Working with Zero Tolerance in GD&T — What It Actually Means and How to Get It Right
What Is The Zero Tolerance Law?
The Zero Tolerance Law, or more accurately zero tolerance in the context of geometric dimensioning and tolerancing, is a concept that shows up on engineering drawings when a feature of size is called out with a tolerance of zero at its maximum material condition. In practice this means you're applying a positional or profile tolerance that starts at zero and grows as the feature departs from its tightest allowable size boundary. People tend to treat it as a blunt enforcement tool when it's really a mathematical relationship tied to MMC and LMC modifiers. Here is the straightforward mechanics. When you specify a positional tolerance of zero at MMC, the actual permissible zone depends entirely on the part size. Take a shaft with a diameter tolerance of 10.00 ±0.05 mm. At the upper limit of 10.05 the positional tolerance is exactly zero — the shaft must land perfectly on the true position. But at the lower limit of 9.95 you have gained 0.10 mm of bonus tolerance. That is the fundamental shift people miss when they first encounter this. The tolerance zone is not fixed. It expands proportionally with how far the feature strays from maximum material condition. To apply this correctly you need to understand the relationship between feature size and the resulting datum reference frame. The datum features themselves should ideally be referenced at their maximum material boundaries so the datum simulator represents the worst case assembly scenario. If your datum references are floating or referenced at virtual condition incorrectly, your zero tolerance callout becomes essentially meaningless during inspection.
Common Pitfalls That Wreck Parts
One of the most frequent problems I see is operators and even some quality technicians treating zero tolerance as an impossible requirement. They assume any deviation is a reject. That is wrong. The feature size tolerance controls the allowable departure, and that departure directly feeds into bonus tolerance. A part sized at the lower limit with a small positional offset is often perfectly functional. Rejecting it immediately based on a literal reading of "zero" wastes material and time. Another issue is using zero tolerance on features that are fundamentally variable in their function. Threads, for example. Calling zero tolerance on a thread positional requirement without considering the pitch diameter versus major diameter creates inspection nightmares. The thread is not a simple cylinder, and the feature of size selection matters enormously for how the bonus tolerance calculates. There is also the problem of over-constraining datums. When all three datum features are held at MMC with zero tolerance positions on secondary features, you effectively lock the assembly into a single hard configuration. Minor surface irregularities on the datum simulators become blocking issues. I have seen entire lots of otherwise good parts get scrapped because the datum pins were worn by 0.02 mm and the zero tolerance callouts left no room for that wear to be compensated.
A Real Edge Case I Ran Into
On a housing plate assembly I was working with, the drawing specified zero tolerance position for a series of dowel holes at MMC, referenced from a machined face and two precision pin holes. The parts measured fine in calipers and CMM readings showed all positions within a 0.08 mm envelope. The inspector rejected the lot because every single part exceeded the stated zero positional tolerance regardless of their actual size deviation. The designer had not accounted for the fact that these holes were sized at 10.02 to 10.06 when the basic size was 10.00, which should have granted up to 0.06 mm of bonus tolerance per hole. The fix was not to change the drawing but to go back and clarify that the positional tolerance was indeed at MMC and that the bonus calculation applied. Once the inspector understood that the zero was at MMC and not a universal constant, the parts that were within functional gauge acceptance were reclassified as conforming. The total rework and scrap on that run came to roughly 400 parts across a batch of 2,800. It cost about three days of delays and a formal corrective action report. After that we instituted a rule that any zero tolerance callout on our shop drawings had to be accompanied by an explicit note stating "at MMC, bonus tolerance applies per ASME Y14.5" so there was no ambiguity on the floor.
Get the Full Details
When Zero Tolerance Makes Sense
This approach works best for features that must align precisely under worst case assembly conditions. Rolling element bearing seats are a classic example. The inner race needs to sit exactly where the drawing says it sits relative to the shaft datum when the shaft is at its largest allowable size. If you allow positional variation at that condition the bearing alignment degrades and you get premature failure. The bonus tolerance still applies when the shaft is undersized, and that is usually desirable because a smaller shaft naturally has more clearance in the bearing fit anyway. It also makes sense for tooling and fixture components where repeatable positioning matters more than individual part variability. A jig plate with zero tolerance positions for drill bushings will produce consistent hole locations across thousands of parts. The key is that the manufacturing process must be capable of holding the size tolerance tightly enough that the bonus tolerance you gain is actually useful and not just theoretical on paper.
When to Avoid It
Zero tolerance positional callouts are a poor choice for cast or forged features where the process variability is inherently high. You are setting yourself up for scrap. The same applies to long slender features where thermal expansion during measurement can shift positions by more than the allowable bonus tolerance band. I once had a 600 mm long aluminum bracket with zero tolerance positions on mounting tabs. Ambient temperature changes of just five degrees between the machine tool and the inspection room shifted the measured positions enough to cause argument between machining and quality. We switched the callout to LMC with a clearly defined bonus tolerance and the arguments stopped immediately. Thin-walled parts are another category where zero tolerance at MMC rarely works well. The part deflects under clamping force during both machining and inspection, and the measured values fluctuate depending on how the part is secured. The solution is usually to specify the tolerance at LMC or to add a non-rigid datum modifier that accounts for the natural flexibility of the part geometry.
Inspection and Verification Approach
Measuring zero tolerance features requires a functional gauge approach more than it requires raw CMM data. A go/no-go fixture that simulates the mating part at maximum material condition will tell you faster and more accurately whether the part is acceptable than a CMM report showing individual positional deviations. The CMM data is still useful for process control and trend analysis, but the acceptance decision should be based on whether the part assembles. If you are using a CMM, make sure the datum reference frame is established correctly and that the software is calculating bonus tolerance based on the actual measured feature size, not the nominal size. I have seen reports generated with the positional tolerance defaulted to zero regardless of the measured size because the inspection programmer did not properly define the feature of size and its associated tolerance in the software. That is a software setup issue, not a math issue, and it costs people hours of rework before they catch it.
Bottom Line
Zero tolerance in GD&T is not a demand for perfection. It is a design intent statement about worst case assembly behavior. The bonus tolerance that comes from size deviation is the safety valve that makes the system practical. Understand the MMC relationship, verify with functional gauges when possible, and do not apply it to processes that cannot hold the underlying size tolerances. The parts will build, the inspections will make sense, and you will save yourself from a lot of avoidable scrap.
