HD3SS3220: Macbook Pro 2017 only USB2.0

Hi Kornel,

Could you provide the schematic for your HD3SS3220 design? I would like to review that and make sure everything is setup properly.

The HD3SS3220 shouldn't have any issues with supporting older USB3.0 ports, so long as there is proper communication between the CC lines and the ID pin is correctly routed to allow VBUS to be sent only when the ID pin is low.

Thanks,

Ryan

Hello Ryan, thanks for your reply.
Here are our schematic.

Hi Kornel,

I don't see the schematic, looks like it's failing to load or did not send correctly:

Could you possible resend the picture or send it as a file?

Thanks,

Ryan

Now?

Hi Kornel,

I can see them now, thank you!

Looking at the USB3 data lanes, I don't see any coupling capacitors on the TX lanes of the mux. Are these on a different part of the schematic? There should typically be 100nF or 220nF capacitors on TX lanes of the device, as the RX lane should already be coupled by whatever is connected upstream from the device.

I see you have the port pin floating for DRP, are you seeing this issue when both set as a UFP and DFP, or just a certain orientation?

For a DFP connection, the ID pin must be connected to something externally, like a voltage/power switch or GPIO of a host in order to enable indicate when VBUS should be supplies during the CC pin communication. If VBUS is supplied at all times during communication, the CC pin negotiation will fail, and cause issues with completing a USB3 connection.

Also, monitoring the DIR pin when you flip the type-C cable can help determine if the CC negotiation is being completed successfully or not, as the DIR pin should change depending on the orientation of the cable.

Please let me know if you have any other questions.

Thanks,

Ryan

Hi Ryan, thanks for your reply.

Our TX lanes are connected to a Microchip USB7206 Hub from there to GL3224 Memory card controller. We have coupling capacitors 0.1uF between the USB Hub and MCC.

We configured the HD3SS3220 on UFP today but it still don't work on the Macbook Pro 2017 Model. On other devices it works just fine as on DRP.

Is there something else we can try?

Hi Kornel,

So the HD3SS3220 works fine with other devices when set as a DRP, but when set as DRP or UFP with the Macbook specifically, you do not see any USB3 connection?

Would it be possible to monitor the CC and DIR pins and see what activity we are seeing on those lines? It's possible there is an issue with the CC negotiation.

Thanks,

Ryan

Hello Ryan,

exactly. Currently we only experienced this behavior on this specific Macbook.

We made some tests today and the DIR is switching correctly but we found that on one CC we only get 0.8V instead of 1.6V. If we unplug the cable at the Macbook and plug it in again it goes up to 1.6V but still no USB3 connection. If we unplug the cable at our device and plug it in again it always stays at 0.8V.

Is there another MUX model from TE with the same layout we could try?

Thanks,
Kornel

Hi Kornel,

It sounds like the CC pin levels aren't setting where we would expect them to. One CC pin should typically be pulled up to a higher voltage while the other is pulled down to ground. Are you at least seeing the CC pins setup like this, or are you seeing the CC pins in weird middle state?

Kornel Berne said:

Is there another MUX model from TE with the same layout we could try?

Unfortunately, we do not have any other muxes that match this exact pinout, as this mux is rather specific in terms of application, being both a mux and a CC controller. 

One thing I am curious about, have you tried this with other Macbooks? Do you see the same issues with them?

Thanks,

Ryan

Hi Ryan,

yes the CC pins have the correct setup e.g at one orientation CC1 is 1.6V (or in the Macbook Pro case 0.8V) and CC2 is 0.0V. Changing the orientation of the cable gives as the opposite results. 

Yes we tried on a Macbook Air M2 and it works alright.

Ryan Kitto said:

Unfortunately, we do not have any other muxes that match this exact pinout, as this mux is rather specific in terms of application, being both a mux and a CC controller. 

Ok I understand. We implemented the HD3SS3220 because we had problems on some Window PCs where only one orientation works. Is there maybe another "simpler" MUX what you can recommend that we can try, because it seems the HD3SS3220 is maybe a little overkill for our usecase.

Thanks a lot for your help,
Kornel

Hi Kornel,

Got it, sounds like this issue is isolated mainly to this one macbook. I can't say for sure off the top of my head what could be causing this, though the CC line being .8V with this specific case does seem to me like where the issue may lie.

Kornel Berne said:

Ok I understand. We implemented the HD3SS3220 because we had problems on some Window PCs where only one orientation works. Is there maybe another "simpler" MUX what you can recommend that we can try, because it seems the HD3SS3220 is maybe a little overkill for our usecase.

The next simplest setup I can recommend would be to use the HD3SS3212, one of our USB3 muxes with the TUSB321, a CC controller. The CC controller would be in charge of monitoring the CC pins and setting the device role, and would control the orientation of the HD3SS3212 via the DIR pin. Unfortunately, we do not have any other devices that can act as both mux and CC controller.

Thanks,

Ryan

Hello Ryan,

We found out that if we plug an external USB-Hub between the Macbook and our board it works correctly. So the negotiation between our board and the USB-Hub works properly and the USB-Hub negotiates correctly with the Macbook.

We checked the schematics of the problematic Macbook Pro model and noticed that other models have additional capacitors on the TX/RX lines.





We also found in the instruction (https://www.ti.com/lit/an/slaae24/slaae24.pdf?ts=1744138419755) - Page 7 - that the placement of the capacitors is relevant.

Since our board has capacitors right before the Multimedia Card controllers, which by default are off, but no capacitors between the USB-C Plug and the MUX nor the MUX and the USB-Hub controller can this be the cause? Should we add additional capacitors earlier?

Thanks and kind regards,
Kornel

Hi Kornel,

It sounds like the issue might be capacitance then, I believe.

Kornel Berne said:

Is this schematic for the working model? It looks like for this one, the main difference is that they have the 330nF caps on the RX lines, which helps with interoperability with devices that might have varying amounts of capacitance on the line, ensuring they stay within the range specified by the USB3 spec.

Could you make something like a block diagram for what the expected setup is for your system? I.E What devices the USB3 signals goes through from host to device? It may be possible that the signal is going through a device, like a redriver or something similar, which intakes the signal but does not recouple the signal. The TX lines of the USB3 signal always needs to be coupled with either 100nF or 220nF capacitors, and will need to be recoupled if it goes through a device which removes the AC coupling.

Just to confirm, does the upstream port connect to the macbook? If you board is connected properly, should the macbook detect a USB3 connection when your board is connected, or does another device connect to the macbook through your board?

Thanks,

Ryan

Yes the mentioned schematic by you is from a working model.

Attached is a block diagram of our setup.

Best Regards,
Kornel

Hi Kornel,

In this segment of the block diagram, going from the TX of the hub to the TX of the type-C port, are there any 100-220nF capacitor coupling the TX lanes? I don't believe I saw any in the schematic, though I could be missing it. If the TX lanes are not being coupled between here, I believe that would explain why the Macbooks with the 330nF caps on the RX lanes have better performance.

Thanks,

Ryan