No Drawings? Reverse Engineering Sheet Metal Parts from a Sample
If you have a sample part and no drawings, the part can usually be reproduced. What cannot be reproduced is the information that only you have: the function, the tolerance that matters, and which features are cosmetic. This page is about what we can measure, what we cannot, and what we need from you to close the gap.
Fulei Metal reverse-engineers simple sheet metal parts from samples, photographs and measurements, in a 2,500 m² plant in Ningbo, China, since 2016.
What we can measure from a sample
| Feature | Method | Accuracy |
|---|---|---|
| Overall outline and plate geometry | Flat-bed scan or measured points | Good — enough to reproduce the shape |
| Hole positions and diameters | Measured from a datum | Good for round holes; poor for irregular cutouts |
| Sheet thickness | Micrometer or calliper | Exact, within instrument error |
| Bend angles | Protractor or surface plate | Good, typically within a degree |
| Bend inside radius | Radius gauge | Approximate — the radius is inferred, not measured |
| Material grade | Spark test and magnetic test in-house; chemical analysis can be arranged where the grade has to be proven | Indicative from spark and magnet tests alone. 304 and 316 cannot be separated without chemical analysis |
| Surface finish | Visual and tactile comparison | Grit and grain direction can be matched on a sample panel |
What we cannot measure, and need you for
| Unknown | Why it matters | What we need from you |
|---|---|---|
| Which dimensions are functional | A hole that is 12.0 mm because it fits a boss must be held; a hole that is 12.0 mm because that is where the last one landed does not | The function of each feature, or the mating part |
| Internal features | Anything inside a closed box or behind a panel is invisible | Section drawings, or the part disassembled |
| Weld type and sequence | A finished weld shows the result, not how it was made | Whether the weld is structural or cosmetic |
| Hardware | Fasteners, inserts and rivets are often not visible | The hardware list, or the mating assembly |
| Material grade with certainty | 304 and 316 look identical; so do several aluminium alloys | The grade from your records, if it matters |
| Tolerance that was originally intended | A worn sample has drifted from its drawing | The tolerance standard you want held |
The three routes, by what you have
| What you have | Route | What we deliver |
|---|---|---|
| Sample part, no drawings | Measure, model, propose a tolerance standard for your approval | A first article measured against the sample, plus a dimensional report |
| Photographs only | Photograph-based reconstruction where the geometry is unambiguous | A first article, with the unmeasurable features flagged for your input |
| Sketch with some dimensions | Treat the dimensions as requirements and the sketch as geometry | A first article to the stated dimensions |
| Concept only, nothing physical | Not reverse engineering — this is design | An honest statement that this is design work, not fabrication |
What we need to quote
The list is the same one that applies to a drawing-based enquiry — see the drawing package checklist. What differs is that some of it has to come from you rather than from the file.
- The sample, or clear photographs from several angles — including any internal features
- The dimensions you already know, and which of them are functional
- Material, if you know it; otherwise we will propose one and tell you how confident we are
- Quantity — sample and annual volume
- Finish, or the sample’s finish to match
- Destination port, so freight can be quoted separately
The boundary that matters
This is the same boundary the DFM review runs up against on drawing-based work: the fabricator can tell you what is manufacturable and what it costs, and it cannot take over the engineering decisions that sit behind the part.
We reproduce parts. We do not certify that a reproduction is identical to a part we have never seen drawn. A first article is measured against your sample and reported, and the features we could not measure are listed as assumptions for your approval.
We also check the provenance question early. If a drawing or sample carries a third party’s name, a “PROPRIETARY” marking, or a title block that is not yours, we will raise it before any metal is cut — that conversation is cheaper at the quotation stage than after delivery.
Frequently asked questions
Can you copy a part exactly from one sample?
For most bent and welded sheet metal parts, yes, to within the accuracy of the measurement. What we cannot reproduce is information the sample does not carry: internal features, the original tolerance intent, and which dimensions were functional.
What if the part has internal features we cannot see?
Then we need either the part disassembled, a section drawing, or your description of what is inside. We will list what we could not verify rather than assume it.
How do you tell 304 from 316?
Not reliably without chemical analysis. If the grade matters — and in chloride exposure it does — tell us the grade from your records. We will not guess between them on a quotation.
Can you match a brushed finish?
Usually. Send a sample panel or a photograph in raking light and we will match grit and grain direction on a sample before production. Matching a finish that has been in service for years is harder, because the original surface has aged.
What lead times apply?
The published lead times are simple samples 3–7 days, complex first articles 15–20 days and production batches 25–35 days. A reverse-engineered first article is quoted against those numbers once drawing review fixes the scope — we confirm the date rather than guess one, because it depends on how much has to be modelled before anything is cut.
Do you sign an NDA before we send a sample?
Yes. Tell us and we will send ours, or we will review yours.
Send the sample and what you know
If you have a drawing rather than a sample, or one that may have come out of an AI drawing tool, the seven failures list explains what we check and why. For staged programmes, see how a prototype-to-production quotation is structured.
Send the sample, or photographs from several angles, with the dimensions you know and which of them are functional. We will tell you what we can measure, what we cannot, and what we would need from you to close the gap — before quoting.