Should I return this Pi?

Yours looks like a  more severe version of Ewan's problem in this thread -- http://www.element14.com/community/thread/19375?tstart=0 .  His mouse and keyboard worked, but the LAN lights didn't come on despite having eliminated Ethernet leads and remote endpoints from the problem space, so it left LAN9512 problems and magnetics/RJ45 issues as candidate faults.

The USB and LAN operation are of course intimately connected, but since your system does actually boot, it does suggest at least one avenue to explore before RMA time.  Try booting without any USB devices nor Ethernet cable inserted at all.  Then, when you get to the config page, plug in the keyboard only.

If there is still no keyboard response, then I'm out of ideas.

PS. You mentioned your Dell SK-8115 keyboard.  I don't have one of those, but I do have a Dell SK-8135 which has many more bells and whistles (and hence probably uses more current), plugged into a different computer.  It's never been tested with the Pi, so your problem offered an opportunity to test two birds with one stone:

• The Dell SK-8135 keyboard works fine plugged into one of the Pi's two USB sockets, both through boot or when plugged in later.
• It also works fine when plugged into a class-compliant self-powered hub connected to the Pi, as long as it is the only device plugged into the hub.
• It does not work when plugged into the hub alongside another wired keyboard or mouse that previously worked fine in that hub.
• Although the SK-8135 "works fine" alone, that "fine" is in the wierd Pi context, meaning that it loses USB events every now and then.

So, that more or less suggests that your SK-8115 should work if it's the only keyboard plugged directly into your Pi.

This doesn't help a lot, but it's a possible data point.

Morgaine.

Thanks Morgaine, in my rush at typing my original post I forgot to mention I'd tried that scenario too. I also tried with my fingers-crossed :-)

To be doubly sure I've done another run through various combinations of keyboards, power supplies, network and powerup cycles including the one you suggested with no joy. In a final step of desperation I also did a scan of my network just to see if an IP address had been served to the Pi even though no LED activity, no signs of life from the network either.

I think I was hoping for inspiration so I didn't have to RMA it but was probably expecting the worse.

Many thanks for your prompt reply!

PS. I added a note about your Dell keyboard which overlapped in time with your reply.

Hey, thanks for taking the time to experiment, much appreciated.

I only had the keyboard connected, so only one USB device at any point.  You're right though, always good to have other data points.

My current view is I'll borrow (or buy) another very basic USB keyboard and give it another try. If that doesn't work then I'll have tried at least 2 different keyboards that others have verified work with the Pi and a third one of whatever spec I can lay my hands on. Past that I think I'll have to do the RMA option and be patient for a replacement.

Thanks again.

Have you looked at this item at the RasPi Troubleshooting Wiki?  http://elinux.org/R-Pi_Troubleshooting#No_USB_device_works.2C_with_known_good_PS.2C_SD_card.2C_KB

I have read a number of reports of problems with the 25 MHz crystal X1 which is under the LAN9512 chip.  If X1 is bad or not soldered properly, USB and Ethernet both fail.  I suspect that X1 is hand-soldered since many crystals cannot survive infrared reflow.  As such, it's more likely to have a bad solder joint.  I suggest inspecting the connections with a strong magnifier.  If you're handy with a small soldering iron, consider reflowing the connections.

Also check capacitors C41 and C51 next to X1.  X1 could misbehave if C41 or C51 is missing, tombstoned, or shorted out.  "Tombstoning" is when only one end of a small surface mount device get soldered, and surface tension causes it to stand erect like a tombstone.  Wonderfully descriptive term!

When I first got my raspi I tried using usb keyboard & mouse, wired and wireless and promptly gave it up. While waiting for the raspi I had been reading everything everyone from this list had been reporting/discussing. I dug out a powered usb hub and never looked back! Maybe it's just dumb luck but everything has worked great. Can't recall if timing had anything to do with my early success - meaning when I plugged the keyboard in to the hub. . . I'm using a noname wireless keyboard and mouse now and everything just works. Have a missed something or have you tried this approach?? My usb hub is also from who-knows-where. . both made in China

wireless keyboard and mouse;

www.perixx.com (power rating 3V 30mA, model PERIDUO-710)

USB powered hub;

www.cyberpowersystems.com , Cyber Power (USB2.0 4-port Hub with Adapter, input 5V 2.6Amp "trusted to comply with Fcc.....)

Best,

Peg

I suspect that X1 is hand-soldered since many crystals cannot survive infrared reflow.

Virtually everything surface mount can survive reflow, defeats the purpose if it can't. Most crystals could easily survive reflow 25 years ago and I'd be very surprised if that trend has reversed recently.

For example here's a roughly equivalent device to X1 where the datasheet specifically mentions what it can cope with for reflow. http://uk.farnell.com/euroquartz/25-000mhz-mq-30-30-40-12pf/crystal-5x7mm-cer-25-000mhz/dp/1640923  I have a datasheet next to me for a RTC crystal that has the full recommended reflow profile..

Tombstoning normally only happens on the top of the board, components on the bottom tend to be glued down, this is either because the board will be reflowed twice, once for each side, or the bottom side goes over a solder wave.

For a small board like this, I wonder if they're using a masking jig and what was called a 'miniwave' in my day to selectively solder just the through hole components. Something like this if you've not seen one before http://www.youtube.com/watch?v=v0y8IKihjXE  the giveaway that they're likely doing it this way is the unpopulated P3 connector that tends to have some combination of holes soldered, but not always all of them.

The overspill from a miniwave process, given the number of connectors around X1 could be what's giving it that hand soldered look.. It's also possible it could exceed the temperature profile of the crystal if left on too long.

Me again! Sorry I neglected to mention that I have used Raspbian Wheezy and AdaFruit Occidentials on my Raspis.. I have 2 Raspi serial #  FN120524140 & E2412RSV1.0B1.1

FN was purchased from Element 14 and the second one was purchased from RS (I received an email from RS saying it was time!) Both units use an LG phone charger power supply that is rated 5.1V=0.7A..  I guess now you know what I mean by "dumb luck"!

peg

Thanks all the replies.  Having tried yet another keyboard I think I'll do a return on this board. In years gone by I'd have tried a few soldering iron interventions but in this case I think I'll save that more more creative activities than fault-finding on the basic board, I'll try my luck with another and see how I do.

Will post back results of using same keyboards/cards with a replacement Pi.

If it's not too late and your Pi is still plugged in, how hot is the LAN9512 device after at least 15 minutes of the board being powered on?  (Square chip nearest the USB host sockets.)

A finger test is sufficient.  Is it so hot that instinctive sub-second reaction jerks your hand away immediately, or can you hold your finger on its surface for some seconds without pain?

This test only works once at any given time, because the device surface has low thermal inertia and/or high thermal resistivity, so immediately applying a second finger will not find it very hot even if the first one did.

The reason I ask is because I'm examining the possibility that excessively hot LAN9512 chips correlate with USB data loss.  There isn't widespread evidence of that, although my Pi is both hot and lossy.

Morgaine.