2

The standard USB 2.0 and USB 3.0 hubs may not get sufficient power through one USB port to power all devices attached to the hub. Therefore, an externally powered USB hub is usually recommended when plugging in many power-hungry USB devices.

USB-C can negotiate for a lot of power, which suggests that a USB-C hub might be able to negotiate for enough power being fed to it from the host computer to enable it to support a collection of more demanding devices. Does a similar requirement for an external power adapter also apply to the newer USB-C hubs?

5

"Standard" USB hubs never can get sufficient power through a single upstream cable, simply because there is only a single connection, which cannot, by very definition, supply four other downstream ports. If somebody wants to have full standard power from downstream ports of a hub to supply all devices, the hub MUST HAVE ITS OWN POWER from a wall adapter. It is not "recommended", it is a must, since bus-powered hubs are formally PROHIBITED from supplying more than 100mA/150mA per port, in accord with USB specifications, and the host system will refuse connection to full-powered devices if the hub reports itself as "bus-powered". If some hub reports fake descriptors (posing as self-powered, while it is not), this will allow to channel multiple full loads from upstream port, and is a brutal spec violation. The Type-C connector is not much different with this regard, since four is still bigger than one.

However, Type-C has a native mechanism of providing higher power over USB ports by means of different pull-up values on CC pin, which can indicate port capability up to 3A at 5V, or even 5A if Type-C cable uses "electronic markers". The Type-C port therefore provides more options for powering downstream ports. In this case the hub must be more intelligent, and change its descriptors dynamically to reflect these additional power capabilities, such that the USB host can manage power budget accordingly. At the moment (oct.2016), I have not heard about such intelligence in USB hubs, nor that this communication is defined in USB specifications.

Additionally, the Type-C connector can support a totally separate standard called "PD", Power Delivery. If a hub does support PD , and the host port does support PD (which are two very big IFs), then theoretically the hub can get up to 100W from the host, and do quite better in supplying power to downstream ports. I have not seen this kind of designs yet, but would wager a guess that in PD case the hub should report itself as full-sized self-powered hub. Only in this case a full-function hub may have no external AC-DC adapter.

The thing is that USB, Type-C, and Power Delivery are three independent and loosely connected standards. So it is left to OEM designers how to combine these features/functions into a product. Having the round-shaped Type-C connector does not automatically imply that it has PD, or USB3, or even USB at all. It could be HDMI, or debug port, or just a power jack.

  • 2
    Your new paragraph is a little confusing. The question asks specifically about USB-C hubs in relation to power (although it isn't explicit about high power). Pretty much by definition, that would require external power (especially if you're talking about PD), because even a USB-C connection to such a hub couldn't support multiple USB-C connections. What is your point in the new paragraph -- that a USB-C hub conforming to the spec doesn't yet exist? can't exist as defined by the spec? – fixer1234 Oct 14 '16 at 20:33
  • 1
    What I tried to say that power delivery and new features of Type-C connector wrt power are intentionally disconnected from USB specifications, for some reason unknown to me. There is no field to describe a port in terms of its power delivery capabilities. As result, there is no consistency in what needs to be reported in USB descriptors to host OS, so it could manage power budget, as it was originally envisioned with distinction between "bus-powered hubs" and "self-powered hubs". So technically a hub with Type-C connector definitely has more capability, just there is no way to know this, yet. – Ale..chenski Oct 14 '16 at 20:36
  • 1
    One more nit pic on a good answer. 1st paragraph, couldn't you have four low-power devices on an unpowered hub (i.e., never say never)? – fixer1234 Oct 14 '16 at 20:45
  • 1
    This is a good pick. I would say that if you have a really USB compliant bus-powered ("mobile") hub, it is very difficult to restrain yourself to "low-power device" only. I have an old Microsoft mouse that reports its power need as 500mA, so it will be rejected by OS. This is very annoying, especially when the MS implementation of power policy is stupid - to limit every port to 100/150mA. It seems clear that the policy can really count power, and if only one mid-power device is connected (while other ports are empty), it should allow it. For some reason this policy was not implemented. – Ale..chenski Oct 14 '16 at 20:58
  • 1
    So, "never" should read as "never with regard to typical USB spectrum of devices", which includes HDD/SSD USB enclosures, web cams, etc. I have a difficulty how to formulate this with brevity. – Ale..chenski Oct 14 '16 at 21:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.