Scott Nixon outboard DAC's, any good?

As far as Scott Nixon's DAC being a "stripped-down version of the 47 Labs," it is just the opposite: the 47 Labs units are more stripped down than the Scott Nixon DAC. Both these units get the timing right and are very direct, but are non-linear, and show alot of triangle-wave distortion in upper frequencies. Neither unit has an adequate power supply, and one of them has a much better pcb layout than the other. Both have jitter levels that can be improved through a-synchronous re-clocking, and both do not have a fully-tuned PLL.
Given all that, I still have to admit that both these units sound better than 85% of the DACs on the market. Scott's KIT is fun to build, priced fair, and a nice learning experience. He is very helpful if you have problems along the way.

Gts87: www.scott-nixon.com.
Nealhood: do you mean the AN DAC KIT 1.1? The AN is more expensive than Scott's DAC, it comes in a large, heavy metal housing, it accepts both SPDIF (on BNC) and AES/EBU (on XLR) at the flick of a switch. Scott's is less expensive, much smaller (2.5 inches square pcb) and accepts SPDIF, but you can also add AES/EBU if you want. Technologically, both Scott Nixon DacKit and AN 1.1 are non-OS, both use the ever-popular CS 8412 as receiver, but whereas Scott Nixon's DacKit uses TDA1543, 16-bit DAC chip, AN uses AD 1865 18-bit ladder DAC. Both passively convert current to output voltage via resistors, but the AN follows this with a CRC filter, using a ferrite-core inductor, and then a common cathode triode stage, delivering extra gain. Scott's DACKIT does not have CRC filter, but he has info about a simple RCR filter for de-emphasis if you need it, and offers a tube buffered version of his DACkit with 6DJ8/6922 in cathode follower mode, with lower output impedance. Output coupling in AN is 0.47uF, whereas with Scott Nixon's DAC it is 100uF. Both have high output impedances, which require some care in system matching, but Scott Nixon's TubeDac and SaruDac+ now offer sinc filters, which offers better system matching flexibility.
Voltage regulation in AN is provided via three discrete regulators, built with op amps, and pass transistors, and referenced from LEDs, whereas Scott Nixon's DAC simply uses two LM78xx-type regs. Whereas AN offers upgrade paths from their basic entry-levels, Scott Nixon has dropped some of the kit options. AN offers an upgrade from Beyschlag metal film resistors to their house-brand tantalum resistors, simultaneous with an upgrade from WIMA MKP/Elna Starget caps to Black Gate/Elna Cerafine for the bypasses. Moreover, AN SIG versions uses output transformer for AC coupling. Scott Nixon uses Black Gate with Oscon or Panasonic FC caps standard. In addition, I just noticed he is offering a 4A transformer for the TubeDAC and added power supply upgrades for his dacs, which makes good sense. It is best to visit Scott Nixon's site, since he has made some additions, and there are even more options than I have mentioned. If you are an experienced builder with proper solder station, get the DACkit, otherwise buy a fully-assembled/tested board as sub-assembly or in the housing.

Suits me: yes, the sn dacs invert phase. Since inverted phase can degrade analog performance in subsequent stages, you might want to correct for this. If you are using an spdif consumer interface, there is a simple and elegant way to correct for inverted phase: connect the signal lead of your digital input to RXN on the cs8412, and the return to RXP. Since the differential line receivers on the cs8412 have input voltage ratings of -12V to 12V, the receiver can accept signals in inverted phase. Although the topic of phase is not explicitly discussed there, look at Appendix A of the Crystal data sheet for extra tips about interface.