HD3SS215: Having issues with 2:1 configuration of this IC. Getting no signal under certain circumstances.

Hi,

The 2k pullup should already be populated on the source side while 47k should already be populated on the sink side. If this is the case, then I will leave R122 and R123 as No Connect.

The connection sequence is 

• Source provides 5V
• Sink sees 5V and drives HPD high when its EDID is ready
• Source sees HPD and reads the Sink EDID

With 5V_HDMI shared between the two sources and the sink, the 5V_HDMI always stays high when switching between A and B and the sink will not toggle its HPD.

If you toggle the HPD after the switching, are you then able to see the video?

Thanks

David

This seems to be related to the source we chose for the test. We tried a raspberry pi and it was giving similar issues even when bypassing the TI chip and connected straight to a monitor/TV. The IC seems to be working as expected for now. 

1. What would be the best recommendation for the 5V_HDMI configuration? 

2. You mention toggling HPD. Should that be controlled by our processor? It looks like it controlled by DX_SEL in the datasheet. Is that what you meant?

Dan

David,

I've had some time to try this suggestion out and it appears to be working. I removed the 0 ohm jumper resistors from the 5V_HDMI signals and injected 5V directly into the J14 5V connection. When I wanted to switch sources, I toggled 5V off, toggled DX_SEL, and then toggled 5V on. This switched the sources as expected. My follow-up question is, for our application can I simply ignore the 5V from the sources, and implement a standalone 5V switch on the sink that our MCU toggles? And then just toggle it prior to switching my sources?

Thanks,

Dan

Dan

How are you going to synchronize the 5V from the HDMI source and the 5V from the power switch? You may want to power off the switch when 5V is not present from the HDMI source.

Please also make sure the 5V from the power switch follow the requirement as listed in section 4.2.7 in the HDMI1.4 spec.

Thanks

David

In our application we are completely controlling the switching of the sources via our MCU, so we'll know when the 5V needs to be toggled. My question was to completely isolate the 5V on the sink from the 5V at the sources. So when we want to use source 1, we would toggle 5V off, send DX_SEL high, and then toggle 5V on. Then when we want to use source 2 we would toggle 5V off again, send DX_SEL low, and then toggle 5V on again. So the only time 5V would be on would be if we want to actively display something on the sink. Otherwise we would keep it low.

Hi, Dan

Thanks for the clarification, then I do not see an issue with your approach.

Thanks

David

David,

So our original circuit up above performs correctly (it switches between sources as expected) with certain inputs. Any insight as to why that is? The configurations are:

1. Our product is source 1, Roku stick is source 2 - This switches as expected with no issues

2. Our product is source 1, Chromecast is source 2 - Our source shows no signal, but the Chromecast displays correctly.

3. Our product is source 1, Laptop is source 2 - This switchs as expected with no issues again

Thanks!

Dan

Dan

I've had some time to try this suggestion out and it appears to be working. I removed the 0 ohm jumper resistors from the 5V_HDMI signals and injected 5V directly into the J14 5V connection. When I wanted to switch sources, I toggled 5V off, toggled DX_SEL, and then toggled 5V on. This switched the sources as expected.

When you did this experiment, is this with the 2nd scenario? 

Did you see HPD toggling (goes low, then high) when you switch between the two sources?

Thanks

David

David,

This is with the second scenario. Here is the process for when it is NOT working (using our original schematic above)

1. Device boots into Input 1 (our product) and is displayed correctly, HPD is high and stays high on Input 1, Input 2 (Chromecast) HDP is low and stays low.

2. Toggle DX_SEL, HPD on Input 1 goes low and stays low, HPD on Input 2 goes high and stays high. Input 2 displays normally.

3. Toggle DX_SEL, HPD on Input 1 goes high and stays high, but there is No Signal, HPD on Input 2 goes low and and stays low.

However, when performing the scenario you just mentioned, when toggling 5V prior to changing DX_SEL, Input 1 displays correctly and its HPD is high. Everything works as it should.

David,

Another note to add. This problem does not occur on every TV/Monitor. Device works as expected on an Insignia TV, HP monitor, and Acer Monitor. Has issues on an LG smart TV.

Dan

Dan

For your source and Chromecast source, is one of them being HDMI2.0 and other one is HDMI1.4?

Thanks

David

David,

Cannot confirm the Chromecast spec, but its a brand new Chromecast so I assume its at least 2.0. Our product is 1.4. I also just experimented with a setting on this LG tv called "Ultra HD Deep Color" that actually fixed the issue with the Chromecast. However its odd that the Chromecast will still work with that setting enabled when toggling the 5V HDMI signal.

Dan

Dan

HDMI2.0 introduces a new set of Status and Control Data Channel Structure (SCDC) registers. One of the new registers is the TMDS Configuration register which specifies the TMDS_Bit_Clock_Ratio. The TMDS_Bit_Clock_Ratio is default '0' for HDMI1.4 and '1' for HDMI2.0. When 5V_HDMI is not present or HDP Detect Pin goes low for 100ms, this register will reset.

So if the Chromecast is HDMI2.0, the LG TV also goes to HDMI2.0. But if the 5V_HDMI power does not go away, the LG_TV will still in HDMI2.0 and not reset back to HDMI1.4, then it will not work with your HDMI1.4 source. 

Thanks

David

This makes a lot of sense now. Would explain why the Roku stick and laptop don't have issues since they're not 2.0 devices.

Dan

So the clean design is to make sure the 5V_HDMI power on the sink connector toggles between the switching of the two sources.

Thanks

David

David,

I agree. On another note I had a question about powering the IC. Since the IC is built into our product and our product is source 1, when it loses power source 2 will not display on the sink. Since this is such a low power device, is it allowable to power the IC using the 5V (dropped to 3.3V) on the other source? Or potentially switch to it in the event of a power loss? 

Dan

Dan

No, the max absolute rating for VDD is 4V. Going to 5V will damage the HD3SS215. The operating condition for HD3SS215 VDD is minimum of 3.0V and maximum of 3.6V.

Thanks

David 

David,

Right, but that can be overcome by implementing a 3.3V regulator. Based on what I've read, the 5V line on an HDMI source is 5V, 50mA, which is more than enough for the 3.3V, 3mA required by the IC. I assume its just not recommended to power anything with HDMI.

Dan