Calibration FAQs: Vaisala Humidity Sensor Calibration, CO2 Sensors, and More

I coordinate rush calibrations for an independent instrumentation service company. In the last six years I've handled 200-plus rush orders—maybe 180, I'd have to check—including same-day turnarounds for pharma clients. When a Vaisala transmitter reads 3%RH off before an audit, I'm the one triaging the problem.

Here are the calibration questions I actually answer, plus one I wish more people asked. Jump to the equipment you care about.

• How often should Vaisala humidity sensors be calibrated?
• Can Vaisala humidity sensor calibration be done on site?
• How do I know if my co2 sensor Vaisala is drifting?
• What's the biggest mistake people make with an oil flow meter?
• Do multimeter probes affect calibration accuracy?
• How to calibrate pipette Eppendorf pipettes?
• Should I rush every calibration that's overdue?

1. How often should Vaisala humidity sensors be calibrated?

Short answer: it depends on your QA system and the sensor's exposure. In cleanroom monitoring, once a year is common. In process environments with chemical vapors, temperature swings, or near-saturated humidity, every six months can be necessary. Vaisala's HUMICAP sensors are known for long-term stability, but long-term is not forever. If you're asking because an auditor is already requesting certificates, that's probably too late.

From the outside, humidity calibration looks like putting a probe in a bag with salt solution and watching numbers. The reality is that a useful calibration needs a stable reference, a known temperature, and enough settling time for the sensor to reach equilibrium. If you only have one point, you can catch a gross failure, but you won't validate linearity.

If you send the instrument out for Vaisala humidity sensor calibration, ask whether the certificate includes measurement uncertainty and traceability. A number without a reference chain is just a guess.

2. Can Vaisala humidity sensor calibration be done on site?

Sometimes, but I'd call that a verification, not a full calibration. Put another way: on-site tells you the sensor is probably within tolerance; a lab calibration gives you a certificate with uncertainty and traceability. If your application only needs a spot check, an on-site reference probe is fine. If an auditor or a quality department asks for calibration history, they usually want something more solid.

I usually recommend sending the probe to a lab for a two-point or three-point calibration. Make sure the lab follows ISO/IEC 17025 if you need a defensible result. Oh, and if the probe has been exposed to silicone vapors or solvents, calibration may not fix it—the sensing element can be permanently damaged. That's a replacement, not an adjustment.

3. How do I know if my co2 sensor Vaisala is drifting?

A Vaisala CO2 sensor in a typical occupied space should settle near outdoor background when it gets fresh air. If it reads 600 ppm in a well-ventilated place, something is off. Some Vaisala models have automatic background calibration, or ABC, that uses a low point over time. That works in many buildings, but not if the sensor spends time in a room with high CO2 and never sees a true baseline. In process applications, use a certified CO2 test gas instead of relying on ABC.

For a formal calibration, the sensor is placed in a chamber or a bag with a certified gas mix and compared against a reference. That gives you the data to adjust or repair it. The drift tells you whether the sensing cell is aging normally or the environment is harsher than the maintenance schedule expected.

4. What's the biggest mistake people make with an oil flow meter?

Treating it like a 'dumb' pipe fitting. An oil flow meter—turbine, Coriolis, or positive displacement—has a calibration factor that shifts with viscosity, pressure, wear, and temperature. If you only zero it and never compare it against a known volume, you can be off by 1-2 percent without knowing. That sounds small, but on a transfer of 40,000 liters, 1 percent is 400 liters.

I remember a facility that saved $200 by skipping flow meter calibration for a year. The next audit found a 1.8 percent error, and the correction required recalibrating three meters and re-billing a supplier. The extra work cost around $900. The 'budget' choice looked smart until the consequences showed up.

5. Do multimeter probes affect calibration accuracy?

Yes, more than people think. If you are checking a 4-20 mA loop from a Vaisala transmitter, a test lead with cracked insulation or a loose connection can add resistance and noise. Worn multimeter probes can also bend or make poor contact on the terminal, giving you a tenth of a volt drop that has nothing to do with the instrument.

For rough troubleshooting, any probe works. For calibration and verification, use clean, quality multimeter probes with tight contact. I once chased a loop fault for an hour. Turned out the probe had a broken solder joint inside the banana plug. The transmitter was fine.

6. How to calibrate pipette Eppendorf pipettes?

The honest answer: unless you have a balance, a thermometer, and a controlled setup, you're estimating, not calibrating. The standard method for how to calibrate pipette Eppendorf-style pipettes is gravimetric: use distilled water, hold the temperature stable, dispense multiple volumes at the selected setting, and weigh each sample. Water density changes with temperature, so this matters.

According to ISO 8655, pipette calibration should be done with a balance sufficiently accurate for the test volume (Reference: ISO 8655-6). For a single-channel pipette, a proper calibration might include 10 to 20 weighings at each volume. That's not a 10-minute job. Should mention: if the pipette reads outside tolerance after calibration, the issue is often a lower seal or tip cone. Replacing a small part is cheaper than a full service.

7. Should I rush every calibration that's overdue?

No. Rush fees are easy to justify when they prevent a shutdown, but they only help if the rush is the bottleneck. If you still have to clear a production line, hire a technician, and wait for a reference standard, paying for same-day turnaround doesn't save time. I have mixed feelings about rush fees: they feel like a penalty, but after seeing the inside of a lab, I understand why they cost more. The real fix is buffer time.

What I calculate now is total cost: calibration quote, shipping, rush fee, and the cost if the certificate gets rejected. The lowest quote is still expensive if the result irritates an auditor and you have to do it again. Schedule calibration like a flight: build in waiting time. That saves more than any discount vendor.