expensive the audio equipment is, it will be no good if the blood is bad.”
(DIR) Post #B6b5HyhyHYAt1zQfiK by tml@mementomori.social
0 likes, 0 repeats
@niconiconi @jernej__s Clearly audiophiles should stop using plain grid electricity altogether and simply use audiophile grade photovoltaics and batteries. Seems obvious.
(DIR) Post #B6b5HysFfKOjXsEsjo by niconiconi@mk.absturztau.be
0 likes, 0 repeats
@tml@mementomori.social @jernej__s@infosec.exchange Yep. Battery banks were known as a Hi-Fi solution for decades, not sure why didn't they use it. But I guess it can be difficult to use in tube amps that need HV bias. Motor-Generator Sets are also used for interference suppression in real labs.