4

When I plug in a Sweex NPUS0480 4-port USB 2.0 hub, it is enumerated as self powered (according to the windows USBView application).

However, it is in fact bus powered. There is no DC connector on the pcb, nor any provision for one.

I was just wondering if this is normal, or could it be a mistake? Or is it just something to do with the type of usb-hub controller chip that is used?

I could not find an answer in the USB 2.0 spec, although Table 11-12 does mention the following option:

This hub is capable of both self- and bus-powered operating modes. It is currently only available as a bus-powered hub.

In addition, the hub is enumerated as a low power device (MaxPower 100mA). If I plug in a high power (500mA) device, the hub still enumerates as 100mA. Is that to be expected?

djvg
  • 173
  • 8

1 Answers1

5

The tactics of presenting a bus-powered hub as "self-powered" is a standard cheating practice.

The reason most manufacturers do this is that a certifiable bus-powered hub must have (a) power switches on downstream ports that don't engage until the hub gets enumerated, and (b) bus-powered hubs must obey power rules and system can enable only "low-powered" (<100 mA) devices, otherwise the power budget of upstream (host) port will be definitely exceeded. (USB3.0 has more reasonable power budget rules, but it is besides the point). The switches add a lot of cost, and 100 mA limit is a major inconvenience for customers.

As result, the Chinese industry invented this kind of cheating hubs, I call them "semi-self-powered hubs". Even if it doesn't have external power, they make its descriptors as it is "self-powered". It can't be USB-IF certifiable as result, and they can do anything with power distribution. I have one USB 3.0 hub that has about 800 uF total capacitance effectively loading its upstream port directly (while specifications allow no more than 10 uF total at the moment of connection).

This cheat also allows system software not to engage the power budget policing, so customers can connect any number of devices including those who consume 500-900 mA, and OS won't object to this. It works up to the point where the upstream (host) port can supply the load, and it might continue to work until the connector melts down or system power supply collapses or port overcurrent trips.

So yes, this brutally violates all good intentions of USB specifications, but it is quite practical.

Ale..chenski
  • 38,845
  • 3
  • 38
  • 103
  • As an example of "user inconvenience" with a hub that meets USB specifications, see this topic, https://superuser.com/q/1350148/620011 – Ale..chenski Mar 26 '19 at 22:57
  • Thanks for both answers. Very interesting. I tried several other "bus powered" hubs I have lying around, and, indeed, all of them make use of the same trick. – djvg Mar 27 '19 at 07:28
  • @djvg, I would hazard to guess than none of these hubs proudly carries the USB-IF certification logo, only maybe a bootleg one. – Ale..chenski Mar 27 '19 at 08:09
  • You are right: they don't. – djvg Mar 27 '19 at 08:14
  • Even AmazonBasics hub does this https://www.amazon.com/gp/product/B003M0NURK/ – endolith Dec 17 '20 at 16:08