Very specific 5v requirements is a curious thing for Leter to say. As far as I know DC is DC and 5 volts of potential is 5 volts of potential. What's so Earth shattering about that?
I thought about stripping the case from my iFi iPower and putting it inside a project box with and IEC on one end and 3 barrel plugs on the other. The iPower is rated at does 2.5 A @ 5v and 1uv of ripple. I think that that along with a plate and barrel plug on the back of the SU-1 would work well and be inexpensive at the same time.
Are there any spears anyone wants to throw in my plan?
Why not disassemble the iPower if you can drop it inside the Singexer SU-1 instead? Can definitely still use the IEC so no need to cover up the hole.
Even though it handles 2.5A, both the USB input power and mainboard will be sharing the same ground so I don't think it would be a good idea. Correct me if I'm wrong. Maybe only one unless there's a way out of ground issues.
But I would totally agree with you. $50 would be a fairly cheap way to improve on those two things and provide 1uV of noise.
I would open the unit, take out the PCB, improve on the heatsink and thermal paste, replace with higher voltage handling capacitors, and use it to power the USB input, mainboard, and ISO Regen. 2.5A should be enough. I bet I can fit both the iPower and ISO Regen into the Singer SU-1 as an all in one solution.
There wouldn't be a need for a 2nd USB cable from the ISO Regen to the Singxer. It would be wired inside with a short wire. I just gotta see how much an ISO Regen would cost first though..
Anyways, I've been listening and comparing S/PDIF to AES/EBU lately. The prior is a 0.2M cable whereas the later is a 0.6M cable. Both sound extremely similar however I feel AES has better timbre, separation, and emphasis on sonic detailing. What I mean by sonic detailing is for example, the bass would be more punchy whereas S/PDIF sounds slightly more thumpy. The decay warps out immediately on the AES whereas S/PDIF slides out sluggishly accompanied with this sense of fuzziness. It wasn't muddy, but because it wasn't as prestine as AES, it didn't sound as fast or lively. There was a 5% difference in favor of the AES connection. My DAC doesn't mount the S/PDIF inputs directly onto the PCB, but use a wire instead so this may cause impedance problems that can cause jitter and thus reflect in its sound quality.
l heard that coaxial cables should be at least 1.5M in length otherwise there will be added reflections due to the signal getting there either too early or too late. These reflections cause additional jitter. I believe the AES cable, although longer, is better quality and balanced in design. The 110 ohm XLR connectors should be less prone to impedance issues like with 75 ohm RCA's. The differential (balanced) from XLR's also offers better common mode noise rejection. Maybe even the extra length helped with the reflections problem as mentioned earlier.
I'm excited to soon get my i2S mod working with my Metrum Pavane and compare that to see if it's better than the AES input. I no use the Holo Audio Spring.
Before when I had the Spring I did the same AES to S/PDIF comparison and prefered the AES but it was only 1-2% better on ONLY benchmark songs (songs I know the most). On the Pavane however, the AES is 5% better than S/PDIF on ANY song. The Spring had it's S/PDIF connectors mounted to the PCB so the downfall in the Pavane is not in the Spring which may explain why the difference wasn't as huge. That or the Pavane, like the SR-009 electrostatic, is just better at revealing the source material as well as other small changes.
. 
