Faulty top USB port (solved)

[Benny Blanco]

Hi all,

I recently received a RPi from RS and when trying it out for the first time last night I noticed my mouse wasn't working but my keyboard was. Did the standard troubleshooting and everything plugged in fine with a power supply I know is capable. Switching the connections around resulted in the mouse working but no response from the keyboard. Got my RPi from the living room which I'm using to run Raspbmc and that worked fine with both keyboard and mouse plugged in.

So basically looks like the top usb port is shot.

Had a quick search around the forum for any similar problems and there was some talk around dodgy soldered X1 pins. However, seeing as the ethernet and bottom USB port is working fine, I don't think this is the problem. Any other suggestions?

I was thinking possibly a dry joint on the actual USB shield pin but am reluctant to reflow the solder for voiding the warranty. I emailed RS about half an hour ago about how long a replacement would take but I'm guessing I could be in for a wait. Anyone had any experience with RPi returns from RS?

Thanks in advance!

[Benny Blanco]

Had a tinker with the multimeter to see if I could trace where the problem is. These are the voltages I recorded.

TP1 -> TP2 @ 4.86V
Bottom USB: VCC -> Gnd @ 4.86V
Top USB: VCC -> Gnd @ 2.93V

So looks like the top USB port is getting power but not enough. Does anyone know what could be causing the voltage drop or where I should look next?

As an aside, I found a non branded powered USB hub to test out the USB data lines which I connected to the bottom USB port. I didn't have the PSU plugged in to the microUSB but to my surprise the RPi powered up through the USB port. Is that normal behaviour?

Any help to solve this puzzle before I send it back to RS would be appreciated.

[dukla2000]

I received 2 Pis in May, 1 from RS and 1 from Farnell, and both had the same/similar problem. I have a wireless keyboard/mouse that uses a single USB dongle and on both Pis it wouldn't work in the top USB port but did work fine in the bottom port. Am not particularly bothered as my USB pendrive works fine in the top USB port on both Pis. Never aimed my multimeter at the ports as in essence can avoid the problem.

[mahjongg]

seems a faulty F1 polyfuse, although a lower voltage than normal should only occur when you are actually drawing current out of that port.
Polyfuses increase their resistance dramatically when they get too hot, which occurs when a short circuit current is running through them. After a few hours, max a few days they should return to a somewhat normal resistance, but probably double its original value of three ohms.

So basically you have a blown fuse!

[Benny Blanco]

seems a faulty F1 polyfuse, although a lower voltage than normal should only occur when you are actually drawing current out of that port.
Polyfuses increase their resistance dramatically when they get too hot, which occurs when a short circuit current is running through them. After a few hours, max a few days they should return to a somewhat normal resistance, but probably double its original value of three ohms.

So basically you have a blown fuse!

Thanks for the response, mahjongg.

Not the news I wanted to hear but good to know. At least I've learnt a bit more about electronics off the back of this, which I suppose is part of the RPi's agenda.

[RaTTuS]

have a look in the USB socket and see if all the mins are there?

[mahjongg]

Rattus probably means all the pin's, as there have been reports of improperly inserted connectors, with one or more pins folded under the connectors body, failing to make contact with the PCB, or causing shorts. If you look closely on the solderside of the connector you should see all eight pins of the double USB connector protruding through the solderholes.

[RaTTuS]

damn my typing ;-p
yes pins*
also ^WHS

[Benny Blanco]

Hi RaTTus,

Yep, all 8 PCB mount pins are there. I took the above voltage readings from them. The USB socket connectors are all present too.

As mahjongg described, I think it's just an unlucky faulty component as I couldn't see any physical board manufacturing defect.

[drgeoff]

The low voltage reading. Was that with something plugged into that USB socket or not?

[Benny Blanco]

The low voltage reading. Was that with something plugged into that USB socket or not?

Nothing was plugged in to the USB or ethernet.

[drgeoff]

The low voltage reading. Was that with something plugged into that USB socket or not?

Nothing was plugged in to the USB or ethernet.

For a polyfuse to drop a a couple of volts (with the only load being your meter) it would have to be very much higher resistance than normal. What voltage when something is plugged in to the USB socket?

[Benny Blanco]

The low voltage reading. Was that with something plugged into that USB socket or not?

Nothing was plugged in to the USB or ethernet.

For a polyfuse to drop a a couple of volts (with the only load being your meter) it would have to be very much higher resistance than normal. What voltage when something is plugged in to the USB socket?

Readings with devices plugged in are as follows. The RPi is sat at the wheezy login prompt:

With MS USB keyboard in bottom port and MS intellimouse explorer in top port (mouse LED doesn't light up but keyboard is working)

Mouse Top Port -> 1.59V
KB Bottom Port -> 4.85V
TP1 TP2 -> 4.86V

With the keyboard and mouse switched over (no response from keyboard but mouse LED does light up)

KB Top Port -> fluctuating between 0.21V and 0.25V
Mouse Bottom Port -> 4.66V
TP1 TP2 -> 4.83V

For reference, here are the same readings with my working RPi using the same PSU and again sitting at the login prompt. Both keyboard and mouse working fine.

Mouse Top Port -> 4.77V
KB Bottom Port -> 4.85V
TP1 TP2 -> 4.86V

KB Top Port -> 4.85V
Mouse Bottom Port -> 4.75V
TP1 TP2 -> 4.86V

Hope that helps.

[mahjongg]

Its unlikely that its just a blown polyfuse! a multimeter has an input of 11.11 mega ohm (yes, its a strange value, but why is another story ), so if even the next to nothing load of only the multimeter can collaps the output voltage then there is more going on than just a "blown fuse", I think the fuse itself is either defective, or simply not soldered to the board right.

The "residual voltage" you see with an USB device connected, is normal, below a certain voltage these devices start to act as "diodes" of a kind.

[Burngate]

My theory would be that the fuse is just not connected - the residual volts you're reading would be coming back from the KB/mouse from the data lines - the only other route for current on the sockets.
If you're into soldering you could try shorting out F1 (or do the same with a croc clip, which would leave no evidence!)
Or go for the full-blown Pi-pass

[rurwin]

Maybe there is a short-circuit inside the USB socket or underneath it. F1 limits the current to 140mA and drops whatever voltage it needs to in order to do that.

... But then I wouldn't expect the voltage to change when you plug something in...

...But something is eating current or, without anything plugged in, you would see the USB voltage as exactly the same as TP1/TP2. I think that proves it is not a problem with the fuse.

It may be worth checking the pins under a magnifier for fine whiskers of solder. If that was the problem you could fix it yourself. Remember to leave it a few hours for the fuse to recover before retesting it.

Otherwise you have a faulty RaspPi and need to send it back and get a replacement.

[drgeoff]

There are only three possibilities:

1. High resistance joint or open circuit between the USB socket end of the polyfuse and the USB socket. This would be confirmed if substantially the same voltage at each end of the polyfuse regardless of what is plugged into the USB socket.
2. Low resistance path from the USB socket end of the polyfuse to ground. With nothing plugged into USB socket, put a current meter across the polyfuse. Just connect the meter for long enough to get an indication of how much if any current is flowing.
3. Duff polyfuse. If no fault found from above two tests, bridge the polyfuse and/or remove polyfuse and test with resistance meter.

[Benny Blanco]

Thanks for the help and suggestions but I put the Pi in the post back to RS this morning. My last test before deciding to pack it up was to bridge the USB VCC pins with a bit of wire and both ports worked fine. Not an ideal solution though so decided to just send it back. Hopefully I won't have to wait too long for a replacement...

[Benny Blanco]

Hopefully I won't have to wait too long for a replacement...

Just a quick update on this for anyone wondering on the replacement timeframe. RS responded to my email saying they hold a small number of Raspberry Pis to cover faulty boards.

Posted back to them on the Monday (1st Class). Fully functional Raspberry Pi arrived on Thursday.

Excellent response from RS.