Teardown: Panlong HDMI audio splitter

[NiHaoMike]

Previously, I did a teardown of a cheap HDMI splitter. Recently, I needed a HDMI audio extractor to debug a HDMI audio issue with a project I was working on. The combined extractors/splitters don't cost very much more than just an extractor, so I bought one.

There's a PI3HDMI412AD HDMI active splitter, EP91A1K HDMI processor, EPF021A microcontroller, and CS4344 DAC.

While I haven't confirmed it, it looks like it might also work as a HDCP stripper right out of the box, based on my understanding that HDCP is point to point (or at least that's what I read) and that the part that does the splitting is just a dumb buffer. I also managed to "overclock" it to work at 4K, but then the two HDMI chips really started heating up so I didn't leave it running for long.

Unlike the plain splitter I tore down earlier, this one forces the input link to retrain if either output is connected/disconnected. It also passes through the EDID from the first display connected. And it does work as a HDMI to S/PDIF or analog audio adapter with no displays connected. Though the link retraining can be an issue as some displays will trigger that on display on/off or even changing inputs. I'll see if there's an easy hack to disable the retraining. I'm thinking of adding a simple circuit to pull up the HPD pin if the link active LED for the port is active, thus masking disconnect events until something else triggers a retrain.

[NiHaoMike]

I just tested it and just like the other splitter, it did in fact strip HDCP.

[LDM]

That's a cool little board. Where are you not seeing HDCP? The chip diagram for the EP91A1K shows it being applied to the repeater output.

I work with HDMI a lot and am always interested in new toys.

[NiHaoMike]

The simple splitting buffer prevents HDCP from working, so they disabled the HDCP on the output. (That's controlled by an I2C register according to the datasheet, and in fact, that's how a similar splitter that does enable the HDCP on the output could be hacked into a HDCP stripper.) I confirmed with a source that requires HDCP and a display that does not support HDCP. With a direct connection, it doesn't display anything at all (expected), but put the splitter in line and it just works.

Now I'll see if there's an easy hack around the retraining issue.