Hey Ron,
I do believe your switching ps is actually a WRI-2731?????
with the following PDF -
http://www.elpac.com/uploads/documents/datasheets/WRI1.pdfI would suggest the following -
1. As switching supplies don't typically like having unloaded outputs, I'd put a 50ohm 5-10watt resistor accross pin 3 and one of the commons. This will draw 100ma of the 3amp rating of the 5v supply, unless you're actually already loading it.
2. Put 2k-10kuf of 25+v electrolytic caps, with the polarity correct, from +12v to ground (common) and from -12v to ground. At those elecrolytic's pins, bypass them with some nekk'd Epcos stacked polyester caps available from Farnell or Maplin. These are very low inductance, just what you need to shunt the hf crap. The fact that Felicia offers little benefit to the ps simply indicates that the switching itself is swamping any benefit provided by Felicia's powerconditioning. The added capacitance should ameliorate much of this switching 'hash'.
The above is really moot, as the Elpac switcher isn't going to give you the perfomance of even a mediocre linear ps. (this is not to say that a switching supply can't be superb, I'm simply saying that this switching supply can't...)
Your best hope is to eliminate that annoying ground loop hum. I'm assuming the linear supplies you've built are Rod's LM317/337 based ones? They're not the ultimate, but are quite servicable. If you're housing it in a metallic chassis, obviously you must ground the chassis to your mains ground. But you do not have to ground the power supply or signal ground (which I assume are tied together) of the crossover and its ps to that chassis, but rather let that ground be established via the signal grounds of your interconnects to the GK-1. [please, please don't tell me that your interconnect's signal ground is a high resistance 32ga wire or some such nonsense] But in reality we really don't know whether the ground loop is caused by ground disparities between your crossover and pre, or your crossover and amps...... I other words, we are stuck in ground loop HELL! And the only rational solution is to adopt the grounding discipline as charmingly extolled by Dan Banquer -
http://www.audiocircle.com/index.php?topic=8780&start=0&postdays=0&postorder=asc&highlight=ground+loopit works.....*
As Davey points out, the quality of regulation issue is totally separate ( to a large extent ) from grounding issues. Assuming you can resolve your grounding issues and can get you ESP regulators to work without hum, again as Davey suggested, you can look at Jung/Didden type low noise regulators. Frankly, implementation of such regulators is beyond me as it requires substantial expertise in layout, ground planes as well as access to a high bandwidth oscilloscope to examine the results of your efforts. But there are regulators available, specifically in GB, Andrew L. Weekes makes a very high quality product.
http://www.pinkfishmedia.net/forum/showthread.php?t=15588http://www.alw.audio.dsl.pipex.com/While I disagree with Mr. Weekes choice of wretched toroidal transformers, his regulators are quite excellent.
But obviously, this is predicated on resolving the hum issue. If you cannot resolve this, you might well get your present switching supply to preform substantially better as described in the first part of the post.
FWIW
*EDIT - Then again, Murphy's Law always rears her head. So you lift the signal/ps ground on your power amp from the chassis ground.... and now you embrace the latest craze of a Squeezedbox directly feeding your amp. And of course, you feed your SB with a 2 pronged ps, so now you've nothing connecting you signal/ps ground to your mains ground and your system hummmmms. So now you've got to reconnect your amps ps/signal ground to the chassis ground which is connected to your mains ground. Its alway sumth'n.
