24
1
I know USB 3.0 is almost entirely backward-compatible, and I know that it introduces a new speed that USB 2.0 devices aren't capable of, but is there any advantage to having a USB 2.0 device in a USB 3.0 port?
Though I'm interested in if it would provide any benefit for any device, I was specifically thinking of a USB Hub that I plug my Bluetooth receiver and flash-drives into.
Nich Del
Posted 2011-06-05T16:02:17.740
Reputation: 378
11
Actually, yes, it will be faster by a small margin. You will only see gain if the device in question can dish out a higher bandwidth over another interface like ExpressCard or PCIe. for instance a modern 7200 hard drive in a external enclosure could more than saturate the USB 2.0 port. If the enclosure is a USB 2.0 device, it will be operate with more of its bandwidth when plugged into a USB 3.0 hub, but not nearly as much as if it was a USB 3.0 to USB 3.0 device to hub link (with a USB 3.0 cable).
At least on my laptop, USB 2.0 external 500 GB on USB 2.0 gives me about 19–23 MB/s and up to 25–32 MB/s when connected to a USB 3.0 express card. So I am getting both a higher minimum speed and ceiling when the same USB 2.0 device is on the USB 3.0 hub. I think the controller is probably more efficient on USB 3.0. When I plugin a USB 3.0 thumb drive on the same ExpressCard USB 3.0 hub though, I get up to 122 MB/s.
So short answer; yes, a small increase, but not nearly as fast as native USB 3.0 links.
ganja_bake
Posted 2011-06-05T16:02:17.740
Reputation: 126
4
This is actually the most correct answer, based on data not (just) theory. There is a measurable difference, but it's very unlikely to be perceptible until your operation is in the tens plus minutes range. Anandtech whipped out the tools and benchmarked reads and writes from/to an SSD to various combinations of USB2-in-3 ports, USB 3.0 Flash Drive Roundup. Best case? Shave 19 seconds off 228 seconds. (They also tested US3-in-2 if you're curious.)
– matt wilkie – 2015-01-03T23:03:56.0431Ditto. This answer backs up a lot of real world experience I have seen. USB 2.0 max speeds are just max speeds; not promised speeds. As controllers get better so do the speeds of those controllers. So it makes sense that a USB 3.0 controller would be more balanced and perform better—overall—than a plain USB 2.0 port. – JakeGould – 2015-02-28T06:04:26.343
"If the enclosure is a USB 2.0 device, it will be operate with more of its bandwidth when plugged into a USB 3.0 hub" Why? – endolith – 2016-02-04T15:33:48.780
@endolith, USB3.0/3.1 xHCI controller does have a better, more software-friendly interface to handle USB pipes/ring buffers, and handles all speed rates uniformly (not like a special UHCI/OHCI/EHCI mixture), and the software is written more recently having less overhead and no legacy bug fixes. So the aggregate bandwidth for USB2.0 devices is somewhat better, which is perceived as "better performance". – Ale..chenski – 2017-02-01T01:36:56.223
20
Since your USB is optimized for USB 2.0, using a 3.0 will see no improvement because it simply cannot operate at 3.0 speeds.
USB 2.0 has a maximum speed of 60 MB/s USB 3.0 has a maximum speed of 625MB/sec
From Wikipedia's Article on the Universal Serial Bus:
Typical hi-speed USB hard drives can be written to at rates around 25–30 MB/s, and read from at rates of 30–42 MB/s, according to routine testing done by CNet.[62] This is 70% of the total bandwidth available.
Based on this, you can see that USB 2.0 devices just are not capable of the speeds 3.0 has to offer.
TL;DR version: You will see no benefits
Simon Sheehan
Posted 2011-06-05T16:02:17.740
Reputation: 8 641
Please note that the maximum USB 3.0 Byte rate is 500 MBytes/s, not 625, all because of 8B/10B redundancy encoding. Each byte takes 10 bit times in serial stream. – Ale..chenski – 2017-07-04T23:11:13.733
3That quote is unrelated to the question; it's about USB-hard-drives specifically, and is saying that even USB 3.0 hard-drives won't benefit from plugging into a USB 3.0 port, because they don't saturate the available USB 2.0 bandwidth as it is. There could very well be devices (such as flash drives) which could otherwise operate at USB 3.0 speeds, but can't simply because they are not USB 3.0-compatible. – BlueRaja - Danny Pflughoeft – 2011-06-05T22:09:43.137
2@BlueRaja it was meant to be taken as an example. – Simon Sheehan – 2011-06-05T22:41:51.190
3+1 for the TL;DR version. Though I had already read the whole answer :/ – Gani Simsek – 2011-06-06T09:38:58.890
1@BlueRaja: It may be saying that USB hard-drives don't saturate USB 2.0, but that would be wrong. USB doesn't allow any single device to use all the bandwidth, and the transfer rates seen by USB disks are very close to the theoretical maximum for any USB 2.0 device (once protocol and overhead are considered). – Ben Voigt – 2012-07-01T00:10:03.853
@BlueRaja Actually, the mechanical hard drive is typically capable of saturating USB 2.0. 70% sounds about right, accounting for overhead. As a general rule of computing, networking, etc., you will never hit the theoretical limit of any interface. In any case, if the HDD interface is USB 2.0, a USB 3.0 port will not raise the theoretical maximum. – Bob – 2013-05-01T05:57:41.307
16
One advantage could be that USB 3.0 can supply more power than USB 2.0.
I have some doubt whether an USB 2.0 device could use that power, as it would be designed for a USB 2.0 port. On the other hand many USB 2.0 devices exceed the specified power and get away with it (mostly external disks when they start up).
starblue
Posted 2011-06-05T16:02:17.740
Reputation: 573
Some USB 2 devices have problems when connected via a hub. Would be interesting to know if they have fewer problems if the hub is plugged into a USB 3 socket with its higher power rating. – pelms – 2011-06-05T22:53:15.990
I'm gonna open this question back up and hope someone has any definite answer about this. – Nich Del – 2011-06-05T23:49:13.823
2This is interesting because technically (according to specs) USB 2.0 can only draw one unit (100 mA) without negotiation, and up to five units (500 mA) upon request. USB 3.0 can draw one unit (150 mA) without negotiation and up to six units (900 mA) upon request. Thing is, many USB 2.0 devices don't negotiate properly. And many computer motherboards just supply power without caring about negotiation. So the correct answer would be that it's more likely that a USB 3.0 port could supply higher amounts of power, but a port fully following standards does not have to supply that amount. – Bob – 2013-05-01T05:56:11.067
8
At least on my computer, the 3.0 port is faster than the 2.0 port. However, that is because it is on a different controller, which is faster than the one built into the chipset.
Basically, if the usb 3.0 is from the same controller as the usb 2.0 port, it's going to be the same. Otherwise, YMMV.
Eli
Posted 2011-06-05T16:02:17.740
Reputation: 802
Emphasis on the word "might". – Jamie Hanrahan – 2017-02-01T01:08:54.647
1This is the only right answer. USB3 hardware tends to be more sophisticated than USB2 (faster microcontrollers, more cache memory, faster PCIe interface) and this might translate to benefits for high-speed devices. – Ben Voigt – 2011-10-09T19:21:30.163
6
None, other than it will work. You gain no performance unless using USB 3.0 devices.
KCotreau
Posted 2011-06-05T16:02:17.740
Reputation: 24 985
0
I believe it is faster, here is why:
I have a Logitech Solar Keyboard. I had the unilink receiver plugged in to the 2.0 port and the keyboard would constantly lag to the point that I was going to return the solar keyboard to the store. I tried everything including contacting support, downloading the SetPoint app, switching ports (USB 2.0), etc... On a hunch I believed that the USB 2.0 port itself was the bottleneck. Sure enough, I plugged the unilink into the 3.0 port instead of the 2.0 port and never had a problem since.
I verified this later on because I have 2 additional monitors connected to my PC via a Diamond (BVU165) through a USB 2.0 powered hub. Whenever I would watch movies or youtube videos on the PC, after 2-3 mins, the hub would crash and I could hear my scanner (also connected to the hub) reboot and go through the warm up procedure (USB powered only). During this crash, the monitors would disconnect (turn black and go to sleep) and reconnect to the PC. I since attached a POWERED 3.0 hub which I connected the Diamond BVU165 (USB 2.0) through the USB 3.0 hub and never had a problem since.
I am not promoting this as a fix all for every device. Using the 3.0 ports instead of 2.0 seems to even out any sort of lag due to bottlenecks dealing with 2.0 technology and gives the devices the best chance of working properly.
Is there any studies to determine if USB 3.0 unpowered is faster than 2.0 powered?
Brian
Posted 2011-06-05T16:02:17.740
Reputation: 1
0
There is a fundamental misunderstanding beneath this question, and many similar questions.
Reality is that when you plug a USB 2 device into a "USB 3 port", you aren't really plugging into a USB 3 port! Not electrically, anyway.
The USB 3 port has all of the pins for a USB 2 port in it. The USB 3 cables have wiring for both. The host controller actually has logic for both USB 2 and USB 3 in it, and enumerates on the host system as two controllers: One USB 2/1.1/1, and one USB 3.
When you plug a USB 2 device into a USB 3 port, the USB 2 device uses the same pins that it always does, and those pins connect to wires that take the connection all the way back to the USB 2 controller int he host. Which does the same things that any other USB 2 controller would, and has the same 480 Mbit/s speed limit.
It is not the case that the hub or whatever takes some of the USB 3 bandwidth and gives it to a USB 2 device if one happens to show up. The controller provides both a USB 3 bus and a USB 2 bus. The USB 2 bus has all the bandwidth available that it usually does, without taking anything from the USB 3 side.
It is also not the case that having USB 3 capabilities implemented in the same controller - even if in the same chip - as the USB 2 controller will make the USB 2 controller work any faster.
So there is no reason from the architecture to expect that a USB 2 device plugged into a USB 3 hub (or port) will work any faster than if it's plugged into a "native" USB 2 port.
As others have suggested, it may be the case that a new USB controller that supports USB 3 as wwell as 2/1.1/1 might be faster at doing USB 2 than an older USB 2/1.1/1 controller. But it isn't inevitable.
Jamie Hanrahan
Posted 2011-06-05T16:02:17.740
Reputation: 19 777
Jamie, you said: "Which does the same things that any other USB 2 controller would". This is incorrect. Every interface has two sides, front end (USB PHY), and back end (USB controller - system interface). And every interface has overhead. The newer xHCI controller specification has (a) faster bus-mastering engine (to satisfy SS rates), and (b) has unified hardware-assisted control over USB pipes and transaction buffers using the same bus-mastering engine. Therefore it is totally understandable why xHCI controller provides less overhead and gives better USB2 performance even over the same 480. – Ale..chenski – 2017-02-01T00:46:03.243
-1
Some of the things stated on here are so wishy washy and guess like, don't you think that a USB 3.0 controller is for USB 3.0 and they simply have a separate controller for USB 2.0, I know my machine has a separate controller chips for both variants I can see it under the hardware listing. Unless you buy one of those cheap motherboards made by a manufacturer that has like 4 ports all feeding off the same controller like a cheap skate which TBH I have never seen a controller with a mixed port setup though that doesn't mean that they don't exist. USB 2.0 is USB 2.0, it cannot run any faster than USB 2.0 that's why it's called USB 2.0, even the wiring that makes the USB port is of a different gauge to that of a USB 3.0, this is evident by the different colour plugs with blue usually signifying USB 3.0, because the wiring is different it wouldn't matter even if they ran off the same chip-set because of the wiring, there are not extra wires or anything as such but the gauge of the wiring is different to accommodate not only for increased speed but voltage if required, otherwise if this wasn't the case it would be pointless to colour code the cables as every cable would support USB 3.0. I mean think about it logically as a manufacturer of cables I could save my self hundreds of thousands in costs as the only thing that would need to be colour coded is the port.
Miles
Posted 2011-06-05T16:02:17.740
Reputation: 1
2This wall of text is a rant not an answer. – DavidPostill – 2017-03-03T23:30:23.507
The text is quite wrong. Yes, some MB do have separate USB2 (EHCI) and USB3 (xHCI) controllers, but the USB2 wires within USB3 xHCI port are handled by a different, xHCI controller, not the older EHCI. The overall transfer speed deviation from theoretical 60 MBps limit is not only in inherent bus overhead, but mostly in system overhead, to prepare data buffers and synchronize raw data with filesystem. The xCHI controller is done better, and its S/W driver has less system overhead. This makes a small difference, as people noted. – Ale..chenski – 2017-07-04T14:23:03.153
3No. The USB device that you are using cannot go faster. – soandos – 2011-06-05T16:06:21.003