Drilling engineering article header
Drilling insight

Overhauling a Liebherr G9508 Gearbox: The 9-Step Checklist I Use Before Sign-Off

Posted on Friday 13th of November 2026 by Reiko Matsuda

Last October, a maintenance team handed me a rebuilt Liebherr G9508 gearbox for sign-off. On paper, everything matched. On the bench, the input bearing had 0.04 mm of axial play against a spec of 0.01 mm. We rejected it. It went back for a second pass at the rebuilder's cost, and the machine sat idle for another nine days.

That story isn't unusual. The failure modes on G9508 rebuilds are almost never a bad part — they're a skipped verification step. I'm a quality and brand compliance manager at a mining equipment company, and I review every component overhaul before it goes back into service — roughly 60 a year, 200-plus over the past four years. In 2024 I rejected about 18% of first submissions. The reasons repeat.

This is the checklist I hand to every rebuilder before the unit comes off the machine. Nine steps. None of them are optional, and at least three are routinely skipped.

When this checklist applies

Use it when a Liebherr G9508 gearbox (or a comparable crawler crane / excavator final drive unit) is coming up on an overhaul rather than a straight replacement. That usually means 8,000–12,000 running hours, oil analysis trending upward on iron content, or a noise complaint that isn't going away.

It also works for the pre-purchase inspection of a used unit — I've used a shortened version of it to talk myself out of two used G9508 assemblies that looked clean on the outside.

The 9-step overhaul checklist

Step 1: Confirm model, serial, and variant before you touch anything

Photograph the nameplate. Cross-check the serial suffix against the Liebherr parts documentation for the G9508 variant you're actually working on — the same base designation ships with different gear sets depending on the installation. I've seen a rebuild stalled for two weeks because someone ordered bearings for the wrong variant. Ten minutes of verification saves a fortune.

Step 2: Pull the oil sample cold — before draining

This is the step that separates a diagnostic overhaul from a guess. Take the sample from the main drain and the magnetic plug before you flush anything. Iron content tells you whether wear is accelerating or whether this is just a long-life unit with normal accumulation. If you skip this, you have no baseline, and the rebuild becomes a parts swap with a story attached.

Step 3: Record baseline measurements before disassembly

Input and output axial and radial play. Rotate the gear train by hand and note any binding or backlash that feels out of character. It takes about ten minutes and creates the reference point every later decision depends on. Honestly, more than half the shops I've worked with go straight to teardown.

Step 4: Label and photograph during disassembly (this is the one people skip)

Gear sets go back together in a specific orientation, and some of them look symmetrical until they don't fit. I mark the non-working tooth flanks with a paint pen, keep a numbered parts tray, and take a photo at every sub-assembly step. The most expensive mistake I've personally witnessed was a gear set installed backwards because it "looked the same either way." The unit ran for six hours before it started whining.

Step 5: Inspect the gears — don't automatically replace them

Look for pitting, micro-pitting, tip breakage, and root cracks. Use a magnifier at the root fillet. If there's a crack, replace the gear set — do not attempt repair or blending. If it's surface wear only, measure the tooth thickness against spec before you make a call. Replacement is expensive; unnecessary replacement is just expensive and slow.

Step 6: Check the housing, not just the internals

Cracks hide under ribs and around bearing bores. Do a surface check — dye penetrant or magnetic particle, depending on the material. A housing with an undetected crack takes bolt preload unevenly, and the rebuild fails months later for reasons nobody traces back to the housing. Most small rebuild shops skip this entirely. It's a deal-breaker for me.

Step 7: Verify bearing clearances and outer-ring fits

Measure the new bearings — don't assume they're in spec because they came in a box. I've been burned by a batch where the outer-ring bore was 0.015 mm over, which would have produced the same axial play we rejected last October. Hold the rebuilder to documented measurements at install, not just at purchase.

Step 8: Replace every seal and gasket — no exceptions

Seals are the cheapest part in the gearbox and the most common point of failure. This is not a place to save. If a rebuilder tells you the original seals "look fine," that's a red flag about how the rest of the job is being run.

Step 9: Bench test and re-sample before it goes back in the machine

Run the unit unloaded, check for noise and temperature rise, and pull a second oil sample after 30–60 minutes of running. The sample should be clean. If iron content is still climbing in a fresh fill, stop and find out why before the unit re-enters service. This step costs an afternoon. A premature failure costs a production week.

Common mistakes (and what I do instead)

Sampling hot oil instead of cold. Hot oil hides the settled metal. Pull cold, or don't bother.

Reusing seals "because they still look good." They aren't.

Using non-OEM or unverified bearings. I have mixed feelings here. On one hand, some aftermarket bearings are genuinely fine and cost 40% less. On the other, I have no reliable way to tell which ones are fine — the sourcing is opaque and batch quality moves around. I've never fully understood why some no-name bearings outlast OEM while others fail inside 2,000 hours. My best guess is batch-level process control. Until I can verify it, we buy traceable parts and eat the cost.

Assuming the torque wrench is calibrated. We had a rebuild lose a cover bolt at 3,000 hours because the shop's torque wrench was out of calibration and nobody checked. Calibrated tools and a documented cross-tightening sequence matter as much as the parts.

One more thing worth saying: what was considered a "good rebuild" five years ago often isn't accepted today. The fundamentals — clean parts, correct clearances, no shortcuts — haven't changed. What has changed is the documentation expectation. Torque logs, oil-sample history, bearing certificates. The rebuilders who embraced that shift have shorter turnarounds and fewer comebacks. The ones still working from memory are the ones whose units show up on my bench.

Eight thousand hours or two rebuilds is my personal threshold on a G9508 — beyond that, replacement starts to look cheaper than another overhaul, especially if the housing has already been through two heat cycles. That's not a rule for everyone. But it's the line I use when somebody asks me to sign off on a third pass.

Reiko Matsuda

Reiko Matsuda

Reiko Matsuda specializes in PSA nitrogen generators, oxygen plants, liquid-nitrogen storage, gas purification skids, compressors, and distribution equipment for mining and energy operations. Her reviews apply ISO 8573-1 feed-air classes while measuring gas purity, dew point, delivery pressure, flow stability, specific power, recovery rate, storage autonomy, and contamination risk. She helps process engineers and purchasers specify gas systems for inerting, oxidation control, instrument service, enhanced recovery, emergency supply, and dependable operation at remote sites.

Leave a Reply