A downloaded STL can look perfect on screen and sit at 96% of true size. Nothing flags it. The model loads, the proportions are believable, you design a chassis around it, and every single member comes out wrong by the same small factor. You find out at the welding table.
Fifteen minutes of measuring up front is the cheapest insurance in this entire workflow. Here's the order to do it in.
Scale first, because everything else depends on it
Find a dimension on the model you can verify independently, measure it, and compare.
Good candidates, best first:
- Bolt-hole spacing and bore centers. These come straight off manufacturing drawings and whoever exported the model almost never touches them.
- Wheelbase or track on a vehicle model, against published specs.
- Overall length of a component you can look up in a catalog.
- A face you can put calipers on, if the real part is sitting in front of you. Always the best option when you have it.
Measure that feature, divide the real dimension by the measured one, apply the result as a uniform scale factor.
If the number lands near 25.4 or 0.03937, you've got a plain unit mismatch — authored in inches, read as millimeters, or the reverse. STL carries no unit information whatsoever, so this is routine rather than exotic. Half the "wrong size" models you'll meet are exactly this.
Now check a second independent feature. If fixing the scale on the first one doesn't bring the second into agreement, stop. The model isn't uniformly scaled — it's been stretched on an axis, or assembled from parts at different scales. No single factor fixes that, and if you proceed you'll get a frame that's correct in two directions and wrong in the third, which is a much worse problem to debug than being wrong everywhere.
Put it where you want it, not where the exporter left it
Scale sorted, place the model deliberately. Downloaded geometry arrives rotated, offset a thousand millimeters from origin, or with "forward" on whichever axis the original author felt like.
Pick a convention and hold it for the whole project. Most builders put the origin somewhere meaningful and repeatable — centerline at the front axle is a common one — with one axis along the vehicle, one across, one vertical. In FrameDesigner, Y is vertical.
A meaningful origin pays for itself three times over: when you apply symmetry, when you're comparing left against right, and when you hand a dimension to somebody else and need it to mean something without a paragraph of explanation.
Reference models stay dimensionally separate from the frame, so scaling, rotating, repositioning and hiding them won't disturb geometry you've already drawn. Use that freely.
Measure what constrains you, not everything
Most dimensions on a model don't matter. The ones that matter are the ones that will stop a tube going where you want it. So hunt for constraints rather than cataloging the model:
- Hard points. Suspension pickups, engine and gearbox mounts, rack mounts, seat rails. Fixed, and the frame has to reach them.
- Swept volumes. Wheel travel, steering lock, driveshaft articulation, pedal throw. A tube that clears a static model will happily foul something in motion.
- Occupant space. Helmet clearance, shoulder width, legroom, and the room to physically climb in and out — which is always more than you think.
- Service access. Can you still get a filter off, reach a bleed nipple, lift the engine without cutting a tube?
Write them down as you find them. They become the skeleton the frame hangs on, and a written list stops you rediscovering the same constraint for the third time on a Sunday.
Decide your clearances now
Measurements off a model give you theoretical positions. Real assemblies move, deflect, and get built slightly off nominal. So pick your clearance numbers per category up front and design to those, not to the surface of the model.
Exhaust wants more room than a static bracket. Anything that moves under suspension travel needs clearance across its whole range, not at ride height. Anything hot needs room for heat as well as geometry — a primary that clears a harness by 8 mm cold is going to melt it.
You're done measuring when
You can answer three questions without reopening the model:
- What scale is this at, and how do I know?
- Where's the origin, and why there?
- Which dimensions is the frame not allowed to violate?
If any of those needs a second look, you aren't finished and the next phase will punish it.
Next tutorial builds a closed structure from scratch and runs it all the way to a fabrication document, so you can see the whole path from geometry to something you could hand to someone at a saw.
Comments
Loading comments…