One PLC, One machine V's One PLC, Two Machines

Mylo

Member
Join Date
Aug 2002
Location
Dublin
Posts
135
G’Day all,



Just would like to get your opinions / experiences on the following, A customer has asked for us to check out the possibility of using one PLC for 2 x machines, typically we use a dedicated PLC for each machine.



Some background:



We are an OEM for the plastics industry mainly Automotive, we have previously and currently built some machinery for an automotive interior application with 4 x variants : Right hand Front machine, Right hand Rear machine, Left hand Front and Left hand Rear, simple so far.. we utilised separate S7-300’s & TP170’s in each machine typical IO count of 192 each machine, the physical layout has been 4 x separate machines. OK The customer has asked to check out the possibility of using one PLC for 2 x machine eg Right Hand Front & Right Hand Rear. My initial concerns are of flexibility, also uptime, if one machine were to go down (Not that they would but if !!!) then the machine in question may have to be powered down to facilitate repairs thus causing downtime on the second machine, god forbid there was a requirement for software mods. I have no real problem with the PLC controlling two operations however my programs have been time sensitive in the past.



Now I would believe the IO count would simply be doubled 192+192= 384 , ok but the savings on a CPU compared to the issues above I don’t believe in our application are worth it, I still believe that 2 x HMI’s would be required as there are a lot of separate parameters and alarms required for each machine, but hey I’m only the designer……

(y)

Any thoughts ?????

Cheers
 
Hi Mylo.

From your description it sounds like that your machines are both standardised and working as standalone units.
From that I would say that it would be madness to integrate the controls for two machines into one CPU+HMI.

It would make sense to try and make a special lowcost version by finding a smaller CPU and a smaller HMI. If you are using for example S7-314 + OP270, then maybe you can cut it down to S7-313C + OP170B (or the new OP77A or OP77B).
But any savings that can be gained from making a special version has to be weighed against the expense for the time you must spend to implement the special version.

Myself, I am standardising wherever I can, and I wont change my standard design unless there are some very good reasons for it. And saving a little bit on hardware is not a good enough reason.
 
You could also use the arguement that the increased application development time will greatly out weigh the cost savins of using one PLC. Using one PLC for two machines will be more expensive.
 
From a business standpoint, why not?

The customer has made a request and is entitled to a thorough Engineering review of the request. Whether we, as programmers and designers, like the idea or not, is irrelevent. If that is what the customer wants, after being shown all the data, consequences, and recommendations, then that is what he should pay for and get.

IMHO, do the evaluation, make your best technical recommendations, point out the apparent ramifications and costs, and then let them decide.

This is THEIR business, after all.

Steve
 
Information is control

Well I am no S-7 expert, but the customer is asking for more than just one PLC control. What they are really wanting is integrated control of several production stations by one central controller. Why, data collection, SPC information, improved troubleshooting, less down time. When all the stations are controlled by one central PLC the cost of information collection will be over all less. As you know the Auto Companies love data. You can easily sell all this as an up grade and it will definitely cost more and they will pay it, because they want information. Just get one of the customers IE engineers involved and you will realize what they are striving for.





Remember the customer is at A they want to get to C, your B.
 
Steve Etter: "...do the evaluation, make your best technical recommendations, point out the apparent ramifications and costs, and then let them decide." I agree, IF you need their business that bad. If you don't, you, as the system integrator, have to choose whether you want to stand behind any issues that may arise from a central PLC setup. Well implemented, this may not be a babysitting job waiting to happen. Poorly done....

Emartell"s "data collection, SPC information, improved troubleshooting" can be done with any number of PLCs.

The end result is to deliver a solution that EVERYONE can easily live with; you and the customer. Bad solutions equal bad advertising, hence lost business. Avoid the pitfalls even if the customer asks you for them.
 
Absolutely, Rube. As a systems integrator (or machine maker, or similar) you always have the option to submit a "No Bid" if you can't live with the request.

Steve
 
Sometimes y'all over complicate a simple request, I reckon thats the Engineer part. I dont see any advantages or disadvantages to doing this. If the machine(s) have already been designed/developed using 2 plcs then the savings would be just in removing 1 plc BUT the program would have to be rewritten and some wiring changes to accomodate possible remote I/O so the savings would probably be eliminated...depending on options it could cost more because of the wiring/devices used.

Actually I dont see a way to make an informed decision on this, the details of the machine are not known, things like timing issues, update times etc etc. Are there critical issues involved that make using one plc impossible.

1 or 10 I dont see how it would improve data acquisition or troubleshooting.

Only the originator of this thread can look at the system(s) and determine if its possible to do or if it will save the "buyer" money which is probably their goal. I think I would talk to the "buyer" to determine "WHY" they made this request.
 
Absolutely, Rube. As a systems integrator (or machine maker, or similar) you always have the option to submit a "No Bid" if you can't live with the request.

True, but do it more than a couple of times and you may not be asked to bid again!
 
From my own experience, I am currently writing up a program to split 1 processor running two separate machines into 2 processors, because whenever we have any type of problem with the rio's on one machine, the other machine goes down as well. There are some other issues contributing to this(poorly sealed cabinets, etc.), but when we do have a problem in this area it takes a lot of work to track down where the problem is coming from.
 
There are several considerations, and as Steve pointed out it is ultimately the customer's money so they get to decide. I generally go to multiple small PLCs. Some of the things to include in the evaluation:

1) Reliability & Redundancy. If you have a single PLC, and it dies, you have two machines dead. With a separate PLC for each the cost of downtime is reduced.

2) Hardware is cheap, at least in my systems. The cost of engineering, programming, and commissioning exceed the cost of the PLC CPU and rack by an order of magnitude. If you use one PLC, you will have to charge for at least some new program development.

3) Data Collection. It is easier with one PLC, but communications nowadays is such that this may not be a big deal.

4) Commissioning. If you have home run wiring to a single cabinet, your test and debug time will increase, plus you have two machines that won't start until you're done. With separate PLCs you can divide and conquer, so commissioning both, but especially the second unit, is faster.
 
Many thanks for all of the post’s, they’ve given me a lot of food for thought,☯



To reply, while our machinery is nearly always a modification on previous standard design’s, I’m always attempting to improve our design. This involves constantly examining our way of thinking, and sometimes re-evaluating the design of the “wheel” ! to see if it is adequate for our current applications if for no other reason than to self check the design. So I never have a problem evaluating for customer requirements, they are the ones paying the bills after all.



However sometimes with a little information and a God like attitude the “customer”👼 does request some damn strange things,,,,,I digress , in this case the customer’s request has reasonable in content, & I do feel it has to be explored hence I came to my best resource available !!!!!!!!:site:



To argue the case from a concept point of view, the demands our application/control puts on a PLC is high in quantity but not really on technical, i.e We have a lot of operations but no high degree of maths/analogue/motion/comms etc the Simatic S7-300 cpu is more than enough for our needs, the only worry in that regard is that our “Process” is very time dependent, & I don’t know if the scan cycle would be more burdened than it already is having effectively twice the program to process, in which case the application would possibly call for a higher spec cpu, i.e S7-400, which would negate the cost saving’s in hardware. I can’t check this until I have some time to write some dummy code out to test.



To argue from a costs point of view, which is primarily where the customers stand, I am pretty sure that the savings would be negated as pointed out with extra development time on a previous design, which in itself is no bad thing, I’m just trying to get a grip on any ways around it. ( as always)o_O



Cheers and thanks,
 
I don’t know if the scan cycle would be more burdened than it already is having effectively twice the program to process
The scan time will certainly increase when having to plough through double the amount of code.
If the increased scan time equals decreased production, then there is really no doubt about it. The loss in production will be effective in the lifetime of the machine.
Plus, if there are problems with the controls, it will affect two machines and not one (the old saying about not putting all the eggs in one basket).
Plus, you can place the controls for each machine as close as possible with an individual PLC+HMI (= shorter cables and easier to troubleshoot).
Plus, you dont have to spend any time with developing a special version.
 
Hi Jesper,

That pretty much sums up this appllication !

Case proven,,,,,,,,,,,,,,,,,,,,this time !
 

Similar Topics

Hi All, How do I set a password to PLC using Proficy machine software 9.5-9.8 or some other way? I as using Emerson CPE305 CPU. Thanks.
Replies
2
Views
157
An outside contracting firm designed a machine for our company. There are several devices connected through Ethernet/IP. This includes a Panel...
Replies
4
Views
182
Hi all, PLC machine vision is barely talked about anywhere yet it is very important PLC and Industrial automation topic. Like you see vision...
Replies
0
Views
1,046
My boss has come up with the idea of renting machines and charging on a per use / per hour basis, so I'm looking into it - just getting my feet...
Replies
14
Views
3,913
I got a problem concerning the change of coins from a vending machine plc based, i use the rs logic starter lite with Rs emulate 500 along with...
Replies
5
Views
2,052
Back
Top Bottom