PLC5 Processor Change

tgoulding

Member
Join Date
Jan 2005
Posts
84
We have a packing machine with a 1785-L60L Extended Local PLC-5/60L controller with a 1785-ENET attached in the main rack and connected from the processor to I/O in an extension rack via a 1771-ALX.
This processor is out of memory and we are unable to carry out any substantial modifications to the machine operation. Yes we have removed redundant code and reduced data tables as much as possible but are still fairly close to the memory limit.
We tried upgrading the processor to a 1785-L80E Ethernet PLC-5/80E(more memory)we had on a redundant process and connected to I/O in the extension rack via RIO and a 1771-ASB. We found when we carried out test runs on the machine it was missing cycles, it appeared to be running slower.
The program scan times for both (i.e. 5/60L and 5/80E) were identical. I know from the specs. that the 5/60L I/O scan appears to be a lot quicker than the 5/80E. I presume this is why the machine does not function as expected. Any suggestions or help as to how we might resolve this problem would be appreciated. Remember lack of memory was our main driver behind the proposed upgrade.
 
The extended local I/O is much faster than remote I/O. RIO links can take up to 125ms to update and are asynchronous to the PLC scan.

It is my understanding that the ALX I/O are synchronous just like local I/O and are updated during the housekeeping portion of every PLC scan.

It may be possible to move the time critical I/O to the local chassis by replacing 16 point modules with 32 point modules or something along that line, then you may get away with RIO for the rest of the less time critical I/O.
 
It may be possible to move the time critical I/O to the local chassis

The part of the machine we ran whilst testing was actually using I/O in the main chassis, so that should eliminate the RIO speed issue for that part of machine.
When looking on Rockwell site at PLC5 specifications I thought it quoted I/O scan times per rack for PLC 5/60L as 0.5ms (ext. local, 10ms @57.6kbits/s, 7ms @115.2kbits/s, 3ms @230.4kbits/s whilst the PLC 5/80E was 10ms @57.6kbits/s, 7ms @115.2kbits/s, 3ms @230.4kbits/s. If this is the case surely the 5/60L is scanning I/O about 20 times faster than PLC 5/80E
 
I think your numbers are only part of the total. With remote I/O, the data is asynchronous to the PLC scan so it will vary.

My experience with a PLC5/40 with two remote racks demonstrated variation in response time up to about 0.25 second with a remote output triggered by a remote input.
Since the I/O you were concerned with was all local, and the scan time did not change, there must be something else going on.
 
Wouldn't the ethernet comm overhead increase scan time also. As before he had a sidecar to handle the ethernet comms.
 
10ms @57.6kbits/s, 7ms @115.2kbits/s, 3ms @230.4kbits/s.
These figures need to be multiplied by the number of adapters on the RIO channel. You then need to add in any block transfer operations. It can get surprisingly long.

Yes we have removed redundant code and reduced data tables as much as possible but are still fairly close to the memory limit.
Have you also eliminated any gaps in the file numbers (both data table and program)? The PLC-5's create 6 words of overhead for every file number between 0 and the highest number you've used - whether or not the files actually exist in the processor. For example, if you have a data table file N999, then there will be 6000 words of "overhead" whether or not you have any files between F8 and N999. This is equally true for program files.
 
The PLC-5's create 6 words of overhead for every file number between 0 and the highest number you've used
There are 26 Gaps (26 x 6 Words = 156 Words), not a lot of saving here. Sorry to keep getting back to the I/O Scan rate but I'm a little confussed by the info on the Rockwell website for the Specifications on PLC5. Does the 5/40L scan local or extended I/O faster than the 5/80E, if so surely the source of my problem is here, unless anybody can tell me differently.
 
PLC5 processors will scan local (in the same chassis with the PLC) and extended local I/O (using the 1771-ALX parallel interface) during every housekeeping cycle after every PLC scan.

The 5/80 does not support extended local I/O, so you switched to remote I/O which is much slower because the ASB module does its I/O scanning on a cycle independent of the PLC scan and because the PLC scans the ASB via a serial link so there are communication delays.

I think your figures for delays are only for the communication portion and don't consider the fact that the ASB has it's own scan delay which will fall randomly in time with the main PLC scan.

You said that the I/O of concern were not in the remote rack, but there is a reason the original designer used extended local in the first place, and it was probably for speed critical I/O.
 
The 5/80 does not support extended local I/O
I know the 5/80E doesn't support extended I/O, maybe I asked the question regarding I/O scan speed incorrectly. In the main chassis (i.e. same chassis as processor) does the 5/40L scan I/O quicker than the 5/80E, from the info on Rockwell website it appears so or maybe I'm reading that incorrectly as well.
My reason for concentrating on the I/O scan issue is that if the 5/80E is indeed slower then this is mostly likely to cause of my problems. As outlined before the scan rates for both processors when observed in the Status files was identical, I would therefore think the speed issue lay elsewhere, i.e. I/O scan rate, or maybe my thought process is way off the mark like the operation of the machine in question.
 
tgoulding said:
The part of the machine we ran whilst testing was actually using I/O in the main chassis, so that should eliminate the RIO speed issue for that part of machine.
When looking on Rockwell site at PLC5 specifications I thought it quoted I/O scan times per rack for PLC 5/60L as 0.5ms (ext. local, 10ms @57.6kbits/s, 7ms @115.2kbits/s, 3ms @230.4kbits/s whilst the PLC 5/80E was 10ms @57.6kbits/s, 7ms @115.2kbits/s, 3ms @230.4kbits/s. If this is the case surely the 5/60L is scanning I/O about 20 times faster than PLC 5/80E

Those are not scan times but references refering to communication.

All the Enhanced PLC 5 will have the same basic housekeeping of 3.5ms (using global flags can increase this), the rest of the scan depends on all that is involved i.e. ladder logic, STI's, etc. etc.

Look at Chapter 6 of the manual: http://literature.rockwellautomation.com/idc/groups/literature/documents/um/1785-um012_-en-p.pdf

The only difference between the 60 and the 80E is memory and Ethernet, so in the process of making the changeover some other overhead must have been added to the scan.
 
Thanks for that. We had to make some compromises in the changeover trial. Channel 1A and 1B were originally set up for DH+ at 230.4K Baud on both to two PV1000e (i.e. a PV connected to each port). For the changeover we had to connect both PV1000e to Channel 1A and use Channel 1B for RIO at 115.2K to second chassis containing 1771-ASB and Digital I/O. Is there a possibility that this communication change added to the overhead.
Also the Ethernet in the original set-up with the 5/40L was via 1785-ENET strapped to the processor, whilst in the new set-up this was via Channel 2 on the 5/80E processor, again was this set-up adding overhead and thus slowing down the processor. Again shooting here from the hip, trying to figure out why and how we might get around problem if at all.
 
I would think attaching the PV's to one channel may double the communication time. I would suspect the changes with comm's are the culprit but not sure how to remedy it.
 
I would think attaching the PV's to one channel may double the communication time. I would suspect the changes with comm's are the culprit but not sure how to remedy it.

My suspicion as well,thanks for help, may need to go back to drawing board on this one.
 

Similar Topics

Hello all, I have been experiencing dropping network connections on the Co Pro quite a bit, 3 times a month. The remedy that we do to bring the...
Replies
11
Views
702
Has anyone had an issue with a PID control after the processor has been upgraded form a L40E to a L80E. It does not seem to be controlling...
Replies
48
Views
5,895
I am working for a company that does chrome plating and the line is VERY old lol So they did a controls upgrade(before my time here) and replaced...
Replies
4
Views
1,762
Hi Friends! How many words are in an Integer in PLC5/04 Processor? In an INTEGER Say N10:0, N10:1, N10:2, N10:3..........N10:x WHAT IS THE...
Replies
4
Views
1,814
Processor PLC5/40E Series E Revision B.2 most his memory, the error log file indicating Error string: "REBOOTp1: hardware fault=0008f at address...
Replies
2
Views
1,434
Back
Top Bottom