An edited thread from a machine-tool retrofit project, the axis that feeds a transfer line. The debate took three weeks and settled on a tooth profile nobody had wanted to start with.
From: R. Almeida, Design — To: H. Weiss, Supplier Applications; cc purchasing
Subject: spline joint RFQ — 45 mm drive shaft, 850 N·m with pulls
Hans, we are replacing the drive coupling on the cross-feed axis of the transfer machine, and the existing rectangular spline is the confirmed failure point, fretting wear on the teeth, visible peaning on the flank after roughly eighteen months of a two-shift cycle. Photo of the worn shaft attached, you can see the contact band is hard against the tip of the tooth on the load flank and there is stepped wear at the root, which reads to me as the hub rocking under load reversal. Duty: 850 N·m steady at the coupling, plus a pull of about 2100 N·m for 0.4 s when the axis decelerates into the indexing stop, roughly four times a minute. Shaft 45 mm, hub bore to suit, speed 300 rpm max, ambient in the machine envelope around 35 to 40 C, splash oil present but not a pressure feed. We would like one joint that lives at least five years on this cycle, and we do not care much about the profile as long as we can get a drawing and a heat-treat spec we can verify on the machine.
From: H. Weiss, Supplier Applications — To: R. Almeida; cc purchasing
Subject: RE: spline joint RFQ — two ways to read the failure
Ricardo, thanks, clear brief. Before I quote, one question that changes the whole answer: is the rocking you see on the tooth from load reversal, or from missing face width? A rectangular spline that is short on engagement width picks up clearance and then the first few millimeters of tooth carry everything, which looks exactly like the photo, tip-loaded, root-stepped. If the machine had been cutting engagement face width to save on the broach, we fix it with the same profile and more width. If it was truly reversal shock, a rectangular spline talks back at you noisily but rarely frets the way yours did.
Interim recommendation so we do not lose a week: measure the clamping bore runout and the tooth engagement length on the old hub, and tell me whether the fretting band was centered or biased to one end of the tooth. That single observation separates a sizing problem from a profile problem, and it also decides whether we quote the DIN 5480 involute style with 30 deg pressure angle, or stay with rectangular and extend the hub. I have attached a quick comparison sheet, but the sheet is generic; your measurement is not.
From: R. Almeida, Design — To: H. Weiss; cc purchasing
Subject: RE: spline joint RFQ — measurements, you win
Measured over the weekend, numbers below. Engagement face width on the old hub: 28 mm on a drawing that called for 34, so someone had already cut the hub down once, probably to clear an interference with the gear case flange. Fret band: biased hard toward the outer end of the tooth, the last 6 mm of engagement, none at the inner end. Clamping bore runout on the hub: 0.07 mm, which is not great and probably explains why the wear was asymmetric around the circumference, it wore worst on the side opposite the radial keyway. So your read is right, it is a sizing and misuse problem, not the profile itself, and the machine is not giving us the width back without a different hub. That is where we are stuck: the axis envelope is fixed.
From: H. Weiss, Supplier Applications — To: R. Almeida; cc purchasing
Subject: RE: spline joint RFQ — the width is gone, so the profile has to earn it
Then we are on the same page, the constraint is: same 45 mm shaft, same hub length that fits the envelope, roughly 28 mm max of useful engagement unless we move the gear case, and we need to carry 850 N·m steady and the 2100 N·m pull, on a joint that fretted at those numbers before. Options, and I want your reaction before I burn hours quoting both:
[Option A — rectangular spline, current size, case hardened] DIN 5463 rectangular, 8 teeth, 36 x 42 x 8, hub face width 28 mm, teeth case hardened 58-62 HRC to 0.6-0.9 mm effective depth, flank ground. At 850 N·m steady this is about 62% of the allowable flank pressure for the size, comfortable. At 2100 N·m pull it jumps to roughly 115% of allowable, which is over, and that is exactly the regime that fretted your old joint, the old one was soft, 40 HRC max at the flank after grinding away the case.
[Option B — DIN 5480 involute, same envelope] DIN 5480 W 45 x 2 x 22, pressure angle 30 deg, hub face width 28 mm, case hardened and crowned flanks, taper crown 0.01 mm. Involute profile gives you a longer effective contact path per tooth and the 30 deg pressure angle pushes the contact load deeper into the tooth than the 45 deg flank of a rectangular spline, so for the same face width it carries roughly 25-30% more torque before the flank pressure limit. At 850 N·m steady it sits near 48% of allowable, and even at the 2100 N·m pull it lands under 90%, inside the envelope I am comfortable with for a cyclic pull.
Cost difference to you is small, the price driver is the case hardening and the grinder on the flank, which both options already pay for. The involute needs a hobbing or generating step we can do, and the rectangular keeps the broach you already own. My honest vote is B.
From: L. Petrov, Purchasing — To: H. Weiss, R. Almeida; cc quality
Subject: RE: spline joint RFQ — the price is not the issue, the lead time is
Before anyone books a gear-hobbing step, check the shelf: invoked DIN 5480 hubs and shafts in the 45 series are standard at several distributors, and the broach we own only feeds the rectangular family, so if we stay rectangular the tooling is paid and dead stock, if we go involute we are buying hub blanks from stock and the shaft becomes the only custom part. Lead time difference is about two weeks in our favor on B, because the shaft is one op and the hub is off the shelf. My only commercial ask: confirm the flank grinding tolerance class and the tooth side clearance class on the drawing now, not after approval, because the inspector who signs the certificate will ask for both.
From: M. Delgado, Quality — To: R. Almeida; cc H. Weiss, L. Petrov
Subject: RE: spline joint RFQ — what we can actually verify
Adding the verification side before you commit, because a spline is one of those features that looks simple on the drawing and is a fight to measure. For the DIN 5480 involute we will check the base tangent length over two teeth with a caliper over pins, and the tooth thickness with a gear tooth vernier on the reference circle, tolerance that we can hold and report is around 0.03 mm on tooth thickness for hardened ground flanks, if you order a class that tight, we fight it. Recommended reading for our shop: DIN 5480 class 7 with side clearance class A5H, which gives clearance around 0.06 to 0.12 mm at the pitch circle for this module, loose enough for the axis to hunt under reversal without binding, tight enough that the 2100 N·m pull does not rock the hub visibly. If you want tight, we can do A6h at about 0.02 to 0.06, but then the hub and shaft are a matched set and a field hub swap becomes scrap on the spot. I would rather you tell the customer they get a serviced matched pair than a field-replaceable joint.
One more: the fretting you had was partly misalignment. Whatever profile we pick, the drawing should hold the shaft-to-hub bore runout at 0.03 mm TIR and the flange face square to the axis within 0.02 mm, and we verify the assembled joint with a dial on the hub face before it leaves the bench. That is the measurement we can stand behind in an audit.
From: R. Almeida, Design — To: H. Weiss, L. Petrov, M. Delgado
Subject: RE: spline joint RFQ — deciding, and why
Decision after the calls today: Option B, DIN 5480 W 45 x 2 x 22, class 7, side clearance A5H, case hardened 58-62 HRC effective depth 0.6-0.9 on the flank, crowned 0.01 mm, and the hub from stock with the shaft as the only custom op. The reasons, so nobody has to rediscover them next year. First, the envelope kills Option A, 28 mm of engagement with a rectangular profile was already the fretting machine, and repeating the same profile with a case to 62 HRC only buys time, the flank pressure at the pull is the same geometry problem. Second, the involute earns a real margin at the pull, under 90% of allowable instead of over 100, and that margin is what the axis needs four times a minute. Third, the off-the-shelf hub cuts two weeks, and quality can certify class 7 with the equipment they already run.
| Parameter | Option A rectangular | Option B involute 30 deg |
|---|---|---|
| Standard | DIN 5463 8x36x42x8 | DIN 5480 W 45 x 2 x 22 |
| Face width | 28 mm | 28 mm |
| Flank pressure @ 850 N·m | ~62% allowable | ~48% allowable |
| Flank pressure @ 2100 N·m | ~115% (over) | ~90% (inside) |
| Hardness | 58-62 HRC, ground | 58-62 HRC, crowned ground |
| Lead time | longer (shaft + retained broach) | shorter (stock hub) |
From: H. Weiss, Supplier Applications — To: R. Almeida; cc purchasing
Subject: RE: spline joint — order confirmed, drawing number attached
Order confirmed, drawing DW-5480-45-22-A5H. Shaft: DIN 5480 W 45 x 2 x 22 x 28, case hardened, crowned flanks, first article with a full report from quality on base tangent length, tooth thickness and cumulative pitch error over the whole tooth count. Hub: stock 45-series DIN 5480, one off the shelf, dimensional report on the bore and the flank. We ship the matched pair with a note on the box that this is a matched pair, field swaps require a quality interval, per Mercedes.’s point about A6h secrecy, we stayed A5H so the pair is replaceable as a set with a normal order. Expected on your dock in 11 days.
Reference data — applied torque 850 N·m steady, 2100 N·m pull 0.4 s x4/min; shaft 45 mm, engagement 28 mm; DIN 5480 W 45 x 2 x 22, class 7, clearance A5H; case 58-62 HRC, crowned 0.01 mm; original failure fretted in 18 months, engagement cut 34 to 28, flank soft 40 HRC.
Glossary — involute spline: a spline whose tooth flanks are involute curves, giving rolling contact and a 30 deg pressure angle; rectangular spline: a spline with straight parallel flanks, DIN 5463 series; case hardening: surface hardening to a shallow effective depth that leaves a tough core; base tangent length: the length measured over two teeth, used to verify tooth thickness; side clearance class: the designed clearance between mating flanks, A5H being a normal running class.
Field note — the question that settles the profile
When a spline fretted, the first question is never the tooth profile, it is the engagement width and the misalignment, because a joint half the width it was drawn with carries twice the flank pressure and sits on a hub that rocks. Measure the face width actually engaged, look at where the wear band sits, tip or root, inner or outer end, and check bore runout on the old hub before anyone re-quotes the profile. If the width is honest and the wear is still tip-loaded under reversal shock, then the profile earns its place, and an involute with a 30 deg pressure angle reaches deeper into the tooth for the same face width, which is exactly the margin a cyclic pull needs. Keep the flank hardness, the clearance class, and the matched-pair note on the drawing; those three lines decide whether the next engineer fixes the joint or just reorders it.
