Analog to digital; now, digital back to analog
When replacing analog with digital, ensure all the costs are considered. In some cases, perhaps remaking analog systems would be better.
Many of us grew up in an analog world: Super heterodyne radios, simple radar, even simpler TV sets. Remember taking the back of the TV set off, being careful not to touch that high-voltage wire from the flyback transformer to the aquadag coating connector on the cathode ray tube (CRT)? You were quickly reminded if you did touch it. Then you unplugged the vacuum tubes and ran over to the TV repair shop to use the tube tester. Most repairs cost less than $20.
It was a world of simple amplifiers, without solid-state devices. Instrumentation was either mechanically driven or pneumatically controlled.
Those old mechanics, they were geniuses. The ability to control a steam boiler, mechanically, was amazing. Steel rods were attached to the fuel valve, the air damper and fan, and the water level control, all controlling that boiler.
Oh sure, we had been making mechanical computers since Charley Babbage in 1833-and the earliest computers using electromechanical devices, like Alan Turing’s Universal Machine of 1931, and like Konrad Zuse’s machine of 1939, which used a 22-bit word. But these were still mechanical analog devices.
Then electronic circuits replaced the mechanical and electromechanical machines. Tommy Flowers, John Atanasoff, and Cliff Berry produced what was called the first automatic electronic digital computer in 1942, still using 300 vacuum tubes. Paper tape programming using Boolean logic, created by Max Newman and Flowers, built the Colossus, which broke the German Enigma encryption machine in 1944. Eniac came next, weighing 30 tons and using 200 kW.
Along came the transistor in 1947. Wow! So we used that little triode (NPN, negative-positive-negative, or PNP, positive-negative-positive) to perform analog functions. It was so cool back then because it was so incredibly easy to understand. We had been using diodes, triodes, tetrodes, pentodes, etc., and vacuum tubes for all our analog devices. Now we were reproducing these same devices in a solid-state form. Very cool!
Then along came an integrated circuit in 1958, created by Jack Kilby and Robert Noyce, combining these solid-state devices into a smaller package. It was a little harder to understand, yet now we had two amplifier circuits in a very small package. Okay, we can handle that.
Pocket calculators began replacing slide rules, progressing from 4-bit machines to 8 bit machines, then 16-, then 32-, then 64-bit machines. So now almost everything we use, from our watches to our airplanes, has digital controls.
Unfortunately, digital controls have several severe drawbacks.
1) Digital controls are very sensitive to electromagnetic frequency (EMF), electromagnetic interference (EMI), and electromagnetic pulse (EMP) radiation.
2) Cyber security must be used to protect digital controls.
Military hardening strengthens digital controls to withstand EMF, EMI, and EMP radiation, but cyber security for digital assets is another matter. So much of our power grid is run with digital assets, and as such, it is susceptible to cyber security attacks.
Our power plants are mostly run with digital assets, and they too are susceptible to cyber security attacks.
What can we do? Imagine going to three electronic manufacturing firms and obtaining bids to reverse engineer existing analog devices, which are currently no longer being manufactured. Along with that bid, ask the suppliers to manufacture the analog devices AGAIN.
But why would we do that? To regress into the past?
Compared costs: New and old
A recent installation of a protection system was initially bid at $25 million to complete all three power plants. After 10 years of verification and validation, at a cost of almost $250 million, the installation was completed.
The standards that our government regulators establish are there to protect and maintain the given infrastructure, be it a power plant or our national grid.
Looking at the baseline costs, the reverse engineering and production costs for analog might cost $10 million for the identical protection systems used in a plant now, so why spend an additional $15 million for digital? The plant would not be buying something different by staying analog; it would be buying like-for-like equipment. The standards would still be met, but at a greatly reduced cost.
Also, the existing training would be identical. Troubleshooting and maintenance would be identical. So to finalize the project, remaking older analog technologies would be better, cheaper, quicker, and more compliant than replacing analog with digital.
– Nick Negoescu is the owner of N-Squared Inc. Edited by Brittany Merchut and Mark T. Hoske, CFE Media, Control Engineering, email@example.com.
Usually when digital replaces analog, other benefits are factored into the return on investment; are other costs factored in, as well? Leave your comments with the online posting of this article, www.controleng.com/archives under September.
More about the author
Nick Negoescu, owner of N-Squared Inc., has owned or operated several security firms. He has worked in power generation for over 34 years and in nuclear power for over 18 years. Negoescu has contracted as a project manager, a start-up engineer, a procedure writer, and a work scheduler/planner. His present contract is with D&Z/NPS at an Exelon Nuclear Plant in Maryland.