Compared to USB 2.0, USB 3.0 is a miracle long overdue. USB 3.0 ups the physical layer rate from the 480Mbit/s to a more reasonable 5Gbit/s – and is on its way to 10Gbit/s, while alleviating shortcomings in power delivery by upping the current limit from 0.5A to 0.9A without the battery charging/supercharging extension. In order to do all of this as cheap as possible without breaking backwards compatibility – they had to introduce a new row of conductors buried inside the A connector and straddling on the side of the various types of B connectors creating cable-compatibility issues – i.e. the USB 3.0 B-connectors will not mate with USB 1.1/2.0 B-sockets
It’s a small price to pay for a drastic increase in performance – no longer are we restricted to seeing rates of 35MB/s and below (sometimes quite significantly lower) when talking to external USB hard drives – we can now attain sequential rates limited by the spinning media in such drives even without UASP – and with UASP to overcome problems with command latency and lack of queueing in the older Bulk-only Transport mode of mass storage, performance improves even more!
But there was always one thing nagging me – while USB 3.0 is obviously great, and the host ports are backward compatible down to USB 2.0 and 1.1, USB 3.0 ports are always rather scarce! I guess we can put this down to cost cutting on behalf of the chipset makers and motherboard manufacturers – high speed USB 3.0 PHYs were not easy to make cheaply, at least initially. But now more USB 3.0 devices can be had at reasonable costs – sometimes even at the same prices as their USB 2.0 bretheren – doesn’t it make sense to have more ports?
To hub, or not to hub?
Before I actually review the device in question – we should probably review whether having a hub is a good idea in the first place.
It’s probably something not often practiced – but you can solve the lack of ports issue with a PCI-E to USB 3.0 adapter. These can be had for about $12 a piece, and if you have free PCI-e x1 slots, it’s definitely a possibility. Unfortunately, these cards usually only offer rear ports, and only two ports – but there’s a big advantage.
A separate controller means a separate USB 3.0 bus! This means stuff attached to the other USB 3.0 ports are not sharing the bus bandwidth. Copying from external drive to external drive can be speeded up quite a bit by attaching them to separate buses – and I practiced this on USB 2.0 quite often – resulting in copies running at 28MB/s rather than a lowly 13MB/s (if attached to the same bus).
Now that you know this “secret” – maybe a hub isn’t the thing for you. If throughput is what matters, look into getting more controllers first!
The other issue is that low quality hubs can often wreak havoc with compatibility with some devices, causing drop outs or “This device can perform faster if attached to a … port” messages. And the bandwidth is shared – so it’s more a convenience thing of not having to plug and unplug devices and being able to address more devices simultaneously.
Also note that bus-powered hubs will often cause problems with devices which are not self powered – as the cable length and traces will cause voltage drop due to resistance – so attaching drives to bus powered hubs is a definitely no-no. Voltage drop will also occur over the back-feed prevention diode (which is sometimes omitted in cheaper hubs).
A self-powered hub (one with a transformer) should be used in this case – but a low quality power supply or poor design of the hub can cause strange device issues or even power back feed issues as well.
In light of this, finding a good hub is important for the reliability of the connection of devices attached to it. But often, good quality hubs can’t be found just by looking at them – some controllers are more tolerant of some timing problems than others. Other hubs can be let down by simple things like poor soldering and poor quality connectors which cause intermittent USB connection.
Hubs can also have an impact on the performance of your older USB 1.1 devices as well. This is because USB 2.0 and 3.0 hubs run a high speed link between the upstream hub/controller and the hub and rely on units called Transaction Translators (TT’s) inside the hub to translate the High Speed (split) transactions into Full Speed packets for USB 1.1 devices. A hub with individual TT’s for each port will essentially give each port’s USB 1.1 bus the full USB 1.1 bandwidth – so older devices can perform to the best of their ability. Some earlier hubs shared one TT amongst all ports to save cost – and on the rationale that older USB 1.1 devices are dwindling in popularity and are unlikely to be connected in large numbers to the hub. Each to their own, I suppose.
The hub in question
The hub in question was purchased from eBay – and seems to be quite popular around eBay. Cost was around $12, and it is a 4-port USB 3.0 hub with a cable. It is bus powered with provision for self-powering, but it does not come with a power supply. Knowing what the quality of some Chinese unbranded power supplies are like – I’d say that might prove to be safer for your devices, but will exclude the possibility of running external hard drives from the hub.
The hub features vertical connection which is rather interesting. The case is black plastic, and there is one LED to indicate the presence of power to the hub (bus or external jack). The plastic shell is held together screwless and gluelessly by internal plastic pins.
So lets open it up.
Taking a look at the PCB, it is already evident that this is a USB 3.0 hub enabled by a Genesys Logic GL3520 USB 3.0 Hub controller with independent transaction translators. Good news, sort of – since Genesys Logic products seem to be fairly decent nowadays, but early on in the USB 1.1 era, I avoided Genesys Logic hubs due to reliability issues.
We also see a Chingis Technology Corporation Pm25ld512 SPI Flash which likely holds the configuration information for the hub. There is also an inductor and a small ChongX capacitor which appears to filter the incoming power – although the capacitor itself is from a questionable manufacturer and seems a bit small for the job. Notably absent are any polyfuses or current measuring shunts – so there is no port over-power protection by the looks of it. But there is an external power backflow prevention diode, and what appears to be a transistor to switch between bus and external power.
There is a QC/warranty sticker sitting over the silkscreen PCB job identifier – so I went ahead and removed it, but am none the wiser as to who is the OEM for this product.
Turning the PCB over – we can see a half decent soldering job with most important pins soldered on well. Unfortunately, it looks like one side of the shell wasn’t soldered down on all ports – it would be wise to solder it down as it provides additional mechanical reinforcement to the ports. Aside from that, there were a few small solder splashes, but I removed the loose solder by hand before taking the photo.
Plugging it in, I found it was functional and identified itself as follows –
- VID 05E3 PID 0610 USB2.0 Hub
- VID 05E3 PID 0612 USB3.0 Hub
This would be expected since USB 2.0 and USB 3.0 are “separated” stack wise, so it’s presenting as two hubs. Self-powered and even low-power devices like USB 3.0 card readers operate flawlessly off the hub – but devices which are a bit more power hungry like my SSD in a box fail to operate properly. Just as expected.
For the price, one cannot complain about it. It does work, although it could do with proper port overcurrent limiting, although this is only a minor safety concern when not powered with external power. The physical construction feels appropriate for the price, although the soldering could do with some help. It does have a backfeed prevention diode, which is good, and it does work with my USB 3.0 ports for low-current-draw devices or self-powered devices. Just don’t expect to power hard drives or DVD writers from it.