Critical issue with non-conforming USB powered hubs.

Any technical questions about the Epiphany chip and Parallella HW Platform.

Moderator: aolofsson

Re: Critical issue with non-conforming USB powered hubs.

Postby mhonman » Fri Dec 27, 2013 12:17 am

Richard, have a careful read over the original post. The problem is that the Parallella can be damaged by use with a hub that violates the USB specification. If the hub follows the USB spec faithfully, there is no problem. The remedial action that Adapteva are taking is thus a bonus for future purchasers - good engineering in action.

The openness about this issue is in striking contrast to the way the PC industry handled the infamous nVidia laptop GPU problem a few years ago...
mhonman
 
Posts: 112
Joined: Thu Apr 25, 2013 2:22 pm

Re: Critical issue with non-conforming USB powered hubs.

Postby vidarh » Tue Jan 07, 2014 8:54 am

richardlich wrote:I don't think, "-Parallella Kickstarter backers will need to use one of the workarounds." is acceptable.

You should send a good product to backers.


If you want to wait for a fix, I'd suggest you contact Adapteva and ask for your place to be switched with someone else - I'd guess that most of us would far prefer to get the board ASAP, given that there are several simple and easy workarounds. I certainly wouldn't want to wait over something as minor as this.
vidarh
 
Posts: 23
Joined: Mon Dec 17, 2012 3:25 am

Re: Critical issue with non-conforming USB powered hubs.

Postby tnt » Wed Feb 05, 2014 9:09 am

Mmm, just seen the proposed fix on the new schematic ... however I'm not exactly sure what's the logic behind it.

Like, overcurrent/thermal will reset the FPGA which will in turn cause the enable to be de-asserted which will shutdown curent flow ? I guess it can work, but it as a pretty serious issue that if you plug something that draws too much current, it will reset your board ... (instead of just shutting down usb power to it)
tnt
 
Posts: 408
Joined: Mon Dec 17, 2012 3:21 am

Re: Critical issue with non-conforming USB powered hubs.

Postby zmc » Thu Jul 03, 2014 10:17 pm

Is this protection fix is already in the boards that just shipped or is it slated for the next release?
zmc
 
Posts: 24
Joined: Thu Jul 03, 2014 10:01 pm

Re: Critical issue with non-conforming USB powered hubs.

Postby ajtravis » Mon Aug 17, 2015 1:38 pm

aolofsson wrote:Hi all,
[...]Potential Immediate Workarounds:

RECOMMENDED:
-Connect a powered USB hub and Parallella to a common power strip to power both devices on/off at the same time.(tested)

OTHER OPTIONS:
-Only use un-powered USB hubs(tested)
-Use a conforming USB powered hub that prevents power flowing back to the Parallella board(untested)
-Modify the USB cable to the powered hub to remove the VBUS line(untested)


Hi, Andreas.

It took me quite a while to figure out why Parallella USB PHY initialisation has been so unreliable on my Parallella boards, yet worked fine on RMA boards that I returned to you (See: https://parallella.org/forums/viewtopic.php?f=50&t=1510#p15439)

I think we can now change this option to 'tested':

-Modify the USB cable to the powered hub to remove the VBUS line(tested)

[...]
Any feedback regarding the severity of this issue would be much appreciated, especially if you have experience with similar issues and workarounds from other projects.

Andreas


Cutting the VBUS (red) wire is not really a solution: it's just a work-around. It may well cause other USB problems and electrical vulnerabilities to the Parallella USB OTG port. In particular, if the Parallella USB OTG port is intended to switch roles by sensing VBUS. In my testing, I've found it necessary to power off both the USB hub and Parallella, otherwise the 'Desktop' kernel complains that it can't re-set the USB port. However, I think it's a BIG step in the right direction that we can now work around the USB problems that I and quite a few others have experienced using Parallella (and RPi) with 'back-feeding' USB hubs. These problems are not confined to 'cheap' USB hubs - I bought an expensive Belkin Hub, which I only recently discovered is on the RPi USB hub 'blacklist'.

According to the USB 2.0 spec.:

Code: Select all
A device (or hub) can only sink (consume) current from its upstream port
A 'self-powered' device is one which does not draw power from the bus


Clearly a non-conforming (i.e. back-feeding) USB hub can source (provide) current to its upstream port, which can cause problems if the upstream port does not prevent it. I'm going to continue testing, and reading about USB OTG :-)

Bye,

Tony.
Minke Informatics Limited, Registered in Scotland - Company No. SC419028
User avatar
ajtravis
 
Posts: 166
Joined: Fri Jul 18, 2014 12:54 pm
Location: Scotland (UK)

Previous

Return to Epiphany and Parallella Q & A

Who is online

Users browsing this forum: No registered users and 20 guests

cron