AMD Framework USB-C charger compatibility issues

[sgilderd]

I don’t think you have read the thread.

There is an issue with the Framework laptop overloading chargers. @patagona has explained in detail what is happening - AMD Framework USB-C charger compatibility issues - #36 by patagona

The issue affects ANY charger, regardless of whether you deem it as from a reputable/“off brand” manufacturer or not. Because it is the PD management in the Framework laptop which has the defect. And the list in that post has some very popular brands including Ugreen and Anker. From my side, I have the same issue with HP, Dell and Anker chargers. As well as a less-than-60W PD supply from a Lenovo dock.

Framework need to fix the PD management - if anything, the laptop overloading the charger and relying on the charger itself to limit current is a fire risk, especially for those brands that are “a nightmare under the hood”

[SoldierSvejk]

@sgilderd

I don’t think you have read the thread

Sure did. UGREEN and Anker reputable? That’s news to me.

The post you linked appears to detail issues with the aforementioned third-party chargers. Have a nice day!

[Mark1]

All of my chargers charge all of my laptops/phones/tablets… except for the framework.
It’s obvious where the issue lies.
Let’s leave it at that.

[MJ1]

It’s understandable that bugs can be frustrating, but this one is not that big, imho. The Framework charger works, and it seems some 65W chargers from other brands might as well. Unless I missed it, the issue is mainly trying to use lower wattage chargers.

@sgilderd

@Kieran_Levin has responded to say that a PD management fix will need to be wrapped with the next BIOS update.

Others have mentioned the time it’s taken for other BIOS updates. Framework is still a small team, under 50 people globally, spread over all the roles. Unfortunately, it will take time. As you say, Framework has responded.

If using a Framework charger is simply unacceptable, perhaps you can contact support and see if you can return your laptop.

[Fraoch]

This is starting to get contentious, please read the Community Guidelines and treat others with respect please.

[Kieran_Levin]

We have tested the following fix on the Embedded controller:

fwk: needs to limit the current during the PD voltage transition (#894)

Which will get into our next bios release.

What we were seeing is that low wattage 20V chargers will advertise PD 2.0 currents using resistor sensing of 3A. Eg 5V3A.
However when we negociate up to 20V, we need to reduce the input current limit on our charger to not go over the 20V current limit, eg 20V, 2A.

So when the charger voltage ramps up, the current would stay at 3A, and trip OCP on the charger.

Now when we transition to 20V, we will limit the input current on our charger.

[Jason_Username_Taken]

Thank you Kieran, you are the best! Looking forward to install the fix on my AMD FW!

(Btw, my Intel FW has extremely good charger compatibility, it will charge with a potato if I want it to. So I assume the EC/PD on Intel FW is already programmed in this fashion?)

[The_Official_GMan]

From what I can tell the intel mainboard uses a different PD controller (so has a different driver). I do however see a similar commit for the intel boards PD driver

hx20: reduce the input current limit when switching the type-c power

Given the commit author (actually same person to author AMD board PD fix), it seems like Framework has had to go back to compal (the board manufacturer, and maybe partial designer? not sure what the full business relationship is there) to have this issue fixed. So that probably explains part of the delay.

Also this too for intel boards

initial work to track device power

Kieran seems to have done a lot of the work (or at least commit authored) the PD firmware for the intel boards

[sgilderd]

@Kieran_Levin

We have tested the following fix on the Embedded controller:

fwk: needs to limit the current during the PD voltage transition (#894) · FrameworkComputer/EmbeddedController@057f6d7 · GitHub

Which will get into our next bios release.

Many thanks @Kieran_Levin - very much appreciate the update and look forward to the BIOS update!

@MJ1

It’s understandable that bugs can be frustrating, but this one is not that big, imho. The Framework charger works, and it seems some 65W chargers from other brands might as well. Unless I missed it, the issue is mainly trying to use lower wattage chargers.

I can only speak for my use case (which is not that unusual for business users whom Framework now supports) - I need to use docks, both at home and at my workplace, which provide power at less than 60W. The issue is triggered and then the port seems to get into a state where it no longer functions reliably. I am hoping that is because the connected accessory cannot negotiate the PD correctly with the laptop and so this will all be resolved with the patch quoted above.

Long story short - I regularly end up with 1 or 2 ports that do not work and then I cannot charge the laptop at all.

[MJ1]

One port for a dock, one port for a 65W charger. Preferably the Framework one.

Prior to USB-C charging, all laptops needed their own charger. And any very power demanding laptops still need this. The new high wattage PD will change this, but they aren’t yet available, last I looked.

[Alan_Pearce]

@MJ1

The new high wattage PD will change this, but they aren’t yet available, last I looked.

But they will be as soon as the FW16 starts shipping.
i.e. in a week or two.

[patagona]

@Kieran_Levin

We have tested the following fix on the Embedded controller […] which will get into our next bios release

Thank you for the fix, I’m really looking forward to the new firmware version!

There are, however also two issues with “dumb” 5V 3A chargers (10k pullups on CC lines):

• starting charging does not work with those either, the laptop pulls 3A for 300-600ms, then just stops pulling anything (the “kickstart” workaround works to get it charging though)
• when the voltage drops too low on the laptop-side (to ~4.6-4.7V), the current starts to oscillate wildly between 0 and 3A (see attached screenshot), possibly the laptop should reduce the current to a value where it can safely charge without oscillating.

I believe both of those might be seperate issues not directly related to the ramp-up current situation, but still need attention / fixing IMO.

[Adrian_Joachim]

@Jason_Username_Taken

(Btw, my Intel FW has extremely good charger compatibility, it will charge with a potato if I want it to. So I assume the EC/PD on Intel FW is already programmed in this fashion?)

Since the AMD does pretty much that with kickstarting(which for the most part seems to work around this overcurrent issue) it likely too will charge off a potato once this is fixed.

[Tom1]

@SoldierSvejk

If these are all off-brand chargers which may or may not conform to the USB-C PD standard, and which may or may not operate at the advertised wattage, does that change your mind?

Is Google on-brand enough? My AMD framework laptop will not charge with my 45 watt google usb type c charger.

[fw13amd]

Just found a bunch of dmesg lines like this:

[ 112.214885] xhci_hcd 0000:c1:00.3: Refused to change power state from D0 to D3hot

My laptop seems to charge correctly, though, but I’m still following this thread with interest.

I’m using a 100W dual output USB-PD charger bought from Amazon a few years back, which has always worked well with all my other laptops/smartphones.

[SoldierSvejk]

I thought I would test my phone charger to see if it worked. Sadly, you can’t buy this one anymore it seems. The unit is here USB Wall Charger Is Compact and Powers Up Your Devices Fast

Advertised as:

Otterbox 30W GaN USB-C charger
Output:
USB-C: 30W PD (5V/3A, 9V/3A, 15V/2A, 20V/1.5A), PPS (3.3-5.9V/3A, 3.3-11V/3A, 3.3-21V/1.5A)

Not sure if this helps for diagnostic purposes. But it charged my FW13 AMD 7840u (prebuilt) without any apparent issue. I only have two expansion cards in at the moment. Win11 Home edition. Is there anything I can do to check if it’s charging improperly somehow?

[Adrian_Joachim]

@SoldierSvejk

Not sure if this helps for diagnostic purposes.

Pretty sure the issue has already been diagnosed and identified, we are waiting for the fix at this point.

[SoldierSvejk]

The fix doesn’t appear to concern 30w chargers (or any <20W)…

@Kieran_Levin

What we were seeing is that low wattage 20V chargers will advertise PD 2.0 currents using resistor sensing of 3A. Eg 5V3A.
However when we negociate up to 20V, we need to reduce the input current limit on our charger to not go over the 20V current limit, eg 20V, 2A.

[Kieran_Levin]

@patagona

• when the voltage drops too low on the laptop-side (to ~4.6-4.7V), the current starts to oscillate wildly between 0 and 3A (see attached screenshot), possibly the laptop should reduce the current to a value where it can safely charge without oscillating.

If the current is below 4.75V, this is below the usb minimum voltage specification.
The current fluctuations could be caused by two things, either under-voltage protection kicking in on the PD controller, or more likely, the charger IC going below the AC present voltage threshold. However this is set to around 4.2V.
Generally I would suggest you use a charger that does not droop so much. Most chargers will output 5.1V or so to account for cable voltage droop.

We are working at refreshing the PD firmware to a newer sdk version in a later firmware update for the system, and we can investigate some of these additional issues at that time.

[patagona]

@Kieran_Levin

Generally I would suggest you use a charger that does not droop so much. Most chargers will output 5.1V or so to account for cable voltage droop.

I have been doing experiments with up to 5.5V and even then, there have been issues sometimes (though, to be honest, I did testing with 2 meter (~6 ft) cables which I have not resistance tested).

@Kieran_Levin

If the current is below 4.75V, this is below the usb minimum voltage specification

While, the USB “5V nominal” is defined by ±5%, so anywhere between 4.75V and 5.25V, that is defined at the port.

Looking at the USB-C spec (https://www.usb.org/sites/default/files/USB%20Type-C%20Spec%20R2.0%20-%20August%202019.pdf, page 140), source-to-sink drop (including connectors) is defined as 250mV on GND, 500mV on VBUS at the cable’s current capacity.
That means, for a standard 60W USB-C cable hooked up to a 5V charger, a worst case scenario of only ~4V actually reaching the laptop (4.75V at the charger, 0.25V loss on GND, 0.5V loss on VBUS).

Of course, this is an unlikely case because chargers usually tend to go more towards 5.25V rather than 4.75V, but even then, the laptop needs be able to charge with around 4.5V realistically, or if it can’t (due to hardware limitations), it should at least reduce the current to restore the voltage to a level where it can charge safely without oscillating.

Edit:
Best case: charger at 5.25V, 100mOhm cable => 4.95V
Typical case: charger at 5.1V, 200mOhm cable => 4.5V
Worst case (still in spec!): charger at 4.75V, 250mOhm cable => 4V

For reference, the official Framework 6ft 100W cable (which is specified to 5A even, instead of just 3A) seems to measure around 160mOhm, which, even with a 5.2V charger, only leaves ~4.72V at the device.