AXILINECorrection weight angle: zero point, direction, and fixed positions
The instrument has calculated it: 6.4 g at 137°. What comes next is where balancing jobs most often go wrong. The zero reference, the direction of the angle, and the “add or remove” switch decide whether the vibration drops or doubles. Below, we go through each point and show how fixed-position mode takes the protractor out of the process entirely.
The zero point: wherever the trial weight was fitted
The instrument doesn't know where “up” is on your rotor. It knows exactly one reference point: wherever you fitted the trial weight. The correction angle is measured from there. The software shows “M1 = 6.4 g, f1 = 137°,” and that 137° is measured from the trial weight's mark. Not from the key. Not from vertical. Not from the tachometer's reflective mark.
That's why the trial weight needs to be marked, not just bolted on. Chalk, a marker, a center punch, a strip of tape — anything that survives the run and doesn't fall off. Once you remove the trial weight, you lose the zero point if you didn't mark it first. There's nothing to recover it from afterward: the instrument's archive doesn't record a “before” angle.
One trick saves time on the next visit. Fit the trial weight at the same angular position as the reflective mark. Then the zero point coincides with the mark, it's always visible, and the influence coefficients can be saved and reused: with saved coefficients, the instrument calculates the correction in one run instead of two.
- The trial weight's location is marked on the rotor before the first run
- The zero point coincides with the reflective mark, if you plan to save the influence coefficients
- The trial weight's actual mass and actual radius are entered in the software
- You've recorded which side of the machine you're viewing the rotor from
The phase F1 the instrument shows isn't the angle of the heavy spot (the location where the extra mass is concentrated). It's the phase of the 1x signal — the vibration at rotation frequency — relative to the tachometer mark. You can't use it to work out by eye where to put the weight. Only the calculation from two runs — baseline and trial — gives you the correction angle.
Direction of the count: with the rotor's rotation
You've got the zero point; now the direction. The Balanset-1A measures the correction angle from the trial-weight location, going in the direction of the rotor's rotation. Fixed positions are numbered by the same rule: Z1 sits where the trial weight was, Z2 is the next position along in the direction of rotation.
“In the direction of rotation” sounds unambiguous, until you've walked around the machine. Clockwise rotation viewed from the drive side looks like counter-clockwise rotation viewed from the impeller side. Half of all mirrored errors are born right here: the angle was measured correctly, but from the wrong end.
- 1
Determine the direction of rotation by eye
Bump the motor briefly, or turn the rotor by hand, and watch which way it goes. Don't trust the arrow painted on the guard: it was drawn during installation and could have been flipped along with the cover, or removed during a repair.
- 2
Draw an arrow where you'll be marking
Use chalk on the rotor disc or the flange, on the side you can actually reach the weight-mounting location from. From then on, measure the angle only by that arrow, and don't walk around the machine again.
- 3
Mark out the zero point and the graduations
Line up a protractor from the zero mark in the direction of the arrow. If you don't have one, measure the circumference with a tape and divide it into equal parts: 30° on a 400 mm diameter is about 105 mm along the rim.
- 4
Say it out loud before you tighten anything
“Zero's here, rotation's that way, 137° lands here.” Thirty seconds of checking is cheaper than an extra stoppage, taking the guard back off, and another run.
The direction has to stay the same for the whole balancing job: when you fit the trial weight, when you fit the correction, and when you add weights after the verification run. If the software explicitly sets the direction in its settings or shows it on the polar diagram, follow that, not habit.
The mirrored error: why the vibration goes up
Let's look at what actually happens physically. A correction weight creates a vector meant to cancel out the original imbalance vector: equal in magnitude, opposite in direction. Measure the angle in the wrong direction, and the weight ends up mirrored relative to the zero mark. The resulting angular error isn't θ, it's 2θ — double the calculated angle.
The weight's mass is correct, the direction isn't. For a linear system, the residual vibration works out to roughly 2 · V0 · sin θ, where V0 is the original running-speed component. This simple geometry shows exactly when the error gets away with it, and when it hits hard.
| Calculated angle θ | Actual error 2θ | Residual 1x with a mirrored fit |
|---|---|---|
| 10° | 20° | ≈0.35 × the original: vibration still dropped |
| 30° | 60° | ≈ the original: as if nothing had been done |
| 60° | 120° | ≈1.7 × the original: it got worse |
| 90° | 180° | ≈2 × the original: maximum damage |
| 150° | 300° | ≈ the original |
| 170° | 340° | ≈0.35 × the original |
Don't try to fix a mirrored fit by throwing on a second weight at random. Remove the weight, return the rotor to its baseline state, redo the layout in the opposite direction, and refit it. The instrument calculates any add-on based on what's actually sitting on the rotor. If the weight is mirrored, the add-on calculation will be off too.
How to tell the mistake was specifically in the angle
The table above explains something unpleasant. A small calculated angle forgives a direction error, and you won't even notice: the vibration dropped, the job “worked.” Angles from 30 to 150° don't forgive it — a mirrored weight there makes things worse than before. The worst case is a calculated 90°: the weight lands exactly on the heavy spot and doubles the original imbalance.
A verification run helps you tell a mirrored error apart from other causes of rising vibration. Look not just at the amplitude but at the 1x phase too.
If 1x has gone up and the phase has shifted noticeably, and the overall vibration (the total level across all frequencies) has moved along with 1x, you almost certainly got the angle wrong, or mixed up the “add/remove” switch. If 1x stayed roughly where it was but the overall level went up, the angle has nothing to do with it: look for loose fasteners, rubbing, a bearing defect, shaft misalignment. Balancing won't remove that kind of vibration.
Separately, check the linearity of the system. The weight can be fitted correctly and the vibration can still behave differently from what the instrument calculated: this happens with resonance, with a flexible rotor, and with supports that change stiffness at large amplitudes. In that case, a careful angle won't save you, and balancing needs to wait until the mechanics have been put right.
Adding or removing metal: drilling flips the angle by 180°
A weight can't always be welded or bolted on. On pulleys, grinding wheels, pump impellers, and wheels under a closed guard, it's often more practical to remove metal instead: drilling, milling, or grinding off a weld bead.
The software handles this with an “add” / “remove” switch (Add / Delete, or Removal in two-plane mode). Switch it to removal, and the instrument automatically flips the angle by 180°. The logic is simple: removing mass at point A is equivalent to adding the same mass at the opposite point.
Forgot to switch it
The instrument gave you an angle for adding, and you drilled by it anyway. That's exactly a 180° error: metal came off the light side, and the imbalance roughly doubled. One of the most common scenarios behind “it got worse after balancing.”
Hole volume for the required mass
Mass equals density times volume. Steel ≈7.85 g/cm³, cast iron ≈7.2, aluminum ≈2.7. A Ø10 mm hole, 10 mm deep, in steel removes about 6 g. If one hole would end up too deep, drill several next to each other instead.
Where not to drill
Into weld seams, into a thin impeller wall, into fillets, or anywhere a hole would become a stress concentrator or open up an internal cavity. If you're unsure about strength, add mass instead of removing it.
The radius changes along with the method
A weight is usually fitted at the maximum radius, while drilling happens wherever there's enough material. A smaller radius needs more mass: m₂ = m₁ · r₁ / r₂. Enter the actual radius, measured with a tape, not the design figure from a drawing.
Removing exactly the calculated mass with a drill almost never works out precisely. Measure the actual depth and diameter of the hole with calipers, calculate the mass removed, enter the actual value, and run the verification. From there the instrument will itself suggest adding a weight or drilling a bit more.
Fixed positions: a number instead of an angle
Now for the main simplification. On most real machines, you can't fit a weight at an arbitrary angle anyway. There are six blades, twelve bolt holes, eight spokes, and those are the only places you can fasten to. A protractor just gets in the way in that situation: you measure out 137°, and the nearest blade sits at 120°.
In the Balanset-1A's settings, this is handled by “Weight attachment method.” “Circum” means the weight can be fitted anywhere around the circumference. “Fixed position” means there's a finite number of mounting points, and you enter their actual count: 12 holes give a 30° step, 6 blades give a 60° step.
After that, the instrument stops talking about degrees at all. On the results tab you see position numbers and a mass for each: for example, Z10 = 4.2 g and Z11 = 2.8 g. Z1 is the position where the trial weight was fitted, and the numbering runs in the direction of rotation.
- No protractor or circumference layout needed: you're counting blades, not degrees.
- The interpolation error disappears. The instrument itself maps the calculated angle onto the real mounting points, instead of you eyeballing “roughly in the middle.”
- A second person can check the layout without the instrument: “tenth blade along from the marked one, in the direction of rotation.”
- The position number goes into the report unambiguously. “Around 137°” can't be reproduced a year later; “Z10 and Z11” can.
One thing still stays on you: the direction of the numbering. Positions have to be counted in the direction of rotation, and fixed-position mode doesn't let you off the hook for that. But it's incomparably easier to check than an angle: the direction of rotation is visible to the eye, and a position number is discrete — you can count it off with a finger and say it out loud. Mark Z1 with paint and write the numbers directly on the rotor while the trial weight is still in place.
Splitting the weight between two neighboring positions
The calculated angle almost never lands exactly on a blade. The instrument does what you'd do with vectors yourself: it splits the required mass across two neighboring mounting points. Two masses at those positions produce the same combined vector as one mass at the calculated angle.
The split is calculated using the law of sines. The closer the calculated angle is to one of the positions, the more of the mass goes there. Exactly halfway between two positions, the masses come out equal.
The split has a side effect that's worth knowing before you head out. The combined mass is always larger than a single mass at the exact angle would be: two vectors at an angle to each other don't add up arithmetically. The overhead depends only on the spacing between positions, and it can be estimated in advance.
- On a fan with 12 holes, the split is almost free — you don't need to worry about overhead
- On a four-blade wheel, check in advance that both weights fit in place and don't foul the guard
- The fastener and the mounting point need to hold the increased mass, not just the calculated one
- If the overhead is a problem, add positions: eight instead of four brings the overhead down from 41% to 8%
| Number of positions n | Spacing between positions | Maximum overhead on combined mass |
|---|---|---|
| 4 | 90° | up to +41% |
| 6 | 60° | up to +15% |
| 8 | 45° | up to +8% |
| 12 | 30° | up to +3.5% |
| 24 | 15° | up to +1% |
The instrument only splits the mass between two neighboring positions. If it gives you a large mass for one blade, don't spread it across three or four neighbors on your own judgment: the vector will drift, and you'll get the same mirrored effect in miniature. It's better to mark out additional mounting points if the design allows it, increase the number of positions in the settings, and recalculate.
Checklist before the verification run
- The trial weight has been removed. If it couldn't be removed, the “leave trial weight” box is checked in the software, and the correction has been recalculated
- The zero point is marked on the rotor and matches the trial weight's location
- The direction of the count matches the direction of rotation, and it's been checked from the same side of the machine you used for the layout
- The “add / remove” switch matches what you actually did to the metal
- The weight's actual mass has been weighed and entered, not taken straight from the calculation
- The mounting radius is the actual one. If a different radius was used, the mass has been recalculated using m₂ = m₁ · r₁ / r₂
- The number of fixed positions in the software matches the number of real mounting points
- The weight is fastened as securely as a permanent one: welding, a bolt with a lock nut, or a retainer
Vibration dropped but didn't come within tolerance — that's a normal part of the process, not a mistake. The software will suggest add-on weights on top of what's already fitted. The zero point and direction don't change in that case: the count still runs from the trial-weight location. And remember that “in tolerance” in the software means one thing: the residual 1x is below the target value you set. The machine's overall condition is assessed separately, from the overall vibration in mm/s RMS over the 10-1000 Hz band. Check the applicable part and edition of the standard for your machine.
If there's no time to work through angles on site
A mistake in the angle reference costs you one extra run if you catch it, and a stripped-down machine if you don't. On an easily accessible fan, you'll pick up the layout and recalculation after a couple of visits. On a unit that only gets shut down quarterly per schedule, up at height, or on a hot process line, there's usually no second attempt.
AXILINE's engineers design and manufacture the Balanset instruments and use them ourselves on site visits. We come to the machine, fit sensors to the bearing housings, first confirm that the vibration really is coming from imbalance and not shaft misalignment, looseness, or resonance, and only then balance the rotor in its own bearings — in one or two planes, with weights or by drilling, by angle or by fixed positions. We hand over the result as a report with the baseline and residual 1x, and a record of where and how much mass was fitted.
If you balance it yourself, consulting support on the instrument and on angle layout comes together with the Balanset-1A. A typical order of magnitude for a fan with a dominant 1x peak: coming down from ~12 mm/s to ~1.5-2 mm/s is realistic. Nobody promises this in advance. The result depends on the condition of the supports and fasteners, and on whether imbalance is the only cause of the vibration.
Sources: Balanset-1A manufacturer specification · Balanset-1A operation manual · ISO 21940-11:2016 · ISO 20816-1:2016 · ISO 13373-5:2020
Frequently asked questions
From which point is the correction weight's angle measured?
From wherever the trial (calibration) weight was fitted. That's exactly 0°. Not from the key, not from vertical, and not from the tachometer's reflective mark, unless it happens to coincide with the trial weight. Mark that spot with paint or chalk before the first run: once the trial weight is removed, there's nothing left to recover the zero point from.
Which direction is the angle counted in?
In the direction of the rotor's rotation, unless the software specifies otherwise. Fixed positions are numbered by the same rule: Z1 is where the trial weight was, Z2 is the next one along in the direction of rotation. Check the direction of rotation by eye and draw an arrow on the side of the machine you're using for the layout. Clockwise rotation viewed from the drive side looks counter-clockwise from the impeller side.
Why does vibration go up rather than stay the same when the direction is wrong?
The weight ends up mirrored relative to the zero mark, so the actual angular error equals double the calculated angle, 2θ. For a linear system, the residual vibration works out to roughly 2 · V0 · sin θ. At a calculated 30° it equals the original level, at 60° it's about 1.7 times higher, at 90° it doubles. At angles under 30°, the vibration still drops, and the error is easy to miss.
Do you need a protractor to fit the weight?
In “Circum” mode, where the weight can be fastened anywhere around the circumference, you need a protractor or a taped-out edge layout. In “Fixed position” mode you don't: you enter the number of real mounting points, and the instrument gives you position numbers (Z1, Z2 … Zn) and a mass for each. You no longer need to count degrees — you count blades or holes from the marked Z1.
The instrument gave two weights for two neighboring blades. Is that normal?
Yes. The calculated angle rarely lands exactly on a position, so the software splits the required vector across two neighboring mounting points using the law of sines. The combined mass always comes out slightly higher than a single mass at the exact angle would: up to +3.5% with 12 positions, up to +15% with 6, up to +41% with 4. Check that both weights fit in place and don't foul the guard.
What do you do if there's nowhere to bolt a weight on and drilling is needed?
Switch the mode from adding to removal (Delete / Removal). The instrument will flip the angle by 180° itself, because removing mass at a point is equivalent to adding the same mass at the opposite point. Forgetting this switch is the most expensive mistake you can make: you'll end up with exactly 180° of error and roughly double the imbalance. Calculate the hole volume for the required mass from the material's density, and measure what was actually removed, not the calculated figure.
Can the correction angle be determined from the 1x phase without a trial weight?
No. The 1x phase is the phase of the signal relative to the tachometer mark, not the angle of the heavy spot. The software calculates the angle only from the influence coefficient, which comes from two runs: baseline and trial. There's one exception: if the coefficients for this machine are already saved, one run is enough. But then the sensors and the mark have to be positioned exactly as they were during the original balancing.
Related content
Mounting correction weights: welding, bolting, riveting, and removing metal by drilling
A weight stays put when it's mounted properly: a continuous weld around the perimeter of a pad cleaned down to bare metal, or a bolt through an existing hole with a lock nut or thread locker. A tack weld, a magnet, and glue are fine for a trial fit only - you can't leave the site with those as the final fix. If there's nowhere to add mass, remove metal instead, by drilling or milling, at a point 180° opposite where the weight would have gone; calculate the volume from the material's density. Use as large a radius as you can, and always enter the actual mass and actual radius into the instrument.
Trial (calibration) weight: why it's needed and how to choose it
A trial (calibration) weight is a temporary weight whose mass, mounting radius, and angular position are precisely known. You fit it to a correction plane, make a second run, and by doing so give the machine a known change in imbalance. From the difference between the before-and-after vibration vectors, the software calculates the influence coefficient — the specific machine's sensitivity to the added weight — and only then calculates the mass and angle of the correction weight. A trial run counts as successful if the 1x amplitude changed by at least 20-30% or the phase shifted by at least 20-30°.
On-site balancing of crushers and shredders: rotors, hammers, knives
Yes, we balance the rotors of crushers, hammer mills, shredders, pelletizers, and disintegrators on site, in their own bearing supports, with no dismantling and no balancing machine. Three conditions apply: the set of hammers or knives is matched by mass, the machine holds a stable idle speed, and there's access to at least two correction planes with the rotor stopped. If a hammer is missing, broken, or the set is mismatched by mass, balancing is pointless: the set comes first, weights come second. If overall vibration (the combined level from all causes) is many times higher than the running-speed component — the vibration at the rotor's rotational frequency, which is what imbalance produces — the cause isn't imbalance, and we'll say so plainly rather than fit weights just to close out a report.
Describe your equipment and the problem
We'll answer your questions, clarify the details, and let you know what's needed for an estimate and a visit.