Three Failure Scenarios, Not One Recipe
Here is the thing I tell every plant manager who calls me when a machine starts acting up: troubleshooting a measurement system is not one workflow. It's three workflows, and the mistake is using the wrong one. When I walk into a facility with a drifting Rice Lake load cell, a testo 174H data logger that will not upload, or a centrifuge 5804 shaking on a bench, I don't grab the same meter. I first decide which scenario I'm in.
The reason this matters is simple: a load cell is a sensor, a data logger is a data collection system, and a centrifuge is a rotating machine. They fail differently, and they get fixed differently. If you treat a logger like a load cell, you will replace a good battery and still have gaps in the data. If you treat a centrifuge like a sensor, you will miss a loose rotor lid.
I don't have hard data on industry-wide failure rates. I can tell you from about 200 emergency site calls that a surprising number of 'bad load cell' reports turn out to be wiring, mounting, or environmental problems. But that is a memory, not a database. Take it as a starting point.
Scenario 1: Your Rice Lake Load Cell Drifts, Won't Zero, or Changes With Temperature
Start with a simple question: does the failure happen all the time, or only at certain times? A reading that slowly drifts upward after 9:00 AM is not the same problem as a reading that jumps randomly.
1. Mechanical first
Check the mounts before you check the cell. Check the bolts. Check for anything leaning against the vessel or the load cell. Then put a multimeter on the signal wires and watch the mV/V output while you push down on the structure. If the signal moves smoothly, the cell is likely alive. If it jumps around, look at the cable and junction box first. A loose shield wire will do exactly that.
2. Electrical second
Measure the excitation voltage at the load cell itself, not at the display. I once spent an hour on a scale that 'died' because the junction box was delivering 6 V instead of 10 V. Oh, and don't trust the color codes. Measure at the connector. Rice Lake's technical data sheets list the expected zero balance and bridge resistance values; use those as your pass/fail threshold, not the display's calibration.
3. Environment third
Temperature is the hidden variable. If the zero drifts after sunrise, place a testo 174H data logger on the frame and record temperature and relative humidity for 48 hours. I once saw a load cell output follow the steam pipe schedule almost minute for minute. The sensor was fine. The pipe was heating the mount. A data logger gave us the proof, and the fix was insulation, not a new load cell. My experience with 174H units is mostly in HVAC and storage environments, not in deep freezers or ovens. If your application is outside the published range, treat the environment reading as a clue, not a verdict.
4. Vibration fourth
Actually, let me correct myself: mechanical, electrical, environment, vibration. Vibration deserves its own check. If there are pumps, fans, or a centrifuge 5804 near the scale, mount a vibration sensor to the scale base and compare readings with the equipment off and on. A modest rise in RMS velocity can make a sensitive scale unstable. Replacing the load cell will not fix that.
I have mixed feelings about telling people to buy a vibration sensor for a single scale. If you have one load cell, a mechanic's hand-held meter is probably enough. But if you are running a plant with a rack of testers, a cheap vibration sensor pays for itself after one false replacement.
Only if the cell fails all of those checks should you suspect the sensor itself. The surprise in these calls is almost never the sensor. It is the mounting screw, the loose junction-box terminal, or a floor plate that flexes. That is why I always quote the two hours of testing before the part purchase.
Scenario 2: Your Testo 174H Data Logger Has Gaps or Won't Upload
Here the problem is not a weight reading. The logger is either logging or supposed to be logging, but the data isn't there when you need it. The troubleshooting path is completely different.
1. Battery and connector pins
Check the battery first—not the software, not the calibration. The 174H runs on a battery, and a dying battery can create partial writes that look like corrupted files. Actually, check the battery first, then check the base station connector pins. A bent pin makes intermittent contact and acts like a dead logger.
2. Testo login and device assignment
Your testo login is the first thing to check if the data is not appearing in the portal. If you are using Testo Cloud, log in and confirm the 174H is still assigned to your account. I have seen a logger stop uploading after an admin changed a user's permissions. That was not hardware. It was a simple permission issue.
3. Know the logger's limits
Check the published operating range for the 174H. Mounting it in a freezer that dips below the rated minimum can cause gaps. The logger may still look active, but the sensor is outside its design range. If you are missing data during off-hours, ask whether your protocol actually requires continuous logging. Not every gap is a failure.
Scenario 3: Your Centrifuge 5804 Vibrates, Makes Noise, or Aborts Runs
This is a different world. A centrifuge is a rotating mass, not a static sensor. The most common causes of vibration are physical: an unbalanced load, a rotor that isn't seated, or a worn part. Calibration does not fix an imbalance.
1. Stop it and check the rotor seat
On an Eppendorf 5804, the rotor lid needs to engage cleanly. If the lid is cross-threaded, the whole assembly runs rough. Remove the rotor, inspect the motor shaft and the hub, and reinstall the rotor. Per the operating manual, the rotor must be locked and the lid closed before starting. That seems obvious, but cross-threaded lids are common.
2. Weigh your tubes
Eyeballing is not balancing. Weigh each tube and each bucket, then load pairs with the same mass. The centrifuge's imbalance sensor can catch big problems, but it will not catch a small mismatch that makes a high-speed rotor wear faster.
3. Run an empty spin and, if you have one, use a vibration sensor
An empty spin tells you a lot. If the machine runs smooth with no rotor but shakes with the rotor, suspect the rotor itself. If it shakes with no rotor, look at the motor mount and the bench. I have mixed feelings about vibration sensors on lab equipment, but in a busy lab, a small vibration sensor mounted to the base can detect bearing wear before the machine aborts a run.
4. Do not adjust the imbalance detector
The detector is a safety device. If it trips, it is telling you something. Bypassing it or changing its threshold turns a diagnostic clue into a hazard. Replace the worn part.
I've only worked with 5804 units and similar lab centrifuges, not large industrial centrifuges. If this is a floor-standing machine with a heavy rotor, call the manufacturer. The principles are similar, but the risks are much bigger.
How to Decide Which Scenario You're In
Ask one question first: is the problem a value, a log, or a vibration?
- If a weight reading drifts, changes, or won't zero, start in Scenario 1. That is Rice Lake load cell territory.
- If you are missing recorded temperature or humidity data, or a logger won't upload, start in Scenario 2. That is testo 174H territory.
- If the symptom is noise, shaking, or aborted runs from a spinning device, start in Scenario 3. That is the centrifuge 5804 territory.
There is overlap. If a centrifuge 5804 sits on the same bench as a Rice Lake load cell or any precision scale, and the scale readings fluctuate whenever the centrifuge starts, treat that as a Scenario 1 environmental issue. Move the scale, stabilize the bench, or use a vibration sensor to prove the path. Do not 'fix' a load cell that was only picking up vibration.
The fastest thing you can do on a failed measurement system is to identify the scenario before touching hardware. The second fastest thing is to change one variable at a time. Put another way: troubleshooting is not about owning the most expensive test gear. It's about not skipping the obvious checks.
One last caveat: if you are dealing with a legal-for-trade scale or a high-speed centrifuge, this article is for basic field checks, not recertification. Stop when local rules require a certified service ticket. The parts are not worth a safety issue, and no blog post should replace a manufacturer's official repair procedure.
If you find yourself in a scramble, that is usually when mistakes happen. Slow down, decide which scenario you are in, and check the simple things first. Not ideal, but workable—and far cheaper than replacing the wrong part.
