Another thing that I would like to mention is the fact that after this change in grounding sound qaulity of RS-485 in my DAC is substantially inferior to this method of connection and it makes me wonder: how on earth can manufactures of very expensive I2S cables write that they sell Hi End cables for I2S when in fact it turns out that balanced i2s can not compete with direct TTL connection?
Balanced I2S lines are for use over longer distances. LVCMOS unbalanced I2S lines are meant to be very short length. So, two different applications.
Other than that, manufacturers have to tell you why you should buy their product if they want to sell anything at all. Most of the time they don't necessarily feel obligated to tell you why you shouldn't buy their products. So its, buyer beware. That's not just for consumer products either.
It comes from transmission line theory. If you have studied RF electronics you should have come across the concept before.
Unfortunately I haven't studied RF electronics. So I haven't come across the concept. But I had a chance to read a few articles on this subject. However these articles did not mention anything aboit i2s wiring and grounding.
Last edited:
It is up to you to recognize that RF is on the I2S bus, so the bus is effectively a set of three RF transmission lines in parallel.
MCLK isn't actually part of I2S bus, but its still a transmission line too.
https://www.nxp.com/docs/en/user-manual/UM11732.pdf
https://www.nxp.com/docs/en/user-manual/UM11732.pdf
Its taken as a total given in high-speed logic (and RF) design - every fast signal needs a return current path in intimate proximity, either twisted pair, ground plane, coax braid, neighbouring ground wires on a ribbon cable. Because if you don't do this the fast circuits misbehave as Maxwell's equations rule at this speed, Kirchoff's laws alone don't cut it.
Its the speed of the edges that matter, not the actual frequency of a clock, so all the I2S signals should be treated the same ideally.