When tyring to optimize the period and phasing of scan blocks to distribute I/O loading of a PLC Gateway/Integrator, I find the following items to be helpful: 1. Reduce the station's BPC to the lowest possible. This depends to a great degree on the turn-around time for the PLC to service the scan query. With AB PLCs I find a setting of 0.5 Sec to be quite acceptable. The lower the BPC, the more scan phases/slots you have to allocate/play with. 2. Start by distributing the 'bigger' scans (those with relatively bigger number of data bytes requested); one to a phase, along the the width of the 'acceptable' overall scan period. For example, with a BPC of 0.5 S and an overall scan period of 4 S, you can scan 8 of such blocks of data. These are usually the analog (AI/AO) data. Suggestions for this exercise: - Turn OFF all compounds with scan blocks and set compound period and phase to default of 1 and 0, respectively - Initially modify the scan blocks' period and phase, without changing those of the receiving I/O blocks; they can be done later - Verify I/O loading 'acceptability' by turning ON only the 'big' scan compounds thus distributed so far - The 'acceptable' overall scan period will have to be 'stretched', if all the 'bigger' scan blocks cannot be allocated a phase/slot each 3. Add the 'smaller' scans to those already distributed; note that only about 80% of the BPC is 'allocated' for I/O scans (or 400 mS in a BPC of 0.5 S). These are usually the digital/binary data. In some cases, I find that I can fit-in 2 of these smaller DI/DO scans with a bigger AI/AO scan. - Add the smaller scans by turning them ON after modifying the scan block period and phase. - If any of the DI/DO scans are 'required' to have smaller scan periods (higher frequency), you can 'think' of distributing one of them like two (or four) scans along the width of the overall scan, separated by the proper amount of phases/slots. For example, with a BPC of 0.5 S and an overall scan period of 4 S, a DI scan required to be scanned at a period of 2 S, can be assigned phases 0 and 4. The latter being equivalent to phase 0 in a 2-second period. 4. If the AB PLC can be reconfigured, and the number of I/O scans is too big for the 'acceptable' overall scan period, try to re-map the points into continuous/contiguous ranges, to reduce the number of scans required. The number of bytes that can be read in one scan has a limit of ???. > -----Original Message----- > From: foxboro-bounce@xxxxxxxxxxxxx@CORP On Behalf Of Corey R > Clingo <clingoc@xxxxxxxxxxxxx> > Sent: Thursday, 03 January, 2002 6:07 AM > To: foxboro@xxxxxxxxxxxxx > Subject: Re: [foxboro] Foxboro and Allen Bradley > > Rob, > > We have 3 Int30s addressing a number of PLCs, mostly 5/15s over DH+. We > use > 1785-KE modules to convert from DH+ to RS-232 (this is the "CAM"). They > work > OK, except for two things: > > 1. We have a problem on one of them where we are unable to change the > phasing > for a compound or block from phase 0; doing so causes that compound to > smurf. I > send my saveall for this one to TAC, but they were unable to reproduce the > problem. I don't know of anyone else who has the problem, so it may be > something maddening with our setup. But I would not have needed to do > this if > it weren't for.... > > 2. low throughput. When I inherited the system, I noticed that our IO > scan on > the station blocks on all 3 integrators was >100%. I really cannot figure > out > why this is happening; in all cases, the amount of data being scanned is > well > under the maximum bandwidth of the serial link. The PLCs aren't > overrunning > either. When I get time I want to investigate this further > > I reduced the scan frequency on many of the blocks and got the IO scan > down > below 100 on two of the integrators, but the third is still close to 200%. > I > was trying to move the phasing around to balance the polling (1 sec BPC, 2 > sec > or greater block processing in most cases) when I discovered problem 1. > > I inherited this system, and am still learning I/A, so I was not aware of > the 8 > PLC/16 ABSCAN block recommended limits. I know we exceed these on at > least one > of the integrators, but these seem awfully low to me. Really, how much > extra > horsepower does it take to drive a serial link? > > Corey Clingo > Sr. Engineer > BASF Corp > > -- No attachments (even text) are allowed -- -- Type: application/ms-tnef ___________________________________________________________________ This list is neither sponsored nor endorsed by The Foxboro Company. Use the information obtained here at your own risk. For disclaimer, see http://www.thecassandraproject.org/disclaimer.html#maillist list info: //www.freelists.org/list/foxboro subscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=join unsubscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave