Experimental Physics and
Industrial Control System

"Martin L. Smith" <[email protected]> · Tue, 29 Mar 2005 07:31:28 -0600

Hello Li Gang,
	The standard VIPC-616 without the -01 option does need to be
modified in order for slot D to work properly.
As delivered by SBS the VIPC610 has IRQ0 of slot D connected to the
VMEbus level 7 interrupt, and IRQ1 of slot D does not cause a VMEbus
interrupt at all. Under vxWorks on the MC680x0 family a level 7
interrupt level may only be connected to a service routine that does not
make any calls to vxWorks routines. Thus with a VIPC610 board slot D
should not be used for modules that generate interrupts. The VIPC610-01
board option is free of this problem.
To modify the VIPC-616 to an option -01:
1. Change the program in U17 to G52211
2. Change the program in U12 to G52212
3. Remove any shunts and wires on E9, E10, E11, and E12.
4. Add wire wrap wires:
   E10-1 to E12-8
   E10-2 to E12-5
   E10-3 to E12-2
   E10-4 to E12-1
5. Add shunts
   E11-3 to E12-3
   E11-4 to E12-4
   E11-6 to E12-6
   E11-7 to E12-7
This option changes the interrupt map on the VIPC-616 to look
like that of two VIPC-310s.
Also, if you need I do have the new PAL code.
I think that I should also mention to you that we at APS have found a
problem with the VIPC-616 board even with the -01 option. The problem
ocurrs only with a PPC and at least 2 interrupters and happens only
during back to back interrupts.
It seems that the VIPC-616 board requires the VME bus IACK* to toggle
when servicing interrupts. Without getting into all the gory details
what happens then is the VME bus gets a bus error (BERR*) and the CPU
wiats for a timeout of about 250us before going on to the next bus
cycle. This may or moay not be a problem for you though; however, you
will probably miss some data from the VIPC-616 when this happens.
I have information on this modification as well but it requires that
a chip be removed from the board, a couple of board traces cut and
a PAL code change in U12.
Future purchases of these board should be the VIPC-626ET instead of
the VIPC-616 as the 626 does not have these problems.
Marty Smith
ANL/APS
Controls Group

Mark Rivers wrote:

Dear Li Gang,

Is your VIPC616 the standard board, or do you have the -01 option? I
believe that the standard board does not support one interrupt, which is
used by the last slot (D). At the APS we use the VIPC616_01, and I
think it works fine with IP-Octal in any slot.

You can convert a standard VIPC616 to the -01 option yourself. You need
to wire-wrap some pins, and replace 2 PAL chips. Marty Smith at the APS
can provide you with the details on how to do this.

Mark Rivers

------------------------------------------------------------------------
From: ligang [mailto:[email protected]]
Sent: Monday, March 28, 2005 3:49 AM
To: [email protected]
Subject: The later four ports of IP-Octal-232 could not work.

Hi Everyone;
 I am using the IP-Octal 232 and VIPC616 carrier board, with
drvAscii/drvSerial/tyGSOctal. There are four slots(A ,B ,C ,D ) in one
VIPC616 and eight ports(a ,b ,c ,d ,e ,f ,g ,h ) in each IP-Octal 232.
The communication is ok between IP-Octal 232 and other devices, when the
IP-Octal 232 is put into slot A, or slot B ,or slot C of VIPC616. When
the IP-Octal 232 is put into slot D of the VIPC616, The later ports(e
,f ,g ,h ) of the IP-Octal 232 doesn't work again, however,the first
four ports(a ,b ,c ,d ) of IP-Octal 232 work well. I have exchange
several IP-Octal 232 and VIPC616 respectively, the same errors will
occur again.
 Is anybody could do help me? any adivce ia appreciated! Thanks in
advance.

Li,Gang

Accelerator Control Group,
Accelerator Research Center
Institute of High Energy Physics (IHEP)
P.O. Box 918-10
Beijing, 100049
P.R. China

Tel:    +86-10-88235040 ext. 808
Fax:    +86-10-88236190
E_mail: [email protected] <mailto:[email protected]>

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System