USB 2 or USB 3

[masster]

HelloI have a simple question to ask.I have this 512 GB flash memory USB 3.0 stick made by Kingston (HyperX Savage). The male USB connector has only 4 pins.From my knowledge the USB 3.0 standard mentions the connector should have 5 more pins for extra signals.My question is: with only 4 pins (as for USB 2.0) will the stick work at full USB 3.0 specifications?

[timgiles]

No, if you have only 4 pins then there is only 1 pair of differential pairs which means USB 2. USB 3 requires a further two pairs of differential pairs, one for receiving and one for return of data.

Strange that Kingston state it is a USB 3.0 device.... most likely their marketing department deciding to stretch the compatibility built in to the USB 3 standard to almost breaking point...

[Lusu]

It should have another row of pins behind the 4 that you see in front, see here for example: http://sc01.alicdn.com/kf/HTB100HuKFXXXXcXXXXXq6xXFXXXS/USB3-0-connector.jpg

If it has only 4 pins, it is USB 2...

"The connector has the same physical configuration as its predecessor but with five more pins.
The VBUS, D−, D+, and GND pins are required for USB 2.0 communication. The additional USB 3.0 pins are two differential pairs and one ground (GND_DRAIN). The two additional differential pairs are for SuperSpeed data transfer; they are used for full duplex SuperSpeed signaling. The GND_DRAIN pin is for drain wire termination and to control EMI and maintain signal integrity." (https://en.wikipedia.org/wiki/USB_3.0)

[masster]

Thank you for the answers.Update: Out of frustration and in order to be 100% sure, I opened the stick and to my surprise the PCB has pads for accommodating 9 pins but only 4 pins  from the USB connector are soldered. Plus, minus, D+ and D-.

I am wondering if the remaining 5 pins of the male connector are derived from the other 4 pins by pulling them to high or low with resistors. I see a bunch of resistors on PCB around those extra 5 pads.

[Lusu]

They are most likely floating, since they are not connected. The 0 ohms resistors on the back probably shunt the extra USB3 lines of the chip to the ground, the chip seems to have the traces for USB3...

[0xdeadbeef]

This sure looks weird and messy. My first idea was that it could be a fake, but the Phison PS2251-08 looks legit and seems to be used in Kingston flash drives.
Actually, on the pictures it looks as if the upper 5 pins were cut on the connector. Also there seems to be a botch wire between the fourth and fifth pad (counting from upper to lower side).
Without that botch wire, I could imagine something like that could also happen in production if they get a defective USB 3 connector. If there really is a botch wire, it must be a repair attempt.
Which raises the question about the origin of this thing. Was it bought new from a retailer or from some guy on eBay?

[EDIT]
Side note: if the upper five pins are not connected, both, the superspeed transmit and receive LVDS channels (full duplex) are missing. This leaves only the D+/D- (half duplex) USB2.0 LVDS lines.

[masster]

Quote from: 0xdeadbeef on December 19, 2018, 10:45:12 pm>

Which raises the question about the origin of this thing. Was it bought new from a retailer or from some guy on eBay?

It is from a reliable dealer.

[tooki]

Is there no chance that the other 5 pins are actually connected via surface-mount pads underneath the blue plastic? Cuz it is categorically impossible for those drives to reach their rated speeds (180MB/sec and up — not megabits, megabytes!!!) using only USB 2.0. So either the extra 5 pins are there and are connected somehow, or the drive is fake (which is absolutely possible even from a reputable dealer, since supply chains can become contaminated).

Have you actually performed any speed tests on the drive?

[fsr]

All the move from parallel to serial interfaces, and now to get more bandwidth they add more "lanes". So, if you transmit two or more bits at a time, is it really still a serial interface?

[tooki]

Others can probably answer this better than I can, but I believe that the key difference is that in classic parallel interfaces, the individual lines must remain in perfect clock sync, whereas in modern multi-lane serial systems, each lane is clocked independently so that minor timing issues between lines don't affect anything.

[Bratster]

it's kind of hard to tell from the poor quality of the first photo, but it looks like that connector has been bent down partially ripping it off the circuit board.

Near the top you can see what looks like a crack in the circuit board where the connector mounting pin meets it, as well as on the bottom you can see that the mounting pin looks like it is torn out part of the through-hole plating along with it.


It does look like there are pins soldered to all nine pads on the board with the super speed pins being broken in half.


Need a higher quality photo to get any more info than that. As well as one from the side showing how bent the connector is.

Sent from my Fi Moto x4 using Tapatalk