MYP Design Jigs, Templates and Quality Control Notes

Repeatability Is a Planning Problem, Not a Skill Problem

1. Making one good part is a craft skill, and making the twentieth part identical to the first is something you design before you start.
2. Every time you measure and mark a part by hand, you add a chance of being half a millimetre out.
3. Those errors do not cancel each other, so twenty hand marked parts spread out into twenty slightly different sizes.
4. The fix is to build the accuracy into a device once, carefully, and then let the device hold it for every part after that.
5. Making that device takes maybe twenty minutes, and it usually pays for itself by about the fifth part.

Jigs, Templates and Formers Take the Measuring Out

1. A template is a flat pattern, usually 3 or 6 mm MDF or acrylic, that you draw round or run a router bearing along so every outline comes out the same.
2. A jig goes a step further, because it holds the work and guides the tool, so the operator no longer decides where the cut lands.
3. A drilling jig for shelf pins is a strip of hardwood with two hardened steel bushes set into it: clamp it to the cabinet side, drill through the bushes, and all forty holes line up.
4. A former is the third of the family, the shape material is bent or laminated against, and the forming article covers how it is used.
5. A jig has to be more accurate and harder wearing than the parts it makes, which is why the guide holes get steel bushes instead of relying on bare timber that a drill bit slowly chews wider.
6. Jigs make work safer as well as faster, since a push jig on a bandsaw keeps your fingers a fixed 80 mm from the blade on every single cut.

Tolerance: How Wrong a Part Is Allowed to Be

1. No process makes a part exactly on size, so a drawing states a tolerance: the range the finished part must land inside.
2. A shelf drawn as 400 mm plus or minus 1 mm is accepted anywhere from 399 to 401 mm and rejected at 402 mm.
3. Tight tolerances cost real money, because holding 0.05 mm needs better machines, slower cutting, more checking and more rejected parts than holding 1 mm.
4. So tighten the tolerance only where the function depends on it, such as the hole a bearing presses into, and leave the outside of the cover loose.
5. Tolerances add up along an assembly, so three parts each sitting 0.5 mm oversize give a stack 1.5 mm longer than the drawing.
6. Put tolerances on your own drawings even when they are loose, since nobody can check a part against "about 40 mm".

Go and No Go Gauges Answer Yes or No

1. A go and no go gauge replaces a measurement with a fit test, so there is no scale to read and no arithmetic to get wrong.
2. For a hole, the go end is made to the smallest size allowed and must enter, and the no go end is made to the largest size allowed and must not.
3. The part passes only when both conditions are true, so a hole that swallows the no go end is oversize and gets rejected.
4. Checking takes about two seconds and needs no training, which is why gauges are used on a line where a part arrives every thirty seconds.
5. You can make one from 6 mm acrylic: cut one slot a dowel must pass through and a second slot it must not, then label the two ends.
6. The weakness is that a gauge gives a pass or fail and never a number, so it cannot show that a cutter is slowly wearing and drifting towards the limit.

Quality Control Checks Parts, Quality Assurance Checks the System

1. Quality control is inspection: measuring parts during and after making, then accepting, reworking or scrapping each one.
2. Quality assurance is everything set up beforehand to stop faults happening: agreed standards, trained staff, maintained machines and written procedures.
3. Control catches a bad part after the material and the machine time have already been spent on it, while assurance is aimed at never making it.
4. A factory runs both, with a documented assurance system underneath and control checks at fixed points along the line.
5. In your own project, control is measuring each part against your drawing, and assurance is the jig, the checklist and the order of operations you wrote down before starting.

Sampling: Check Some, Trust All

1. Measuring all 10,000 parts in a batch would cost more than the faults it finds, so batch production checks a sample instead.
2. A typical scheme pulls one part in fifty at random, measures it and writes the result down.
3. If a sample fails, the line stops and everything back to the last good sample is checked, which is why the checking interval sets how much you risk.
4. Some tests wreck the part, such as loading a bracket until it bends, so sampling is the only possible method for those.
5. Plotting sample measurements on a chart shows a slow drift towards the limit while the parts are still good, so the cutter can be changed before anything is scrapped.
6. Sample size is a straight trade off: fewer checks cost less and let more faulty parts slip through to the customer.