Originally Posted by stevekale
- there's plenty of the feedback and reports you seek at Computeraudiophile.com and on diyaudio.com you will see a number of implementations. The latter is a more interesting forum as the discussion is more technical. It's where the stubborn and educated speculators meet whereas CA is more "novice".
I am getting tired of this formal education bashing on these forums. So because a person has taken the time to earn a college technology degree and/or has the equivalent industry experience, they are closed minded, untalented, in your words "stubborn and educated"?
Did it ever occur to you that the reason we don't just run out and "try" something we see on an audiophile forum is because we have the experience and education to realize it isn't going to work? What's the point of any formal education or expereince if we as a society eschew the work of those before us and re-invent the wheel every generation. Where would we be today?
I and others over on the audio DIY forum have raised very similar technical questions as to the viability of a linear PC power supply within the context beiing discussed here. Nobody on your side dares to challenge those questions but rather hides behind the "well have you tried it" wall. That's not the way any engineering is done and hasn't been for at least 100 years. We take the accumulated knowledge of our parents generation and build on it. That's how technology advances.
So you did just measure the current draw with a DMM! OK, you have also found the transient and loop gain functions of LTspice (the free version). Do you have the mathematics knowledge to utilize these functions. Since you obviously didn't measure the current draw properly, that is under AC conditions, how are you going to model a load in Spice to test your circuit design?
Let's not BS the EEs here!
And those Swenson power supplies you show are for MacMinis. They are not for ATX PC power supply replacements. I find the whole idea of the MacMini linear power supply rather silly. There is a big old switch mode regulator inside the MacMini. So what's the point of the outboard linear supply? The sales carp form Swenson and his side kick are just that. Let's invent a problem we can sell a solution for. I see absolutly no engineering review or specifications on the performace of thse power supplies.
This is not about you or me. My goal here and I do think we have both clearly explained our views and objectives, is this:
1) There is no documented engineering evidence that a linear power supply in a media PC will improve audio quality over a generic switch mode power supply. To be specific we are talking about media servers that output a digital signal stream, AES,SPDIF,USB, Ethernet, Wifi, ect.
2) Showing a lower noise floor from an isolated resistively loaded linear power supply does not satisfy the requirement for determining the total system EMI generated inthe computer. In layman terms, the computer circuits themselves generate prolific levels of EMI and RFI. The power supply is a small contributor to the total system noise floor.
3) Digital signal processing systems are immune to electrical noise below the logic voltage threshold. Now it's true that digital cables can carry this noise into sensitive outboard analog circuits. It is also true that noisy power rails do aggravate clock circuit jitter. However the important thing to keep in mind here is that merely replacing the PC power supply will not negate these problems because the source of the contamination problem is not the power supply.
4) As I outlined above, power supply systems for high speed digital system are complex to design. Your home PC is a highly complex digital system. If a home built or incompetently designed commercial after market power supply does not have the needed transient response or current reserve, the computer will be unstable. That means crashing as a well as other serious operating system errors. IOW, bad audio will be the least of your problems.
5) To spite many incompetent claims on audio forums, digital audio is not considered a high speed data medium. AES or SPDIF timing requirements are pale compared to some of the timing tolerances on your mother board. Do you think that 1gbs serial PCIe data transfer is not sensitive to timing errors? If your system timing was as sloppy as some audiophile's claim the computer would never even boot up.
6) Linear power supplies are subject to over voltage under failure conditions. That means your 5volt rail could suddenly become 10 volts or more with full current. Good linear power supplies have protection built in to try and protect against this. Switchmode power supplies by the nature of their design cannot overvoltage. There is no risk with switch mode power supplies on the primary power rails.
It is my professional opinion as well as others you can check out on the DIY audio thread I referenced that replacing your PC power supply with a linear power supply from a niche unknown vendor with unverified technical credentials is asking for trouble - possibly destruction of your entire media server. And all for what - where is the scientific evidence that a linear power supply in this context offers any improvement.
That's my warning. At least readers here can weight the opposing views and decide for themselves.