2

USB Hubs don't get the reviews that most other PC parts do. What should I look for in a 2.0, 3.x, or -C hub to make sure it won't cause problems, won't be a bottleneck, and give me hassle free, awesome performance?

I'm not referring to obvious features like if the cable can be removed, an included power supply, number of ports, or things like that. This answer refers to TTs, of which I was unfamiliar with until I read it. I know that if I get a $3 or $4 USB 2.0 hub, I won't expect much for performance. But I have no way to gauge the quality or performance of a more expensive hub.

2 Answers2

1

Performance of hubs does not depend on quality of plastics or number of ports. Once a hub gets proper electicals from host port and signal link does not break, its data transfer performance doesn't depend on anything (multi-TT and FS behavior with MIDI devices is an exception). This performance is ensured by certification of silicon ICs that go into the hub. However, for high-speed connectivity the quality of electrical routing (PCB and cables) does matter. A hub either work, or doesn't, except when behaving flaky, in narrow cases.

Unfortunately, for end user, there is only one meaningful criteria to make sure a hub won't cause problems. It is to look for USB-IF (Implementer Forum) Certification Logo on the device itself (or its packaging), like this one,

enter image description here

This logo is awarded to USB devices (and hubs!) if they successfully pass all logic and physical layer tests developed by USB consortium specifically to ensure inter-operability between products coming from different manufacturers.

An ordinary user cannot make the assessment of USB product quality, because it requires very expensive test equipment (to evaluate compliance of physical transmitters/receivers to USB standard), and specially configured host PCs with USB protocol analyzers to evaluate compliance to all logical protocols during connect-disconnect, all kind of data transfers, and all layer of aggressive power management of USB links.

For USB 2.0-only hubs this equipment must include a 4-channel high-speed oscilloscope (2GHz at least) with a special processing software, and with a set of specialized test fixtures, plus a dedicated protocol analyzer. The cost of all test equipment goes at about $100,000 and up. For USB 3+ certification testing the cost of equipment goes to $500,000 and above, because of the need of expensive BERT (Bit-Error-Rate-Test) instruments, 12+ GHz oscilloscopes and USB 3.1 protocol analyzers. Testing for functionality of Power Deliver options carries additional cost.

The USB-IF set of tests are fairly comprehensive, but the governing body is lenient and can grant a manufacturer several exceptions if something does not meet some required parameters. So if a device does not have the USB-IF certification logo, it means that something is brutally deficient with the device.

For example, USB requires all devices to have no more than 10uF capacitance on upstream cable upon initial plug-in (there could be effectively more caps, but they should engage later, and smoothly). This requirement serves to protect upstream (host) ports from excessive inrush currents that can disturb inner functionality of host. Most so-called "portable hubs" totally ignore this important parameter, and hook up all downstream port caps (100-200uF per port) directly to the input VBUS rail. I have a hub (TP-LINK) that has 880uF of capacitance on its VBUS. Plugging this kind of hub into a laptop essentially crowbars the USB port. No wonder this hub doesn't have any certification logo.

Unfortunately, the USB-IF certification process does not have any legal means to be enforced, so unscrupulous manufacturers skip on tests, and most USB devices on consumer market lack the certification.

Ale..chenski
  • 13,235
0

These links may be of help. (I just googled "benchmark USB hub")

http://thewirecutter.com/reviews/best-usb-hubs/

http://www.pcworld.com/article/2455205/test-the-speed-of-your-usb-drives.html