Tabs are free on the machine and expensive off it
A tab does not slow the cut down in any way that matters. The controller lifts the Z axis for a quarter of an inch and puts it back, and on a job with a couple of hundred inches of perimeter the total is a rounding error against the cutting time. The cost is entirely in what happens after the sheet comes off the table: someone cuts each tab, snaps the part free, and sands the stub back to the line.
Twenty seconds a tab is not a pessimistic figure for that on a part you care about the edge of. Twelve tabs on forty parts is 480 tabs, and 480 tabs at twenty seconds is two hours and forty minutes. Set against a machine time of perhaps forty minutes for the same nest, the hand work is the job.
The lever is the count, not the area
Here is the useful asymmetry. The holding a part gets is the tab count times the tab width times the tab thickness — an area. The cleanup time is the tab count times the seconds per tab, and the seconds per tab barely move with the tab width, because most of the twenty seconds is picking the part up, finding the tab, positioning the chisel and putting the part down. Cutting a quarter-inch stub instead of an eighth-inch one adds a second or two at most.
So halving the tab count and doubling the tab width leaves the holding cross-section exactly where it was and cuts the cleanup roughly in half. Six tabs at half an inch wide hold as much as twelve at a quarter, and cost half the labour. Whether six anchor points distributed around the shape actually restrain the part as well as twelve is a different question — it is about geometry and leverage, not about area — and it is the reason nobody should take this as a blanket rule. But on any part where the twelve were chosen out of habit rather than out of a test, it is the first thing to try.
Thickness is a different lever again
Doubling the tab thickness doubles the holding area for no extra cleanup operations at all, but it makes each one harder: a tab left at forty thousandths snaps with a light twist, while one at an eighth of an inch has to be sawn. That is why the useful range for tab thickness is narrow, and why leaving the tab at more than about a third of the stock thickness usually costs more in the cleanup than it buys in security. Above the material thickness it stops being a tab at all and becomes an uncut section, which this page flags.
What the numbers here are not
The cross-section printed above is an area of material, and nothing on this page turns it into a force or a verdict. What holds a part depends on the cutting force the tool generates, which is set by the depth of cut, the feed, the material and the tool geometry, and on how well the sheet under the part is held down in the first place. A part that is well vacuumed to the table barely needs tabs. A part in the middle of a warped sheet on a screwed-down spoilboard needs more than the arithmetic suggests. The only way to know is to run one and watch it.
Where this sits
The hold-down underneath the tabs is on the vacuum table hold-down force calculator, which is the other half of keeping a part where you put it. The cut time the tabs interrupt is on the CNC router job time calculator, and the layout that sets the perimeter count is on the sheet nesting yield calculator and the cut list optimizer. If the cleanup is going to be a sanding job rather than a chisel job, the sandpaper grit progression calculator is where that time comes from.
Questions people ask
How thick should a tab be?
That is a question this page will not answer, because it depends on the material, the cutting forces at your feed and depth, and how the sheet is held. What can be said is what each choice costs. A thin tab is quick to break and quick to sand but gives you less cross-section per tab, so you need more of them. A thick tab is secure and slow, and past roughly a third of the stock thickness the cleanup stops being a snap and becomes a cut. The practical route is to run one part at a thickness you think is conservative, watch whether it moves, and then work downward on subsequent parts until you find the point at which you stop being comfortable.
Do interior cutouts need tabs?
Usually more urgently than the outside profile does, because a small drop-out in the middle of a part has nowhere to go except up into the cutter. Whether you tab them or accept the drop depends on the size: a slug small enough to fall through into the spoilboard groove is often left alone, while one large enough to lift and be caught by the tool is not. If you do tab them, add the interior perimeter into the perimeter figure on this page so the tab count and the cleanup time both include them, because the cleanup on an interior tab is generally slower than on an outside edge.
Why does the spacing I actually get differ from the spacing I asked for?
Because the tab count has to be a whole number and the perimeter rarely divides evenly. A 36 inch perimeter at 5 inch spacing is 7.2 tabs of arithmetic, so the page rounds up to eight and then reports the spacing eight equally distributed tabs actually produce, which is 4.5 inches. Rounding up rather than down is deliberate: rounding down would give a spacing wider than the one you asked for, which is the direction you did not intend. Most CAM software does something similar, though not all of it tells you what it did.
Can I get the cleanup time down without changing the tabs?
Sometimes, and it is worth trying before adjusting the tab scheme. Cleaning tabs while the parts are still in the sheet is faster than picking each part up, because the sheet holds the part for you. A flush-trim router bit or an oscillating tool takes the stub off faster than a chisel on some materials and much slower on others. Cutting all the tabs first and snapping all the parts second beats doing both operations part by part, for the ordinary reason that batching does. And on parts where the edge gets sanded anyway, the tab stub costs almost nothing extra, which means the twenty seconds is really only twenty seconds on parts that would otherwise have come off the machine finished.
What is an onion skin and how does it compare?
Instead of discrete tabs, the whole profile is cut to within a few thousandths of the bottom and the part is held by a continuous thin membrane. The holding is far more even and the part cannot rock about a tab, which matters on small or intricate shapes. The cost profile is completely different from tabs: there is no per-tab operation, but there is a continuous edge to release around the whole perimeter, so the labour scales with perimeter rather than with a count. This page models discrete tabs, so it will not describe an onion skin correctly. What it will tell you is the cross-section your tab scheme leaves, which you can compare against the perimeter times the skin thickness to see which leaves more material.