Eight hundred brackets, all 6061-T6, all from the same bar lot, all with a clear Type II anodize callout in the title block. On paper it was a clean job. On the assembly line it was a dead stop: the M4 screws would not start in the tapped holes.
We had shipped the first two batches before anyone on the customer side tried to assemble one. When the third batch landed, their line rejected all of them. That is how we found out that a finish nobody thinks of as a dimension had eaten our threads.
What the drawing asked for, and what it did not say
The note read “ANODIZE PER MIL-A-8625 TYPE II CLASS 2, CLEAR, 0.0003 in MIN”. That is a normal, complete-sounding finish callout. What it did not carry was a single word about the threads. There was no masking note, no over-tap allowance, and no post-coat thread class. The tapped holes were dimensioned 8× M4×0.7 − 6H and left at that.
Here is the part that gets missed. Anodizing is not a paint. It is a conversion coating grown out of the aluminium itself, and it grows in both directions: into the part and out of the surface. On Type II sulfuric anodize you typically get 10 to 25 µm of coating, split roughly two-thirds inward and one-third outward. On hard anodize, Type III, you can see 50 µm and more, closer to an even split.
On the outside of a bracket that growth is invisible. Inside a tapped hole it is not. The coating builds on both flanks of every thread, and the flanks are what set the pitch diameter.
What we measured
We pulled 40 brackets back and gauged every hole. The go gauge would not enter. To get a number rather than a feeling, we measured pitch diameter over thread wires on 20 holes and compared them with 20 identical holes from the same lot that had never been anodized.
| Measured on M4×0.7 internal thread | Bare, uncoated | After Type II anodize | Change |
|---|---|---|---|
| Coating thickness, average | — | 18 µm | — |
| Pitch diameter, average | 3.612 mm | 3.574 mm | −0.038 mm |
| 6H pitch diameter, minimum | 3.545 mm | — | — |
| Go gauge, 20 holes | 20 pass | 0 pass | fail |
A 0.038 mm loss on a thread with only 0.118 mm of pitch-diameter tolerance is a third of the entire window, spent on a finish. Our bare threads sat comfortably in the middle of 6H. After coating they were below the minimum, and the go gauge told the truth the drawing had not.
The arithmetic is not subtle. Coating grows on two flanks, so pitch diameter shrinks by roughly twice the coating thickness. Eighteen microns per flank is about 36 µm of closure, and we measured 38. The finish is doing exactly what the physics says it will do. Nobody had asked it not to.
The rework, in hours and dollars
Scrapping 800 brackets was not an option, so we re-tapped. Every hole got a 6H bottoming tap run through it by hand at the bench, with the tap backed out and cleared every few turns so it did not load up and snap. That is slow work, and it cuts coating off the flanks, which leaves bare aluminium inside the hole.
Eight hundred brackets at eight holes each is 6,400 holes. At 22 seconds a hole that is 39 hours of bench time, and at USD 45 an hour that is USD 1,755 before inspection. We broke taps in 34 brackets and scrapped them at USD 6.80 of floor cost each, another USD 231. Add re-gauging and repacking and the job carried about USD 2,400 of pure rework that never appeared on any quote.
We also had to tell the customer that the reworked threads now had exposed aluminium in them. In a dry indoor enclosure that is a manageable risk. In a washdown environment it is a corrosion site, and we would not have shipped it that way without a written sign-off.
Four ways to spec it so this does not happen
The cleanest answer is to mask the threads. A silicone plug in each hole keeps the coating out entirely, so the thread stays 6H as tapped. It costs a few seconds a hole and some handling discipline, and on a bracket with eight threads that is a real line item, but it is far cheaper than re-tapping 6,400 holes by hand.
The second answer is to over-tap before coating. You cut the thread slightly oversize so that after the coating closes it down, the finished thread lands in 6H. This only works if you know the coating thickness and control it, and it needs a thread class callout that says so. A plain 6H on the drawing does not authorise your supplier to do it.
The third is to specify the thread class after coating. Writing “THREADS TO 6H AFTER ANODIZE” puts the problem where it belongs, on the finishing step, and it forces a real conversation about masking or over-tap instead of a silent assumption.
The fourth is to ask whether the thread needs to be anodized at all. On many brackets the tapped holes take fasteners that are themselves plated, and a masked, bare thread is the right engineering answer. If the part sees salt or washdown, the answer may be different, and that is a decision to make with the finish schedule in front of you rather than after the brackets are coated.
The same problem runs the other way
Internal threads close up. External threads grow. A coated M4 stud gains roughly 36 µm on the major diameter and can stop entering a nut it used to thread into freely. Any time you anodize a part that has threads, male or female, the finish is a dimension and it needs a number on the drawing.
Type III hard anodize is worse. The coating is thicker, so the thread closure is larger, and hard coat is also harder to mask cleanly. If a print calls for hard anodize on a threaded part and says nothing about threads, expect a problem.
Our anodizing cell masks threads by default now, and we raise the thread question before the parts are coated rather than after. If you are weighing finishes, the comparison of anodizing against powder coating covers where each one belongs, and powder coat has its own thread-growth problem for the same reason. The dimension side of it sits in our thread specification guide, and the alloy behaviour is on the 6061 aluminium page. Both bracket runs above came off our CNC milling cells.
If your print has a finish callout and a threaded hole on the same sheet, send it through instant quote and we will flag the conflict before we cut metal. The team confirms within 1 business day. Any material figure in this post is a budget estimate, not a live rate, and firm pricing comes from a human review of your drawing.