The Grumpy Old Tech's Guide to Instrumentation (Wisdom You Won't Find in a Manual)

 The Grumpy Old Tech's Guide to Instrumentation (Wisdom You Won't Find in a Manual)

Welcome back, survivors. If you've made it through eleven articles of this series, you've earned the right to call yourself a true instrumentation nerd. You've wrestled with pressure transmitters, argued with digital buses, tuned PID loops until your eyes bled, and explained to operators why their pH probe isn't actually possessed by demons.

But now we're doing something different. We're putting away the textbooks, the datasheets, and the calibration procedures. We're sitting down with the grey-haired, sun-beaten, grease-stained veterans who've been doing this since before HART was a thing. The techs who can diagnose a stuck valve by the sound it makes. The engineers who've seen every possible failure mode—and a few impossible ones.

This is the collected wisdom of the craft. The shortcuts. The workarounds. The "we tried that in '98 and it nearly blew up the plant" stories. The advice that isn't in any manual, but should be.

Consider this your mentorship session. Grab a notepad. Or better yet, just remember it. Because this is the stuff that separates the competent from the truly excellent.

1. If it ain't broke, don't calibrate it.

I know, I know—the schedule says annual calibration. But if the transmitter has been stable for three years, and the process is running fine, and the "as-found" data is always within spec... stop wasting your time. Calibration is not a religious ritual. It's a diagnostic tool. If the instrument is working, leave it alone. You're more likely to introduce errors by messing with it than you are to find a problem. The best calibration is the one you don't do.

2. The most expensive tool in your kit is your ears.

Listen to the plant. A healthy control valve has a smooth, consistent hiss. A sticky one has a stuttering, chattering sound. A pump with bad bearings has a high-pitched whine. A leaking steam trap has a distinct gurgle. You can diagnose more problems by walking the plant and listening than by staring at trends on a screen. Your ears are free, they never need batteries, and they work in any weather. Use them.

3. Never trust a cable you can't see both ends of.

Cable faults are the most common, most frustrating, and most time-consuming problems in instrumentation. A single nick in a cable, a single loose terminal, a single corroded ground—and you've got a ghost in the machine. The fix? Label everything. Use proper cable management. And if you're troubleshooting a mystery fault, start at the instrument and work your way back to the control room, checking every connection along the way. The problem is almost never at the ends—it's always in the middle.

4. Water is the enemy. Water is always the enemy.

I said this in Article 1, and I'll say it again: moisture kills instruments. It corrodes terminals, it shorts out electronics, it freezes in impulse lines, and it makes calibration impossible. If you're installing a junction box, seal it properly. If you're running cable, use drip loops. If you're in a humid environment, use desiccants. The best instrument in the world is useless if water gets inside it. Keep it dry. Keep it sealed. Keep it alive.

5. The manual is a suggestion, not a commandment.

I'm not saying ignore the manual. It's a good starting point. But the manual was written by an engineer in a clean office, based on ideal conditions. Your plant is not ideal. Your process is dirty, your power is noisy, and your operators are... let's say "creative." Sometimes, you need to deviate from the manual. A different calibration point. A modified installation. A workaround that the manual doesn't mention. That's fine. Just document it. And for the love of all that's holy, tell the next shift.

6. If the operator says the instrument is wrong, they're probably right.

Operators aren't stupid. They may not know the technical details, but they know the process. They know what it sounds like, what it smells like, and what it feels like. If they say the pressure reading is wrong, it's probably wrong. Listen to them. Work with them. They're your eyes and ears on the process when you're not there. Treat them with respect, and they'll save your bacon more times than you can count.

7. Spare parts are not optional.

You will need them. You will need them at 3 AM on a Sunday during a blizzard. If you don't have spares, you'll be down for days. If you do have spares, you'll be down for hours. It's that simple. Keep a stock of the most common failure items: transmitters, positioners, I/P converters, fuses, relays, and cable. And for the love of all that's holy, keep them organized. There's nothing worse than having a spare transmitter buried under a pile of junk.

8. Always check the simple stuff first.

Is the power on? Is the loop closed? Is the valve open? Is the line plugged? Is the manual override engaged? Is the operator actually running the right recipe? These sound like stupid questions. But I've seen highly paid engineers spend four hours troubleshooting a transmitter that was simply turned off. Start simple. Work your way up. The obvious answer is usually the right one.

9. Document everything. Then document it again.

If it's not written down, it didn't happen. Calibration records. As-found data. As-left data. Maintenance logs. Fault reports. Modification requests. The paper trail is your protection. It's your evidence that you did the right thing. It's your record of what worked and what didn't. And when the next guy comes along—and there will be a next guy—they'll thank you for leaving clear documentation. Or they'll curse your name. Your choice.

10. Don't be afraid to ask for help.

Instrumentation is a broad field. No one knows everything. If you're stuck on a problem, ask a colleague. Ask the vendor. Ask the operator. Ask the internet (but verify everything you read). There's no shame in not knowing. There is shame in pretending you know and making things worse. The smartest people I know are the ones who are constantly asking questions.

11. Calibrate your calibrator.

You use your handheld calibrator to calibrate transmitters. But when was the last time the calibrator itself was calibrated? A drift of 0.1% on a 20 mA loop is significant. If your calibrator is out of spec, your calibration is out of spec. Send it out for annual certification. And keep a backup. Because calibrators, like all instruments, can fail.

12. Always zero the transmitter at operating temperature.

If you zero a pressure transmitter cold, and then the process heats up to 100°C, the zero will drift. The housing expands, the diaphragm changes stress, and the reading shifts. This is called "thermal zero shift." To avoid it, let the transmitter warm up to process temperature before you zero it. Or use a transmitter with good temperature compensation. But the real trick? Just let it warm up. Patience saves a lot of headaches.

13. The best tool is a sharp pencil and a clean notepad.

Fancy analyzers and diagnostic software are great. But sometimes, the simplest solution is to write down what you see. Draw a sketch. Make a timeline. List the symptoms. Write down what you tried and what happened. There's something about putting pen to paper that clarifies your thinking. Your notepad is your most powerful troubleshooting tool. Use it.

14. Test, don't guess.

If you're not sure what's wrong, don't start replacing parts at random. Test. Apply a known pressure. Simulate a temperature. Inject a 4-20mA signal. Isolate the problem. Eliminate variables. The scientific method is not just for labs. It's for the plant floor. Guesswork leads to wasted time and wasted parts. Testing leads to solutions.

15. Always keep a spare tube of anti-seize compound.

And a spare roll of Teflon tape. And a spare set of o-rings. And a spare bottle of thread sealant. The little things matter. A stuck fitting can turn a 10-minute job into a 2-hour ordeal. A leaky thread can bring the whole plant down. Preventative maintenance isn't just about big things. It's about the little things that keep the big things running smoothly.

16. Your multimeter is your best friend. Treat it well.

Check the battery. Check the fuses. Check the leads. A dead multimeter is a useless multimeter. And a multimeter with blown fuses is dangerous. Keep a spare set of leads. Keep a spare set of fuses. And for the love of all that's holy, always check the voltage before you check the resistance. You don't want to blow up your meter because you forgot the loop was live.

17. If it's working, don't fix it.

This is the corollary to Rule #1. I've seen techs "improve" a perfectly working system and break it. If the plant is running fine, and the instrument is within spec, and the operators are happy—walk away. Don't tweak. Don't adjust. Don't "optimize." Leave it alone. You'll thank me later.

18. Never trust a single source of information.

Your DCS says the pressure is 100 PSI. Your handheld calibrator says 102 PSI. Your gauge on the pipe says 101 PSI. Who's right? You need a third reference. A second gauge. A different calibrator. A different transmitter. Always cross-validate. The truth is somewhere in the middle. The only thing worse than a wrong reading is a wrong reading you believe is right.

19. Treat your operators like humans, not obstacles.

I said it in Article 10, and I'll say it again: operators are your partners, not your enemies. They're the ones who have to live with your instruments day in and day out. If they don't trust them, they'll override them, bypass them, and eventually break them. Build trust. Explain what you're doing. Show them the data. Listen to their concerns. A happy operator is a safe operator. And a safe operator is a plant that stays running.

20. At the end of the day, it's all about reliability.

The best instrument in the world is the one that never fails. The best calibration is the one that never drifts. The best design is the one that works for years without any intervention. Reliability is the holy grail. It doesn't matter how fancy the instrument is if it breaks once a month. It doesn't matter how accurate the calibration is if it drifts every week. Simplicity, robustness, and repeatability—that's what matters. The rest is just marketing.

And there you have it. Twenty years of hard-won wisdom, distilled into a single article.

Instrumentation is a craft. It's a blend of science, art, and intuition. It's about understanding the physics, mastering the tools, and—most importantly—learning to listen. Listen to the plant. Listen to the operators. Listen to the instruments themselves. They'll tell you what's wrong, if you're paying attention.

The industry is changing. Digitalization, AI, IIoT—they're all transforming the field. But the fundamentals remain the same. The laws of physics haven't changed. The need for reliability hasn't changed. The importance of human judgment hasn't changed.

So keep learning. Keep questioning. Keep your tools sharp and your mind sharper. And never, ever forget that the most important instrument in the plant is the person holding the screwdriver.

---

That's a wrap on The Instrumentation Insider series.

We've covered:

1. 20 Burning Questions (The Basics)

2. Calibration, Logic, and Gremlins

3. The Digital Shift (Fieldbus, Profibus, Wireless)

4. Control Loop Tuning (PID and Beyond)

5. Analyzers (pH, Conductivity, GCs)

6. Vibration and Motion (Machine Health)

7. Weighing and Force (Load Cells, Belt Scales)

8. Valves and Actuators (The Muscle)

9. Safety Systems (ESD, SIS, HIPPS)

10. The Human Factor (Operators, HMIs)

11. The Future (IIoT, Digital Twins, AI)

12. The Grumpy Old Tech's Wisdom

I hope you've enjoyed the ride, learned a few things, and maybe even laughed a little.

Now go out there and make your plant a better, safer, more reliable place. The instruments are counting on you.

Stay safe. Stay curious. And never stop learning.