A batch of small steel washers in the incoming condition, edges sharp and not uniform part to part
Small Parts Deburring: How to Choose the Process and Count the Batch September 05 , 2026

Small Parts Deburring: How to Choose the Process and Count the Batch

Small parts deburring fails on three questions, and none of them is "will the machine remove the burr". They are: how many parts the batch will actually hold, whether the media can reach the burr without the parts damaging each other, and how the parts come out of the media. Answer those on the drawing and the machine choice is almost forced; skip them and a clean trial turns into a batch that needs hand work.

This page gives the process families and their limits, the media sizing rule, the batch-count arithmetic, and the point at which small parts should not be tumbled at all. The numbers are a starting point for a trial — a published cycle assumes a material and a removal that may not be yours.

What you see on the floor when a small-parts process is set up wrong

Five symptoms cover most of the failures, and each points at a different variable:

What you see What it usually means
A hand-picked trial looks perfect; the full production batch comes out matte, dented or uneven The trial ran at demonstration load. Media coverage per part fell as the part count rose
Outer edges are clean but the cross-hole, slot or thread start is still sharp Media too large to enter the feature — a sizing problem, not a cycle-time problem
Burrs go, but thin sections bend or dimensions drift Grade too coarse for the stock, or a cycle stretched to chase one stubborn burr
Parts are deburred but the operator still touches up a share of them The remaining work is marks and handling, not burrs — a loading and separation problem
Cost per thousand stops falling as volume grows The bottleneck moved off the machine and onto separation, drying or inspection

The last two are the ones buyers miss. A small-part line can be technically correct and still not pay, because the machine was never the constraint.

Why small parts behave differently from large ones

Two effects pull against each other, and both scale with part count.

Reach. The media has to get to the burr. On a small part the burr is often in a feature of the same order of magnitude as the media itself — a thread start, a slot, a cross-hole exit — so the usable media size collapses, and with it the cutting rate.

Cushioning. Parts are deburred by media, but they also hit each other. Each part needs a surrounding volume of media to separate it from its neighbours, and that volume does not shrink when the part does. Load twice as many parts and you have not doubled throughput; you have halved the media between them.

The practical consequence: a 60 L bowl typically holds a few hundred small components per batch, but the honest count comes from the part envelope and the media size, not from a catalogue figure. A part with a deep cavity occupies far more room than its outer dimensions suggest, and a batch that looks fine with simple geometry can drop by half once a feature traps media.

Check these seven things on the drawing before selecting anything

  1. Where the burr is, and how strongly it is attached. A loose burr at a drilled cross-hole and a rollover at a stamped cutoff are different jobs. Extending a gentle cycle to remove a heavy burr rounds every good edge first.
  2. The smallest feature on the part. Smallest hole, narrowest slot, tightest thread pitch, smallest internal radius. This sets maximum media size.
  3. The worst-case envelope, not the average part. Overall length, diagonal, protrusions, hollows, and anything that can nest with a neighbour.
  4. Section thickness. Thin walls fail by distortion before they fail by over-cutting.
  5. Which surfaces must not change. Cosmetic faces, sealing faces, thread flanks, bearing bores, datums.
  6. Family mix per batch. Parts of different lengths shield each other or interlock; families with different acceptance criteria should not share a charge.
  7. What happens downstream. Plating, anodising, assembly, a thread gauge, a leak test. Acceptance should be written from that, not from brightness.

Choose the machine family before the media

The machine families differ in how the abrasive reaches the part, and that decides which features are reachable at all:

Family How the media reaches the part Fits small parts when Main limit
Vibratory bowl media flows past the work in a circulating charge general-purpose small parts; process can be watched and adjusted the charge needs volume to circulate, so very small loads run poorly
Vibratory tub same flow, longer open channel long pins, small shafts, heavier small blanks long parts can bridge and tangle
Centrifugal barrel the whole charge rotates in small compartments at high energy compact precision parts where a short cycle matters higher intensity, so loading and thin sections need control; a 60 L machine (LX60B) runs 0–170 rpm with 15 × 4 kg of compartments and 3.0 + 0.75 kW
Centrifugal disc the part is held against a rotating abrasive tool a controlled, local cut on one part at a time throughput is per held part, not per batch
Magnetic finishing a flexible needle mass is drawn around the part burrs in slots, cross-holes and thread roots; light burrs light stock removal; retained needles must be checked

Two things worth knowing before a machine is quoted. A centrifugal barrel's advantage is intensity, not capacity — it shortens cycles for compact parts and is wrong for fragile points unless the load is controlled. And magnetic finishing is sized by the workpiece envelope and the required removal, not by the part weight, so its litre figure cannot be converted into a parts-per-batch number.

If you want the machine-side detail rather than the application view, the small-capacity vibratory finisher line, the centrifugal disc range and the magnetic finishing range carry the platform data.

Set the media size, then the grade

Media should never be larger than about one third of the smallest feature on the part. A part with a 3 mm blind hole needs media smaller than 3 mm, which means a fine grade and a longer cycle, and there is no way around it. A useful field check: put the part in a bag with a sample of the media. If the media will not enter the features by hand, it will not enter them in the machine.

The working bands are 3–10 mm for ceramic and 10–25 mm for plastic. For small parts the ceramic end is usually the usable one — brown fused alumina triangles run from 3×3 mm upward, ceramic balls from 2 mm, and high-frequency porcelain cylinders as small as 1.3×3 mm for tight features. Plastic in its larger sizes (8×8 mm and up) is generally too coarse for small-part features, though the 5×5 and 6×6 mm cones are useful where a softer cut is needed.

Grade matters as much as size, and the wet and dry scales are not the same:

  • Wet grinding: VR / R / N / NF / F / FF
  • Dry grinding: VR / R / M / MF / F / FF

For a small part, the grade is set by how much stock has to leave and what the surface must look like afterwards — those are two different jobs. A light deburr runs 10–20 minutes; a bright polish runs 2–8 hours. A small-part vibratory cycle usually sits between 15 minutes and a few hours, and the media and compound decide which end you land on.

Media is a consumable, not a purchase: a charge lasts 20–200 operating hours depending on material and cycle. The signals that it is spent are the same everywhere — the finish stops improving, a matte grey haze appears, and the cycle needs roughly 20% longer for the same result.

Count the batch before you commit to a machine size

Work the arithmetic on the worst-case part, in this order:

  1. Parts per batch. Start from the honest figure for the family you are considering, then reduce it for the worst-case envelope. Do not scale up from the trial count.
  2. Cycle time for the stock to be removed. Deburring and polishing are separate cycles.
  3. Parts per hour = parts per batch × 60 ÷ cycle minutes. This, not the batch figure, decides whether the machine or the operator is the bottleneck.
  4. Add the non-machine steps. Loading, separation, rinsing, drying, and any inspection the drawing requires. On small parts these often exceed the machine time.
  5. Overlap what you can. A cycle that runs unattended overnight changes the economics; one that needs an operator watching does not.

Record all of it in one place. A trial record carrying the part number, media grade and size, load and measured result at each interval is what makes the second batch repeatable — the sample trial record for small precision parts is the format we use, and how to measure burr height, edge change and Ra after deburring covers what to measure so that "better" becomes a number.

Clean, dry and separate are part of the process, not the aftermath

A small-part cycle is not finished when the burr is gone; it is finished when the parts can be separated, rinsed and dried without creating a second defect. Separation should be decided at the specification stage, because it constrains the media: the window between part size and media size has to hold for new media and for worn media, and worn media is smaller than the grade you ordered.

This is a subject in its own right — see how to separate small parts from tumbling media. The purchasing consequence is what matters here: an internal screen deck, an external separator and a dryer are separate lines on the quotation, and a process that is proven but cannot be separated at production rate is not a proven process.

A batch of small steel washers in the incoming condition, edges sharp and not uniform part to part
Small parts at the scale where separation, not finishing, is usually the real bottleneck. Read what makes them hard: the edges are sharp and not uniform from part to part, the parts are small enough to pass a screen that was meant to retain media, and they nest. The finishing cycle is the easy half.
The same small washers after finishing
After finishing. The result an operator is judging is not "are they brighter" but are the edges consistent across the batch — because on small parts a batch is judged as a batch, and one part with a rolled edge and one without is a rejected lot even if the average is acceptable.

When tumbling small parts is the wrong answer

Be specific about the limits, because these are where small-part projects stall:

  • Threads under a tight pitch tolerance. Media locks into a fine thread or rounds the thread start. Either a smooth small media, masking, or accept that thread forming comes after finishing. Fastener geometry is covered in how to deburr small fasteners without damaging threads, points or recesses.
  • Blind holes and deep pockets close to the media size. Media enters and does not leave. Fix it by media choice, not by adding a wash step.
  • Long thin parts — pins, springs, needles, thin clips. They align, bridge and tangle, and they bend before they deburr. Compartments or a different process.
  • Heavy burrs from a cutoff or a coarse stamping. Mass finishing is a finishing process, not a repair process. Tool wear and die condition set the incoming defect; if the burr is a fold rather than a loose edge, correct it upstream and finish what is left.
  • Thin sections and soft alloys. The part distorts or dents before the burr leaves. Reduce energy, change media, or change machine family — do not simply shorten the cycle.
  • Mixed families with different acceptance criteria. One charge, one standard. Two part numbers with different gauge requirements are two processes.
  • A batch too small to fill the charge. Below a certain load the media cannot shear properly and the result is soft and uneven. That is a case for a smaller machine, a magnetic machine or a disc machine rather than a large bowl run nearly empty.

What to ask before approving the quotation

  1. A trial on the worst-case part, not the average one. Send representative parts including the worst, the material, and both the finish you need and the finish you will accept. Our trials return the finish with the cycle time and the media consumption measured.
  2. Acceptance defined before the trial. Burr height, edge condition, which gauge or assembly test passes, which faces are protected. A trial without a written acceptance standard only tells you what it looks like.
  3. Batch count at production load, in writing. The number that matters is the one at full bed depth, not the demonstration load.
  4. The media specification in writing, as a grade, a size range and a shape — not as "ceramic media". A 2-to-4 mm range covers a spread that matters on a fine feature.
  5. Media consumption per 100 hours. With a 20–200 hour life, a supplier who cannot state consumption in kilograms per 100 hours is not quoting a usable consumable figure.
  6. What ships with the machine. Standard packing includes the spare parts kit, a media and compound starter set, tools and manuals, and every machine is run with a media load before it ships — so the first trial can run on arrival.
  7. Lead time against your plan. A standard machine ships in 25–35 days from the deposit; a built-to-order configuration takes 45–60 days, and adding a separator or a dryer extends it again. Media shipped with the machine adds about a week, and sea transit from Xiamen to Europe or North America is roughly 30–40 days. Minimum order is one machine; payment is 30% deposit with the balance against the shipping documents.
  8. The separation plan. Screens, decks and the dryer belong on the same quotation as the machine, because they turn a cycle into a production rate.

FAQ

How many small parts will a 60 L machine deburr per batch? A 60 L bowl typically holds a few hundred small components, but the count is set by the part envelope and media size, not the litre figure. A part with a hollow or a deep feature takes disproportionately more room. Test with your worst-case part, at production load.

Can several small part numbers share one batch? Only when they share an acceptance criterion and a media size, and their geometries do not shield or interlock. Mixed lengths and section thicknesses finish at different rates, so one part is over-run while another is under-run. High-mix work is better served by a machine that changes over quickly than by a bigger charge.

Is magnetic finishing better than vibratory for small parts? For burrs in slots, cross-holes and thread roots it often is, because a flexible needle mass follows the contour instead of touching at one point, and it does not round a sharp edge the way a tool would. It is a light-removal process, so it suits fine burrs rather than heavy edge removal. Deep blind features remain outside its reach.

Do small parts get dented or tangled in the machine? Yes, if the load is too high for the media to separate them. Direct part-to-part contact is the usual cause of dents, bent edges and inconsistent brightness, and the fix is more media coverage and a lower part concentration, not a longer cycle.

What media size do I need for a small hole or slot? Media should not exceed about one third of the smallest feature. For a 3 mm blind hole that means media under 3 mm, a fine grade and a longer cycle. Confirm it with a bag test before ordering the charge.

#+86-592-2381506

Email : info@surface-polish.com

Headquarters address : No. 31, Xinchang Road, Xinyang Industrial Zone, Haicang District, Xiamen

click here to leave a message

Leave A Message
If you are interested in our products and want to know more details,please leave a message here,we will reply you as soon as we can.

Preparing secure enquiry form…

Home

Products

whatsapp

contact