|
Computer Audio Asylum Music servers and other computer based digital audio technologies. |
|
|
|
Computer Audio Asylum Music servers and other computer based digital audio technologies. |
|
In Reply to: No troll, no snark, just curious posted by Jaundiced Ear on September 20, 2014 at 20:09:36:
Hi,
The first "SMPS" I listened to was actually a mechanical chopper inside an all tube (the solid state, even diode, count in that thing was the same as the winter temperature in Siberia - close enough to absolute zero as not to matter) battlefield radio-set of russian origin. It was meant to remain fully functional after a nuke strike (re. EMP). I am happy I never had to test that one in practice.
Just as I long gave up counting the number of girls, so I did count SMPS's. Of course, not all or even most of them were PC aimed ATX Types.
Past that, I will say that the technology of SMPS's has matured tremendously in recent years and we now have really exciting tech available. I see no longer a reason to insist on "linear", given that modern switched mode tech can be made to be substantially superior, with savings, if not in cost, at least in bulk.
How this tech is used is another story of course.
If you look inside "junk grade" stuff (e.g. most what is sold out of china on e-pay) you might be forgiven thinking we are still in the last millennium , except a lot of the parts that would have been used in the last millennium to reduce noise leakage both ways have been "economised" out, to save costs.
Ignoring the horrible noise levels and build quality, they often also suffer serious electrical safety issues.
If you look into "good quality commodity" stuff, such as you might find with products from Apple, Asus, Dell etc., they do include the new tech, but focus on deploying it for cost/size/weight reduction, so these "commodity" type units perform as good or as bad as ever (say mid 90's), but are lighter, smaller and cheaper, often dramatically so.
Finally you have what one might call "High End" or "Enthusiasts" products. There has been a strong drive towards lower noise (mechanical and electrical) in conjunction with "Overclocking" enthusiasts (no overclocked system brooks bad/noisy/unstable power). So in this group these new developments are used with gusto in the latest products (but the industry is conservative and slower to change to mainstream) and without attempting to be cheaper/lighter/smaller. In fact, in this market "more expensive", "heavier" and "bigger" seems more like it, so benefits of new tech tend to be realised.
However it is worth checking reviews, especially those who have real objective tests of both electrical and acoustical noise and one often finds that "Big Brand Name" products are light on engineering and big on branding/marketing, while often less expensive latest generation products of the latests generations from the major OEM's (of which Seasonic is arguably one) perform well.
Something to think about for all the "SMPS Doubters".
The main controlling element for the output voltage of almost any SMPS out there is a chip called a "431" (many different prefixes and suffixes). People in audio discovered them in the 90's as very simple, low cost and most crucially, low noise and wide bandwidth "Shunt Regulators" (I still use them on occasion, call them "old faithful", you know what you get and it usually is good enough).
At the time they tended to beat the stuffing out of most alternatives for audio (including all sorts of after market super/hyper/mega/brazillion regulators) and I still have no issue powering a DAC Chip or sensitive analogue stage directly of a 431 used as shunt regulator (and getting superb results).
So in principle, ANY SMPS should be able to turn in a similar audio band noise performance as such a shunt regulator and any of the switching noise can be knocked flat using a suitable LC Filters of moderate size and cost (often there is even space for these on the PCB). It just needs some basic, solid engineering.
For fun (or rather for necessity), I recently had to "fix" the design/manufacturing on a bunch of wall warts we got to power a number of test/burn in devices at the factory.
Stock they were hopeless. We could have send them back, buy different ones. But likely we would only have gotten worse or at best more of the same. So I got a junior to open them up and to test and then messed around a little myself. I had to fix a few design mistakes which meant the PSU's were oscillating at 1...2KHz with around 2V Peak-Peak. We hand-cut slots into the PCB's with a dremel to make sure there was enough isolation between mains and the bit that is supposed to be "low voltage".
After we ordered a bunch of the parts the original manufacturer had felt were unimportant and uprated parts for other positions, I then got the junior to fix all the Wallwart's we had on hand.
The resulting SMPS's using appx. mid 90's SMPS Tech, the original circuit boards with slots, original transformers (we did a hi-pot test, they passed ok) and switching IC's are now quiet enough in noise on the output as well as on the input (leaking into the mains) to beat many (but not all) linear supplies. No magic, no special tricks, just plain old fashioned solid engineering from 20 years ago. Took me maybe 30 Minutes to figure what to do, the junior spend around 15 min per wallwart in the bulk rebuild.
So, really there is no reason for any SMPS to be much worse than a "cooking" linear powersupply (say torroidal transformer, bridge rectifier, big cap + 108X Chip regulator and vestigal RFI filtering), if they are, you bought "Junk" grade and not "Enthusiast" grade.
Thor
At 20 bits, you are on the verge of dynamic range covering fly-farts-at-20-feet to intolerable pain. Really, what more could we need?
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- RE: No troll, no snark, just curious - Thorsten 08:43:59 09/22/14 (1)
- RE: No troll, no snark, just curious - fmak 11:54:03 09/22/14 (0)
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2024, Audio Asylum, all rights reserved ]
