Tactics and Strategies for Updating Obsolete Systems

I once had a customer ask me for a National Instruments DIO module to replace one that had gone bad. The problem was that the module had not been in production for over 10 years. It was an ISA bus module. I didn't even really know what ISA was. This system was old and had been working correctly, so no updates had been made. If it ain't broke then don't fix it. The problem is they were unable to fix it when it did break. Because they had to completely start over, there was a high cost to recreate the test system.

Endigit presented a webinar on strategies for updating obsolete systems. Below is a quick summary. Check out the full presentation for more detail.

The first step to updating an obsolete system is recognizing there is a need to update.There are several red flags that indicate it is time to do an update. These red flags can be with software, hardware, or operating systems. It is important to review the red flags and have a plan to remove or mitigate the risk of the red flags.

Once you have decided to update a system there are hardware design decisions that need to be considered.

  1. Evaluate the true needs of your system.
    What were the original requirements for the system? Is there anything it should be doing that wasn't possible when it was first made? 
  2. Don't try to match instrumentation specifications. 
    Select hardware that best matches what you need to accomplish. Chances are high that the instrumentation available now is better than was available at the time of the original design. There also could be instruments that are more geared for your application but may be less optimal on a different, less relevant specification. I have seen people not choose an instrument that was better for their application because there was one thing the old instrument did better. This is a mistake.
  3. Don't redesign around your old workarounds.
    During development of the old system there were likely some design decisions you made to solve a problem that there was not a clear solution for at the time. You maybe had to create a workaround because sampling rates were not high enough. It is important to recognize those workarounds and not design them into the new system. Identify what your system actually needs and design around that. 
  4. Don't select tools and hardware based only on cost.
    There are a lot of options for engineering solutions. I have seen people who select a tool or part because it can mostly do what they want and costs less than an option that can do everything needed. The cost and headache of engineering around something that is "cheaper" far outweighs the cost of purchasing the right tool. If you need a motor that has 5000 ft-lbs of torque then don't get one that almost has that much torque but costs less than the one that does.
  5. Don't try to force equipment to be what it isn't.
    If you need equipment that performs a specific function then get the tool that does that function. Don't try to use a general purpose analog input card for measuring thermocouples. Measuring thermocouples requires cold junction compensation. Just because an analog input module can have a small enough range to measure the signal doesn't mean that it will be accurate. Get the equipment that is right for your needs. You can use a screwdriver to pound in a nail, but a hammer will get the job done a lot better.
  6. Don't try to save money by using equipment you have laying around the shop.
    There can be a temptation to use things you have laying around to try to save money. Unless the equipment is in good condition, up-to-date, and it is the right equipment for your system then don't use it. If it is just laying around the shop then chances are good that it is old too. Replacing something obsolete with something else that is or will be obsolete soon does not make any sense. Get the right equipment for your needs. 
  7. Use commercial off the shelf (COTS) products as much as possible.
    COTS equipment is easy to maintain. If there are problems you can get repairs or replacements.
  8. Select tools and interfaces that will last.
    When updating a system it is important to choose technologies that will be supportable and available. 
  9. If applicable, create a simulation system.
    If you are replacing a complex control system that needs to continue running then tearing the old system out and throwing the new system in untested could lead to disastrous results. Creating a system that simulates the physical signals or software behavior of what you need to control allows for development without stopping the current system from functioning. It also allows for more complete testing in situations that could be dangerous. 
  10. Create great documentation.
    This may be the most recommended and least implemented strategy in engineering. Writing documentation can feel like a burden. Understanding what someone else did and why is a greater burden. Save yourself and others the future headache by taking the time to document your systems.

Endigit has experience updating obsolete systems. We recently completed a project at Idaho National Laboratory (INL). The system we updated was running Assembly code on 8087 processors. The new system is using new COTS equipment with up to date software. We have a plan in place to evaluate yearly and make updates before things become a problem. 

We would love to help you update your old systems. If you have a system in need then let us know on our contact page and we will help get your systems updated and maintainable.

 

Bryan Heslop

Vice President - Sales and Marketing, Staff Systems Engineer

Bryan first discovered he was a megageek in a college physics class when he got way too excited about tesla coils (most people figured it out much sooner). Bryan worked as a machinist programming and operating CNC lathes while attending Brigham Young University. He earned a degree in Computer Engineering with a Business Management minor in 2008. After graduation Bryan moved to Austin, Texas to work at National Instruments. He worked as an Applications Engineer in the Engineering Leadership Program for 2 years then spent 1 year in Albuquerque, New Mexico as a Field Sales Engineer. He then moved back home to Utah and worked as District Sales Manager for National Instruments where he managed NI’s large accounts in Utah like Hill Air Force Base, Boeing, Northrop Grumman, OrbitalATK and L-3.

At Endigit Bryan is responsible for sales and marketing. As a Certified LabVIEW Architect, he also works as a Systems Engineer doing LabVIEW development.

Bryan enjoys everything about water (boating, swimming, rock skipping, drinking, etc.), riding ATVs, comedy writing, exploring the mountains around his home and dreams of going to space.  He also serves as a volunteer firefighter and emergency medical responder in his community.

Add new comment