[RESPONDED] AMD Batch 1 Guild

[nrp]

@Alessandro_Renzi

In this context “back” refers to the ports closer the the hinges/monitor or closer to the user while facing the open laptop?

Back is the side closer to the hinges. Front is the side closer to the user.

[Shiroudan]

@nrp

We’re not aware of any theoretical fix for this in firmware. We are investigating some hardware solutions for it, including a productizable version of essentially what this community member did: Passifying the USB-A Card

Passifying the USB-A Card

Just a note on this overall. If you’re running the latest BIOS (and if you use HDMI or DP, you modify the HDMI or DP Expansion Cards following the new instructions), you should no longer see higher power consumption with USB-A plugged in.

I thought this was fixed in this BIOS update? Or was that another problem.

[blkhawk]

@nrp

We’re not aware of any theoretical fix for this in firmware. We are investigating some hardware solutions for it, including a productizable version of essentially what this community member did: Passifying the USB-A Card

Hey - I am surprised that got mentioned here but sad to hear that the issue is non-fixable. That probably means the signal from the card must go to something that is not switchable by firmware without compromising sleep. AKA its a hardware bug on the main board - ouch.

Meanwhile the mod has been working for me for a couple of months now without issue.

[a4955]

Very glad to see the amount of updates now, great stuff

[Alex_Shpilkin]

@nrp So just to be sure—that mod (productized or not) will not work to connect things like Yubikeys and some flash drives that only have a USB port’s four contact pads but not the outer shield? That… is not the end of the world, I guess, but I hope this is not going to become the only option, seeing as I use an old Yubikey NEO and am thinking of also ordering a Feitian in a similar form factor.

[nrp]

@Shiroudan

I thought this was fixed in this BIOS update? Or was that another problem.

This was on Intel platforms which use Intel retimers. Note that we have gotten feedback from the community on there still being scenarios on 11th and 12th Gen that result in higher power consumption with USB-A, but we have confirmed that the solution on 13th Gen works as expected. We have additional investigation to do on 11th and 12th Gen.

[nrp]

@blkhawk

AKA its a hardware bug on the main board - ouch.

Not a hardware bug, but actually inherent behavior on the Kandou retimers we use on the AMD platform. As far as we are aware, there are no USB4 retimer options currently available for AMD platforms that would resolve this issue, and any hardware fixes would instead be working around it from the Expansion Card end.

[nrp]

@Alex_Shpilkin

@nrp So just to be sure—that mod (productized or not) will not work to connect things like Yubikeys and some flash drives that only have a USB port’s four contact pads but not the outer shield? That… is not the end of the world, I guess, but I hope this is not going to become the only option, seeing as I use an old Yubikey NEO and am thinking of also ordering a Feitian in a similar form factor.

It’s too early to say since we don’t have a prototype, but our design goals would be to come up with solution that works with all USB-A devices. I agree that a scenario where shell-less devices don’t go into a lower power state would probably still be a reasonable compromise.

[stiV]

@stiV

As I said - I am more than okay with delays, even into Q4. Provided there is communication from framework. We don’t know how many laptops are in Batch 1 and/or Batch 2, we will see how this turns out.

I just wanted to say thank you to the framework team - the communication of the last few weeks is exactly what I personally wished for. e-mails to customers with information and a more in-depth discussion in the community forums.

Maybe (!) a little earlier would have been nice. Even if that would have meant less details early on.

But this is still near perfect communication imo!

[Ddd]

Is there any statement from Kandou about their product’s issues?
(and where did my previous post go?)

[Peter_Wren]

@Ddd

where did my previous post go?

I think someone didn’t like that you used the word buggy.

[Matt_Hartley]

@stiV

I just wanted to say thank you to the framework team - the communication of the last few weeks is exactly what I personally wished for. e-mails to customers with information and a more in-depth discussion in the community forums.

Maybe (!) a little earlier would have been nice. Even if that would have meant less details early on.

But this is still near perfect communication imo!

We appreciate the kind words.

[Adrian_Joachim]

@Peter_Wren

I think someone didn’t like that you used the word buggy.

Just deleting posts without a trace is a little iffy

[Aquatica]

@nrp

It’s too early to say since we don’t have a prototype, but our design goals would be to come up with solution that works with all USB-A devices. I agree that a scenario where shell-less devices don’t go into a lower power state would probably still be a reasonable compromise.

Thanks for the transparency. Do you expect to have “fixed” USB-A ports by the time Batch 6 ships? If you don’t, do you plan on replacing expansion cards with high power usage with ones that are OK?

[Matt_Hartley]

@Adrian_Joachim

Just deleting posts without a trace is a little iffy

Going off topic and calling products buggy without a specific point isn’t happening in this thread.

There is constructive, detailed criticism worthy of discussion here. We’re aiming to keep this thread readable and on topic.

Continues going no where, it will be closed. That’s what happened.

[Matt_Hartley]

@Aquatica

Thanks for the transparency. Do you expect to have “fixed” USB-A ports by the time Batch 6 ships? If you don’t, do you plan on replacing expansion cards with high power usage with ones that are OK

Not commenting on this at the moment, but questions like this are welcome and on topic. Thank you for the asking.

[Aquatica]

@Matt_Hartley

Not commenting on this at the moment, but questions like this are welcome and on topic. Thank you for the asking.

Thanks Matt. Can you then comment on how bad the impact of the “high” power usage is? How much of the laptop’s battery life is getting lost?

[Matt_Hartley]

@Aquatica

Thanks Matt. Can you then comment on how bad the impact of the “high” power usage is? How much of the laptop’s battery life is getting lost?

I’ll see if I can tap an engineer Tues when I get back to my desk. Great questions.

[Aquatica]

@Matt_Hartley

I’ll see if I can tap an engineer Tues when I get back to my desk. Great questions.

Thank you!

[Bohden_Landis]

If idle/sleep power draw from certain expansion cards is substantial and is caused by retimer behavior that can’t be changed then would it be possible to reduce power draw with changes to power delivery controller firmware? I see that the CPYD6227 power delivery controller that is listed on 12th gen mainboard schematics has a programmable cortex m0 in it and is connected to the retimers. The limited Kandou retimer documentation I can access also talks about “initialization and configuration operations” through a connection to the “Power Deliver (PD) microcontroller”. Would it be possible to program the power delivery controller to simulate a usb disconnect or otherwise turn off the retimers when the device is in a low power/sleep state?