CNC Workholding on a Small Router: Tape and CA Glue vs Clamps vs Vacuum

The job that taught me to respect workholding wasn't a complicated one. A flat plaque, straight pockets, nothing heroic. I had it held with double-sided tape, because that's what was on the shelf and the part was thin and I didn't want clamps in the cut path. Two-thirds of the way through the last pass, the edge lifted. Not dramatically — no projectile, no ruined spindle, just enough peel at the leading edge that the bit took a bite out of the rim I'd spent an hour dialing in. The scrap bin got the part. The lesson was cheaper than it could have been, and it's the reason I now think about workholding as a system with its own failure modes rather than as an afterthought bolted on before I hit cycle start.

On a small router, that thinking matters. What you do have is a small bed where the work is often flat sheet goods, and flat sheet goods are exactly the case where the cheap methods work well — if you understand what each one is actually doing.

I want to lay out the three methods hobby owners reach for, what the sources I could verify actually say about them, what they cost at list, and where each one fails. I'll be candid about what isn't verified, because a lot of the numbers floating around this topic are folklore.

Tape and CA glue: what it is and why it holds

The method is straightforward and one hobby CNC publisher, DesktopCNC Forge, describes it in detail. You press a layer of painter's tape onto the back of the workpiece and a matching layer onto the spoilboard or fixture plate. Thin CA glue goes on one layer and accelerator on the other. You press the part down and the joint cures in seconds.

The part that surprises people is why there are two layers of tape. The publisher explains it: the double layer exists so the CA bonds the two tape backings to each other rather than to the part or the fixture plate. That keeps the part and plate clean, and the tape-to-surface bonds peel away afterward. In practice this is the whole trick. You are not gluing wood to wood; you are welding a sacrificial interface that releases cleanly.

The publisher claims two advantages over plain double-sided "carpet" tape. First, the CA-welded tape stack behaves more like a rigid bonded joint and "doesn't creep." Second, the hold is distributed across the whole footprint rather than concentrated at clamp points, so thin parts stay flat instead of bowing. That second point is the one I've come to believe matters most on a small machine. A thin panel clamped at two edges will lift in the middle under cutter load. Tape over the full footprint won't, because the hold is everywhere the tape is.

The stated failure modes are worth memorizing. Weak tape adhesion is named as the most common failure of the method, so the tape must be pressed down hard and burnished. The publisher also advises using thin CA, keeping the glue bead modest, and gluing one tape layer while accelerating the other — not mixing accelerator into the glue bead. The stated limits are flat parts only, thick or heavy stock that takes loads the tape bond can struggle with, and very deep or long-running aggressive cuts in metal, which belong in a vise.

I'll flag the sourcing honestly here. DesktopCNC Forge is a hobby blog with affiliate links and no measured test data, and its page carried a publication date in the future, which means treat it as an unreliable-dated secondary source. It gives no hold-force numbers and no consumable prices. So I use it for method steps and stated failure modes, and I don't quote it as gospel.

What the tape and glue actually cost at list

Prices move, so treat these as list prices at time of access from Taylor Tools. 3M ScotchBlue 2090 Original Blue Painter's Tape in 60-yard rolls starts at $3.99. 3M Scotch 233+ Green High Performance Masking Tape, 60 yd/180 ft, starts at $7.99. FastCap Speedtape double-sided pressure-sensitive tape starts at $11.99 and Taytools Double Stick Tape at $12.99.

On the glue side, the Starbond Medium CA Glue Pro Kit, 2 oz, is $14.79 on Amazon sold by Starbond — $7.40 per ounce — with Subscribe & Save at $13.31 or $12.57. Note the viscosity: the tape method calls for thin CA, and this listing is medium, with a stated 10–15 second set time.

What I cannot give you is cost per hold. Nobody verified how much tape or glue a given part consumes, or how many parts come out of a roll or a bottle, so any per-part arithmetic I did here would be invented. I'd rather hand you the shelf prices and let you run your own numbers against your own part sizes. The honest framing is that tape-and-CA has a low entry cost, and both tape layers cover the whole back of the part.

Clamps: the method I can't fully cost

Clamps are the default. But I have to be straight with you: I could not verify a single clamp price or clamping-force figure for this article. The workholding collections I tried returned nothing usable. So I can't put numbers on clamps, and I won't pretend otherwise.

What I can say is structural rather than numeric. Clamps concentrate force at the points where they contact the work. On a thin panel, that means the hold is strong where the clamp is and absent in between, which is precisely the bowing problem the tape method claims to solve. Clamps also live in the cut path unless you plan around them, which on a small bed with a small work envelope is a real constraint.

I've come to believe a small router is best served by treating clamps as the method for stock that is thick enough to resist bowing and simple enough that clamp positions clear the toolpath. For anything large and thin, you're fighting geometry.

Vacuum: the levels, and the shop-vac confusion

Vacuum on a router holds the work by dropping pressure between the part and the fixture. Busch Vacuum, a pump manufacturer, states that CNC router clamping uses 150 to 500 hPa (mbar) absolute. For scale within their own figures, handling and lifting uses 300–900 hPa and dust cleaning uses 800–900 hPa. Busch also distinguishes two hardware families: a vacuum table, with a flat surface of evenly distributed suction holes and a large suction area that holds down an entire wooden panel, often used for engraving; and a vacuum clamping system, using several repositionable suction cups or pods matched to the panel shape. Busch notes vacuum clamps leave no pressure marks or other damage to the wood.

Here is where the hobby conversation goes wrong. Those are industrial, dedicated-pump figures. Busch does not state holding force in pounds and does not address hobby shop vacs. So I cannot tell you what a shop vac will pull or hold, because that wasn't verified, and I'd rather say so than repeat a number I can't defend.

Matching method to part instead of to habit

Designing with intent, in this context, means picking the hold from the part rather than from what's nearest the bench. A few rules I work to:

• Large, flat, thin sheet: tape and CA over the full footprint. The distributed hold is the point, and the tape costs are modest.
• Thick or heavy stock, or metal: clamps, or a vise for aggressive metal cuts. The tape bond has stated limits here.
• Parts that need re-fixturing: clamps, because there's no consumable to burn through and no cure to wait on.

The failure mode that ruins a cut isn't dramatic in any of these. For tape, the publisher is explicit that weak tape adhesion is the most common failure, so burnishing is not optional. For plain double-sided tape, the creep-and-peel at edges is the stated weakness. For clamps, it's bowing and collision. Same root cause in each case: the hold doesn't match the load the cut puts on it.

What I'd verify before trusting any of this further

I'd rather close on a next action than a summary. Before you commit a method to a job that matters, run one scrap part of the same stock and the same thickness and cut it with each candidate hold. You're not measuring pounds of force; you're watching for the edge lift, the creep, the bow. That's the test that matters on a small machine, and it costs you a few minutes and a piece of scrap. What the brief couldn't give me — per-hold cost, measured hold force, clamp prices, and what a shop vac can actually pull — is exactly the ground I'd want tested before I sized a fixture around it.