1. Bench test. When working with a controller, it's important to know its capabilities and if it meets the needs of your project. There's nothing worse than being halfway through the project, only to find out that your controller isn't capable of doing something. If the project includes something you haven't done before, always bench-test the process before committing it to the project.
2. Don't over-specify. Don't select a specific controller too soon in the design phase of a project. Selection should occur only after the initial machine or process automation conceptual design us completed. Far too often an over-specified controller unnecessarily adds to the automation project's costs. Or even worse, an under-specified controller may result in redesign efforts, additional purchases and schedule delays.
3. Clarify interfaces. If multiple people program parts of the PLC code, spend some time initially to make sure that the interfaces are clear and everyone understands them. Dumping a diagram on someone is not good enough.
4. No canned code. When planning for PLC-based controls, if your in-house techs will be involved in maintaining and upgrading the new system for future needs, don't bring in "canned code." Ensure that your techs have input as to the structure and nomenclature of the program. Bugs and changes are the norm with any new control system, and you will find your facility rebounding much quicker after such a major change in equipment.
5. Can we talk? A primary concern when implementing any PLC is communications. This has been minimized with the advent of ODVA, but can still be a problem when dealing with some sophisticated sensors or peripherals. Choosing a PLC family that has all the device type communications modules is paramount. Most PLCs on the market can communicate on device buses, but not all can accommodate every type of Ethernet. Do your homework and choose a PLC that can support all your buses without a lot of pain.
6. Not all controllers or brands are the same. Consult with people who are knowledgeable in many different hardware platforms before specifying or selecting a certain make and model or even a platform. Too many times what is specified will not perform to the customer's expectations. That leaves gaps in what is bid and will require changes in the project.
7. Key check points. Whenever upgrading a system or changing from one controller to another, always remember to check the controller's power supply, memory, I/O type, size and availability in the panel. Double-check the purpose for which the controller is intended, and always carefully check the model number. Make sure the controller has the basic features you want, such as online bypass, online programming, alarm management or closed loop control.
8. HMI first. When programming your project, it's best to start with the HMI first. If you write your program code first, quite often you'll have to re-write your code to accommodate your HMI.
9. CPU Load. It is important to consider the CPU load if communication handling is vital to the application. Running at the maximum cyclic load will result in poor capacity and response times for peer-to-peer and OPC Server communication. Peak cyclic load should be kept below 65%, and static cyclic load below 60% under all conditions. Even lower cyclic loads (30-40%) are desirable if high communication throughput is important.
10. Get out of the code. Ask the people at the plant why they are running the way they do. Sometimes what you might think is bad programming is really needed to optimally run a process. On the flip side of the coin, after examining the code you often find people think a system is running in a particular way, but the reality is that it is not.
11. Document software. Use of programmable controllers offers significant advantages over analog devices. However, it brings a new set of issues. First, consider the control of multiple function block parameters. It is a good practice to maintain the software configuration control document that summarizes all programmable/tunable parameters in one place. You will use this document for any disaster recovery event, including cybersecurity issues.
12. Think ahead. Don't go for the cheapest option when installing a PLC. Think five years ahead to accommodate factory integration, peer-to-peer networking and data exchange between controllers. The modern factory environment is becoming a network of integrated controllers.
13. Don't over complicate your solution. Keep it as simple as possible. Don't use a PC-based system when all you need is a little advanced I/O control.
14. Prior experience. Do not use a controller for critical applications in your plant without a successful prior use experience with a like product with the same revision level components, firmware and software.
15. Are micro PLCs an option? Micro PLCs with flexible, 'just-enough' control may enable OEMs to differentiate their equipment, particularly in stand-alone machines. By developing a range of stand-alone machines using the same controller platform, OEMs can reduce design time and lower their costs. Look for micro PLCs that include:
- Flexible hardware configurations, like USB, up to 6 serial ports and Ethernet for communications.
- Up to 3-axes of embedded motion.
- Plug-ins and 2085 expansion for I/O USB.
- Single programming software package that eases installation, configuration, connectivity and maintenance.
16. Single control platform. Customizing your controller selection for each application may seem like a smart move, but sometimes it pays to narrow down to a single control platform. Survey your requirements and select a platform capable of handling all current and near-future needs, from the simplest to the most complex ones. Selecting a single control platform for all of your automation control requirements, whether it is motion, robotics, numerical control, C program or sequence control, can reduce spare parts inventory, unify programming methods across all machine types and simplify training and maintenance efforts.
Speed up controller response
Shave milliseconds off controller response times by optimizing subroutines, exploiting interrupt programming and leveraging peer-to-peer review. See how by viewing this white paper. http://awgo.to/025
Source: Rockwell Automation
Flexibility, support, warranty, total cost
There are four key factors to consider when selecting a PLC for a project:
- First, the controller must meet your performance needs today, but consider that the system may operate over the next 5 to 10 years. During this time, changes to the system may require modifications and choosing a flexible system upfront can help reduce costs in the future.
- Second, consider what type of support is available to you if you have questions on the hardware or software. It can be both frustrating and time-consuming if your PLC vendor does not provide in-depth support in a timely manner. Find out what type of telephone support and local support you can leverage from the manufacturer. Be sure you are comfortable with your access to help if needed.
- Third, think about quality. Everyone wants to know that his or her system is reliable. Ask your vendor about the warranty. When vendors are confident about their quality, they stand behind it with lengthy warranties.
- And finally, consider cost. Today every company, large or small, is looking for ways to become more cost efficient. Consider the total cost of ownership for the PLC. You must always consider hardware and software costs, but think about ancillary costs. Are there annual software licensing fees? Are there annual technical support fees? What is the cost of training? All of these costs can greatly increase the total cost of ownership of a PLC.
What are your goals?
When selecting a control platform to automate a machine, you should first decide your end goals. Ask yourself:
- Do I want increased productivity? Better repeatability? More consistent accuracy?
- Do I need servo or stepper motors? Servomotors will generally have better acceleration and top speed characteristics over a stepper. Stepper motors do better when holding a position without dithering.
- What type of controller? Will a PLC work or do I need a more dedicated motion and machine controller? For basic material and conveyor control, a PLC is a good choice. If high-speed electronic gearing, product registration or more complex motion is required, a dedicated motion/machine controller is a better choice. If data handling is required, then a more dedicated motion/machine controller would do better.
- Will the controller do the job? Best to spec in a controller that can handle more axes of control than you expect as you may need to add these capabilities later. Be sure the controller has the processor power to perform all the needed functions. It's common to 'run out of gas' when performing multiple tasks simultaneously. Look for a fast processor, enough user memory and a lot of connectivity.
Liked this article? Download the entire playbook here