0

I am currently troubleshooting a single port in a two-port USB 2.0 bracket (example here) that does not work. A USB bracket is a simple device (consisting only of cables and connectors) that connects the pins of a USB header (on the motherboard) to two external USB ports.

The symptoms of the problem are as follows:

  1. It is only a single port of a two-port bracket that does not work.
  2. Whenever a keyboard or mouse is plugged into the port, it is only working once. That is, if I unplug then re-plug it, it is not detected until I reset the computer. None of these happened if I plug them into the other port.
  3. Whenever any other, more complex device is plugged into the port, they are behaving differently yet none of them is detected. When I plug a bus-powered external drive, the power light goes on and the drive spins. When I plug a cell phone, the phone connects briefly then disconnects. None of these happened if I plug them into the other port.
  4. Inside Windows, although the device is not detected, Windows is playing "Device Disconnect" sound when a device is connected, and playing "Device Connect" sound and showing a "USB Device Not Recognized" balloon when a device is disconnected.
  5. All the above symptoms also happen both in BIOS/POST and inside Windows (not detected = not shown in POST device list = not working in UEFI BIOS setup).

The troubleshooting steps I've taken are as follows:

  1. First, I am certain that it is the bracket (not the motherboard) that is having problem because no matter to which header I plug the bracket, it is always the same port in the bracket that is having problem.
  2. I've done continuity tests between both external ports' pins and the bracket connector's pins and they are all good (all less than 1 Ω).
  3. [Update #1] I've done continuity tests between all pins of the bracket, and found that the two Vcc pins are shorted to each other, and the two GND pins are also shorted to each other. All shorted pins' resistance are less than 1 Ω, everything else is "OL".

So, what could the problem possibly be? As the USB bracket is a custom part deep inside an old computer case, if at all possible I want to repair instead of replacing it.

2 Answers 2

1

I can't fully explain all the behaviour, however I may be able to help you a bit.

Q1. Have you checked that their are no shorts between wires in the port of the USB bracket that is failing?

As you know, their are three different USB electronic signalling systems which all share the same old USB 2 plug and socket.

When a USB device is plugged into a USB host, the host USB port 'checks' which of the two devices' data wires is being pulled high (with a modest resistor), and chooses to communicate as low-speed or full-speed, at about 3.7 volts (IIRC). That usually covers mice and keyboards. After that the host might need to do a bit more work to discover a High-Speed USB device (e.g. disk drive), which runs the data signals at a much lower voltage, and so would exhibit somewhat different symptoms. This process is called enumeration.

When a plug is removed from a socket, the 'default high' data wire will be pulled down by resistors on the hosts side. So the host's USB port will recognise a disconnect. I think that disconnect could be triggered by a data wire being partly-shorted to ground.

The point is, every time a device is plugged in, the electronics in the USB host port should trigger software enumeration to start, and whenever the data wire is pulled out, and hence stops being pulled up by the devices resistor, the host USB port should trigger the host software to believe USB device is unplugged.

So, some of the symptoms you have found, sound like enumeration may be being triggered to the host, but not completed, or disconnect has been signalled to the host without removing the cable.

All I can think of is either there is something wrong in the external USB lead, or their is an intermittent break or partial-short within the unreliable bracket socket or the wires to the motherboard, involving a data signal.

You could tell if there is a partial-short, with some resistance, by measuring the resistance of every wire against every other wire. It should be huge. However, if it is a few k ohms, it might explain a lot of the behaviour.

A related, but different possibility might be, again a partial-short, but this time causing the external device to look like it is taking too much power, and causing the host to shut down power on that port.

I apologise if you have already been through all of these, but I thought they might help.

4
  • Thanks for the answer. No, these are all new to me as I haven't checked for any shorts. I will test the ports and report back here soon.
    – user52282
    Aug 30, 2014 at 2:48
  • I am in the UK, and off to bed. So best wishes.
    – gbulmer
    Aug 30, 2014 at 2:57
  • I've done continuity tests between all pins of the bracket, and found that the two Vcc pins are shorted to each other, and the two GND pins are also shorted to each other. All shorted pins' resistance are less than 1 ohm, everything else is "OL".
    – user52282
    Aug 30, 2014 at 8:26
  • Okay. Have you checked the ground 'shield' of the USB socket for continuity with the other wires? After that, my last thought is the problem is in the socket inside the bracket, and/or the problem is activated when things are plugged and unplugged. This might be too difficult to repair, depending on the way the bracket is made. How have you been doing the continuity test? To detect that it is the socket, you need to do the continuity tests via a USB lead. I think that needs a USB cable to be disassembled or cut. You might not want to chop up a cable, and this might not find the problem.
    – gbulmer
    Aug 30, 2014 at 11:04
0

Well, I finally solved the problem. The cause is probably dust buildup. I disassembled the case down to the USB bracket assembly and lo and behold, they are covered with thick dust and grime. The dust and grime are so thick that no amount of blowing and dusting will clear them. At this point I do not bother with measurements anymore as everything but the ports (i.e. the PCB, screws, solder legs and tracks) are so covered with dust and grime. Despite the dust and grime, the components look okay and there are no burn or scorch marks. I then cleaned them with Q-tip and isopropyl alcohol. After I cleaned them and reassembled the case the problem does not appear anymore.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .