USB 3.1 Gen 2 drops to 5Gbps on M1 Mac mini

Please check this article and specially the replies. Is all in there. These are new and better ports but not 100% retro compatibile. Apple should have explained better this. Even if you call in for assistance they don’t know about it.

https://forums.macrumors.com/threads/usb-on-m1-macs-isnt-actually-10gb-s-also-definitely-not-usb4.2269777/

I’ve read the thread, it’s just more people reporting the issue and testing various chipsets. This isn’t “retro compatibility”, USB 3.1 Gen 2 at 10Gb/s is a modern and current USB interface. It was plainly advertised, and still is, on Apple’s website and marketing material. This is a bug in their hardware or firmware, plain and simple. To fault the user in this case is absurd.

This issue is related to the new M1 chip and apples to all Macs running this hardware. It needs to be fixed by Apple. My NVME drives run 50% slower on my MBA M1 than on my iMac intel machine on the 3.1 gen. 2 ports. Thunderbolt drives are OK

No reason? Sure reason! The M1 is USB 4/TB, which means speed is negotiated from USB 4 downwards. Looks to me like certain chipsets are up to the task whereas others get confused and drop to 5. It ain’t an Apple problem yet.

> Looks to me like certain chipsets are up to the task whereas others get confused and drop to 5.

Equally, the M1 SoC may not be up to the task of properly negotiating the link speed, gets confused and drops to 5 Gbps.

The jury is still out.

trevoz wrote:

> Looks to me like certain chipsets are up to the task whereas others get confused and drop to 5.

Equally, the M1 SoC may not be up to the task of properly negotiating the link speed, gets confused and drops to 5 Gbps....

The M1 does not negotiate the connections. There are "brand new" Intel

thunderbolt4/USB4 chips that do the actual communication and negotiating

in all of the M1 Macs and the M1 chips interface with them..

It is possible that it is these chips that are the culprit. Unfortunately,

so far Apple is the only one "in the wild" that are using these Intel chips at

the moment. It is also possible that it is the firmware that is not configuring these

chips properly or there is a bug in the Intel configuration process.

Personally, while a "one interface fits all approach" is desirable, it also increases

the complexity exponentially in terms of making everything play together

making debugging a bit difficult, especially with huge backward compatibility/interoperability

(USB4, USB 3.2, USB 3.1 Gen 2, USB 3.1 Gen 1, USB3, USB 2, USB 1, TB4, TB3, TB2, TB1,

Power Delivery, DisplayPort 1.4, DisplayPort 1.2, DisplayPort 1.0) and the infinite number

of types of devices that can connect to each to the USB-C port these standards.

> The M1 does not negotiate the connections. There are "brand new" Intel

> thunderbolt4/USB4 chips that do the actual communication and negotiating

> in all of the M1 Macs and the M1 chips interface with them..

You are mistaken - Apple is not using the Intel chips in its M1 computers. It is using its own USB4 silicon in its M1 SoC. This has been confirmed by multiple technical sites, but you can do so yourself by using the macOS System Information utility on any M1 system.

Additionally there is no such thing as "Thunderbolt 4". USB4 _optionally_ includes Thunderbolt 3.

IIRC, TB4 is an instruction set change, not a change to the cable. It enables new features with the same cable. It matters to the chipset at either end of the connection, but the cable is exactly the same as TB3.

I would like to point out again that this issue is not limited to the new USB4 implementation, as the issue is present on my Early 2020 Intel Macbook Air. It could be firmware, or there could be a specific chip/chipset present in that model that is also in the new M1 machines. There were only months release and likely had some overlap in development behind the scenes

Well, there’s no debate that the M1’s Thunderbolt/USB 4 implementation is different from prior MacBooks and also (therefore?) restrictive. In fact, the spec says so. Never mind that it’s a curious combination given that Thunderbolt is a connection to a shared Bus whereas in point-to-point USB, speeds are granted solely on what the connected peripheral can achieve (although USB can be hubbed/enclosed/caged and given bus-like features).

Meanwhile, broken ports being blamed on 11.2 look increasingly triggered by connections to unsupported docks with 11.2 blameless.

I’d say the two are related: there’s code in the mini’s firmware that’s essentially playing bouncer in a member’s only club, enforcing the dress code. If it was put there for that reason, it won’t just go away with an update.

So instead of waiting for an update, I’d get on board with the bigger goal: making USB 3.x much-awaited history and with it, it’s various cages, enclosures, cables, dongles, docks etc. aka underworld, history, and begin beating the drums for USB 4. Starting with USB4 drives, hopefully by fall this year.

hcsitas wrote:

....

Meanwhile, broken ports being blamed on 11.2 look increasingly triggered by connections to unsupported docks with 11.2 blameless.

...

FWIW, there have been numerous reports with the M1 MacBook Airs

that have their TB/USB 4 ports going dead in terms of powering

them which could likely be related to the same issue being seen

on the dead Mini ports. Nearly all of them seemed to point

to docks that were attached to those ports.

Some thought this was an issue with the USB-C power delivery,

but if the USB-C functionality (which USB-C PD is part of) has been

borked somehow by the docks, well, they wouldn't function.

So, there is perhaps the same issue in both products but manifests itself

more "catastrophically" in the Airs by rendering them unusable and

in the Minis simply killing the TB/USB-C ports.

Wasn’t aware, who would’ve thunk! My only source for M1 Airs is you and I’m happy to note that you’re a happy camper (last I checked). But it does gives more credence to the dock theory, as does another update by the OP of the USB-ports-broken thread confirming he too had one. My efforts there to give up on the 11.2 theory have been met with, err, mixed success shall we say.

Separately, I must note that my oft-repeated advice on several threads to “Chuck the hub”, for a variety of unrelated issues which, while also roundly ignored, is now proving to be quite a gem😁.

As for power delivery, we know for a fact that Apple machines are picky about providing power. Add to it, most ultra high speed SSDs are dependent on bus power which they extract to the maximum but if that isn’t sufficient, with a drop in performance and possibly reduced life. I’m certain external SSDs with their own power supply don’t exist which I don’t understand, but they should.

So if you add all that up, you can’t get these SSDs to perform fully without powering them outside the Macs, which means a self-powered hub or enclosure, and depending on how that’s done and by what/who, it can create the perfect storm to kill the ports, and with it, possibly the warranty too depending on the stories these dead ports tell. Think Apple will roll out a fix for it? Think not.

For me at least the lessons are clear: (a) avoids docks, hubs, cages, enclosures (b) avoid external ultra speed SSDs without their own power supply even if they perform flawlessly because their MTBF is likely to be shorter (c) await USB4 to become mainstream (displays in particular).

I spoke with Apple about this. They still won’t acknowledge officially the problem but they know about it. Currently they are trying to understand if is a problem linked to older firmwares of theirs party devices or kernel. They say is definitely software and will be solved.