"I find it hard to believe that you have an EE degree and you are designing software protocols for computer chips. It does not make since to me."
Then you obviously have no experience in embedded software. Embedded software engineers tend to come from either an EE background or computer science. Embedded software is actually a mix of EE and computer science. The best embedded software team will comprise members with EE degrees and members with computer science degrees.
The power transfer of USB is, as Herman points out, irrelevant. The USB power is transferred on a separate pin, carrying 5vDC. I would not expect any quality USB dac to use this power supply, partly because it is polluted with noise, and partly because it can only supply a couple of watts, which is insufficient. I suspect that it is left open circuit (not connected) on most all USB DACs, and that the DAC has its own separate power supply.
The high speed packetized nature is COMPLETELY RELEVANT, because it means that a USB DAC MUST buffer and reclock the data stream to the DAC, unlike an SPDIF based DAC which can attempt to regenerate its DAC clock from the SPDIF datastream. We all know that using SPDIF for the DAC master clock makes the DAC subject to the jitter in the transport, and is generally regarded as an outdated design. Since USB must reclock this weakness is removed.
So the computer can read data with zero errors, USB will transfer with zero errors, and the data is buffered and reclocked before being presented to the DAC chips ... it sounds like a formula for excellent digital sound. Of course it can be ruined in the execution, but then so can any one box or two box CD player.
Then you obviously have no experience in embedded software. Embedded software engineers tend to come from either an EE background or computer science. Embedded software is actually a mix of EE and computer science. The best embedded software team will comprise members with EE degrees and members with computer science degrees.
The power transfer of USB is, as Herman points out, irrelevant. The USB power is transferred on a separate pin, carrying 5vDC. I would not expect any quality USB dac to use this power supply, partly because it is polluted with noise, and partly because it can only supply a couple of watts, which is insufficient. I suspect that it is left open circuit (not connected) on most all USB DACs, and that the DAC has its own separate power supply.
The high speed packetized nature is COMPLETELY RELEVANT, because it means that a USB DAC MUST buffer and reclock the data stream to the DAC, unlike an SPDIF based DAC which can attempt to regenerate its DAC clock from the SPDIF datastream. We all know that using SPDIF for the DAC master clock makes the DAC subject to the jitter in the transport, and is generally regarded as an outdated design. Since USB must reclock this weakness is removed.
So the computer can read data with zero errors, USB will transfer with zero errors, and the data is buffered and reclocked before being presented to the DAC chips ... it sounds like a formula for excellent digital sound. Of course it can be ruined in the execution, but then so can any one box or two box CD player.