|
Matt,
The bent backplane pin is usually associated with the connector on
the module. If you look at the module that came out of the slot with
the bent pin, you will see the hole for the pin to go in is collapsed.
Once the pin gets bent, it jabs the module, and instead of going into
the open hole, it jabs the sidewall of the hole, crushing and collapsing
it.
Then, the module becomes the cause of the problem. If you move that
module into a different slot, the backplane pin wont be able to
slide into the collapsed module hole and then that backplane
pin gets bent. It is a viscious cycle now.
I do not know if the issue has been addressed since I last knew
of the problem.
Dennis
|
|
Hi,
Below you will find some hints/tips I recieved from Tom Flemming
( AYR Plant), Some info will be known by now, but for some less experience
engineers it can be a advantage to have these tips.
Cor
Hello to my MCS TurboLaser colleagues...
THIS IS PROBABLY THE MOST IMPORTANT MAIL THAT I'VE SENT YOU TO DATE.
***** Please take a few minutes to scan through it. *****
(Don't hestitate to comment back!!)
CONTENTS
========
o MOST COMMON ISSUES, CORRECTIVE ACTION, & SUGGESTIONS
o Your FEEDBACK & our RESPONSES
MOST COMMON ISSUES:
-------------------
1) TLSB >> PINS BENT :
Pins on the TLSB Backplane can be bent if either the module
connector/socket is damaged or if the module is not carefully
installed either in MFG or in the field. We suspect that our Module
Test "QV" equipment has been a contributing factor here in the
early period causing module connectors to get damaged. Although
machines are extensively tested in MFG, some bent pin connections
would seem to get through. You will understand it is not possible
to visually inspect for bent pin after the module is in position.
ACTIONS:
(i) QV stations to be upgraded (will include stronger guides
more suitable for continous MFG use etc.)
(ii) Ongoing visual inspection of modules and backplanes in the
production areas
(iii) Mail 26th May "Installing/Replacing TurboLaser Modules"
(COmm. No. 4) advised you on the subject.
(iv) We have relayed our concern to Engineering.
2) DOORS NOT CLOSING (UPGRADES):
Doors may not close because they come in contact with ..
a) The TLSB module levers (2 cases reported this week)
b) Double bar doors on Lasers shipped approx Feb-Mar 1994
c) TZ78-VA if installed in BA655 in the FRONT of the cab.
ACTIONS for a,b,c.
a) Engineering plan to change the TLSB mounting brackets at
front and rear to give an extra 4mm but this may not happen
till nearer Q2. (Looks like you will have to do some local
adjustments to the door for current problems! )
b) Remove 1 bar or get new door as per my mail 21-Jun (Comm no.10)
c) Move BA655 to the rear as MFG do today for such orders.
(see KENT::TURBOLASER notes file #218)
3) CONSOLE CD / LFU distribution / System firmware
My mail of 10-Jul "AlphaServer 8200/8400 Software" covers the
supply of the Console CD ... in the operating system "H-kit"
(when ordered, or supplied by MFG) and the CD is not supplied
with the system as in Laser.
Machines from MFG have the latest LFU revisions available at time of
shipment. The CD utilities may not be just as current and so I am told
you should be careful in its use.
4) KZPSA-BB : (REF notes file 270)
Installing KZPSA's ...when you do "show config" some are not seen.
At least 2 cases have been reported in Europe and we have also
seen it in Ayr. The most common solution I'm aware of now is to move
them to different slots in the PCI. Engineering are investigating
the issue.
5) 8200 POWER CABLE: (not being ordered)
This item has caused a lot of "hassle" on many orders. If you know
of an 8200 being ordered please highlight to the Acct team/Admin
the need for a cable to be shipped WITH the machine and NOT
"partialled" from the system. In MFG we ship supported configurations
of what is on the order. It is not our role to RE-tech_edit the
orders. (However.. see next issue).
6) ORDERING ISSUES:
Most configuration issues have been caused, it seems, by the people
responsible for the order, NOT USING THE SOC (Sales and Options
Catalogue)....missing floppy drives, power & other cables.
We will endeaver to collect the most common Order/Tech Edit issues
or pitfalls that we see or that YOU make us aware of, and work
work with Marketing/Sales/Admin to get these HIGHLIGHTED.... if not
already being done by the TL Program Mgt. (I am aware of a draft copy
of a planned Q1 Marketing communication due 1st August, which covers
many of the issues)
SUGGESTIONS:
------------
Most of you are probably aware of the following ( or the suggestion
may be normal practice or what we would like to see!)
- Try to get a local MCS Eng or the Acct. Resp. to check the
order with Order Admin (exactly what is being delivered & when)
and the site (eg especially for Upgrades.. Cab doors, BA655-AA etc)
preferably 2 to 3 weeks before we are scheduled to ship!
- ONLY TL trained Engineers (ie you) or Eng's that you have given
hands-on training to, should install machines, do upgrades or
change modules. I have had a report of a Customer changing modules
.. and, guess what?.. bins were bent. MCS have suitably rapped that
customer's knuckles!.
- AlphaServer 8200/8400 Service Manual "EK-T8030-SV.A01" has been
recommended to me as a very useful document for any TL Eng.
- Use the KENT::TURBOLASER notes file
TURBOLASER PROBLEM FEEDBACK & ACTION
=====================================
I am taking the liberty of sharing all the problems that you have
reported to me so far with the full TL_MCS.dis dist. and giving a
response to each issue raised for ALL to see. The more recent
problems are not fully investigated yet.
Country/ MCS/
Customer/Date
Dec No / Ref. PROBLEM COMMENT / ACTION
-------------- ----------------------------- --------------------------------
US Seed / Upgrade TLSB box, packaging Final packaging not released or
Gary Kushner poor. (consolidation box) required for seed units.
11-May New packaging parts now being
planned by MFG.
Originally Upgrades were not in
the plan for Q4 shipment.
Belgium Seed TLSB backplane pin bent MFG taking action.
SBA 19992 & module damage. MCS (Europe) informed 24-MAY
L.R.Pirlet Also seen/experienced ENG (=Design Engineering) advised
18-May on JGO Repair unit. of issue (See No 1 above)
RZ28-VA with FIS OSF replaced. Drive DOA. RZ28 problems being
Drive not seen on SCSI bus. addressed by the Ayr COE
(The COE deals with vendors)
UNIVERSITY MS7CC-DA FCC strip damaged Straightened on site.
MAINZ,Germany Reviewing handling practices
950098192 8400 in MFG. Info passed to ENG
(700302176) to see what can be done.
MFG visit
G.Leroy Memory Diagnostic not sizing Details being sought from MCS
30-May correctly on test being run. (can anybody
remember what test involved??)
Stand-offs on Memory module Issue passed to Memory Design
preventing install of last Group. Install instructions
2 simms. are in the manual.
Local fix = remove stand-offs
2 Licenses missing Hardware shipped ahead of S/w.
as per customer request. Copies
faxed to site to solve issue.
Valbonne 8400 Needs LFU distribution or Mfg checking status on this.
951652459 pointer to it. We don't have an official source
P.Alliez for this. (Does anybody?)
31-May
PCI PIU availability Not available for this seed unit.
Swiss PTT PCI PIU cannot be seen after Box tolerance changes, and
8400 x 2 power-on/off, both systems. assembly changes worked by ENG.
950225161/2 Reseated the regulator card. Fixed now.
M.Schweizer
1-Jun SCSI terminator on iosp2 It seems care is needed to
not properly plugged in, plug in some of these cables.
2 pins bent. No other occurance reported but
Straightened and OK. MFG will watch out for it.
Upper screw of Y cable Not seen before. MFG will watch
connector on iosp0 was not for it.
fully home because screw bent
CD's for firmware updates not Common issue NO 3 above.
present. Part of "H-kit" & not ordered.
Some module handles are rusty Design issue. Now being phased
over to stainless steel.
ATLAS, Bremen Installed TZ87-VA in front. Advised to order a BA350 pedestal
Germany, 8400 & could not close door and this was done. Customer did
950298065 not want 'rear' install solution.
H.Klimmeck (This is design issue/ #2 above).
2-June
10 days later had functional Rebuilt OSF disk and did some
problems... CPU changed but checks/changes. System ran OK.
was not the problem. It's not clear what exactly
corrected the problem.
Parts for ethernet connection Thinwire need DETTR-AB & BN26L-07
not ordered with system. Thickwire.... DETTR-BB & BN26L-07
KFTIA-AA has 2 twisted-pair, .. MFG intend to list the most
customer needs thinwire or common "ordering" pitfalls and
thickwire ethernet. advise SALES/TechEdit etc.
UK, University Addon TLZ07 installed in Design issue. However a new
of Nottingham front but SCSI cable loops cable routing is being introduced
8200 950284746 in front of tape drive. in MFG to alleviate the issue.
MFG visit (DK)
6-June Addon RZ28 x 4 installed MCS recommendation to add
into available slots at rear. another shelf but customer
room for 3 only. concerned at extra cost.
(These RZ's were from a 3rd 4th drive can be installed but
party & not installed. in MFG I understand it is difficult to
replace.
1 obscured by ribbon cable. This cable will be recessed in
later version of box.
System would not "ping" over "Auto-sensing" setting checked.
to internet address System needed to be power cycled
to carry out the command ... due
to cabling being added after
system had been booted.
Germany, #1 PCIA ...contact problems. Same problem as at PTT above.
Uni. Mainz, Needed reseating. The design has been improved.
2nd 8400
14-June #2 IOP Hose connector damaged We are not clear which connector
950098192 on outer lower right edge was involved.
(700302175) Bent (straightened?) the outer MFG will check ongoing that
A.Khan metal, then HOSE OK came up. connector are OK.
#3 DEFPA, new type of connector. This type of information is in
Most customers don't have the SOC (Sales & Options Catalog).
the right cable to connect We intend to highlight the most
to FDDI ring. Recommend that common ordering issues to our
SALES should be informed. colleagues in Sales & Marketing.
#4 IOP module springs are brown This is a copper coating. Now
(look rusty) stainless steel will be used.
(ENG change)
#5 Why are many diagnostics Preliminary release of console.
missing? How do I test a Future releases will correct
twisted pair ethernet in problem.
external loopback,
run mp-tests, or a system
exer with cpu,memory &
devices?
#6 FIS OVMS OK but customers The full kit will not be provided
want to upgrade existing ver. on FIS. Operating S/w added to
Can we not supply the kits on packages for Week50/51/52
the FIS disk too? Are the shipments. (See my 10-Jul mail)
'mandatory patches' included? P1 missing kits from Galway if
still required (earlier ships).
FIS includes latest patches
available at production time.
#7 OpenVMS=IO AUTOCONFIGURE Could be caused by KZPSA looking
takes minutes on each boot? for non existent drives, adding
..might think system hung! a max of 80 secs to boot cycle.
Minimum boot is fast. Known problem to ENG.
#8 FIS S/W did not configure FIS Enginering have found that
the cluster noe properly. a "qualifier" was missing in
(see detail in Azfar's report)OpenVMS V6.2, which FIS normally
uses. FIS fixed 10-Jul.
(OVMS will be fixed next release)
#9 While editing on the console, I just talked to Azfar now and
we had a few error exits from he will try to get hard
the editor. (Azfar will evidence next time he's on site.
follow up on this himself.) HAS anybody else seen this??
AEROSPACIALE, MS7CC-UC kit module handle A few early UPGRADE shipments
Upgrade, and covers missing were missing these parts.
France, MCS in Europe were notified
950056094 (21-June) to place a priority
J-E. Thomas replacement P1 order with Ayr
19-June for the parts required.
With TLSB card cage installed Some Feb/Mar 1994 LASER ships
the cab. front door can't be had 2 bars. MCS in Europe were
closed. TLSB contact with notified about the potential
one of reinforcing bars. problem 21-June. Solution either
\/ (see No 2b above) remove excess bar or P1 a new
/\ (common issue ) door or swap with another Laser.
BOUYGUES TLSB box pins bent Known issue in MFG.
TELECOM in 4 slots MCS (Europe) were advised 24-May
France to check for this potential
Upgrade problem.
950050343 (See most common issue No 1)
(700294087)
26-June
MFG visit PCI would not go fully home A BA655-AB is required (it allows
by J.V. in slot. BA655-AA at back the PCI to fit in correctly).
prevents it. "-AB" P1 ordered from Ayr.
This restriction is not Product Mgt (US) has been
documented for Sales/field informed. Issue being worked.
True, this restriction was not
in the SOC. We will check that
it is added. A WW marketing/sales
flash due 1-Aug (yes not till
1-Aug!) highlights this item.
KZPSA to PCI connection very "Show Config" problem known
loose. "show config" indicated to ENG. Being worked by PCI
only 2 of 4 after install. design ENG. (see No 4 above)
KZPSA min rev is D01 (FW A07) MFG missed out on the special
1 of 4 had earlier FW (A06). upgrade of this single module.
All boxes were marked D01 Sorry!. Nevertheless A06 FW rev
works fine in TL. For other
system types (Mikasa/Demi/Sable)
A06 caused problems and "Hold".
TELEDANMARK Empty carton 99-09150-03 The QZ kit was not available
Denmark (Part of QZ-00RAB-GZ kit) to MFG for the earlier shipments.
950039399 with no CD. The kit was added to Ayr BOMs
8400 22-May for Galway S/W to supply.
26-June The DECevent CDs were added to
F.Thomsen the Galway BOM 1st June. This
kit was shipped from Galway
between the 2 dates (ie no CD's).
Field should place a P1 "short
ship" (order type 99Q) with ESC
Galway for missing CD's.
No console CD with system. See my mail 10-July and common
issue No 3 above!
BANCO DE Power regulator (30-41523-10) JD requested to send module
COMERCIO... transistor exploded! through normal Logistic process
Spain, 8200 Regulator replaced. Runs OK. to JGO (with error detail) marked
950318235 for attn of Component Engineer.
28/29-June If repaired in JGO, JGO will
J.Delgado feed back failure info to the PFT.
(PFT=Product Focus Team)
PCI PIU not factory installed The PCI PIU was "partialled".
No written info about mech. Manuals not available from ENG.
installation procedure. We are trying to establish the EK
availabilty schedule for options.
CPU module installed in 6th MFG go by XCON, it said 6th.
slot, moved to 4th Manuals were written early on.
(manuals say 4th) Issue put to ENG about manual.
(It should work in any slot)
PCI PIU power. Lack of info Manuals not available from ENG.
about it cabling etc. MFG gave specific response to
JD as to how to do it.
Could not test the CPU module Later version of console
will include this.
KFTIA test... says no disks Same as above. (maybe August)
found.
Customer did not like no door This is the current design.
access to rear disks Fixed disks are only put at back
...frequent access not required!
One rear disk is blocked by Design issue. During Q1 new
big connector to TLSB B/plane metalwork will be issued.
DEPT. of No tail lift on truck Ayr will follow up through
AGRICULTURE Distribution channels why
Northern this happened.
Ireland
950274685 Power connection from PCI (red and black twisted pair)
(700303266) to floppy not connected at New metalwork on module to
8200 PCI. Cable connection at PCI be used. This will fix the
Mfg visit (TF) not firm. Suspect that the 2 problem.
25/26 June solder joints on PCI were
cracked after multiple use.
Order partialled with 4 other Common problem. Field & Customer
"BUN no's". Requires multiple agrees "partials" due to material
deliveries & visits to site availability issues in MFG/Corp
particularly for end of year/qtr
revenue & customer needs.
Minor dent in PCI guide rail MFG will watch for these items.
(but was fully operational). "1 off" items.
3 marks (exposed metal) half
inch long each at base bottom
edge of rear cover.
PCS-PFUNDNER Floopy not ordered but Ordering issue been seen before!
(MEINDL) essential for this RAID Action: Highlight this common
Austria 8200 config. missed item to Sales etc.
950004744 Problem overcome locally
G.Spittersberg/ together with a fast ship
H.Putzenlechner from MFG.
4-Jul
KZPSC hangs on a warm start No solution yet. Copied to ENG.
under OSF.
AEROSPATIALE Missing 3 CPU's. These was agreed "Partial".
France, 8400 Shipped 11 days after system.
950310941
(700308770) Wrong cables supplied Wrong substitution made in MFG.
P.Alliez (Required 3 x BN31L1E ) (correct cable shipped later.)
7-Jul
Require LFU pointer Does any one have an official
pointer? (not known in MFG yet.)
CAP PROGRAMATOR Customer manuals & inst Oops! MFG will investigate...
STOC., Sweden, manuals missing., only a
950213577 8200 service manual for VAX3100.
(700294700)
G.Enge CK-KFTIA-AA Y-cable BN21W-0B Not investigated yet.
7-Jul not delivered. (was it the BN that was missing?)
TRADEKA-GROUP OY 2 Pins on backplane bent Most common issue (NO 1 above)
Finland, 8200, behind 2nd blank module (It's good to have the exact
950326086 from the right. detail like this in case
(BUN 700307378) Pins straightened a trend develops which we
J. Nyholm & system runs. can highlight)
6-Jul
TELECOM SA Problem experienced with Request for more detail from SA
South Africa KFTIA module on 1 system.
950301396
L.Barnes/ KZPSA's had to be reseated REF: common issue No 4 above
B.Fendick after devices disappeared.
7-Jul
2 x 8400 E2054-AA (ITIOP) AY51510887 Have requested LB/BF to return
did not see internal devices module through normal channels
connected to port 0. for attn of JGO TL Component Eng.
DEXCEL BN20P-4E missing MFG error. Now changed the
South Africa (was ordered) process to add it in "Stage III"
950301659
S.Fraser Thumbscrews on Y-cable to Seen before in MFG. MFG will
(MFG visit) ITIOP badly chewed. Looks bad. investigate what action to take.
8-Jul
Devices could be seen on the MFG will investigate if test
bus but under VMS but they process can catch this.
could not be initialised.
One of the Y-cables had a
bent pin (top pin on
connector from ISP0.
(config. in detail report)
CENTRE_FILE (Laser) Front door won't close. Known issue to ENG.
UK Upgrade TLSB module levers in contact Ref: No 2a above
950260821 with horizontal rib at the
D. Bazley bottom of the plastic log
10-Jul panel. Local solution .. file (Final F/back report to
a notch in door & plastic follow)
logo holder for each lever!.
SEED UNIT Memory Motherboard (E2036) Known issue on approx first 6
Japan Module handle & covers Upgrade units shipped worldwide
Upgrade missing. MCS (Europe) were notified to
K.Yamaguchi place (via Admin) a P1 order
12-Jul with Ayr for the missing bits.
(Ref: my 21-Jun mail)
CPU board failure (1 of 2) KY will send more details
intermittent TLSB error. for MFG to follow up.
(E2056-DA AY51506005, L05)
Door will not close on due Same issue as CENTERFILE.
to TLSB handle ENG solution gives 4mm more
(Door 70-31071-01 A02) for clearance. (Ref: No 2 above)
Distribution:
Not printed
|