Standard cable or ordinary cable: what does your DAC receive?
The same S/PDIF signal goes through two cables. Set the length, the sample rate and the receiver: the simulation shows what comes out at the end of each one, using the same calculations as the full version.
What the DAC receives
Set the length, the sample rate and the receiver, then run the comparison.
- AGotham GAC-1 10070:
- BNon-standard RCA "spaghetti":
The charts could not be loaded. The figures in the table still apply.
Figure 1.The received signal compared with the sent signal
The start of a subframe (32 cells, that is 16 bits) as it reaches the receiver: grey for the sent signal, dotted blue for the Gotham, dash-dot orange for the RCA lead. The two three-cell plateaus at the start form the preamble that marks the subframe; they never occur in the data.
The more closely a coloured trace follows the grey one, the more intact the signal. Over a few metres, the small step on the RCA trace just after each edge comes from echoes caused by its off-standard impedance. Lengthen the cable: the edges round off, and the RCA lead's signal flattens, then blurs.
Figure 2.The signal eye at the receiver input
AGotham GAC-1 10070
BNon-standard RCA "spaghetti"
The chart overlays successive slices of the received signal, as an oscilloscope would: the more open the eye, the more reliably the bits are read. The dashed rectangle is the mask set by the standard: 200 mV high over half a cell, that is a quarter of a bit. It is the smallest eye a compliant receiver must be able to read. To judge compliance, the simulator also subtracts from the eye height the noise margin that keeps errors to one in a trillion bits, so an eye that seems to clear the rectangle can still fall outside the mask.
Lengthen the cable: the RCA lead's eye closes until it leaves the mask; the Gotham's stays open.
Figure 3.The jitter that reaches the converter
Jitter is the timing irregularity of the clock that drives the converter; it is computed here after the receiver, as an RMS value. The scale is logarithmic: each gridline is ten times the previous one. The shaded band covers the published audibility thresholds, from 10 ns (full-scale 20 kHz sine, the most demanding case) to 500 ns (random jitter on music). The value includes the receiver's own jitter: when both cables show the same figure, the cable's share is negligible.
The further left the dot, the less jitter. With the ASRC, the WM8805 or the word clock, both cables give the same value: the receiver's clock sets it, not the cable.
Table 1.The numbers
Both cables carry the same test signal (1 kHz sine) in the electromagnetic noise of a hi-fi living room.
| Measure | AGotham | BNon-standard RCA |
|---|---|---|
| Impedance (standard: 75 Ω) | ||
| Cable loss | ||
| Reflection at each end | ||
| Usable eye height (200 mV required) | ||
| Jitter at the converter (RMS) | ||
| Misread cells (half-bits) | ||
| Longest reliable length | ||
| Effect on listening |
A cell is half a bit. The echo that returns to the receiver is the square of the reflection (8% for 28%).
Key points
- A matched 75 ohm cable produces virtually no echo, whatever its length.
- As long as the eye stays within the mask, both cables deliver exactly the same bits.
- As long as the eye stays within the mask, the jitter at the converter stays at least 20 times below 10 ns, the lowest published mean audibility threshold.
- The ordinary lead leaves the mask at about 20 m at 44.1 kHz and 13 m at 192 kHz, mostly because of picked-up noise: an order of magnitude, since the noise model is approximate.
- At 100 m, the standard cable still meets the mask at every sample rate on this page; the ordinary lead no longer carries any sound.
Go further
This version already simulates the electromagnetic noise of a hi-fi living room and every echo in the cable. The full version offers 21 cables, including AES/EBU and optical links, and lets you choose the electromagnetic environment. It adds AC coupling, transmitter jitter, a matrix showing which parameter weighs on which measurement, and a cable ranking.
Open this scenario in the full version