Pick and place code that was uploaded...

[Chris Elston 78]

I downloaded that pick and place code that was uploaded. I always wanted to know how others might use that Squencer output instruction, but I struggle to see any saving of logic programming time. It seems to me that you still need to set up some kind of a file that has all your prox switches to verify your location of your device. I dunno about you other guys, but we like to use prox switches for every motion to check and make sure the output or cylinder made it to the position we told it to go to before we tell it to move somewhere else. Or in the example of a pick and place. If I tell a cylinder to retract, I want to make sure the cylinder retracted before I tell it to transfer. So if the cylinder fails or something jams, I don't transfer will the cylinder in the DOWN position or mid stroke position. Anyone else think this squencer instruction is kewl to them? My point of view right now is that I am going to have just as much programming time coming up with input "checks" for the sequencer compare file that I would have if I just programmed it all in bits, like the Index PDF file that was uploaded. I guess what I am trying to make sense of, is this worth my time trying to figure out a way to use a sequencer output rather than programming bits or "step logic" like I have for the past 5 years? Anyone have any horror stories with squencer output instructions? I have heard they are dangerous and unsafe was my main concern I sought after... :-3 Thanks for the upload on the pick and place.RSS file..I enjoyed it.  :)

[Guest Jerry Desaulniers]

To sequence or not to sequence? That is the question! In reply to your statements the AB sequencer is a Great function in my opinion, great with a capital G. Your right it does not save a ton of time in many programming application, but when a ladder gets overwhelming I prefer to look at a data table.  Also another major benefit of sequencing is that when someone decides they want to add something(sensor or extra actuator) it takes only minutes to change a few bits in the data table. As for safety there is no difference a zero is a zero and a one is a one. Troubleshooting is another story. When I submitted the P&P program I intentionally left out a concept I would normally include in a P&P program. I would normally sequence only bits and then in a dedicated section of the program I would list all the bits related to I/O in order and clearly label this section of the program as the troubleshooting section, this way a maintenance person could have a page where they could watch their inputs and outputs. I know this might not be a clear explanation but I'm out of time and must ready myself for work, if you need the supporting code just ask and I will be glad to supply it for you. Also the more motions that need to be controlled the more advantagious it is to use sequencers. Of course the opposite also applies.

[BobLfoot 301]

I know that this is an old thread, but I just ran into a machine in our facility that uses not one but four sets of three SQO's each to run a sanitization process. Never mind that for the last five years corporate policy has declared the SQI, SQI & SQL as well as JMP, MCR, TND, UID and UIE are illegal instructions and shall not be used. These PLC 5's pre-date that edict and are critical enough no one wants to risk breaking a working system. Now comes my dilema they want to change time length of certain steps. Think I've found the correct integer's to manipulate. I'll post my observations here as I progress. Sorry I can't post the code {proprietary and all}.

[Smoke 0]

We build many new machines every year. Everything is is ran with a SQO. To add to the confusion every motion uses the same sequencer. We have a user defined JSR that plugs the required input data into the standard sequencer. eg. the standard is like A_dir or B_dir. the JSR would plug in Press_up or Press_Down. To make trouble shooting more complicated for the average troubleshooter the is a mask involved to ignore sensors that are not needed to step the sequence. Now the JSR passes out the command to actually turn on the required output. I wish I could show you. You wouldnt like it. I do like to sequence though. I use a bit shift left, in GE I use a bit sequencer (its about the same) It is very simple to troubleshoot if the step is true but the status wont come on all you have to do is search the the status bit and what ever contitions to step the BSL are there. I will export the logic and post it since it is my own. BSL_32_bit_sequencer.zip I would like to here comments thanks.

[Gerry 13]

Hearing about these sort of edicts really gets my hackles up. It indicates arrogance and ignorance and perhaps incompetence. "I'm the boss and I don't understand these instructions, therefore they're illegal" or "I'm the boss and although I understand these instructions, lesser mortals are too stupid to use them and therefore they're illegal"

[jstolaruk 9]

Customer specifications that rule out sequencer instructions are very common in the machine tool business and I believe it doesn't come from unknowledgeable bosses or supervisors but from the fact that it can be very difficult for maintenance electricians to quickly trace and resolve problems. From what little I've experienced, sequencer logic can only withstand a small variation in timing and once it gets too far out of whack it can be difficult to get back.

[Gerry 13]

The last time I programmed using SQI and SQO was on a PLC-2 over 20 years ago. Probably, I haven't used it since because it wasn't warranted for the applications or maybe the novelty wore off. But I use MCR and JMP instructions often. I have come across other edicts banning the RTO and TOF which I also make use of. While I have no specific application in mind where I would chose to use sequencers, I would object to having my hands tied. Not sure what you mean about timing - the sequencers should advance on events, not time intervals.

[BobLfoot 301]

You are correct the sequence is advanced by the Step Done Bit Technically, but that is driven by the Step Length Timer Done bit. Each Step in the sequence has a predetermined length in time and the timer trips the next step.

[jstolaruk 9]

I was thinking to the timing of the machine verses time intervals. I witnessed a situation where a pipefitter had changed the flow control on a cylinder so that its extend time finished later than some other movements and the sequencer locked right up, like the inputs were in a state that confused the sequencer. Thankfully, I wasn't the controls guy on that machine because he had his hand full there for a bit trying to figure out what happened. I understand the frustration with the feeling of having my hands tied but in some cases there is a good reason for it. I had a lot of time getting bloody fingers on T3 terminals plugged into PLC-2s back in the 80s and thankfully never got into the sequencer instructions. In fact, I recently replaced a 2/30 with a PLC-5 and there was a T3 dedicated to it. Sure brought back some memories. Edited 10 Dec 2006 by jstolaruk

[TimWilborne 108]

I have to disagree here. When timing is an issue it is better attributed to a poorly designed sequencer. I have found many maintenance personel find the sequencers faster to troubleshoot than other methods It is all in how you write them. BUT I think the real reason that these instructions are frowned upon by many companies is that they are not standard instructions with all PLCs. So using these instructions require special training. Just my two cents

[BobLfoot 301]

I am enjoying this ongoing debate. I must say that I have written a number of sequencers based of bit shifts and multiplies but this is the first machine where I've see SQO used.

[jstolaruk 9]

That must have been the case in the few that I've known about and the end-user had such a terrible time with them.

[Peter Nachtwey 15]

I think I feel as strongly about this as Gerry does. All these instructions are just tools. I also agree with TWControl about the timing issues and the quality of software makes a lot of difference in how easy the system is to debug. A good program should have a significant part of it used for diagnostics. I was told a long time ago that a quality program makes it easy to recover from foul ups. Often times a machine must meet a piece rate specification. Because of physics it is often very hard to make one machine run faster than another. What often gives a machine a significant advantage over the next is its ability to run consistently and recover from difficulties more quickly. In saw mill applications this may require backing up the shift register data so a piece in the machine can be cleared. Now back to Gerry's point. The world is becoming a competitive place. The winners will seek every advantage and realize that the workers they have are the ones they have to win with. Sometimes management needs to be like a coach. A good coach does not limit his play book, he trains his team to execute the play book. This allows the coach flexibility to meet the challenges of the other teams. You have to train your team to meet the changes.

[jstolaruk 9]

Yep, all good points. As with any tool, if its not used properly the results can be less than desireable. Customer's have specifications because they believe they know what they want and how its supposed to be served up. As a vendor its my job to make sure I stay in the guidelines because I want to be assured of getting a chance to quote the next project. I posted some pics in my album here of a machine that uses (4) servo actuators instead of tried and true hydraulic cylinders to move around a square transfer in a high pressure deburr. Perfect application for ControlLogix (imbedded motion instructions) and SERCOS interface. But the customer wouldn't budge on allowing anything but a SLC 5/04 or a Siemens controller; her electricians didn't have the training on the hardware and software so they couldn't support it. Their sister plant 3 miles away would allow it but not this plant. So, I follow his spec. because I want to remain on the bid list for future projects. "The customer is always right, even when he's wrong." Edited 10 Dec 2006 by jstolaruk

[TimWilborne 108]

Know what you mean. I know one plant that insist on using PLC5s on their machines. Now there is nothing wrong with PLC5s but these machines all have 2 axis motion control on them. Much easier with the Controllogix but everyone there is familiar with the PLC5

[Smoke 0]

I am sure here what happened here was when the flow was slowed down another done bit turned on and then the done and enable bit couldn't equal up and advance the sequence. I have seen this before. this is where you start getting into masking. The e-techs should just look at enable and done bits but they always end up searching and getting to the SQI and SQO then they freeze and through there hands up. That is what have against these drum sequencers.

[Ken Moore 23]

I think it's partially due to the cost of another software package. In my plant we have drug our feet on going from 5's to CLX because of that. It's hard to justify the software cost for a small project. If Rockwell had some kind of upgrade option, we would have moved on a long time ago. I do have a very large project coming up 1st qtr 07, that will use a CLX because of the SIL-2 requirements. Once I have the software, I can then start to use Compact and Control logixs instead of 5's and SLC's.

[BobLfoot 301]

Don't forget Ken that if implemented wisely with ControlNet I/O that you can convert multiple PLC 5's into a single control logix and not sacrifce your 1771 racks or hardware. Just change the CPU's to ACNR-15 cards.

[panic mode 246]

"Controls Olympics" should quickly show how champions do it...

[jimdi4 2]

Don't forget Ken that if implemented wisely with ControlNet I/O that you can convert multiple PLC 5's into a single control logix and not sacrifce your 1771 racks or hardware. Just change the CPU's to ACNR-15 cards. Woow!! Can you elaborate on this please?? Are you saying you can get rid of all of your PLC-5 processors and keep your 1771 racks? Are these ACNR-15's a CL 5k processor disguised to fit into the place of a standard 1771 processor slot? Does that mean you have to then change your programming software to RLogix 5k or does this ACNR-15 allow you to program with the RL-5 software? Edited 13 Dec 2006 by jimdi4

[Peter Nachtwey 15]

Sign me up. I wouldn't use a PLC though, I would use a motion controller. A pick and place is should be very easy unless there are many more axes than two and it requires trig in three dimensions. But then SQOs and SQIs would be the least of the worries. The basic go here and wait till done then go there and wait till done is easy.

[paulengr 44]

No reason to do it in Controlnet either. You can just as easily do it in DH+. Simply buy DHRIO cards for the Contrologix. Set the processor into scanner mode. Then extend the DH+ run back to the Contrologix DHRIO card. The PLC-5 becomes I/O (rack 0) on the DH+ port of the Contrologix. The rest of your remote I/O and the 1771 rack itself (and the wiring) stays in place. You can do something similar with Controlnet but since we already have everything on DH+, I didn't investigate that option.

[BobLfoot 301]

There are a couple of options available each with pro and cons. Option 1 Description If you have typical PLC 5 system with a main 1771 rack where your CPU resides and additional remote racks which are connected by 1771-ASB modules. You can convert that system to ControlLogixc by using a 1756-Rack containing a 1756-LXX CPU and a 1756-DHRIO card set to run DH+. You simply replace your PLC 5 processor with a 1771-ASB and "port" you RSLogix 5 code over to RSlogix 5000. Changing the address scheme accordingly. Option 1 Pros This works well for converting one PLC 5 to ControlLogix. It is great for testing and debug because you simply module swap PLC5 CPU with ASB to test controllogix and can return to PLC 5 control for production until ControlLogix program is fully debugged. Option 1 Cons RIO has a speed limit of 230Kbaud and therefore bandwidth limits. HMI stations get an equal timeslice with I/O stations. IF trying to convert multiple PLC5 to single controllogix may not have enough RIO bandwidth. This can be overcome with multiple DHRIO cards. Option 2 Description If you have typical PLC 5 system with a main 1771 rack where your CPU resides and additional remote racks which are connected by 1771-ASB modules. You can convert that system to ControlLogixc by using a 1756-Rack containing a 1756-LXX CPU and a 1756-CNB card. You simply replace your PLC 5 processor and 1771-ASB cards with a 1771-ACNR cards and "port" you RSLogix 5 code over to RSlogix 5000. Changing the address scheme accordingly. Option 2 Pros ControlNet has a higher speed and bandwidth over RIO. ControlNet supports scheduled and unscheduled communications allowing priority to be given to time critical i/o while non-criticals and HMI's can update at a slower rate. Can provide redundant comm paths which RIO does not. Option 2 Cons ControlNet does require coax cables and new RSNetworx software that RIO does not. There have been some issues with 1771-Analog Block Transfers and ControlNet. In a high Analog situation a mix of RIo and CNET may be necessary. Hope this makes it clearer.

[Ken Moore 23]

Sometime last year, don't remember exactly when, I looked into replacing the 5's with a CLX or two. I seem to remember that there are a few tricks you have to use if you have a lot of block transfers, which I do. The process in question uses three 5/80E's. The plc's are networked via DH+ and Ethernet is used for HMI's and programming. Two of the 5's also have Pro-soft ModBus cards that connects to dedicated modbus networks for each plc. There isn't a lot of data passed between processors, around 100 words each, mostly permissive and status bits. The analog/discrete mix is about 40/60, with many drives and weigh scales connected via RIO as well. Can one CLX replace three 5/80's? Each one has about 15-20% free memory, with lots of ladder. This is a batch processing application based on ISA-S88, to keep the scan time respectable, I have lots of subs that only get scanned as needed. Average scan time is around 40-50 msec, but can sometimes jump to around 100 msec. Since this is process control and not machine control, scan rates are not as critical. There are also 4 5/04's and two panel views on the DH+ network that handle the packaging machinery. In this case scan time is critical, thus the dedicated SLC's. Since 5/80's cost about the same as a good used car, once we use up our spare, CLX will have to be considered. We can repair our 5's and get refurbished ones, but this is not a valid solution for the long haul.

[BobLfoot 301]

Let me start out by saying this is just my $0.02 and not what Rockwell would claim or recommend. I am quite comfortable that a 1756-L63 version 15 could do the trick. Will it be a walk in the park. Not in the least. Would I advise engaging the services of an SI that has done several of these before. Definitely. Now my initial impressions. Move the PLC 5's to the CLGX one at a time. Debugging the system at each move. Done Correctly you can debug on a weekend and put the 5/80 back in for several weeks of production before shifting over the the CLGX for Control. 1756-A13 Chassis as follows: Slot 0 - 1756-L63 Version 15 CPU Slots 1 & 2 - Replace the PLC 5 Prosoft Modbus with the Controllogix Modbus cards in your ClontroLogix Chassis. This may prove easier than Blcok transfers to the modbus cards over RIO. Slot 3 - 1756 DHRIO Card - Port A is DH+ network #1 talking to the SLC network which we leave alone for now. Port B is DH+ Network which will connect to any existing PLC 5 not yet moved to controlLogix. SLot 4 - 1756 DHRIO Card - This is Configured 230K RIO Both Ports and Picks up the CPU Chassis and all RIO components of PLC 5-1. SLot 5 - 1756 DHRIO Card - This is Configured 230K RIO Both Ports and Picks up the CPU Chassis and all RIO components of PLC 5-2. SLot 6 - 1756 DHRIO Card - This is Configured 230K RIO Both Ports and Picks up the CPU Chassis and all RIO components of PLC 5-3. Slot 7 - 1756-EN2T or ENBT card for communication with the HMI's. Slots 8 thru 12 as spare for anything we forgot. Unless you ahve redundancy in you field IO -- Don't get lead down the Redundant ControNet or worse yet Redundant Processor Primrose path.