Rocket IO failure after power cycle.

Do you have a question? Post it now! No Registration Necessary

Translate This Thread From English to

Threaded View

In the system I work on, we use Rocket IO to communication
between the different boards.   I see some
Rocket IO failures after power up, so I wrote some
scripts to do power cycle testing to charaterize the problem.
I see Rocket IO communication failures after power cycle in about
3-5% of time.  On some system the failures rate can be
as high as 40% of time.   With this particular system, the
failures seem to related to how long I power down the system.

When I power down the system for 1, 5 seconds between bootup,
I see very few failures 3-5%.  When I set the power down->up time
to 20, 60 seconds, I see the failure rate go up to 40%
to 50% very consistently.

Once the Rocket IO enters this failurs state, the only
way to recover is to power cycle the system.   Reload the
FPGA bit file from the sysace did not help at all.


When it fails, the serial/parallel loopback on the failure
board works fine, our HW engineer probe the RX of the failed
channel and the signal looks ok (signals are clean and
no issue.)


Anyone see this kind of problem before?   Any pointers on
what direction to look and any idea on how one might
recover from it.



Thanks in advance.

-Tony


Re: Rocket IO failure after power cycle.
Could this be a DCM initialization problem?  You might want to bring
out your various clocks and see how they look in the failure mode.

On 8 Jul 2005 20:43:44 -0700, " snipped-for-privacy@gmail.com"

Quoted text here. Click to load it


Re: Rocket IO failure after power cycle.
Marko napisał(a):
Quoted text here. Click to load it
Marko napisa�(a):

 > Could this be a DCM initialization problem?  You might want to bring
 > out your various clocks and see how they look in the failure mode.
 >
And look at power supply voltage behavior on digital scope.

--
furia

Site Timeline