Strange Legend failure mode? 2

[altaphon]

We're installing a Merlin Legend system (R7 V12, 008GS/LSID, 408 GS/LS/MLX, 007MLM, 012) and it was going well when the GS/LSID card stopped working... no action on trunks 801 through 808, though the four lines on the 408 card seemed to work OK. Tried moving the 408 card into the slot where the 008 GS/LSID card was, and now the MLX phones don't come up at all. We can get into the processor with WinSPM just fine, everything looks normal there. Voltages on the power supply test ports look OK. Even with just one MLX plugged in, no action on the extensions. No smoke seems to have been released... what happened?

 

[merlinman]

Sounds to me like slot 1 is a bad slot, based on how I think I understand your trouble shooting steps so far.

I would suggest a close visual inspection of the back plane, even to the point of taking the carrier apart and looking at it the back of the Printed Circuit Board that makes it up.

Or, grab another carrier and see if you still have the failure.

OR, (ASSUMING YOU ARE STUCK ON A SATURDAY NIGHT) and considering you only have 4 modules, put SOMETHING in slot 1 and busy it out and then put every thing in starting with slot 2.

 

[dagwoodsystems]

How exactly did you test the "no action" condition on the trunks? Are these ground-start lines? Did you do a board renum after relocating the card? Were the station ports working on the 408 board before you moved it to a different slot?

I would hesitate to say that "everything looks normal" in the processor without looking at the error log in detail. I'm quite positive that you will find something in there that's of interest. Is the red status LED on the processor lit up?

Help me...help you. Help me, help you! --Jerry Maguire

 

[altaphon]

Thanks for the tips ... we're a little closer, and it's not hardware, I think. I have probably driven something off into the weeds in the processor and this is part of the in-field training process, right? Is there any way I can default the whole processor?

I took the backplane apart and found some corrosion or flux residue possibly bridging a couple of pins, but nothing ugly. Cleaned and touched up a couple of solder joints, put all back together, same symptoms.

A new clue - tried an older processor board (CKE3 like the primary one but R5 v10 loaded) and the system seems to work fine. Back to the R7 processor, see what we can find out. Yes, I did a board renumber each time I swapped something around.

Earlier, the red light on the processor had been out. Now, it stays on. The only error log entries are "Command Buffer Full" on slot 02, port 00, which at that time was the MLM, and "Pool M-Busy exceeds 50" on slot 00, port 01.

I found that the 008 GS-ID module (which is now in slot 04) was in Maintenance Busy. I restored it, and the trunks worked again. Some minutes later, it went into maintenance busy mode again. I restored it again and it seems to be working now. The processor red light is still on, but there are no new errors in the log.

Maybe some part of the flash has become corrupted? Do I need to find a R7 upgrade card somewhere?

 

[dagwoodsystems]

merlinman tells you how to do a "frigid" start (full default) in thread689-910447.

Let's review the symptoms immediately after you've done this procedure.

 

[altaphon]

Hm, not many symptoms to report. Thanks, guys, frigid start cleared all the errors (error log contains only the time/date reset and frigid start entries), and everything seems to work OK, it's a fairly simple system in key mode (and the AA in voicemail remembered what it had been told before).

How long will these 1 Farad capacitors keep system programming up after a power fail? Do they ever need to be replaced?

And the inevitable, what are all the other jumpers and switches on the CKE3 board for?

 

[merlinman]

The jumpers near the "SUPER CAPS" are to "KILL" the battery back-up, and there is another one that makes the system SQUARE/KEY No matter what.

As far as how long do they keep translations, the book says 4 days, but I have seen them last for weeks.

Need to replace, never seen that, at least not with the SUPER CAPS.

 

[tdaugirdas]

Well - I have one CKE4 that has a "bad" super cap or both - it retains memory for about 5 minutes - if I'm lucky.

But Merlinman is right - I have some processors that have retained programming for several weeks or longer.

Tom Daugirdas,
President
STCG, Inc.
stcg.com

 

[altaphon]

What could have caused this failure in the first place? I have read about power glitches causing the Legend processor to crash. We were running more extension wiring just before this happened, could a momentary short across some of the MLX ports have caused this? The system is on a solid UPS, and is backed up to a file in WinSPM and to its own flash card. What else should we be doing to ensure reliability?

 

[HiDesert]

It is possible the processor software is corrupt..aquire
yourself another processor...make sure its been upgraded
to either legend 7.0142 or magix 4.71 and program your
system from scratch...this way no corrupt software will
interfere with service....just a thought!!!!