Re: Problems with large numbers of VME58s.

[jemian@arev.uni.aps.anl.gov (Dr. Pete Jemian)]

> From lavender@sparky.csrri.iit.edu Fri Feb 23 10:02:06 1996
> Bill Lavender
> lavender@sparky.csrri.iit.edu
> 
> Hi Pete.  I seem to remember that a while back you were finding problems
> when a large number of VME58s were in one VME crate.  Since IMCA will be
> buying a goniostat and CCD detector package from a commercial vendor,
> most of our initial use of EPICS will be for devices upstream of the
> experimental hutch.  This leads us to believe that we can get by for now
> with one VME crate for the whole beamline with about 5 or 6 VME58s in it.
>
> Are we likely to see any problems with that many VME58s in one VME crate?

Probably not.  It seems that OMS has fixed the problems we were
having.  Be sure to use the version 1.8 EPROM firmware.  Lower versions
have one flaw or another, same for some higher versions.  Best to check
with Joe Sullivan or Tim Mooney at APS if you are uncertain.

'What was the solution to our problem?' you may ask...
Well, the answer is not satisfying in that OMS was never able to
reproduce our problem, even with a 60 MHz MC68040 crate processor
(mv177 board).  However, they did find that some pin on the carrier
(socket) for chip U67 (one of the big fat square ones) was bent.
Probably making trouble.  That was the only thing they fixed; now the
board works great.  They told me that a third party had stuffed some of
their boards and those particular boards had a higher rate of returns
for rework.  Could be that I got a couple of them.

The other day, I 'torture tested' my crate with as many VME 58 
boards as I could lay my hands on at the time.  By the way,
the crate processor was a mv162 (25 MHz mc68040).  I successfully
was able to use, simultaneously in the same crate, these motor cards:
  VME 8-8  at  0xFC00  (8 stepping motors, no encoders)
  VME 58-8 at  0x5000  (8 stepping motors, no encoders)
  VME 58-8 at  0x6000  (8 stepping motors, no encoders)
  VME 58-8 at  0x7000  (8 stepping motors, no encoders)
  VME 58-8 at  0x8000  (8 stepping motors, no encoders)
  VME 58-8 at  0x9000  (8 stepping motors, no encoders)
  VME 58-4E at 0xA000  (4 stepping motors, 4 incremental encoders)
That's a total of 52 motors.  They all responded to 'caget' in the
appropriate fashion.  I'm now satisfied that UNICAT can run the
motors we need from one crate.  Also, there were several other
cards in the crate during this test.  These were:
  mv162 crate processor
  mv162 HiDEOS processor
  VSC8 (Joerger) scaler
  AVME 9440-i-L Digital I/O board
They all worked correctly.

Regards,
-------------------------------------------------------------------------
			   Pete R. Jemian
			  UNICAT  Controls
-------------------------------------------------------------------------
UNICAT, Sector 33, Bldg. 438D                  office:   708 - 252 - 0863 
Advanced Photon Source                            lab:   708 - 252 - 0810
Argonne National Laboratory                       fax:   708 - 252 - 0862
9700 South Cass Avenue                          email:     jemian@anl.gov
Argonne, IL  60439                              voice:          Hey, you!
-------------------------------------------------------------------------
Education is the one thing people are willing to pay for and not receive.
-------------------------------------------------------------------------