- Sheet
- SVC-001
- Station
- SVC Service
- Logged
Holding Tolerance by Hand in a Small Workshop
How a small workshop holds tolerance by hand: torque, feel, and the bench tools that keep a Porsche 944 repair true, with checks you can repeat.
1,190 wordsReading 5 minSources read 1

Holding a tolerance by hand means controlling three things at once: the force you apply, the angle you apply it at, and the way the part is held while you work. A mechanic knows a fastener is tight enough when the specified torque is reached with a calibrated wrench, and not too tight when the wrench clicks or the angle stops advancing before the metal yields. In a small workshop, the tools that repay their cost are the ones that make those three things repeatable rather than lucky. The bench holding tools that keep a small piece still are the same idea in a smaller scale, and the principle travels: if the part moves, the reading is worthless.
What does it take to hold a tolerance by hand?
A tolerance is a permitted range, not a target. The 944 front wheel bearing, for example, is set by a rolling torque figure rather than a simple tightness, because the spacer and the nut together decide the preload. Holding that by hand requires a torque wrench that reads low values, a way to turn the hub smoothly, and a surface that does not move while you read the scale.
The same logic runs through the whole car. A control arm bushing is pressed to a depth, a ball joint nut is run to a figure, a steering rack is shimmed to a drag. Each of those is a number plus a method. The method is the part that a small workshop owns, and it is built from holding, measuring, and checking again.
The bench habit that carries over from fine metalwork is simple: hold the work so both hands are free, then measure before you commit. A bench peg, a vise with soft jaws, or a fixture that locates the part does the same job in a garage that a peg does at a jeweller's bench.
How does a mechanic know a fastener is tight enough and not too tight?
Three signals, used in order. First, the specification: the workshop manual gives a torque figure and often a sequence, and that figure assumes clean, lightly oiled threads unless it says otherwise. Second, the wrench: a click-type or beam-type torque wrench that has been checked against a known mass or a calibration service. Third, the feel: the last few degrees before the click should be smooth, and any sudden softness means the thread is pulling, not tightening.
Angle-controlled fasteners, such as some cylinder head bolts, are tightened to a torque and then turned a further angle. A protractor or a marked socket is the tool here, not the wrist. Stretch bolts are single-use for a reason: they have already passed their elastic limit, and reusing one removes the clamp load the designer counted on.
For fasteners without a published figure, the useful reference is the bolt grade and size. A grade 8.8 M10 bolt and a grade 12.9 M10 bolt do not share a torque value, and a torque table from a fastener standard is a better guide than memory. When in doubt, stop at the low end of the range and check the joint for movement after a short run.
Which hand tools repay their cost in a small workshop?
A short list, in the order it earns its place.
A torque wrench that covers the low range, roughly 5 to 50 Nm, because most of the small fasteners on a 944 live there. A second wrench for the higher range, or a calibration check on the one you have.
A set of flare nut wrenches for brake and fuel lines. An open-end wrench rounds a soft fitting; a flare nut wrench spreads the load across five flats.
A dial indicator with a magnetic base. Runout on a hub, end float on a shaft, and backlash in a gearbox all read as a number on a dial, and the number is the tolerance.
A beam-type or split-beam torque wrench for checking, because it reads continuously and shows drag rather than only a click.
A set of pin punches, a brass drift, and a dead-blow hammer. The brass drift moves a bearing race without marking the steel, and the dead-blow moves a part without bouncing.
A thread pitch gauge and a set of taps and dies. Chasing a damaged thread before assembly is cheaper than repairing it after.
A good vise with soft jaws, mounted at a height where your forearm is level. Holding is not a tool you buy once; it is a setup you build for each job.
Where does feel stop and measurement start?
Feel is a warning system, not a gauge. A bearing that feels rough when turned by hand is telling you to measure it, not to declare it good. A fastener that feels tight may still be short of its figure, and one that feels loose may already be at the yield point.
The practical rule is to use feel to decide when to pick up the measuring tool. Spin the hub, listen, then mount the dial indicator. Turn the fastener by hand until it seats, then bring the torque wrench. Feel narrows the search; the instrument closes it.
Temperature belongs in this conversation. Aluminium expands more than steel, so a torque figure taken on a cold engine block and one taken on a hot block are not the same joint. Most workshop manuals specify a cold figure for that reason, and the note is not decoration.
How do you check a repair before the car leaves the bench?
A short, repeatable sequence. Re-read the specification for every fastener you touched. Confirm the torque wrench was set to the right unit, because Nm and lb-ft are not interchangeable and a mis-set wrench is a silent error. Check the witness marks you made before disassembly, if you made them.
Then run the system by hand where possible. Turn the steering from lock to lock and listen. Rotate each wheel and feel for a tight spot. Press the brake pedal and hold it, watching for a slow drop that would point to a leaking seal.
Finally, record what you did. A notebook entry with the date, the figure, and the tool used turns a one-time repair into a known baseline. The next time the same joint comes apart, the note tells you whether the last setting held.
What does tolerance work look like over a season?
Tolerance is not a single event. A car that is driven hard will move slightly at every joint, and the workshop that checks a few figures at each service catches the drift before it becomes a failure. Pick the joints that matter: wheel bearings, suspension pivots, steering, and the fasteners that clamp the drivetrain.
Measure the same way each time, with the same tools, at the same temperature where the manual allows it. Consistency is what makes a small workshop's records useful, and it is the only way a hand-held figure becomes a trend rather than a guess.
The work is quiet and mostly unglamorous. It is also the difference between a repair that stays true and one that comes back.


