For the past six months I have been working on an article for audioXpress that would try to make some sense out of all that was happening with the popular USB interface and the ongoing transition for an updated version of the protocol's specification (USB 3.2 as it was standing) and the more visible evolution toward the ubiquitous USC Type-C (USB-C) universal connector. This included my latest conversation with the heads of the USB Implementers Forum (USB-IF) at CES 2019. But the more I would look into it, the more questions would come up and the more complicated the puzzle seemed to get.
|
There's no official logo for USB4 yet. This is what audioXpress created while we wait for the specification and have to talk about it...
|
I've seen many Internet articles published by tech websites capitalizing on the typical click-bait strategy of "Everything Explained" promises, when in reality everyone seems to have the same or more doubts than I have - and worse, many articles are just contributing to the confusion. But in the latest weeks, the web exploded with anecdotes regarding how confusing the whole USB strategy was getting, which was a clear indicative that things had gone astray.
And since audioXpress is committed to look at technology from an audio developer's perspective, the more I approached manufacturers and industry experts, the more they would also confide in me that they were feeling confused and increasingly frustrated.
Back in December 2016 and January 2017, I wrote two articles for audioXpress on the topic of USB Type-C and Thunderbolt 3, at a moment when the (then new) SuperSpeed USB 10 Gbps (USB 3.1 Gen 2) interface, the new USB Type-C connector specification, and the subsequently published USB Audio Device Class 3.0 specification, which standardizes audio over USB Type-C, had been published. And because I already felt that the topic was so closely linked with the related Thunderbolt 3 technology, which also adopted the USB Type-C connector and introduced updated specs at that time, I decided to cover that as well.
|
audioXpress
discussed "USB 3, Type-C, and Thunderbolt 3 - Interfaces and Implications on Audio" in its Standards Review article series in December 2016. Now available online.
|
Boy, was I early... but that article hit the problem right on spot, and for that reason, I decided it should be made available online for everyone to access it, because it touched on so many important aspects, and because it pointed out what was great and what was not so great with the whole thing.
During subsequent months and years a lot happened, with work focusing on the USB Power Delivery specification, leading to the introduction of the Certified USB Charger Compliance and Logo Program, Certified USB Fast Chargers, and more recently (January 2019) the USB Type-C Authentication Program, an optional USB security protocol that defines cryptographic-based authentication for USB Type-C chargers and devices.
Power charging has indeed been the main focus for the USB Implementers Forum (USB-IF) and the USB 3.0 Promoter Group, given this was a major market requirement on the mobile front in particular. Other problems, which were deserving attention, as I originally pointed out in my Standard Review articles, was cable length, leading the USB-IF to announce a new active cable specification and also a specification update to allow charge-through USB Type-C digital audio products, to create an optimal method for charging a smartphone, while also using USB Type-C based headphones. All familiar territory (and many headaches) for manufacturers struggling to offer solutions first.
Fast forward to March 4, 2019, the USB Promoter Group once again surprised us with the announcement of the pending release of the USB4 specification, a major update to deliver the next-generation USB architecture that complements and builds on the existing USB 3.2 and USB 2.0 architectures, adopting an architecture based on... the Thunderbolt 3 specification contributed by Intel Corp. As detailed - and there's not much more we can say - USB4 doubles the bandwidth of USB to 40 Gbps and enables multiple simultaneous data and display protocols, increasing compatibility among USB Type-C connector-based products.
|
How many of the existing USB Type-C products, including these new Neutrik mediaCON cable and panel connectors, will support USB4?
|
Why did things evolve that way?
First and foremost, as my original 2016 article pointed out, the coexistence of multiple USB 3.0, 3.1 Gen1 and Gen2, and the later fragmented 3.2 specifications, plus Thunderbolt 3, all sharing the same connector but using different cables and allowing different possibilities, created confusion in the market.
Second, the latest announcements from the USB-IF generated an unsustainable situation regarding the logic of the whole process, with USB 3.0 interfaces in computers not supporting the speeds allowed by SSD storage devices, confusing data rate support depending on cable length, restricting key applications such as automotive systems, which did not support USB 3.1 Gen2 but could support 10 Gbit/s bidirectionally with USB 3.1 Gen1., and finally the USB 3.2 specification (published in September 2018) introducing two-lane 10Gbit/s operation and a new nomenclature according to connection speeds:
General nomenclature: Gen X Y - (Speed x Lanes)
Enhanced SuperSpeed Gen 1×1 - (5 Gbit/s)
Enhanced SuperSpeed Gen 2×1 - (10 Gbit/s)
Enhanced SuperSpeed Gen 1×2 - (5 Gbit/s*2 =10 Gbit/s)
Enhanced SuperSpeed Gen 2×2 - (10 Gbit/s*2 =20 Gbit/s)
This could seem extremely logical for the engineering groups involved, but considering that the specifications also needed to be updated for critical applications like display support - and of course the audio device specification, while dealing with a completely unsustainable communication problem to explain all that to manufacturers themselves, not to mention consumers - a nightmare situation was created. The more the USB promoters worked to solve the problems, the more complicated the whole thing was getting.
Third, Intel - which was basically handling most of the workload on both standards - confirmed it had contributed its Thunderbolt protocol specification to the USB Promoter Group, enabling other chip makers to build Thunderbolt-compatible silicon, royalty-free. This paved the way to move forward and avoid the "Generation" conundrum created.
Now, these are the doubts I - and I'm sure a lot of other people - have.
First, how to avoid freezing the market while waiting for the USB4 specification to be released and products to reach the market? What will be the most sensible option for developers who need to launch products in the short term? Use USB 3.1 Gen2 or Gen1 products available today? Still plan to use USB 3.2 and wait for Gen 2×1 or Gen 2×2 controllers to be released this year? Go directly to Thunderbolt 3?
With semiconductor companies knowing that such a major change is happening, will it pay to consider designing with "outdated" IC controllers, voltage regulators, etc., to handle Power Delivery plus all the engineering time to learn something that - while still staying compatible - will be obsolete soon? And what about the implications over price of chips?
And what are the key differences and implications with USB4?
Of course, as the Thunderbolt consortium promoted, Thunderbolt 3 can do everything that USB 3 Type-C can do and more. It offers improved latency for audio due to the time synchronization protocol use in its transport layer, it supports PCIe 3.0 allowing the use of external graphics cards, supports DisplayPort (up to) 1.4 alternate mode, and supports delivery of up to 100 W of power, among many other things.
And Thunderbolt runs in USB Mode, but they are different protocols. Thunderbolt 3 defines a superset of capabilities - it is bi-directional with four lanes of PCI Express Gen 3 and eight lanes for DisplayPort with application-specific protocol stacks - and also uses different modes to deal with legacy devices like USB 2.0, 3.0 or 3.1, which is very different from the Thunderbolt 3 Mode or the Thunderbolt Networking Mode.
Well, long before this discussion was going on, some people like myself were using Firewire. Because Firewire 800 hard-drives were much faster, because video cameras used Firewire and basically because all Firewire audio interface were vastly superior to anything on USB. So, much in the same way I have managed to completely ignore the pains of Windows XP for many years and moved from Windows 2000 Professional to Mac OS, USB 2.0 for me was basically something used for connecting a mouse and to download pen drives with press kits. When Thunderbolt arrived on the Mac, I could continue to use all my Firewire devices as before, simply using a basic adapter, until today.
|
Universal Audio Thunderbolt 3 cards. These were probably the best investment that users and audio manufacturers could have made, compared to anything USB.
|
I mention this, because there lies some of the key differences between Thunderbolt and USB. Focusing on professional applications, Thunderbolt was designed to support the PCI Express (PCIe) standard but also network interfaces (e.g., Fibre Channel), storage devices such as external SATA (eSATA) and Serial Attached SCSI (SAS), and FireWire peripherals. Also, Thunderbolt supports Ethernet, theoretically up to 10 Gbps. Those are important things to keep supporting.
While USB uses a "tree" topology and "master-slave" protocol for addressing peripheral devices, where a host cannot "broadcast" signals to all peripherals at once, Thunderbolt 3, on the other hand, can address multiple end-device types dynamically. It uses a tunneling architecture designed to take a few underlying protocols, and combine them onto a single interface, so that the total speed and performance of the link can be shared between the underlying usages of these protocols - whether they are data, display, or something else. More important, Thunderbolt supports up to six Thunderbolt devices via hubs or daisy chain.
So, basically, if USB4 is implemented as Thunderbolt 3 - as is - we have basically a superior solution that should support all legacy USB devices and much more. But will it be exactly the same? As the USB Promoter Group describes, USB4 will need to be a superset solution that supports "multiple data and display protocols to efficiently share the total available bandwidth over the bus," while maintaining "backward compatibility with USB 3.2, USB 2.0 and Thunderbolt 3". So, this is not as simple as "rebranding" Thunderbolt as USB4.
What will happen to all the work with USB Power Delivery and how can that be carried forward to USB4 without disrupting the ongoing transition in mobile devices? Will the mobile industry eventually decide they are fine with USB 3.1 or 3.2 for the foreseeable future and leave USB4 to the same market that already adopted Thunderbolt 3 (laptop, desktop, and workstation computers)?
The process will also open access to all manufacturers that are not part of the USB Promoter Group and can now benefit from the royalty-free license to both the Thunderbolt protocol and the new USB4 specification. But Intel already confirmed it will retain its role certifying Thunderbolt-compatible devices. So, can we expect to have both Intel and AMD processors with USB4 support in early 2020?
|
This is a graphic published by Intel that already mentions USB4 and defines it as a subset of Thunderbolt 3.
|
As always, I'm certain many of the answers will be soon provided by Apple. After all, Apple is the company that could benefit the most with the transition, finally adopting the same specification for cables, connectors, and protocols for connection, communication, and power supply in all its devices. Most important, Apple did contribute significantly to the Thunderbolt and USB Type-C specifications and USB4 looks pretty much like the type of solution Cupertino would have developed in the first place.
I really would appreciate hearing what the industry has to say. That's why I decided to write about this topic in The Audio Voice. Any contributions welcome. Feel free to email me here.
The USB4 specification is expected to be published around the middle of 2019 and holds the potential to finally define a possible end to the reigning confusion. It's important that this happens fast. Given the product development cycles, USB4 is not going to make a significant impact before 2021
.