My first PLC replacement took place in 1979 with a machine control manufacturer. The PLC was made by Reliance. We were developing a new generation of control for this company. Reliance wanted our business so badly, that they made a prototype at their cost with our application programmed into their PLC. They flew us out in their private jet to Cleveland. The hardware was impressive but the software was understandably a little shaky.

The control consisted of:

• 32 Optically isolated Digital Inputs
• 48 Optically isolated short circuit proof 24 VDC Outputs
• 2 RS-422 Serial Interfaces (1 for the HMI and one for Networking)
• 4 Optional Analog Inputs (0-5 VDC)
• 2 Optional Analog Outputs (4-20 ma)

I was pushing for developing a custom control for two reasons:

1. The custom control would cost about $200 per unit less ($1,000 versus $1,200)
2. The $60,000 development cost would be quickly recovered since we were selling about 300 units per year.
3. We would have control of our destiny - not be dependent upon the next release of Reliance's hardware or software. Or their pricing! We were concerned that the price they were giving us was a teaser. It was not what we could buy the unit for "off the shelf."

After shipping over 6,000 of these, this custom control system is still being manufactured. Even though it was replaced by a newer design in 2000, customers who have scores of these units, still want the older units for obvious reasons (experience base, known reliability, etc).

What would have happened if we took the Reliance alternative? Reliance made a business decision to obsolete the 1979 version PLC sometime in the 80's. Their PLC line merged with Allen Bradley and Rockwell. The hardware and software would have been scrapped and redesigned a number of times since 1979 (at least 3 times that I know of).

A company was using a custom embedded control system which was built upon a poorly designed hardware platform using our well designed software. The hardware was creating significant problems for their customers. After several years of floundering, management decided to go to a PLC manufacturer and replace the embedded control with a more reliable proven design.

The control consisted of:

• 48 Digital Inputs
• 64 Digital Outputs
• 10 Axis DC Servo Controls
• 1 Ethernet network connection
• 1 RS-422 HMI

It has taken the PLC company more than 24000 man-hours to replace the existing custom control software using the high level PLC language (Structured Text or ST). They are replacing code that was custom developed for 6000 man-hours. They are still not done - but are shipping "production units." Although the hardware is much better, the software is a mess and still not done.

I am convinced that this is going to be a failed effort in the long term. The PLC software infrastructure had to change to meet the custom demands. The PLC manufacture will not be able to sustain this level of customization over the long haul. The ST language does not support modern software engineering paradigms that ease re-use and lower costs of software development.

Check in again in two years. We will provide an honest assessment.

A small machine manufacture came to us with a request to replace an existing PLC control with:

• 6 Digital Inputs
• 8 Digital Outputs
• 4 Analog Inputs (PID loops)
• 1 Touch Screen Interface

They ship about 20 per month of these systems

Their PLC hardware costs were running about $1100 from Automation Direct (www.automationdirect.com)