Linear power supply

[AlliPlatt]

Hi....Several USB powered D/A converters you have tested have benefited from hub-fed power as opposed to port power have they not? An SMPS is arguably less likely to have stronger mains harmonic spurs depending on how well it was designed, but it will gain noise at the other end of the spectrum.
And in terms of reliability, there is no comparison whatsoever when it comes to low power, low cost SMPSs (like the majority of ones used for small desktop devices you review en-masse here) vs an equivalent linear supply.

medical pcb

[hazelnut]

The easiest source of USB linear supplies is audio-targeted ones, typically DAC use, however since they're aimed at the golden ears market you'll end up paying audio woo-woo prices for them.  I managed to pick up an LHY Audio linear USB supply relatively cheaply, this isn't a brand as such but a generic label that various companies sell clones under.  It's pretty decent, Talema sealed transformer, LM1085 LDO regulator, Philips electrolytics, etc.  If anyone's interested I can hook a scope to the output and see what the noise is like, but I'd expect it to be minimal.

[hazelnut]

As a followup to the previous post, I decided to check the regulation on the power supply at various loads using the same mini USB cable I use on the System Blue, which is the least bad of a collection of mini USB's I have.  Unloaded the voltage was 5.03V, at 250mA this dropped to 4.98V, at 500mA it was 4.91V, and at 1.1A it was 4.78V.  This is... not terribly good given that it claims to be a 15W supply, but I wasn't going to take it above that current in case they'd substituted an LM1086 or skimped on heatsinking or one of the other typical corner-cutting measures, and also because in various places you see 15W, 2.5A, and 2.1A as the maximum output for the supply so who knows what it really is.

As I mentioned in another post, this reference of different USB cables mentions the problems that occur with voltage drops on many types of cables.  So when you're powering your Blue it's best to run a short DC cable from the power supply and a longer cable to get AC to the power supply to deal with cable-induced voltage drops.  However this means you're bringing your AC hash closer to the Blue, although presumably that shouldn't upset it too much since you've still got the E and H-field sensors some way away.

A better option would be to get a 12V linear power supply and run 12V to the Blue, then use a small linear regulator to drop whatever's left of the 12V down to 5V to power the Blue, with no AC getting close to it.  Normally you'd use a UBEC for this but that's the wrong type of regulator for this instance.

[hazelnut]

Some more on using a 12V supply, you can get prebuilt modules that'll do this, one lot based on the ubiquitous LM317, the other on the much more recent AMS1117, both costing a few dollars.  Both are also bare-bones, copied-from-the-data-sheet designs.  However, what this means is you could run your System Blue from a 12V, or whatever's left of 12V when it gets to it, supply and get a nice solid 5V supply sent to the Blue, you'd just need to add a barrel jack connector and one of those two types of modules heatsinked to the inside of the case.

[dupreezd]

Hazelnut, on the internal forum there are various discussions about which power supply to use.

Suffice to say that a lot of us early members use this power supply as Cutty mentioned in one the posts above.
https://www.powerstream.com/AC-1898.htm

It needs a bit of modification to splice in a USB plug that fits the controller.

Dries

[hazelnut]

Ah, in that case I'll request to be added, I've currently got my recently-finished Blue offline so I'm not sending in garbage data while waiting for the ferrites to arrive, but I assume I can still get access to the forums...