USB Issue Troubleshooting

Hardware related problems and workarounds

Re: USB Issue Troubleshooting

Postby ArunD » Fri Jun 19, 2015 6:30 pm

If I do not even connect a USB hub, at least the root hub onboard should be recognized. The "lsusb" command should list the Linux root hub as the only USB device present. Following was as seen on one occasion:
root@parallella:~# lsusb
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub


And just after the above successful detection of the onboard USB root hub I power cycled the Parallella. The "lsusb" command now displayed:
root@parallella:~# lsusb
unable to initialize libusb: -99


Doesn't this mean my Parallella board USB has a problem? The onboard root USB hub is apparently not recognized!
ArunD
 
Posts: 34
Joined: Sun May 24, 2015 12:33 pm

Re: USB Issue Troubleshooting

Postby gruzija » Sat Jun 20, 2015 12:11 pm

I have the kickstarter boards and have the exact same problem. A simple USB flash drive is not recognised and have the same error for lsusb. I tried shorting U33 and CR10, but to no avail.

I'm kinda disappointed since its not really usable

Regards:)
gruzija
 
Posts: 2
Joined: Fri Jan 02, 2015 7:55 am

Re: USB Issue Troubleshooting

Postby ajtravis » Sun Jun 21, 2015 12:34 pm

gruzija wrote:I have the kickstarter boards and have the exact same problem. A simple USB flash drive is not recognised and have the same error for lsusb. I tried shorting U33 and CR10, but to no avail.

I'm kinda disappointed since its not really usable

Regards:)


Hi, gruzija.

This is a depressingly familiar experience for people who bought the 'desktop' version of the Parallella only to find the USB so unreliable that it's practically impossible to use as a 'desktop' with a keyboard and mouse or USB storage attached. I've tried to be positive about this in my responses to messages on the Parallella forum and I've acknowledged the efforts that Adapteva have made to fix the problem. In view of the number of people reporting USB problems on the Parallella forum I'm very surprised how few people have mentioned the USB problems in Adapteva's recent survey https://www.parallella.org/2015/06/12/responses-from-parallella-user-survey/.

I realise that it might not seem cost-effective for Adapteva to invest yet more time getting to the bottom of the USB PHY initialisation problem but if Adapteva want to know why people are not using Parallella instead of R-Pi or other ARM SBC's I think that USB problems forcing a requirement for a serial TTY or SSH to communicate with the board explains a lot. Dropping the price on Amazon looks like a desperate attempt to move stock that is not selling to me. I think it might have been better to resolve the USB problem instead. Parallella is head and shoulders above the competition except for this one critical USB problem that makes it look unattractive as a 'desktop' ARM SBC like the R-Pi and similar systems popular at the moment.

Telling users, and potential users, of Parallella that USB problems don't exist or are caused by the users is not an answer. We, as a community, need to help Adapteva get to the bottom of the problem. Complaining that what you bought is not useable is not an answer either: This was a Kickstarter project brought to us at cost by Adapteva and I, for one, think it is a fantastic product let down by one or two technical problems. BUT, to get into the mainstream market for R-Pi and compatibles and democratise education about parallel computing Parallella MUST work as a 'desktop' PC requiring ONLY a USB keyboard/mouse and HDMI TV/monitor to work. My attempts to use Parallella to teach bioinformatics have crashed and burned because of the USB issue.

Bye,

Tony.
Minke Informatics Limited, Registered in Scotland - Company No. SC419028
User avatar
ajtravis
 
Posts: 162
Joined: Fri Jul 18, 2014 12:54 pm
Location: Scotland (UK)

Re: USB Issue Troubleshooting

Postby ArunD » Sun Jun 21, 2015 4:59 pm

The perspective offered by Tony is correct. I like the Parallella product idea & architecture, and makes sense help to Adaptiva resolve this in ways practical/possible.

In my case the problem is seen without any external cable and/or devices connected. I have not noted any consistency in the rate or sequence of the occurrence of the problem. With this much I have no more ideas to gather more information. So posted this problem here.

One choice, if possible, is to send the board to Adapteva directly for the problem evaluation. The other is I send it back to Amazon, as I have had it with me for less than a month... and let the return process send it to Adapeva - and hoping a replacement will not have this problem. What can I do now to help resolve this for myself as well as for Adaptive, what is the more effective choice?

As an individual buyer, I have put my own money towards this purchase, and as a work around, getting the embedded version is a out of my budget at this time.
ArunD
 
Posts: 34
Joined: Sun May 24, 2015 12:33 pm

Re: USB Issue Troubleshooting

Postby xilman » Mon Jun 22, 2015 10:49 am

ajtravis wrote:This is a depressingly familiar experience for people who bought the 'desktop' version of the Parallella only to find the USB so unreliable that it's practically impossible to use as a 'desktop' with a keyboard and mouse or USB storage attached. I've tried to be positive about this in my responses to messages on the Parallella forum and I've acknowledged the efforts that Adapteva have made to fix the problem. In view of the number of people reporting USB problems on the Parallella forum I'm very surprised how few people have mentioned the USB problems in Adapteva's recent survey https://www.parallella.org/2015/06/12/responses-from-parallella-user-survey/.
I'm one of the few who mentioned the issue in the survey.

I've given up trying to use my systems as anything other than headless boxes.
xilman
 
Posts: 80
Joined: Sat May 10, 2014 8:10 pm
Location: UK

Re: USB Issue Troubleshooting

Postby ajtravis » Mon Jun 22, 2015 12:48 pm

xilman wrote:[...]
I'm one of the few who mentioned the issue in the survey.

I've given up trying to use my systems as anything other than headless boxes.


Hi, xilman.

I'm quite surprised that only 300 people (including you and me) responded to the survey when Adapteva have announced that 10,000 Parallella boards shipped by Sep 2014 https://www.parallella.org/2014/09/19/parallella-the-crowd-funded-open-source-research-platform/. More boards have, presumably, been shipped since then. If the majority of people bought, say, four boards each this would still amount to about 2,500 Parallella users by Sep 2014. I would be interested to know what the TRUE extent of the USB problem is and I urge Adapteva to RMA any faulty boards quickly to resolve any USB problems on a 'minority' of faulty boards, which is what Andreas has said on this forum. Personally, I believe it is a case of marginal design affecting most Parallella boards.

I've said this before and Adapteva did RMA my boards, but the SKU SKU101020 replacements they sent me have exactly the same USB initialisation problems as the boards that I returned - The boards do sometimes work, from a cold-start, but most of the time they don't. One SKU P1601-DK03 that Adapteva sent me initialises USB more reliably, but even that one is very particular about what is connected to the USB and doesn't initialise it if NOTHING is connected!

AFAIK, a USB OTG interface should NOT require anything to be connected at power on. This is the common interface used on millions of mobile phones and it would be ridiculous to ask phone users to make sure they have a powered USB hub connected before switching on their phone. I think it's equally ridiculous to expect users of 'desktop' Parallella boards to have to experiment with different combinations of USB hubs and devices just to get the USB PHY to initialise.

BUT, as I said in another message on the forum, I think we as a community of Parallella users need to help Adapteva fix the problem and promote the board as a viable platform for teaching parallel computing or it will disappear without trace.

I was involved in the development of a 'massively' parallel SIMD image processing system called CLIP4R with a 9,216 processor array designed by Michael Duff and Terry Fountain at University College London (UK) over thirty years ago http://www.computer.org/web/csdl/index/-/csdl/proceedings/afips/1978/5086/00/50861055.pdf. Few people will have heard of it now because it was very difficult to program and was built using custom VLSI processor chips. However, the processor array was controlled by FPGAs in a similar way to Parallella and the host computer was a VAX750 mini computer - Not exactly the low-power consumption of Parallella, but employing quite similar ideas. My point is not that either CLIP4R or Parallella are bad designs, but simply that Parallella may share the same fate as CLIP4R unless it enters the main stream of low-end commodity computing like the R-Pi has: THIS is what 'disruptive' technology is all about - Doing something that used to be expensive and complicated in a MUCH cheaper and simpler way. THAT, I understood, was the primary objective of the Parallella Kickstarter campaign.

What's this got to do with the anecdotal evidence that the Parallella USB interface is unreliable?

Well, I had to bang some of the CLIP4R processor cards in the VME rack with the blunt end of a BIG screwdriver to get them to work (on the expert advice of the electronics design engineer I was working with). Sometimes, it feels horribly familiar trying to get the Parallella USB interface to work, but I've not (yet) tried banging them with a BIG screwdriver handle!

Bye,

Tony.
Minke Informatics Limited, Registered in Scotland - Company No. SC419028
User avatar
ajtravis
 
Posts: 162
Joined: Fri Jul 18, 2014 12:54 pm
Location: Scotland (UK)

Re: USB Issue Troubleshooting

Postby ajtravis » Mon Jun 22, 2015 12:59 pm

ArunD wrote:[...]
What can I do now to help resolve this for myself as well as for Adaptive, what is the more effective choice? [/u][/b]

As an individual buyer, I have put my own money towards this purchase, and as a work around, getting the embedded version is a out of my budget at this time.


Hi, ArunD.

You can do what Adreas asked and report your USB problems in a standard way, so Adapteva can get a more accurate picture of the USB problem: https://parallella.org/forums/viewtopic.php?f=50&t=1650&start=60#p10185.

If anyone else also has USB problems, but has not yet reported them to Adapteva, please post a message here using the format Andreas suggested.

Thanks,

Tony.
Minke Informatics Limited, Registered in Scotland - Company No. SC419028
User avatar
ajtravis
 
Posts: 162
Joined: Fri Jul 18, 2014 12:54 pm
Location: Scotland (UK)

Re: USB Issue Troubleshooting

Postby tnt » Mon Jun 22, 2015 1:24 pm

Has anyone tried different power supplies with the "faulty" boards ?

I mean one possible cause would be that the proper initialization is dependent on the rise time / curve of the power being applied. This may explain why some boards returned as "faulty" turned out to work great when adapteva tested them themselves.

I checked the schematic and data sheet of the chips involved and I can't find anything wrong with it. There is even a proper voltage monitor that will ensure a good reset pulse and it should get released at the same time as the PS reset from the zynq. To those that have the issue, have you ever looked at the reset pulse going to the PHY ?
tnt
 
Posts: 408
Joined: Mon Dec 17, 2012 3:21 am

Re: USB Issue Troubleshooting

Postby tnt » Mon Jun 22, 2015 1:56 pm

nevermind, reading the history I see that the problem is somewhat identified.
tnt
 
Posts: 408
Joined: Mon Dec 17, 2012 3:21 am

Re: USB Issue Troubleshooting

Postby ajtravis » Mon Jun 22, 2015 2:41 pm

tnt wrote:Has anyone tried different power supplies with the "faulty" boards ?

I mean one possible cause would be that the proper initialization is dependent on the rise time / curve of the power being applied. This may explain why some boards returned as "faulty" turned out to work great when adapteva tested them themselves.


Hi, tnt.

Well, I originally used the Parallella-compatible PSU's supplied by Adapteva then, at Fred's request, I also tested the RMA replacements with a new PSU supplied by Adapteva - Made no difference to USB initialisation...

I checked the schematic and data sheet of the chips involved and I can't find anything wrong with it. There is even a proper voltage monitor that will ensure a good reset pulse and it should get released at the same time as the PS reset from the zynq. To those that have the issue, have you ever looked at the reset pulse going to the PHY ?


If you've not seen it already, take a look at this message and the rest of the thread:

https://parallella.org/forums/viewtopic.php?f=50&t=1510&hilit=usb+reset&start=70#p9341

Bye,

Tony.
Minke Informatics Limited, Registered in Scotland - Company No. SC419028
User avatar
ajtravis
 
Posts: 162
Joined: Fri Jul 18, 2014 12:54 pm
Location: Scotland (UK)

PreviousNext

Return to Troubleshooting

Who is online

Users browsing this forum: No registered users and 4 guests

cron