[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference netcad::hub_mgnt

Title:DEChub/HUBwatch/PROBEwatch CONFERENCE
Notice:Firmware -2, Doc -3, Power -4, HW kits -5, firm load -6&7
Moderator:NETCAD::COLELLADT
Created:Wed Nov 13 1991
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:4455
Total number of notes:16761

2812.0. "FDDI 'sensitive' to power glitches after upgrade to 4.0.2 " by DECPRG::LUFT () Fri Sep 29 1995 12:24

Hello,

The customer have the following problem : when he switch on/off the monitor on
one of the connected DEC2100, or other inductive device on the same power
distribution ( i.e blower in cabinet ), the FDDI ring is broken.

The green LED on the PMD goes to amber and the ring is unavailable for about
30 seconds. After that period everything is returned into normal, the green LED
goes on. But all the network connections are broken.

DEChub 900 in following configuration :

DEChub 900 Multiswitch HW=F, RO=V1.1.6, Firmware 4.0.2

slot 1: DS900TM, SW V1.3, ROM:V5.1-0, LAT:V5.1
slot 5: 900TX, desbf100.bin SW:1.1.0, HW:V1/2, RO:V0.4.
slot 6: 900MX, def6x301.bin, SW:V3.0.1, HW:V2, RO:V1.2
        2 PMD's DEFXU-AA Rev A01, 2 used to connect AXP2100 systems over twisted
        pair FDDI,

slot 8: same as slot 5

two 140W PS, one 160W.

The error log on HUB shows following :

   [5] Dump error log

   entry        = 4
   time stamp   = 0 34600
   reset count  = 65
   trap @ 885 in console_hccp.c


   entry        = 3
   time stamp   = 0 5043000
   reset count  = 64
   trap @ 885 in console_hccp.c


   entry        = 2
   time stamp   = 0 0
   reset count  = 11
   catch VO=07C SR=2714 PC605db4
 
 

   entry        = 1
   time stamp   = 0 0
   reset count  = 10
   SW V3.0.0 -> V4.0.2 ; config retained

   **********************************************************
 
   [3] show current settings :

   DEChub 900 Multiswith
   =============================================
   Hub900 Multiswitch HW=F,RO=V1.1.6,SW=V4.0.2
   sysuptime                    :3 days, 11:13:39 70 resets
   SNMP Read/Write Community    :public
   SNMP Trap address            :not configured
   Out-of-band management       :disabled

   interface    ip address      subnet mask     def. gateway    other info
   OBM port                                                     speed 9600
   hub slot 8   200.200.1.15    255.255.255.0                   active

 
   The customer claims, that the problem become after the upgrade to 4.0.2
   and he have the problem on two completely different sites.

   Bedrich.

T.RTitleUserPersonal
Name
DateLines
2812.1NETCAD::DOODYMichael DoodyFri Sep 29 1995 18:286
    Using Hubwatch, double-click on the PMD that had the problem and
    see what the error counts say. Sounds like you got an LEM error maybe?
    
    Are your FDDI UTP cables run alongside your power cables?
    
    md
2812.2Maybe EMI?NETCAD::THAYERMon Oct 02 1995 14:3021
I would question if the problem were coming through the power lines only.
The other end of the cable in the network is plugged into the system which 
could be more susceptible to the power glitch than the HUB itself. 

We saw this 
kind of situation in testing DEFHU in stand alone setup, when appication of ESD 
pulse to the module (DEFHU) could disrupt the FDDI ring in DEF6X, which was 
part of the ring. Another guess is that transient could couple into the cables 
and cause reset if the cables were positioned to pick up a lot this way by 
accident.

We test power supply design on Power Line Transients and Electrical Fast 
Transients, and this behavior of the equipment would not allow to pass the 
test. Is there enough info on the setup, such as revision levels, etc.?
Is there any way we can simulate the problem here?

Regards,

Boris Shusterman

(entered by John Thayer)
2812.3utp-fddi problems on Decconcentrator 900MXHLFS00::BEEK_MMon Sep 02 1996 20:0314
    Hello,
    
    I seem to have the same problems with a Decconcentrator 900MX .
    Almost every time the lights in the computer room are turned on/out
    an Alpha1000 connected with utp-fddi (defpa) crashes.
    Symptoms: The green led turns to amber for about 20-30 seconds,
              turns out for a few seconds and then turns steady green
              again. This takes to long for the Alpha1000 and the node
              decides to leave the cluster.
    
    I replaced the defpa-utp module with a defpa-fiber version and i
    installed a fiber pmod in the concentrator.
    Now the problems are gone.
    But this is a temporarily solution. Does anybody know a better one ?
2812.4Signal Integrity between systemsRWCVAX::COULSONRoger Coulson - RSE DTN 223-6158Tue Sep 03 1996 12:4022
    RE: .0 & .3
    
    Some basic signal integrity issues sound like the problem here. 
    Network products are not my work but they are products that I use in 
    my day to day work.  
    
    A fiber connection isolates the two ends of the FDDI connection so 
    any electrical transients do not effect the link.  This is ideal where
    the two ends are on different power sources.  With a UTP connection the 
    same considerations need to be made with regard to both ends of the 
    connection as would be done in a Thinwire Ethernet connection.  The 
    common ground or earth bus needs to be in place.  The UTP connection is 
    much cheaper that the fiber connection but there is a price to pay and 
    it sounds like you are both experiencing it.
    
    I suspect that the two ends of your FDDI UTP connection are not connected 
    to the same immediate power and ground bus.  I would recommend that you
    follow the signal integrity guidelines for a Thinwire Ethernet and the 
    problem should go away.
    
    	Roger
    
2812.5Is all your FDDI utp cable Category 5?NETCAD::BATTERSBYTue Sep 03 1996 17:515
    Also make sure that all your UTP FDDI cable is CAT 5 cable.
    CAT 5 UTP cable has additional shielding that non-CAT 5 cable
    (wired & pinned out the same), may not have I believe.
    
    Bob
2812.6still utp-fddi problemsHLFS00::BEEK_MMon Sep 09 1996 15:2911
    Thank you for your replies,
    
    I checked grounding and cables, all seems to be o.k.
    We are now checking the environment. We saw that datacabling and power
    cables were together in one big bundle of cables.
    
    The customer is going to seperate data from power cables.
    When he is finished doing that, i will go back onsite.
    
    Mario.