The alignment step you're probably skipping

Most alignment jobs look good when you're done.

The numbers are tight. Everything's dialed in. You torque the bolts, document the readings, and move on to the next job.

Then a week later, the vibration alarms start going off.

Here's what happened: you skipped the post-torque soft foot check.

I've seen this pattern on all types of coupled rotating equipment—compressors, pumps, turbines, gearboxes. Single-train setups, multi-train configurations. Any driver and driven equipment connected with a coupling. You align everything perfectly, torque the hold-down bolts, and assume you're done.

But torque changes everything.

When you tighten the hold-down bolts, you're applying force to the base if there's any soft foot condition—angular or parallel—that forces the frame to distort.

The movement isn't always obvious. Sometimes it's vertical. Sometimes horizontal. Sometimes angular. It depends on the severity of the soft foot and the field conditions—thermal growth, induced pipe strain, and base stiffness.

But here's what I typically see: when I recheck after torque, the readings have moved. Not always by much—sometimes 0.0005" to 0.001", which is acceptable. But when there's a real problem, I've seen movement over 0.020" or higher.

That's not a rounding error. That's a misaligned machine that's going to eat bearings and seals.

If you torque bolts randomly instead of using the proper sequence (opposite corners, cross-pattern), you're making the problem worse. Uneven torque creates uneven loading, which distorts the frame even more.

This is why I always follow OEM specifications for bolt torque and sequence. It's not optional. If you don't have OEM specs, use a bolt sizing torque chart. Guessing doesn't work.

Always use stainless steel pre-cut shims or OEM-patented shims. I've seen people use mild steel or whatever's in the shop, and it corrodes. Once that happens, your readings are unreliable because the frame is sitting on corroded material that's compressing unevenly.

Use the right shims. Keep them clean. No debris, no burrs between shims and surfaces.

Here's the procedure I follow every time:

Use the proper sequence (cross-pattern, opposite corners). Make sure every bolt is at the specified torque before you start checking.

With all other bolts tight, slightly loosen one bolt—just enough to break tension, not enough to visibly lift the machine. Watch for movement. If the foot "springs" or lifts, you've got a soft foot.

After loosening the bolt, insert a feeler gauge (0.002" or 0.003") between the foot and the base at all four corners. If the gauge slides in easily, there's a gap. Measure the largest thickness that fits.

If the gap is more than 0.002", you need to correct it.

Add shims equal to the measured gap thickness. Retighten the bolt to spec. Move to the next foot. Repeat for all four feet.

5. Confirm repeatability and shim stack integrity. After correcting all feet, torque everything again and recheck. Make sure you didn't introduce new gaps by correcting the previous ones.

Use as few shims as possible—typically no more than 3-5 shims per foot. Too many thin shims in a stack can compress unevenly and create instability.

Always shim for full-foot contact. Only use stepped shims when you have an angular soft foot that requires correction—otherwise, full-foot shims prevent frame distortion.

I've responded to failures where post-torque movement was the root cause. Seals fail first. Then bearings. Couplings usually fail last because they have larger tolerances.

How long until failure shows up? Depends on the equipment. Higher RPM means faster failure. Sometimes it's weeks. Sometimes months. But it's coming.

The post-torque check takes 10-15 minutes. It catches problems before they become unplanned shutdowns.

I accept no more than 0.0005" to 0.001" of movement after torque. If I see more than that, I'm correcting it before I walk away.

Some people say this is overkill. Maybe. But I've also seen too many seal and bearing failures that could've been prevented with one extra check.

Check after torque. Confirm your numbers. Document both readings—before and after.

That's the job.

What's your process? Do you recheck after torque, or do you rely on the initial readings?

See you next week.