Cypress Announces USB 3.2 & USB4-Ready Controllers: EZ-PD CCG6DF & CCG6SF
by Anton Shilov on March 3, 2020 8:30 AM EST- Posted in
- Peripherals
- Thunderbolt
- Thunderbolt 3
- USB
- USB 3.2
- USB4
Cypress has unveiled a pair of upcoming USB Type-C controllers for next-generation desktops and laptops. The new EZ-PD CCG6DF and EZ-PD CCG6SF controllers are single-chip solutions that support USB 3.2, DisplayPort, and Power Delivery with fault protection. In addition, the controllers are ready to work with USB4 devices.
Cypress’s EZ-PD CCG6DF and EZ-PD CCG6SF are highly-integrated dual-port and single-port USB-C controllers. Both controllers are based on a 32-bit Arm Cortex-M0 processor, with 64 KB of eFlash to enable enhanced programmability as well as firmware upgrades to gain new functionality or deploy bug fixes. The controllers connect to their host using a PCIe interface and support a variety of protocols, including USB 3.2, DisplayPort, and PCIe.
Meanwhile, the controllers also offer a limited degree of support for USB4; however just what this entails is somewhat murky. According to the company's press release, the new controllers support "firmware upgrades to support the new USB4 standard", however we've been told sperately by the company that the chips are not USB4 controllers. Further still, the company does use the USB4 logo on the product page. What this may come down to is data rates: USB4 offers both 40Gbps and 20Gbps rates, and while both use the standard's fully encapsulated packet format, the 20Gbps data rate isn't any faster (or using any additional wires) than what today's SuperSpeed USB 20Gbps (Gen2x2) devices can already do. So if Cypress is waiting to enable USB4 support at a future date, it could very well be that they will (eventually) be usable as USB4 20Gbps controllers.
Moving on, one of the other important capabilities of the new chips is an integrated VBUS provider load switch, which allows the chip to support 100 W Power Delivery with fault protection. The integration of switch reduces design footprint and BOM costs, which is helpful for getting wider-spread support for high capacity USB power delivery.
Cypress uses an unknown process technology to build the EZ-PD CCG6DF and EZ-PD CCG6SF controllers and will offer them in 96-pin BGA and 48-pin QFN packages to support various requirements.
Cypress has started sampling of its EZ-PD CCG6DF and EZ-PD CCG6SF controllers with major OEMs. The company expects to start volume production in early Q3, so we may see consumer products by the end of this year.
Related Reading:
- USB-IF: USB4 Products Coming in Late 2020
- USB4 Specification Announced: Adopting Thunderbolt 3 Protocol for 40 Gbps USB
- Cypress and Zhaoxin Have USB 3.1 Gen 2 USB Controllers
- Infineon Set to Acquire Cypress for $10 Billion
Source: Cypress
Facebook
Twitter
RSS
17 Comments
View All Comments
edzieba - Tuesday, March 3, 2020 - link
AIUI, "USB 4" encompasses USB 3.1 Gen II + Thunderbolt 3 under one standard, with supporting both being the superset and supporting USB 3.1 Gen II being the minimum support. Similar solution to USB 3.1 Gen II's Alt Mode handling, but with more stringent enforcement to avoid the current 'confusion' in Type C support due to companies half-arseing their implementations (at best).jeremyshaw - Tuesday, March 3, 2020 - link
Sadly, no. Vendors can still pick and choose. TB3 integration is not guaranteed. It's why Intel is using TB4 (yes, 4) as a way of identifying which USB4 ports have TB3 support (more importantly, PCIe passthrough).ksec - Wednesday, March 4, 2020 - link
Has this been confirmed? That is what I thought they would do but it has never been officially confirmed by Intel.rahvin - Tuesday, March 3, 2020 - link
USB-IF rolled over to the manufacturers marketing divisions. USB 4 is as meaningless as USB 3.1. You need to look at the actual spec sheet to find out of it actually supports all the feature or only part of them.Valantar - Wednesday, March 4, 2020 - link
Yeah, this sounds like those meaningless GPU "refreshes" that are just the same part with a new name as laptop OEMs want a new name to put on their stickers. Seems like this has been the way of things since after 3.0, given how dumb naming has been since then.repoman27 - Wednesday, March 4, 2020 - link
Although USB4 is derived from Thunderbolt 3 and builds on USB 3.2, it really is its own thing.USB4 is a tunneling architecture designed to combine multiple protocols onto a single physical interface, so that the total bandwidth can be dynamically shared. USB4 host interfaces must support USB3 tunneling and provide muxes allowing native USB3 signaling (10 Gbps, Gen 2 x 1 minimum) as well as a USB 2.0 bus in parallel. DisplayPort tunneling and signaling via DisplayPort Alt Mode are required for USB4 hosts, however, PCIe tunneling and interoperability with Thunderbolt 3 systems via Thunderbolt Alt Mode are both optional. USB4 supports 20 Gbps (Gen 2 x 2) and 40 Gbps (Gen 3 x 2) operation, but the PHY differs materially from those implemented by either USB 3.2 or Thunderbolt 3.
It's quite possible that TBT3-Compatible USB4 devices prove to be the exception, rather than the rule. However, the requirements for USB4 hubs and docks to support interoperability may help to tip the scale. Unless sites like Anandtech get on top of the situation, the current confusion is only bound to intensify as USB4 products reach the market.
quiksilvr - Tuesday, March 3, 2020 - link
Why can't USB4 just be forced to have four lanes 40Gbps, 100W power delivery and all the specs of Thunderbolt 3? Why can't it be 100% required?name99 - Tuesday, March 3, 2020 - link
So you want your phone to have USB4?AND be able to deliver 100W to a plugged in device?
Good luck with that!
Some of the compromises are just dumb. But the power delivery compromises are physics!
Valantar - Wednesday, March 4, 2020 - link
... A phone doesn't have room for a TB3-class USB4 controller anyhow, so that point is moot.lmcd - Wednesday, March 4, 2020 - link
I'm just going to point you toward the Librem 5Yes it's niche but regardless, a future niche phone could support the bandwidth but not the power usage.