If you've ever accepted a delivery and then discovered the instrument wasn't performing to spec, you know the particular kind of headache that causes. The vendor says “it's calibrated, we guarantee it.” You sign off. Three weeks later, a validation run fails, and you're stuck with the cost and the delays.

I'm a quality compliance manager for an analytical testing company. I review roughly 150 items each year before they go into service—Shimadzu HPLC systems, portable X-ray machines, clamp-on flow meters, RF spectrum analyzers, multimeters, you name it. I've rejected about 8% of first deliveries in 2024 for a few recurring reasons: documentation mismatches, performance drift, or simply missing accessories.

Here's the five-step checklist I use for every instrument that comes through my lab. If you're the person responsible for accepting instruments at your facility, take it and adapt it.

Step 1: Lock down acceptance criteria before you order

This is the step that saves the most pain downstream, and it's also the one most of us rush through. The conventional wisdom is that your vendor's standard specification sheet is enough. My experience with 200+ orders suggests otherwise—the standard spec sheet tells you what the instrument can do under ideal conditions, not what it must do for your application.

Before I even issue a PO, I define, and write into the purchase order:

  • Critical performance parameters. For a Shimadzu Prominence HPLC system, that's pump pressure stability, retention time reproducibility, and detector baseline noise.
  • Minimum thresholds. We set retention time reproducibility at ≤1% RSD across six injections, based on the system suitability guidance in USP Chapter <621>.
  • Required documentation. Calibration certificate, QC report, and the exact firmware and software version.

The third time we received an instrument where the actual performance didn't match what had been discussed on a phone call, I finally added a clause that ties final acceptance to documented performance verification. Should have done that from day one.

Step 2: Check the physical delivery before the driver leaves

This sounds basic, but you'd be surprised how often it gets skipped. We didn't have a formal receiving inspection at every location, and it cost us when a $22,000 instrument arrived with a damaged detector cell—we didn't find it until the vendor's installer had already left.

For every instrument, regardless of vendor, I verify:

  • Serial number on the unit matches the packing list and the PO
  • All accessories, cables, and consumables listed on the packing list are present
  • Physical condition: dents, scratches, fluid leaks, bent connectors
  • Factory calibration certificate matches the serial number

It takes about ten minutes, and it's saved us way more than ten minutes of headaches. Twice this year alone, the serial number on the instrument didn't match the certificate. Both were caught during receiving instead of during validation.

Step 3: Run your own verification against reference standards

Here's the step most people ignore: the vendor's calibration certificate is a useful starting point, but it's not proof that the instrument works in your environment. Basically, a calibration cert says the instrument was accurate at the vendor's facility, under their conditions, with their operator. That's not the same as being accurate in your lab. I say this from experience. In my first year, I made the classic newbie mistake of approving an instrument based entirely on the vendor's factory calibration certificate. It cost us a $600 revalidation when the unit failed our internal check two weeks later.

For a Shimadzu Prominence HPLC system—or any Shimadzu HPLC, since the workflow is the same—I run a system suitability test with our own reference standard mix. I use a certified reference mixture that covers our main analytes, and I run it under the exact method parameters we'll use for routine tests. Same column, same mobile phase, same flow rate. That way, the verification conditions match real operating conditions, not the vendor's ideal conditions. It's roughly 30 minutes of run time—or rather, block out closer to an hour once you account for equilibration and setup.

  • Pump pressure stable after a 20-minute equilibration
  • Retention time RSD ≤ 1% across six injections
  • Column efficiency above the minimum theoretical plates spec
  • Detector baseline noise below our SOP threshold

For a Shimadzu portable X-ray machine, I use our reference thickness samples to confirm measured values fall within our acceptance criterion of ±2% of the known standard. I don't run the vendor's demo file—I use something my team can trace back to a certified reference material.

The same principle applies to non-Shimadzu instruments in our lab:

  • Clamp-on flow meters: verify the zero reading on a static pipe, then compare against a known flow rate from a calibrated reference meter.
  • RF spectrum analyzers: check frequency accuracy against a signal generator with NIST-traceable calibration.
  • Multimeters—technicians love debating HIOKI vs Fluke—I check DC voltage accuracy against a known reference source before approving the unit.

The point isn't to distrust the vendor. The point is to establish a baseline that belongs to you. If something drifts six months from now, you need a reference point from the day the instrument entered service.

Step 4: Repeatability is the real test

Single-run verification tells you the instrument works—that once. It doesn't tell you it's stable. I didn't fully understand this until a supplier sent a demo unit that performed beautifully on the first run and then degraded substantially by the third. The issue turned out to be a thermal management problem, but we only caught it because we repeated the test.

For any instrument that will be used in routine analysis, I run the verification at least three times, ideally on separate days. If you're pressed for time, at minimum:

  • HPLC: three consecutive injections of the same standard
  • Portable X-ray: five exposures on the same reference sample
  • Clamp-on flow meter: repeated zero checks over 30 minutes to observe drift
  • RF spectrum analyzer: repeated frequency measurements to check reported accuracy

If I remember correctly, we caught a multimeter drift issue two years ago by repeating a measurement sequence that had passed on the first run. Nothing dramatic—a few millivolts—but enough to trigger an investigation and reveal a faulty reference circuit. The vendor replaced it without argument once we showed the data.

Step 5: Create the baseline document before you release the instrument

This step feels like paperwork, but it's the difference between owning an instrument and renting a problem. The baseline document becomes the point of comparison for every future calibration, service event, or repair.

Here's what I include:

  • Instrument serial number, firmware version, and software version (check the software at acceptance; it's a pain to verify after the fact)
  • Verification results from Steps 3 and 4
  • The specific reference standards used, including their calibration expiry dates
  • Any deviations from the acceptance criteria, even small ones
  • Photos of the instrument, its connections, and any accessories

For example, when our Shimadzu Prominence HPLC systems were installed in early 2024, each one has a baseline folder on our shared drive containing the system suitability data, the software version, and the installation report. If a column pressure issue comes up next week, we can check it against the baseline and know immediately whether something changed. This also helps during audits. When an auditor asks how we know an instrument was performing within spec when the data was generated, the baseline document is the answer.

Common mistakes I still see

If you take nothing else from this, remember these three:

  1. Don't assume “calibrated” means “fit for your purpose.” Calibration is a point-in-time measurement. Verification is an ongoing conversation.
  2. Don't skip the receiving inspection for trusted vendors. The vendor's quality usually isn't the issue—shipping, handling, and customs are. A 30-second serial number check can prevent a cascade of problems.
  3. Don't skip software version checks. The instrument can physically perform perfectly, but if the software version doesn't match what your IT security team approved, you'll be redoing validation later. It happened to us, and the revalidation cost far more than the ten minutes it would have taken to check.

Every instrument's first day tells you what the relationship with that vendor will look like. The ones that pass through this checklist smoothly are the ones we keep. At least, that's been my experience with labs running routine compliance testing. Because in analytical services, your instrument park is part of your brand. Clients may not know the difference between a Prominence and another HPLC platform, but they know when your data is consistent and defensible. When it isn't, the questions start with the instrument and end with your reputation. The checklist isn't about making the vendor's life harder—it's about protecting your lab, your data, and your name. Take it from someone who has reviewed over 600 instruments at this point in their career.