HDMI inter and intra pair skew clarification

[Upballoon]

I'm working on a KVM and have a 1:2 HDMI switch. This is the first time I'm laying out HDMI and was wondering what the tolerance for skew was. I found a design guide by TI that says inter-pair skew needs to be within 2.67ns. Since they're differential pairs and ignoring the top solder mask, the signal travels at the speed of light, this gives me a skew of 662mm. This seems to be a lot. So I can getaway with just not length matching it?

Similarly for intra-pair skew the design guide says 178ps. which would be 44mm. Again this seems to be a lot. Am i missing something?

Design Guide: https://e2e.ti.com/cfs-file/__key/telligent-evolution-components-attachments/00-138-01-00-00-10-65-80/Texas-Instruments-HDMI-Design-Guide.pdf
HDMI Switch: https://www.ti.com/lit/ds/symlink/hd3ss215.pdf?ts=1646088521310&ref_url=https%253A%252F%252Fwww.ti.com%252Fstore%252Fti%252Fen%252Fp%252Fproduct%252F%253Fp%253DHD3SS215IZXHR

[DavidAlfa]

Also take in count the velocity factor, which is usually half in a pcb, so now the skew limits are half your calculations.
That TI article seems for older HDMI standards. Modern revisions are much faster, have a look here:
https://ez.analog.com/video/w/documents/687/hdmi-layout-guideline

[nctnico]

I'm working on a KVM and have a 1:2 HDMI switch. This is the first time I'm laying out HDMI and was wondering what the tolerance for skew was. I found a design guide by TI that says inter-pair skew needs to be within 2.67ns. Since they're differential pairs and ignoring the top solder mask, the signal travels at the speed of light, this gives me a skew of 662mm. This seems to be a lot. So I can getaway with just not length matching it?

Similarly for intra-pair skew the design guide says 178ps. which would be 44mm. Again this seems to be a lot. Am i missing something?

On a PCB you get a delay of about 6ps per mm (which is far slower than the speed of light). Still I'd try to length match the pairs to within a few mm. As you don't know how much of the slack you have is already eaten away in the devices that connect to your KVM box, better stay on the safe side.

Also match the skew within the pairs so the phase of the signals stays aligned.

[Upballoon]

I'm working on a KVM and have a 1:2 HDMI switch. This is the first time I'm laying out HDMI and was wondering what the tolerance for skew was. I found a design guide by TI that says inter-pair skew needs to be within 2.67ns. Since they're differential pairs and ignoring the top solder mask, the signal travels at the speed of light, this gives me a skew of 662mm. This seems to be a lot. So I can getaway with just not length matching it?

Similarly for intra-pair skew the design guide says 178ps. which would be 44mm. Again this seems to be a lot. Am i missing something?

Design Guide: https://e2e.ti.com/cfs-file/__key/telligent-evolution-components-attachments/00-138-01-00-00-10-65-80/Texas-Instruments-HDMI-Design-Guide.pdf
HDMI Switch: https://www.ti.com/lit/ds/symlink/hd3ss215.pdf?ts=1646088521310&ref_url=https%253A%252F%252Fwww.ti.com%252Fstore%252Fti%252Fen%252Fp%252Fproduct%252F%253Fp%253DHD3SS215IZXHR

See that's the thing, since these are differential pairs, on the top layer of the board, the velocity would same as the velocity of light (since the wave would go from the DP to DN). So the 6ps would be true for microstrips and inner layer TL line would it still be true for top layer differential pairs? Also wouldn't using the speed of light give me a more conservative number?

[jonpaul]

Beware of the HDMI DRM, even in a simple KVM

Jon

[NiHaoMike]

Beware of the HDMI DRM, even in a simple KVM

Not relevant for a dumb switch. It only becomes relevant if you want to decode the data stream.

[jonpaul]

Have seen HDMI switches that indded get a DRM issue on some material, but that was long ago.

Think it depends on the details of the switch

Jon

[ejeffrey]

Even for differential pairs on the top layer the velocity will be slowed relative to c since about half the field will be in the substrate.

[Upballoon]

Beware of the HDMI DRM, even in a simple KVM

Jon

Why would there be DRM? Do you have to buy the IP from the HDMI foundation if you ever want to implement HDMI?

[jonpaul]

There have been HDMI switches that fail on some sources and sinks with certain media.

The implimentation of BR, DRM and use if online enforcement is apt to have false positives.

The latest versions of HDMI TX and RX and latest HDMI versions may have fixed this preerenial problem.

After you finish your project, test on evcery availabe source and sink device and type of media.

Jon