2

I am trying to get around my head on the three questions below.

  1. Is SPDIF protocol is really driving 0.5 V pk-pk without having any DC offset?

enter image description here

  1. If yes, HDMI1.4 mentions ARC protocol which is adhered to SPDIF standards. ARC line is 75 ohm impedance controlled ?

  2. If I get 0.5 V pk-pk without any DC offset from SPDIF inputs, how are TI inverter ICs doing the conversion to CMOS?

enter image description here

Null
  • 7,448
  • 17
  • 36
  • 48
  • the first inverter is used as a form of amplfier - R102 and R103 bias the inverter into a semi linear region. The technique is also seen with crystal oscillators. The second inverter squares it up. – Kartman Dec 15 '21 at 10:43

2 Answers2

3
  1. Yes. That's SPDIF protocol data, sent over coaxial physical interface.

  2. The logic fails here. Only the protocol data is SPDIF. The physical inteface is what HDMI ARC defines. It is not 75 ohm coaxial, but a differential pair with differential, common mode and single ended impedance and voltage levels defined by HDMI standard.

  3. That is an unbuffered CMOS inverter being used as linear amplifier. It's a bit of a trick, but quite old trick used in various places. The 10k resistor biases the DC point where Vin=Vout to approximately VCC/2. So small signal fed in via 100 ohm resistor is amplified accordingly. The capacitor blocks DC which allows the inverter to bias itself. The second inverter just amplifies or squares the signal back to digital logic signal.

Justme
  • 127,425
  • 3
  • 97
  • 261
3
  1. Yes. The intention is for the SPDIF signal to be AC-coupled with a video-like voltage swing and impedance, so that it is roughly compatible and won’t cause damage if mis-connected.

SPDIF and AES/EBU (its pro cousin) employ a kind of balanced ‘dc-free’ coding, called biphase-mark or differential Manchester code. Signals using this coding can be AC coupled using use capacitors or transformers in the cabling chain.

This same encoding is used for ARC, but electrically different as explained below.

TOSLink (optical) also uses the same coding.

  1. No, ARC is a bit different than SPDIF coax. Basic (non-HEAC) ARC is 50 ohms, vs. 75 for SPDIF coax. Voltage swing is similar however (+/- 250mV from common mode).

HDMI 1.4 defines a previously unused, reserved pin (pin 14) as ARC. A cable that supports ARC will have this line at 50 ohm single-ended and use a coax to carry it.

So in the simplest case, ARC is mostly like SPDIF, save for its impedance. If only things were this simple...

Enter HEAC. The ARC (pin 14) and HPD (pin 19) signals are redefined as a diff pair for Ethernet+ARC, as HEAC+ and HEAC-, respectively. When full HEAC is in use, ARC is sent as common mode on both ARC and HPD (HEAC+/-) pins, while Ethernet is sent/received differentially.

The ARC and Ethernet signals are combined and teased apart by the HDMI chips at each end, and further, by a hybrid circuit that separates the Ethernet duplex.

My take: HEAC is a dumpster fire, since it is an electrically-closed system, and is not widely used as a result. (More here: Can one have ethernet connection through laptop's HDMI port?)

Nevertheless an HDMI 1.4 or later cable needs to support it to be compliant. HEAC+/- pair are 50 ohm single ended / 100 ohm differential.

  1. Back to the SPDIF receiver circuit. WHat does it do? It is an AC-coupled amplifier which takes the dc-balanced SPDIF signal and converts it to logic levels.

How? The circuit uses the LVC2GU04 chip, a dual, unbuffered (single-stage) inverter.

  • The first stage is wired as a gain amplifier
  • the second, an open-loop inverter

The fact that the LVC2GU04 is unbuffered matters, as buffered types connected as amplifiers can oscillate (although that point is debateable.)

More here: Logic inverter feedback resistor, SPDIF Input/Ouput circuit

Anyway, here's a quick sim of the whole input path as shown (simulate it here):

enter image description here

What's going on? The first stage behaves as a gain of -100 (roughly) amplifier, with a bias of Vcc/2. The amplifier self-biases in its linear region due to the feedback. The second stage buffers the amplifier and drives the signal to the rest of the system.

Could this circuit be used for recovering SPDIF data from ARC? Yes, with two caveats:

  • doesn't work if HEAC is in use
  • impedance needs to be changed to 50 ohms

In reality, if you're building an HDMI source or switch that supports ARC or HEAC, you'll likely be using an integrated solution that has the ARC receiver built-in.

hacktastical
  • 49,832
  • 2
  • 47
  • 138
  • Not entirely correct. The ARC most certainly is not carried by coax but a differential pair. Impedance not 50 ohms either. And I wonder if any HEAC devices ever got to market. If there are, it will only be used if both devices negotiate it. And same with ARC, it won't come out of TV unless it is negotiated - so the circuit could be used, if the TV would output ARC to it. – Justme Dec 15 '21 at 23:50
  • Half-true. HPD and ARC get bundled as a pair (I fixed this), that is 50/100 ohms. It’s definitely not 75. And a TV can choose to implement ARC and just output it as a static setting, it doesn’t have to be negotiated. That only comes up with HEC. – hacktastical Dec 16 '21 at 17:46
  • Nope. HPD/UTIL pair has 55 ohms single-ended, 100 ohms differential, and 30 ohm common mode impedances. And it is strictly forbidden to enable ARC or Ethernet signaling unless it is negotiated that both devices support it. – Justme Dec 16 '21 at 17:57
  • Thank you so much for such a detailed explanation. One more question, do you think above inverter circuit will work at 6MHz with SN74LVC2 ? I am trying to run simulation but very bad output wave looking. – Atharva Upadhye Nov 16 '22 at 17:23