Many coating failures that surface in year two were decided at the wash line, days before the part ever reached the booth — in the sheet metal pretreatment and conversion coating stage that leaves no visible trace on the finished enclosure.
Key takeaways
- The conversion layer is invisible on the delivered part, so decide on the drawing whether you are buying looks or corrosion life.
- Declare every metal in the assembly, including installed hardware, before a chemistry is chosen.
- Ask which substrates the proposed conversion coating covers, and which it does not.
- Require bath and rinse checks logged per run, not visual approval after coating.
- Name the acceptance test and frequency; require non-drainable geometry flagged before first piece.
- A salt spray figure quoted without the pretreatment identified is marketing, not evidence.
The enclosure left the shop looking perfect: colour matched, film build in range, adhesion clean at first article. Eighteen months later a telecom rack is bleeding rust along the seam where two sheets overlap, and the film lifts off in one ribbon. Nobody changed the powder or the cure. What differed was the clean, the conversion coating and the rinses before the booth: never specified, never logged, never designed around that seam.
You cannot inspect that layer on the delivered part. You can specify it, require evidence for it, and ask a few questions that reveal more than any certificate — which is the entire job of a sheet metal pretreatment specification.
Why sheet metal pretreatment decides whether the coating survives
Two mechanisms do the work, and they fail differently.
The first is mechanical keying. A conversion layer is not laid on the metal, it is grown out of it, and where it is crystalline it gives the powder a profile to lock into. Without it the coating rides on a surface too slick to hold it, and the signature failure is peeling after an impact or a thermal cycle.
The second is under-film corrosion inhibition. A scratch, a cut edge, a corner where film build thins: each is where an electrochemical cell can start. The conversion layer slows that cell, deciding whether a scratch stays a scratch or becomes a front creeping sideways under an intact film.
Enclosures are the worst case for both: a bracket fails visibly and cheaply, an enclosure is deployed and sealed, so the failure is found in year two at a site visit.
One spoiler before the chemistry. The most common enclosure failure described in this article is not a bath that drifted. It is a folded seam that took solution in during the clean and dried to salt — and no dryer setting on the line fixes that.
So the question is not "do you pretreat" — everyone does. It is which conversion coating, on which substrate mix, verified how, logged how often. For where this sits in the wider set of sheet metal surface finishing options, pretreatment gates everything downstream of it, and specifying corrosion-resistant coatings only pays off once this layer is right.
The stages of a pretreatment washer, and what each stage is actually for
Most washer lines are built from the same five functions, whatever the stage count on the nameplate. Read the table as a checklist: for each function, ask which stages your coater runs and what they control.
| Stage | Function | What the coater controls | First check when it drifts |
|---|---|---|---|
| 1 · Alkaline clean | Remove forming oils, drawing lubricant, mill oil, metal fines | Concentration, temperature, contact time, oil loading, impingement | Water-break test fails: surface not hydrophilic |
| 2 · Rinse | Strip spent cleaner before it reaches the conversion bath | Overflow rate, water make-up, nozzle condition | Carry-over destabilises the conversion bath over time |
| 3 · Conversion coating | Grow the conversion layer on the substrate | Chemistry family, substrate match, concentration, temperature, coverage | Patchy or absent layer, then peeling months later |
| 4 · Final rinse, often deionised | Remove loose crystals and dissolved solids before they dry in place | Rinse water quality and freshness | Spotting, then blistering at those spots |
| 5 · Dry-off oven | Remove free moisture before the booth | Part temperature, hold time, cavity drainage, air movement | Moisture carried into cure: outgassing, pinholes, craters |
Cleaning is not cosmetic
What has to come off a fabricated enclosure is more than finger marks: drawing lubricant, cutting fluid, weld smoke, heat tint, cut-edge oxide and seam dust. Alkaline cleaning handles the organic load, not a thermally produced oxide or weld heat tint. Mechanical preparation of weld zones and cut edges is a separate step with its own acceptance criteria, which is why weld condition entering the paint line belongs in the same conversation as bath chemistry. Decide whether your drawing requires it before the wash, and who signs it off.
Rinsing is the most under-designed stage
Rinsing is where lines quietly degrade. In a counter-flow rinse the cleanest water meets the part last and flows backwards stage to stage, so a line that looks generously rinsed may finish on water that has already done duty upstream. Nozzles are the cheapest component on the line and the most neglected: a partially blocked one leaves a strip of under-rinsed surface no inspection catches. Ask what the stages overflow at, when nozzles were last pulled and checked, and whether that is recorded.
Final rinse water quality is where mineral deposits are created
The principle explains a failure mode that gets misdiagnosed for years. Whatever is dissolved in the final rinse water stays on the part when the droplet evaporates: that residue sits under the coating, on the cleanest surface you will ever have, and it is hygroscopic. There is no universal number to quote, because the limit depends on chemistry, line and substrate. Ask what your coater controls in the final rinse, what their limit is, and whether it is logged.
Dry-off and flash rust
Between dry-off and the booth there is a window in which bare prepared steel re-oxidises. Thin gauge cools fast, humidity shortens the interval, and a part holding water in a pocket will flash rust although the exposed face looks clean. That window is a line-discipline variable, not a constant: ask what the maximum interval is on their line, and what happens to a part that misses it.
Choosing a conversion chemistry: iron phosphate vs zinc phosphate vs zirconium
Table first, then the judgment. This is the chapter most guides stop at, and it is the least decisive part of the decision.
| Iron phosphate | Zinc phosphate | Zirconium, chrome-free | |
|---|---|---|---|
| Typical use case | Interior, low-exposure steel | Outdoor, corrosion-critical service | Mixed-metal lines, chrome-free replacement |
| Substrate range | Mainly steel, weaker on aluminium and zinc-coated surfaces | Mainly steel | Steel, aluminium and galvanized in one line |
| Layer nature | Amorphous, thin | Crystalline, larger crystals, high surface area | Ultra-thin oxide, often with a silane component |
| Coating weight, typical published range only — confirm against your chemical supplier's technical data sheet | 0.3–1.0 g/m² | 1.5–4.0 g/m² | Not comparable; reported as film thickness in tens of nanometres |
| Process complexity | Lowest; some clean and convert in one step | Highest; more stages, tighter control | Low to moderate, often near ambient |
| Environmental load | Moderate; generates sludge | Higher sludge, heavy-metal waste treatment | Lowest; no heavy-metal stream, no Cr(VI) |
Iron phosphate
Its advantage is real: some formulations clean and convert in one step, removing tanks, floor space and a rinse from the line. Finishers who still run it argue fairly that it is not outdated, it is proven, for the case it was designed for. Decide by service case, not habit.
Zinc phosphate
A heavier crystalline layer, the conventional choice for outdoor and corrosion-critical service, and the most expensive to run. The cost is not the chemistry but the stage count and the control burden: published descriptions disagree on how many stages a zinc phosphate line needs, which tells you the answer depends on the line and the chemistry supplier. Ask what the extra control buys on your substrate, and whether your coater runs it or a truncated version.
Zirconium and other chrome-free systems
The mechanism is genuinely different: there is no crystal layer. The deposit is an ultra-thin oxide, often with a silane component that forms its own network and bonds to the coating chemically. One consequence follows from that. Because the layer is orders of magnitude thinner than a phosphate, it cannot deliver the same mechanical interlocking, yet adhesion on zirconium conversion coatings is routinely excellent, because the bond is chemical rather than mechanical. It is also less forgiving of a poor clean, since there is no crystal reservoir to compensate.
One contra-indication most vendors will not volunteer
Zirconium is not universally better, and the evidence says so on one substrate. The EU ENABLE project report found that zirconium-based systems generally pass the requirements for hot-rolled steel, galvanized material and aluminium, but that results vary a lot on cold-rolled steel and depend on the specific alloy. If your enclosure body is cold-rolled steel, that is why you should ask for corrosion evidence on your actual alloy rather than accept a general claim of superiority.
Our default position, stated plainly: on a mixed-metal workload, a single chemistry that tolerates the whole mix is usually the better engineering answer. On a single-substrate cold-rolled steel enclosure, zinc phosphate still earns its cost, and switching away from it should require alloy-level corrosion evidence rather than a preference for the newer process.
Chrome-free, trivalent, and the no-rinse trade-off
Hexavalent chromium is restricted under RoHS and REACH, which is why chrome-free conversion coatings exist. Trivalent chromium and zirconium systems both avoid Cr(VI); the regulatory detail belongs with chrome-free conversion and Cr(VI) exposure rather than on a supplier's marketing page, and the scope of your own obligation belongs with your compliance owner.
One variant deserves a warning. A no-rinse conversion coating drops a rinse stage: solution goes on and stays on, so fewer tanks and less effluent — but the concentration applied is the concentration that dries in place, demanding tighter control of bath and drainage. In a high-variability shop that trade looks like a saving and behaves like a risk. Ask which of the three your coater proposes, and why, against your substrate mix.
Where sheet metal pretreatment actually goes wrong on an enclosure
Everything above describes flat coupon behaviour. Your enclosure is not a coupon, and this is where pretreatment failures on fabricated parts originate.
Folded seams, hems and lap joints
A 180 degree hem or an unsealed lap joint is a capillary. It takes solution in during the clean, holds it through the rinses, and does not give it back in the dry-off. Spray cannot reach inside it, so coverage in the seam is whatever the capillary delivered.
The remedy is a design decision made before the part reaches the line: seal the joint with a continuous weld or sealant so there is no cavity, or design it as a wept joint that drains and dries. The mechanical options are set out in the guide to hem bends, folded edges and corner joints. Decide on the drawing, for every hem and lap, whether it is sealed or wept.
Why drying does not save it
This is the point most specifications miss, and it kills the most common proposed fix. Drying does not solve entrapment. Practitioner discussions on the finishing.com hotline put it plainly: solution trapped in a crevice dries to a salt residue, the salts are hygroscopic, they reabsorb humidity from the air, and the crevice bleeds again. No amount of extra drying fixes that, because the problem is not water, it is salt.
The chain runs: solution enters the seam, dries to salt, absorbs atmospheric moisture, wicks out of the lowest edge, then corrodes under the film at exactly that line. The two cases usually cited are a telecom equipment rack rusting at the exit seam of two overlapping steel surfaces, and hardware rusting where two surfaces sat one to two millimetres apart. Neither was a coating defect. Both were geometry.
So when a longer dry-off is proposed as the fix for seam bleed, ask what happens to the dissolved solids. If the answer is only about water, the failure has not been understood.
Box sections and internal pockets
Closed tube, internal stiffener channels, deep drawn boxes and double-wall doors all share one problem: no flow, no drain and no spray impingement inside. The sealed volume you designed for ingress protection is the same one that cannot be rinsed or inspected, and a double-wall door is the classic case — two skins, an air gap, and a cavity that holds cleaner until it dries.
There is no free fix, only a declared one. A drain or vent hole is an opening in the enclosure, so it has to be reconciled with the sealing design rather than added after it: the gasket and seam strategy in sealed enclosures and gasket sealing is what decides whether a drain path costs you the rating you specified. Where a cavity can be sealed instead of drained, seal it. Where it can be neither, write it down as an accepted risk, name who accepted it, and put it in the qualification record rather than discovering it in year two. Ask which internal volumes your coater considers unrinsable, then decide, for each one, whether it is sealed, vented, or accepted in writing.
Blind holes and installed hardware
Blind tapped bosses, studs and standoffs already pressed in, captive nuts, a hinge barrel: each is a reservoir and a shadowed crevice in the conversion spray. A boss that traps cleaner bleeds into the coating at the worst location, a threaded joint someone will later torque. The design logic already used for racking has to be applied one stage earlier: hang points, vent holes and drainage through the coating line covers the coating-stage version. Decide which hardware is installed before finishing, and require a drain path for every feature that is.
Mixed-metal assemblies
This is the least discussed and most common failure on real enclosures. One workload, four substrates: a cold-rolled steel body, an aluminium subplate, galvanized brackets, installed hardware. A chemistry tuned for the steel body underperforms on the aluminium and the zinc-coated parts, and on a single-pass line there is no way to treat each separately. The options are to split the run by substrate, which costs scheduling and handling, or to choose a chemistry that tolerates the whole mix. That, and not novelty, is the real argument for zirconium on enclosures, and why the substrate mix belongs in the RFQ, starting with material selection for the substrate mix.
Put one line on your drawing: list every metal present as coated. It costs nothing at the design stage and it is the single field that most often decides which chemistry the line can run for you at all.
Bath control: the one question worth asking
You will not audit titration technique, and you should not try. You need one question and a look at a log.
Every bath is controlled on the same variables: concentration, temperature and contact time. Which drifts first depends on the chemistry. Ask the question that discriminates: what were the values at the start and end of the last shift, and can I see the log. Some defence finishers advertise chemistry monitored and adjusted every shift with logs per run; a coater who logs by shift is a lower purchasing risk than one offering only visual approval after coating. Rinse and nozzle checks follow the same rule — ask for them on the same log, not as a separate promise.
Verifying adhesion before you approve production
The water-break test
Free, takes seconds, and the best gate you have: clean metal holds an unbroken sheet of water, while residual oil makes the film bead and break. Run it after cleaning and again as a final gate, on the actual part rather than a coupon beside it. Require it on first article and record the result.
Cross-cut adhesion
Two conventions cover the same idea, ISO 2409 and ASTM D3359. Pick one and name it on the drawing. What a cross-cut tells you is whether the film is bonded at that moment at that spot; it says nothing about resistance to under-film creep. The classic case is a batch that passed cross-cut at first article and on every production sample and still crept from the scribe in the field, because adhesion at time zero and corrosion resistance in month eighteen are different properties.
Where the demonstration tests stop
Corrosion evidence is the piece that gets skipped, because it is slow. Neutral salt spray to ASTM B117 or ISO 9227 is the default, and hours is the most quoted number in this subject. There is a documented problem with using it to choose a pretreatment: the EU ENABLE project report found that cyclic test methods and neutral salt spray do not rank pretreatment systems in the same order. If two accepted methods disagree on the ranking, then a ranking cannot be read off a single one of them, and that is the whole difficulty. The mechanics and limits are set out in neutral salt spray testing — and what the hours don't tell you.
So make it a rule rather than a preference: treat a quoted hours figure as unverified unless the pretreatment is named in the same sentence and a cyclic method has confirmed the ranking on your substrate. Ask for the method, duration, substrate, pretreatment and the creep evaluation convention.
Failure, likely cause, first check
Our powder coating defect troubleshooting guide is where to start when you already have a defect on the bench. That post starts from the symptom and works back; this one starts from the process and works forward. The table covers failures originating upstream of the booth.
| Symptom seen after coating | Likely pretreatment origin | First check |
|---|---|---|
| Sheet or edge peeling after impact, bend or thermal cycle | Missing, thin or patchy conversion layer; residual oil | Cross-cut and water-break result for that batch, not the line average |
| Blistering after cure or after humidity exposure | Entrapped moisture, or dissolved solids from the final rinse | Was there a drainage path out of every hem, boss and pocket |
| Blistering concentrated along one seam line | Dried salt bleeding from a crevice that did dry | Re-dry the part; if the stain returns, seal the joint |
| Craters or fish-eyes | Silicone or cleaner residue, on the part or in the rinse | Rinse stage condition and the incoming contamination source |
| Pinholes clustered over weld zones | Outgassing from a porous or contaminated weld | Weld condition and mechanical preparation before the wash |
| Uneven colour or film build across one part | Inconsistent conversion coverage across a mixed substrate | Was it run as one chemistry despite two or three metals |
Specifying pretreatment on your drawing and in your RFQ
Name the service case before the chemistry: state where the enclosure will live and what it will face, then let the coating operation propose a process against that. Specifying a chemistry you have not validated against a substrate you have not declared is how projects end up with a defensible specification for the wrong part.
PRETREATMENT REQUIREMENT — [PROJECT / PART NUMBER]
1. Service case: interior / outdoor sheltered / outdoor unsheltered / washdown — state which
2. Substrate mix: list every metal present as coated, including hardware already installed
3. Pretreatment: process family proposed by the coating operation against items 1 and 2, in writing before first piece
4. Process control: bath and rinse checks performed and logged per run; logs available on request
5. Acceptance: cross-cut adhesion per ISO 2409 or ASTM D3359, at first article and at [stated batch frequency]
6. Geometry: coating operation to flag any feature it considers non-rinsable or non-drainable before first piece
7. Corrosion evidence, if required: method, duration, substrate, pretreatment identified, creep evaluation convention
8. Sealed or wept: every hem, lap and internal cavity declared as one or the otherItem 6 is the cheapest ask in the document and the one enclosure projects die on: a coater who raises a drainage problem before first piece has saved you a field campaign. The remaining commercial fields are covered in what to put in an RFQ. Send the drawing with that block filled in and we will return a quote within 24 hours; minimum order quantity is 1.
Six questions that tell you more than a certificate
- Which conversion coating do you propose, and on which of my substrates is it qualified? A good answer also says what it does not cover.
- What do you log, and how often? A good answer lists bath and rinse checks by shift.
- What is your maximum interval between dry-off and coating? A good answer gives a number and a rule for a part that misses it.
- Which features on this enclosure would you call non-drainable? A good answer arrives before first piece and cites your geometry.
- Show me corrosion evidence for this pretreatment on this substrate, and name the test method. A good answer names the pretreatment alongside the hours.
- Show me an enclosure you rejected for non-drainable geometry. A supplier who has never rejected one has never looked.
Where the decision sits
Specify the service case, declare every metal, require the logs, name the acceptance test. Those four decisions are yours, because only you know where the enclosure will stand for the next ten years.
Coating and chemical selection should be governed by the performance requirement you specify, not by a house default. SanJun builds the enclosure and supplies an evidence pack with the build: material records, inspection results, and the documentation we control. We do not issue coating certificates and we do not make compliance declarations on the coating's behalf.
If you want the geometry reviewed before it becomes a specification, send the model for a free CAD design and DFM review. We will flag the hems, blind bosses and mixed-metal interfaces that will not rinse, and quote within 24 hours.
Frequently asked questions
Is zirconium always better than zinc phosphate?
No, and the answer is substrate-dependent. Zirconium systems are reported to perform well on hot-rolled steel, galvanized material and aluminium, but the EU ENABLE project found that results vary a lot on cold-rolled steel depending on the specific alloy. If your enclosure body is cold-rolled steel, ask for corrosion evidence on your alloy before accepting a switch.
How many washer stages do I need?
Ask for the number your chemistry supplier specifies for your substrate mix and service case, rather than a figure copied from an article. Published descriptions of zinc phosphate lines disagree on stage count, which tells you the answer depends on the line and the chemistry supplier, not on a rule of thumb.
Can I skip pretreatment and rely on the powder?
No. Powder is a film sitting on the substrate, and without a conversion layer it has neither the mechanical key nor the under-film inhibition that makes it durable. The predictable outcome is peeling after impact or thermal cycling, and creep from cut edges and scratches.
Why did my coating pass cross-cut but fail in the field?
Because the two tests measure different things. A cross-cut per ISO 2409 or ASTM D3359 tells you the film is bonded now, at that spot, and says nothing about resistance to under-film creep, which is governed by the conversion layer and by whatever is trapped at seams, bosses and pockets. The usual finding is that adhesion was never the problem; corrosion starting at a crevice was.
Do we need chrome-free conversion for compliance reasons?
Chrome-free conversion coatings exist because hexavalent chromium is restricted under RoHS and REACH, so the principle is to reduce Cr(VI) exposure rather than to treat one chemistry as automatically compliant. The scope of your obligation depends on your market and your product, so have your compliance owner confirm the wording and ask the coating operation which chemistry they run.