# On-Site Balancing of Driveshafts and Propeller Shafts > A driveshaft doesn't behave like an impeller. It has two universal joints that on their own generate vibration at the second harmonic — twice the rotational frequency — a splined joint with alignment marks, a thin-walled tube, and often a centre support bearing partway along. That's why on a driveshaft we check the assembly and clearances first and bring out the instrument second. If imbalance is confirmed, we balance in two planes at the ends of the tube and hand over a report with before-and-after numbers. **In short:** Yes, we balance driveshafts and propeller shafts on site, assembled on the machine, without removal. Four conditions apply: the shaft is assembled to its alignment marks, there's no play in the U-joints or splines, the shaft reaches stable running speed, and the running-speed component 1x — vibration at the rotational frequency, which is what imbalance produces — dominates the vibration. We correct in two planes, at the ends of the tube near the yokes. If the vibration sits at the second harmonic from the working angles, or if there's play in the joints, we'll say so before any trial runs: what's needed there isn't a weight but correct assembly and repair. Source: https://axiline.pt/en/equipment/on-site-balancing-driveshafts-propeller-shafts/ Publisher: AXILINE · Vila Nova de Gaia, Portugal · +351 931 831 229 · axilinegeral@gmail.com ## Symptoms: how a driveshaft specifically vibrates A driveshaft gives itself away by having vibration tied to speed rather than load. The drive runs smoothly, then over a certain part of the speed range a hum and a slight tremor through the frame appear, calming down again above and below that range. A narrow range is already a clue: it's either resonance in the supports and brackets, or the shaft has approached its bending critical speed — the speed at which it starts to bend noticeably. A second typical scenario shows up after a repair. A U-joint cross was replaced, the slip joint was taken apart and reassembled, a new yoke was welded on, or the tube was replaced. The shaft went back on the machine, and vibration got worse than it was before the work. On these call-outs we more often find assembly out of alignment mark rather than imbalance. - [x] Hum and tremor through the frame or drive housing appear in a narrow speed range and disappear above and below it. - [x] Vibration increased right after replacing the U-joint crosses, taking apart the splined joint, or repairing the tube. - [x] A clean patch with no paint or rust is visible on the tube: that's where a factory balance plate used to sit before it came off. - [x] The centre-bearing bracket creaks, the bolts keep coming loose again, the rubber cushion of the support bearing is worn out. - [x] The U-joint crosses run hot, grease squeezes out from under the needle-bearing seals. - [x] A clunk is audible on start-off, reversal or throttle release, and on coast-down the shaft visibly runs out. - [x] The shaft tube is caked in mud, bitumen, ice or wound-on material, and the layer sits unevenly. - [x] The splined joint's seal is torn, and grease has been squeezed out of the slip joint. > What correction weights actually reduce is covered in our article on overall vibration, 1x and phase. For a driveshaft this rule applies more strictly than usual: it has its own inherent source of second-harmonic vibration, and it has nothing to do with mass distribution. ## Construction: where imbalance comes from on a driveshaft Imbalance on a driveshaft is almost always induced from outside. The factory balanced the shaft and welded on the plates, and from there repair, impact, corrosion and caked-on deposits take over. Let's go through it by component, because the component determines how we'll fix it. ### Thin-walled tube Most of the shaft's mass is the tube, with a wall usually 1.5 to 3 mm thick. It's flexible, easily dented by a stone, and deforms under careless slinging. A dent, a weld build-up, or localized corrosion shifts the centre of mass at the maximum radius, so the effect ends up noticeable. ### Yokes, U-joint crosses and flanges The yokes are welded to the tube, and the U-joint crosses sit in the yokes on needle bearings. A new cross differs in mass from the old one, and a welded-on yoke almost never sits in exactly the same position. Flange bolts, locking plates and caps also count toward the balance. ### Splined joint The slip joint compensates for length changes from suspension travel and thermal expansion. Both of its halves carry relative-position alignment marks. Assemble it out of alignment, and the yokes at the two ends of the shaft end up rotated relative to each other. ### Factory balance plates Thin steel plates are tack-welded to the tube near its ends. A plate comes off from corrosion or impact, and the shaft loses its balance instantly. Look for a bare patch under the paint and remnants of weld tacks: that's direct evidence. ### Centre support bearing On long drivelines the shaft is split into two sections, with a support bearing on a bracket with a rubber cushion between them. The support introduces a third point and its own stiffness. A worn-out cushion generates vibration on its own, and a worn bearing makes the balancing result unstable. ### Foreign mass on the tube Mud, bitumen, ice, wound-on film, straw, a broken-off piece of cable. The layer sits unevenly, holds poorly, and changes the imbalance from shift to shift. There's no point balancing such a shaft: the correction will disappear along with the piece that falls off. > A shaft that was originally balanced doesn't go out of balance on its own. So during the inspection we look for an event: what was done to the shaft, what came off it, and what got stuck onto it. ## Angles, yoke phasing, and the second harmonic A universal joint working at an angle rotates non-uniformly. The driving yoke turns at constant speed, the driven one speeds up and slows down twice per revolution. The larger the working angle, the stronger the effect. In the spectrum you see not the running-speed component but the second harmonic, and correction weights don't remove it. The designer cancels this non-uniformity out with the second joint. For that, the yokes at the two ends of the shaft must lie in the same plane, and the input and output angles must be equal. Then the second joint gives back what the first one took away, and the output runs smoothly. Assembly out of alignment mark breaks exactly this compensation. Rotate the slip-joint halves relative to each other, and the yokes stop being in phase. The non-uniformity from the two joints no longer cancels but adds up. Vibration appears without any imbalance at all, grows with speed, and sits at the second harmonic. There's only one fix here: disassemble and reassemble to the marks. We also check the angles separately. The engine mounts have sagged, the unit's seating has changed, the centre-support bracket has shifted, and the angles are no longer equal. Compensation becomes incomplete. That's a matter of restoring the components' fit, not a job for weights. - 1x dominates and the second harmonic is small: that's imbalance, and balancing is appropriate. - 2x dominates and grows with speed: check the yoke phasing and the working angles. - 2x together with noticeable axial vibration at the connected units: that's shaft misalignment, and shaft alignment is needed. - A raised noise floor and high-frequency peaks: U-joint crosses, needle bearings, the support bearing. - A narrow peak at a specific speed with a fast phase shift: resonance, or the shaft approaching its bending critical speed. > Spectrum interpretation and the signs of resonance are covered in separate articles. For a driveshaft, keep one thing in mind: the second harmonic on a driveshaft is more often geometry and assembly, not mass distribution. Sources: [ISO 13373-3:2015](https://www.iso.org/standard/40840.html) ## What we check before balancing: alignment marks and play first, weights after This order isn't a formality — it's a condition for the method to work at all. The correction calculation relies on the system responding to added mass predictably and repeatably. Play in a U-joint cross or in the splines makes it nonlinear: the clearance gets taken up differently on every start, the phase drifts, and two identical starts give different numbers. You can't get an influence coefficient — the shaft's response to a trial weight, on which the whole calculation is built — from data like that. Next, the alignment marks. If the shaft is assembled incorrectly, the vibration sits at the second harmonic, and you'd chase it with weights all day for zero result. Checking the marks line up takes minutes, and it saves the whole visit. That's why we start with a wrench and a dial indicator, and bring out the trial weight last. And safety. A weight on a worn driveshaft doesn't restore remaining life — it only masks the vibration level. A worn-out U-joint cross on a tube spinning under the machine is a risk of the shaft coming apart at speed. We don't balance a shaft like that, and we say plainly that it needs a repair. - [x] The marks on both halves of the splined joint line up, and the yokes at the ends of the shaft lie in the same plane. - [x] Radial and angular play of each U-joint cross, checked by hand-turning the yokes in both directions. - [x] Circumferential play in the slip-joint splines, seal condition, presence of grease. - [x] Flange bolt tightness, condition of the locking plates, signs of flange slippage. - [x] Tube and flange runout with a dial indicator at several cross-sections, checked at slow rotation. - [x] The centre support's bearing, its rubber cushion, bracket tightness and cracks. - [x] The joints' working angles, sag in the engine and unit mounts, the relative position of components. - [x] Presence of all factory balance plates, and signs of any that are missing. - [x] The tube's cleanliness along its whole length, including the area near the yokes. - [x] Alignment of the connected units. Misalignment is fixed by shaft alignment, not balancing. > Play is a contraindication, not just an inconvenience. Until the clearances are eliminated, any calculated correction will turn out to be random, and on the next start the level will drift right back. Sources: [ISO 281:2007](https://www.iso.org/standard/38102.html) ## How the on-site work proceeds 1. **Inspection and assembly check** — We check the splined joint's marks, look for play in the U-joint crosses and splines, check for balance plates, flange tightness, and the condition of the centre support. We check tube and flange runout with a dial indicator at slow rotation. We decide here, before turning on the instrument, whether to proceed. 2. **Sensor setup** — A driveshaft has no bearing supports of its own, so we mount the accelerometers on the bearing housings of the units at the ends of the shaft, and on the centre-support bracket if there is one. We clean the mounting spots, fasten rigidly, and measure radially, perpendicular to the shaft axis. We stick the reflective tape onto a flange or the tube, for the laser phase sensor. 3. **Baseline measurement** — We bring the drive up to running speed and take the overall level, the running-speed component, phase, speed and spectrum. We calculate 1x's share of the overall level and separately check the second harmonic: it decides whether this is a balancing job or an assembly issue. 4. **Choosing planes and the attachment method** — We determine where mass can physically be placed: the tube near the ends by the yokes, the flange joint, the locations of the factory plates. We agree with you where welding is acceptable and where it isn't, and confirm that the tube won't be drilled. 5. **Trial weight in the first plane** — We place a weighed mass at a known radius, fasten it as securely as a permanent one, and run the shaft. We consider a response of at least 20-30% in amplitude or 20-30° in phase valid. If the response is weak, we increase the mass and repeat, rather than calculate a correction from noise. 6. **Trial weight in the second plane** — We move the weight to the other end of the tube and take another run. The software obtains influence coefficients for both planes and calculates the correction mass and angle for each. 7. **Correction and confirmation run** — We attach the correction masses at the stated positions, remove the trial weight, and take a confirmation run. If there's a small gap remaining to the target, we add small trim masses, usually within a single attempt. 8. **Post-work measurement and report** — We repeat the measurement at the same points, same direction, same operating point. We record the before-and-after numbers, the masses fitted, radii and positions. We save the influence coefficients so the next trim balance on this machine goes ahead without trial runs. > We work with the Balanset-1A instrument: two vibration channels, a laser phase sensor reading the reflective tape, a laptop with software, single- and two-plane balancing, fixed-position mode, and tolerance calculation by grade G. On a flange joint, fixed-position mode is especially convenient: instead of an angle and a protractor, you get a hole number and a mass. Sources: [Balanset-1A operation manual](https://vibromera.eu/balanset-1a-operation-manual/) ## How many planes, and where they are on a driveshaft A driveshaft is an elongated rotor. The length-to-tube-diameter ratio is usually above ten, so a single mass won't solve the problem: it would lower the level at one end and raise it at the other. We balance in two planes, at the ends of the tube near the yokes. That's where the radius is largest, that's where the factory puts its plates, and that's where access is least often a problem. We don't treat a two-section shaft with a centre support as a single rotor. Each section has its own ends and its own support, so we work section by section and watch how the correction on one affects the neighbouring support. A three-support configuration behaves like a multi-support rotor, and there's a separate article on that logic. Length sets a limit on speed. The bending critical speed falls as tube length increases, so on a long shaft the working range runs right up against it. Such a shaft can no longer be treated as rigid: it bends at speed, and a correction found at one operating point works worse at another. We balance at the actual running speed, and when the operating range is wide, we take data at several speeds. | Shaft | Configuration | Correction planes | | --- | --- | --- | | Short driveshaft between units, up to 1,500 rpm | Single span, flanges at both ends | Two: at the ends of the tube or at the flanges | | Drive shaft, up to 3,000 rpm | Single span with a splined joint | Two: on the tube at both yokes | | Two-section shaft with a centre support | Three supports, support bearing in the middle | Two per section, working section by section | | PTO shaft on machinery | Large working angles, speed fluctuates | Two, but angles and yoke phasing come first | | Long transmission shaft | Flexible rotor, running speed close to critical | Two, measurements at several speeds | > The rule itself for choosing the number of planes is covered in a separate article. For a driveshaft the conclusion is short: two planes almost always, and two channels are needed precisely so you can see both ends at once. Sources: [ISO 21940-12:2016](https://www.iso.org/standard/50429.html) · [ISO 21940-11:2016](https://www.iso.org/standard/54074.html) ## Attaching weights on a thin-walled tube A driveshaft tube forgives less than a drum's end disc. The wall is thin, speeds are high, and any stress concentrator on the tube eventually becomes a crack. That's why we choose the attachment method to suit the construction, not for the convenience of fitting. The standard method mirrors the factory one: a thin steel plate, bent to match the tube's radius, seated flush over its whole area and tacked with short weld points near the end of the tube, close to the yoke weld. We weld with low current and short passes, so as not to burn through the wall or anneal the metal around it. - We place mass on the tube near its ends, close to the yokes, where possible at the same radius where the factory plates sat. - We bend the plate to the tube and tack it with spot welds. We don't run a continuous circumferential weld: that's a ready-made stress concentrator. - We don't weld over the yoke's weld seam or in its heat-affected zone. - We don't drill the tube. A hole in the thin wall is not acceptable at a driveshaft's running speed. - We don't weld near the splined joint or the seals: heat ruins the grease and the dust boot. - We remove metal only from the flange, and only where thickness allows it, with your agreement. - We clamp the welder's return lead to the same part as the electrode. Current passing through the U-joint crosses, needle bearings and unit bearings destroys them. - As an alternative to welding: a clamp-on weight on the tube, or balance washers and eccentric bolts at the flange joint. - We fasten the trial weight just as securely as a permanent one. At these speeds a mass that comes loose becomes a projectile, so we guard the rotating zone and clear people away. > We agree the attachment method before the trial runs, not after. If welding on your shaft isn't allowed at all, and the flanges don't give the needed radius, it's more honest to remove the shaft. General methods for attaching correction masses are covered in a separate article; here we've covered only this construction's limitations. ## When on-site work won't succeed, and when the shaft is better off removed Declining to balance is also a result of the visit, and it saves you a shift. Below are the cases that come up most often on driveshafts. - Play in the U-joint crosses or splines. Worn parts get replaced first, then a repeat measurement, and only then a decision about balancing. - The shaft is assembled out of alignment mark. We disassemble, reassemble to the marks, and check again. The visit often ends right there. - The second harmonic from the working angles dominates. It's fixed by restoring the components' fit and equal angles, and it isn't removed by weights. - The centre support's bearing or its cushion is worn out. We replace the support, otherwise the result will be unstable. - The tube is deformed, with a dent or runout. Balancing comes after repairing or replacing the tube, not instead of it. - The shaft doesn't reach stable running speed, or speed drifts across a wide range. - There's no safe access: the shaft is in a tunnel, under a non-removable guard, or it can't be spun up without risk to people. - The running speed sits near the bending critical speed. A shorter shaft, an additional support, or a revised operating regime is needed. - [x] Assembled on the machine, we see the shaft together with the flanges, yokes, real angles and support stiffness - that is, in the system where it actually works. - [x] On a lathe the shaft is balanced separately, on its own trunnions or in V-blocks. That's the right path when the shaft is already removed, or when it can't be spun on the machine. - [x] The lathe option comes with a condition: the shaft must be reassembled exactly as it was when balanced. - [x] Before disassembly, mark the relative position of the flanges and the splined-joint halves, and on reassembly put the bolts back in the same holes. Otherwise a different rotor goes back onto the machine. > A comparison of an on-site visit versus shop work is in a separate article. On a driveshaft, the choice is more often decided by access and the ability to spin the shaft safely, not by the accuracy of the method itself. ## What you get, the price, and how to order The main result is numbers you can put on the table. We hand over before-and-after measurements at the same points, at the same operating point, in the same frequency band. Plus a finding on assembly and clearances: it's often exactly that which determines what to do with the shaft next. vibration diagnostics with a report costs 300 EUR per unit, balancing adds from 250 EUR, minimum invoice for a visit 500 EUR. The calculator on the website gives an exact figure, accounting for the number of rotors, travel, and scope of work. Diagnostics without balancing is worthwhile too, because you find out exactly what to repair before ordering parts. We're the engineers who design and manufacture Balanset instruments, and we do the on-site balancing ourselves. We're based in Vila Nova de Gaia, near Porto, and travel throughout Portugal. - Shaft type: drive shaft, intermediate shaft, PTO shaft, single-section or with a centre support. - Working speed and drive method, and whether the shaft can be spun up unloaded. - Approximate shaft length and tube diameter, to assess the planes and the speed limit in advance. - Photos of the whole shaft, the flange joints, the splined joint, and the centre support. - Measured vibration level, if a reading has already been taken, and the measurement points. - Constraints: where welding isn't allowed, site access requirements, the shutdown window. - [x] A report with overall vibration and the running-speed component at each point, before and after the work. - [x] The correction masses fitted, radii, angles or fixed-position numbers, and the attachment method. - [x] A spectrum with an assessment of the second harmonic, the condition of the U-joint crosses and the support bearing, and any signs of misalignment. - [x] An assembly finding: whether the marks line up, play in the U-joint crosses and splines, tube and flange runout. - [x] An assessment against the applicable part of ISO 20816 as a reference, and, if needed, a residual-unbalance calculation by balance quality grade. - [x] Saved influence coefficients for this shaft: the next trim balance will go ahead without trial runs. - [x] Recommendations: what to replace before the next shutdown and what to watch for in vibration monitoring. > We apply the ISO 20816 zones and grade-G balance quality classes as a reference and state the applicable part and edition of the standard, because there are exceptions by power, speed and support type. Preparation cuts the visit almost in half: access to both ends of the shaft and to the centre support, cleaned spots for the sensors, a clean tube, and the ability to reach running speed and stop the required number of times. More detail is in our article on preparing equipment for balancing. Sources: [ISO 20816-1:2016](https://www.iso.org/standard/63180.html) · [Balanset-1A manufacturer specification](https://vibromera.eu/product/balanset-1/) ## Frequently asked questions **Can a driveshaft be balanced without removing it from the machine?** In most cases, yes. We work on the shaft as installed: sensors on the bearing housings of the units at the ends of the shaft and on the centre-support bracket, reflective tape on a flange or the tube, correction masses on the tube near the yokes or on the flange joint. Removal is needed when the shaft can't be spun up safely, there's no access to the ends of the tube, or all welding is prohibited on this construction. **The shaft started vibrating right after the U-joint crosses were replaced. Is that imbalance?** More likely not. First check whether the marks on the splined-joint halves line up, and whether the yokes at the ends of the shaft lie in the same plane. If they were rotated out of alignment during assembly, vibration will appear without any imbalance at all and will sit at the second harmonic. Replacing a U-joint cross does change the component's mass too, but that's secondary: assembly is checked first, because its error can't be compensated with weights. **What actually happens if the splined joint is assembled out of alignment mark?** The mutual compensation between the two joints breaks down. Each joint working at an angle rotates non-uniformly, speeding up and slowing down twice per revolution. When the yokes are in phase, the second joint gives back what the first one took away. Rotate the slip-joint halves out of alignment, and the non-uniformities stop cancelling and start adding up instead. In the spectrum you'll see a second harmonic that grows with speed. There's one fix: disassemble and reassemble to the marks. **There's play in a U-joint cross and we have nothing to replace it with right now. Can it be balanced now and replaced later?** No, and that's not caution talking, it's the physics of the method. The correction calculation requires the shaft to respond to added mass the same way from start to start. The clearance gets taken up differently each time, so the phase drifts and the influence coefficient comes out unreliable. Plus a weight on a worn component masks the level and doesn't restore remaining life, and a rotating shaft coming apart is dangerous. U-joint crosses first, balancing after. **Where do you place the weights, and will you damage the shaft's tube?** We mirror the factory solution: a thin plate bent to match the tube's radius, tacked with short weld points near the end of the tube, close to the yoke. We don't run a circumferential weld, don't weld over the yoke's seam, don't drill the tube, and don't weld near the splines and seals. We clamp the welder's return lead to the same part, so current doesn't pass through the U-joint crosses and bearings. If welding is ruled out on your shaft, we use a clamp-on weight or balance washers at the flange, and if there isn't enough radius, we'll tell you the shaft will have to be removed. **We have a two-section shaft with a centre support. Is it balanced as a whole?** No, we work section by section. Each section is a separate rotor with its own ends, and the correction on one shows up at the shared centre support, so we measure and monitor both. We separately check the support bearing itself and its rubber cushion: a worn-out support generates vibration on its own and makes the balancing result unstable.