People who just don't listen to troubleshooting.

[Stonent]

I had an issue last night at work where a specialty device we had that converts USB to single or multimode fiber and back to USB again had apparently failed. The USB  ports on the device were not providing power to anything plugged in.  We took the USB devices and carried them back to the PC which was about 200 feet away and the devices powered up fine on USB.  Just as a precaution, we replaced the USB device that was plugged into this fiber to USB box and it still wouldn’t power up.

We took the unit and power supply back to my desk and I did some testing with a multimeter.  The power supply is 5V 3A and it was putting out 5.5V with no load. So that seemed ok.

I cut a USB cable open and probed the power and was getting .5, .69, 1.0, and .05 V on the 4 ports on the front of the unit.  I plugged in an optical mouse and got no power at all.  There were what appeared to be some diodes on the +5 line on the USB ports.  I probed both sides and got less than a volt. I found a polyfuse and probed both sides and it was working fine.

I opened it up and found a TI TPS2044BD that was very close to the USB ports.
http://datasheet.octopart.com/TPS2044BD-Texas-Instruments-datasheet-149034.pdf


Here’s what the datasheet says about the part.

The TPS204xB/TPS205xB power-distribution switches are intended for applications where heavy capacitive
loads and short circuits are likely to be encountered. These devices incorporates 70-m? N-channel MOSFET
power switches for power-distribution systems that require multiple power switches in a single package. Each
switch is controlled by a logic enable input. Gate drive is provided by an internal charge pump designed to
control the power-switch rise times and fall times to minimize current surges during switching. The charge pump
requires no external components and allows operation from supplies as low as 2.7 V.
When the output load exceeds the current-limit threshold or a short is present, the device limits the output current
to a safe level by switching into a constant-current mode, pulling the overcurrent (OCx) logic output low. When
continuous heavy overloads and short-circuits increase the power dissipation in the switch, causing the junction
temperature to rise, a thermal protection circuit shuts off the switch to prevent damage. Recovery from a thermal
shutdown is automatic once the device has cooled sufficiently. Internal circuitry ensures that the switch remains
off until valid input voltage is present. This power-distribution switch is designed to set current limit at 1 A
typically.

So as I understand that it takes 2.7 to 5.5V and outputs 5v and current limits it.

I probed the ground and input line and got 5.5V again.  The output lines were all around 2V to 3V which is much less than I was expecting.  The output lines all disappeared into some vias so I couldn’t tell where they went.

We did not have a spare of this item as it is around $900 so I let everyone know that the ports were not putting out but less than 1V.

This morning I get an e-mail. That they reinstalled it and said “Oh it was just a PC COM port setting”

Last night this thing had no power and the device manager did not list any of the devices that were plugged in.

So today the person that reinstalled it said "Oh it was just a com port issue on the PC, I think you just need to drop it, because it is working now. I don't see what your deal is."

My deal is this thing is flaky and has cut out 3 or 4 times in the last week where I've had to power cycle the whole unit (it is mounted in an industrial cabinet) not knowing where the issue was.

 

[ConKbot]

If you can blame the software guy, do it

[c4757p]

Yeah, come on, don't you know that yet? "It's the software guy's fault" is rule numero uno of having software guys.

[Bored@Work]

It would be a shame if something unexpected happens to that device, wouldn't it?

[Stonent]

Ok well I heard back from the vendor. They said some PCs can't provide enough current to the box inside the building. (I didn't realize it until today that the inside building unit is USB powered) If it can't get enough current, the signal cuts out and causes the unit on the other end that the USB devices are connected to, to try to boost its signal to the fiberoptic cable and that causes a power drop to the USB devices.

I'm not 100% sure that's the issue because the external device was putting out less than a volt.

But he had me take some voltage measurements and here's what I got.

USB Voltage without device: 5.5V
USB Voltage with device: 5.2V
Current draw: 280mA.

To me that looks ok. USB is supposed to be able to feed 500mA.

[Experimentonomen]

USB is limited to 100mA unless a handshake is made between the pc and the device.

[ConnorGames]

But almost nothing actually enforces that.

[ejeffrey]

That sounds pretty dodgy.  The remote unit can boost the optical signal to the point that its external power supply can't power it and the connected USB device?  That is insane!

It isn't 100% clear to me what the setup of devices is here, but while it is true that few USB hosts enforce the 100 mA until negotiation for high power, if you try to draw more than is actually available, you will have problems.  For instance if you connect multiple devices that unconditionally draw 500 mA to a bus powered hub, you are going to have a bad day.

[bilko]

Just unplug it and test it with a megger 

[Stonent]

Well they offered to send us an add-on power supply for the USB powered half, hopefully that will fix it.  I also shortened the USB cable from 12 feet to about 6 or 8 feet.  That should save a little bit.

They've also said they've had several firmware updates since the product came out and gave me a link to Digikey to buy an FTDI cable so I can program it. I thought that was funny.