Influence of Grinding on Surface Roughness in Train Axle Journals
This page is for process engineers and grinding cell planners working on EA4T and similar axle steels. It walks through wheel selection, dressing, depth of cut and spark-out in the order they should be set, and shows where the influence of grinding on surface roughness stops being worth chasing.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
What actually moves the Ra number
How the influence of grinding on surface roughness actually works
A ground surface is the sum of two things: the groove pattern left by grit and the plastic flow pushed sideways by dull grit. Sharp grit cuts clean grooves, so Ra tracks the depth of those grooves. Dull grit ploughs, and the metal piles up at the edges instead of leaving the cut. That is why two wheels of the same size and the same machine settings can land 0.3 μm apart on Ra.
On an EA4T axle journal, the grit depth is set by three variables at once: infeed per pass, wheel surface speed and the effective grit protrusion after dressing. Cut deeper per pass and the groove gets deeper, but chip load per grit also rises, so the wheel can start shedding or loading. Cut too shallow and the grit only rubs.
The rubbing regime matters more than most shops expect. Once the specific cutting energy climbs past the point where the grit bites, the surface work-hardens and the Ra starts climbing again even though you reduced the nominal cut. You get a worse finish from a gentler setting. That is the counter-intuitive part of the influence of grinding on surface roughness.
Residual stress runs on a different track. Compressive stress at the surface is what you want on a rotating axle, and it comes from the mechanical action of sharp grit, not from heat. If the zone under the grit gets hot enough, the surface layer goes tensile and you can measure it with X-ray diffraction even when the Ra looks acceptable.
- 1Sharp grit cuts, dull grit rubsInspect the wheel face under a loupe at 10× before blaming the parameters.
- 2Shallow cuts are not automatically betterBelow about 0.01 mm per pass, ploughing and burn take over.
- 3Ra and stress are separate targetsYou can hit the Ra number and still leave a tensile layer.
Picking the wheel before you touch the parameters
For axle steels in the 300 to 350 HB range, a vitrified bond aluminium oxide wheel in the 46 to 60 grit band is the usual starting point. Grade J to L works for most journals. Harder grades hold form on long shafts but glaze faster on interrupted cuts. Softer grades stay open but lose size control over a long journal.
Grit size dominates the Ra ceiling. A 60 grit wheel can reach Ra 0.4 μm with proper spark-out. A 46 grit wheel usually stops around Ra 0.8 μm unless you spend a lot of spark-out time. If the drawing calls for Ra 0.2–0.8 μm and the journal is long, the 60 grit is the safer pick even though it loads sooner.
Porosity and bond type decide how much coolant reaches the contact zone. A more open structure runs cooler and holds Ra better on wide contact lengths, which matters when you grind a full axle journal on a cylindrical machine. Closed structure gives finer finishes on short, stiff parts but is more prone to burn.
Do not select the wheel from a catalog table alone. The same grade from two suppliers can behave differently after dressing because the abrasive grain shape differs. Run one test journal, measure Ra at three positions, and keep the supplier that holds the number across the length.
- 146–60 grit, J–L gradeStandard window for 300–350 HB axle steel journals.
- 2Open structure for long contactFull-journal grinding needs coolant to reach the arc.
- 3Test journal before full releaseOne shaft, three Ra readings, then commit the wheel.
Heat, burn and the stress that Ra does not show
Burn shows up in stages. First you get temper colours on the journal, then a mottled patch, then a dark band that you cannot polish out. By the time you see the dark band, the surface layer is already tempered and the residual stress has flipped to tensile. Ra may still read in spec, which is exactly why a roughness gauge alone is not enough for an axle.
Coolant has to reach the contact arc, not just the wheel. A nozzle aimed at the top of the wheel does almost nothing for the grinding zone. Direct the flow at the nip with a coherent jet, and check the flow rate at the nozzle, not at the pump. Most burn problems on long journals come down to a split or misaligned jet.
Wheel speed is the other lever. Higher surface speed thins the chip and lowers Ra, but it also pushes more heat into the part per unit time. On a 600 mm journal, going from 30 m/s to 45 m/s can drop Ra by 0.1 to 0.2 μm and still leave burn if the coolant does not keep up.
If the axle drawing requires a compressive surface layer, plan a stress measurement on the first article and again after any wheel or parameter change. A Nital etch on a scrap journal is a cheap daily check that catches burn before it reaches a finished part.
- 1Jet at the nipCoolant aimed at the wheel top does not cool the cutting zone.
- 2Watch the surface speed30–45 m/s is the usual working range for these steels.
- 3Etch a scrap piece dailyNital etch spots burn long before visual inspection does.
Measuring Ra so the number means something
Ra is an average, so it hides the peaks. On a ground axle journal, the peak-to-valley Rz often matters more for fatigue than the Ra value. Cut the cutoff length to the right value for the surface: 0.8 mm cutoff for Ra 0.2–0.8 μm finishes, 2.5 mm if the finish is coarser. A wrong cutoff gives you a number that does not match the drawing.
Take readings at three positions along the journal, at least 90° apart, and record the direction of the traverse relative to the grinding direction. A reading taken across the lay will differ from one taken along it, sometimes by 0.1 to 0.2 μm. Note the orientation on the inspection sheet so the next shift can repeat it.
Calibrate the roughness tester on a certified specimen before every shift. The stylus tip wears, and a worn tip reads low on fine finishes. If the shop has a cut-off filter setting stored for a coarser job, a quick operator change can leave it wrong for the axle job.
Roundness and taper belong in the same inspection report. Grinding can hit Ra while the journal is slightly tapered because the wheel wore over the length. Measure diameter at both ends of the journal and log the difference before you sign the part off.
- 1Report Rz with RaFatigue life is driven by peaks, not the average.
- 2Fix the cutoff to the finish0.8 mm for fine, 2.5 mm for coarse.
- 3Log traverse directionAcross-lay and along-lay readings differ.
Step-by-step grinding setup for a train axle journal
Run these in order. Changing two of them at once makes the Ra result impossible to read.
- 11. Mount and indicate the axleSupport on centres or steady rests, indicate the journal to within 0.02 mm TIR before grinding. A part that moves under the wheel will show taper no matter what you do to the parameters.
- 22. Balance and true the wheelBalance to grade G1 or better and true the face with a single-point diamond at 0.02 mm per pass, two passes. Skip this and the first journal will show chatter marks that no parameter change can remove.
- 33. Dress for the finish you wantFor Ra 0.4–0.8 μm, dress at 0.01 to 0.02 mm per pass with a 0.2 to 0.4 mm per rev cross feed. Coarser dressing cuts cooler; finer dressing lowers Ra but loads sooner.
- 44. Set wheel speed in the 30–45 m/s bandStart at 32 m/s for a first article. Raise in 3 m/s steps only if Ra is short of target and burn checks stay clean.
- 55. Set work speed and infeedWork speed 25 to 40 m/min for a 600 mm journal. Rough passes at 0.02 to 0.04 mm, finish passes at 0.01 to 0.015 mm. Never finish below 0.005 mm per pass, that is where rubbing starts.
- 66. Add spark-out passesTwo to four passes at zero infeed after the last cutting pass. This is the cheapest Ra gain on the whole setup and adds almost no heat.
- 77. Check the first articleMeasure Ra at three positions, Rz, taper and roundness. Etch a scrap piece for burn. If Ra is short, adjust one variable at a time and re-dress before the next trial.
- 88. Lock the proven recipeRecord wheel, dress, speeds, infeed and spark-out on a setup card. Re-dress at the interval proven on the first article, not at a default interval.
Parameter comparison for axle journal grinding
Values are starting points for EA4T and similar 300–350 HB axle steels. Adjust one line at a time.
| Parameter | Finishing window | If you go too low | If you go too high |
|---|---|---|---|
| Wheel speed | 30–45 m/s | Ra rises, wheel loads | Burn risk climbs fast |
| Work speed | 25–40 m/min | Heat builds in one spot | Chatter and spiral marks |
| Infeed per pass | 0.010–0.015 mm | Rubbing, work hardening | Deeper grooves, higher Ra |
| Spark-out passes | 2–4 passes | Ra stays above target | Cycle time with no gain |
| Dress depth | 0.010–0.020 mm | Glazed, dull grit | Fast wheel wear, poor form |
| Coolant aim | Jet at the nip | Burn and tensile layer | Splash, poor arc coverage |
| Ra target | Ra 0.2–0.8 μm | Not achievable in one setup | Needs finer grit and dress |
Set the wheel and coolant first, then chase the Ra number
If Ra is off target, fix wheel grade, dressing and coolant aim before you touch speeds and infeed. Parameter tuning alone cannot rescue a glazed wheel or a jet that misses the nip.
Questions engineers ask about axle grinding
Can we hit Ra 0.2 μm on an axle journal in one setup?
Rarely on a long journal. Ra 0.2 μm is achievable on short, stiff parts with a 60 grit wheel, fine dressing and four spark-out passes.
On a 600 mm journal, the wheel wears over the length and the taper limit usually forces you to accept Ra 0.4–0.8 μm as the practical target. If the drawing demands finer, plan a second light pass after a re-dress.
Why did Ra get worse after we reduced the depth of cut?
You dropped into the rubbing regime. Below about 0.01 mm per pass, the grit stops cutting and starts ploughing, which work-hardens the surface and raises Ra even though the nominal cut is smaller.
Go back to 0.01 to 0.015 mm per pass and take the finish from spark-out passes instead.
Does a good Ra number mean the residual stress is acceptable?
No. Ra and residual stress are set by different parts of the process. A glazed wheel with poor coolant can leave a tensile surface layer while Ra still reads inside the drawing limit.
Check burn with a Nital etch on a scrap journal and, where the drawing requires it, confirm stress by X-ray diffraction on the first article.
How often should the wheel be dressed on an axle run?
Set the interval from the first article, not from a default. Grind one journal, check Ra and taper at both ends, and note where the readings drift.
As a working start, re-dress after every two to three journals on 300–350 HB steel with a 60 grit wheel, then tighten the interval if taper or Ra drift appears.
What coolant flow should we run at the grinding zone?
Aim for a coherent jet that reaches the nip, not a wide spray over the wheel. Flow at the nozzle matters more than pump pressure, so check it there.
Most burn problems on long journals trace back to a split or misaligned jet rather than to low flow.
Can grinding replace turning on the whole axle?
No. Grinding is a finishing operation for the journal and seat surfaces where the tolerance and finish matter.
The body of the axle is normally turned or milled first. Grinding the whole shaft would cost far more cycle time and introduce taper risk for no functional gain.
Send the axle drawing and get a grinding plan back
Upload the journal drawing with the Ra and stress callouts. We reply with a quotation and a DFM note within 12 hours, and parts ship in 3–5 days once the setup is approved.
12-hour quote100% inspectionNDA on request
More grinding and machining notes
We publish setup notes, tooling trials and inspection data from the factory floor.
