Take the PAIN out of PAINT

By Doug Smilor, Principal Consultant, Smilor Consulting LLC

Picture a Monday morning call from a customer. Parts coated three weeks ago are failing adhesion in the field. You pull the records. Pretreatment titrations were in spec. The oven temperature chart shows every zone at setpoint. The inspector signed off on film build and appearance. By every document you have, those parts were good.

Then someone remembers that second shift turned up the conveyor speed to make up for a downtime event. Oven air temperature never moved. Time at temperature did. Nobody connected line speed to cure, because nothing in the control system was designed to.

Every finishing manager has heard someone call it PAINT. Most of that pain is self-inflicted, and most of it is avoidable.

The fix isn’t a sharper inspector. It’s three quality planning tools automotive suppliers have used for decades (APQP, PFMEA, and the Control Plan) and a control system that actually carries them out. If you know these tools, I’d bet the gap described below is familiar too. If you don’t, this may be the most useful framework in manufacturing nobody ever showed you.

Why paint is different

Coating is a special process. Most of what makes a finish good (cure, adhesion, corrosion resistance, conversion coating weight) can’t be verified by looking at the part. It’s latent, or it takes a destructive test. An under-cured part looks exactly like a good one until it fails in the field.

That means final inspection can’t carry the load. Quality in a paint shop is made by holding the process inside a proven window every minute the line runs. The industry knows this; it’s why AIAG publishes CQI-12, the Coating System Assessment, and why automotive customers audit their coaters against it. The tools for doing it are just as well known, at least inside automotive. Outside of it, most finishing operations have never been introduced to them.

Where these tools come from

APQP, PFMEA, and the Control Plan are three of the AIAG Quality Core Tools, developed in the early 1990s when Chrysler, Ford, and GM agreed on a common set of expectations for their suppliers. They’re now built into IATF 16949 and most OEM supplier manuals, well beyond cars: agriculture, heavy equipment, powersports, anywhere a customer wants evidence instead of promises. AIAG released the 3rd edition of APQP and the first standalone Control Plan manual in 2024; the new Control Plan manual adds guidance for highly automated manufacturing.

Strip away the acronyms and the core idea fits in one sentence: before the line ever runs, decide what has to go right, what could go wrong, and exactly how you’ll know. Each tool handles one part of that.

What APQP, PFMEA, and Control Plans are supposed to do

APQP (Advanced Product Quality Planning) is the overall framework, a phased, gated process from planning through launch. It connects what the customer needs (OEM specifications, salt spray hours, gloss, film build, appearance) to the process characteristics that produce it, before launch instead of after the first escape. Along the way it identifies special characteristics, the handful of features that affect safety, regulatory compliance, fit, or function, and flags them for heightened control. It also produces the process flow diagram, load to unload, that anchors the next two documents.

PFMEA (Process Failure Mode and Effects Analysis) walks that flow one step at a time and asks what can go wrong, what it does to the customer, and why it happens: chemistry drift, rinse carryover, oven profile changes with load density or line speed, booth humidity, reclaim ratio, gun output, grounding. Each cause is rated for Severity, Occurrence, and Detection. Older practice multiplied those into a Risk Priority Number; the 2019 AIAG & VDA FMEA Handbook replaced that with Action Priority, which weights severity first so a serious failure can’t hide behind a favorable score elsewhere. Either way, effort goes where the risk is.

The PFMEA also separates two kinds of controls, and that distinction drives the rest of this article. Prevention controls stop the cause from happening. Detection controls catch the failure, or its cause, after it happens but before the part leaves. A conveyor interlocked so it can’t outrun the oven’s cure window is prevention. An alarm that flags loads that saw too little time at temperature is detection. An end-of-shift cure test is detection too, just very late. I have personally worked in organizations where lower RPN scores itself was justification for CapEx investment (Poka Yoke)

The Control Plan is where it becomes operational. For each critical characteristic, it defines the limits, how and how often you measure, who owns it, and what happens when it goes out: the reaction plan. It covers both product characteristics (film build, adhesion, gloss) and process parameters (zone temperatures, bath concentrations, line speed), and it matures in phases, from prototype to pre-launch to production. A single line for one oven characteristic might look like this:

Example: one Control Plan line for a powder cure oven

Illustrative only. Actual limits come from the coating supplier, the customer specification, and your own validation.

Done right, the three form one unbroken chain. A customer requirement becomes a special characteristic, the PFMEA identifies what threatens it, the Control Plan defines how it’s held, and the reaction plan defines what happens when it isn’t. Step numbers match across all three, so anyone can trace an item end to end, and the documents stay living: every escape and process change flows back through them. That’s how a plant moves from detection to prevention. You stop relying on the inspector at the end of the line and start knowing, in real time, that the process stayed in the window.

How it usually goes

If you’ve been through a few PPAP submissions, you know the reality often looks different:

• The PFMEA is written the week before PPAP, by one quality engineer working alone, and filed.
• Detection gets scored generously. “Visual inspection” earns credit against under-cure, even though you can’t see under-cure. Honestly rated, most cure, pretreatment, and adhesion failures belong at the top of the action list.
• The Control Plan says “operator checks oven temperature once per shift, recorded on log sheet” for a variable that can drift in minutes.
• Nobody updates them after launch. Line speed goes up, part families are added, chemistry changes, and the PFMEA still describes the line on PPAP day.

None of that happens because people don’t care. It happens because the documents and the equipment were never connected.

The gap between the “paperwork binder” and the PLC

Here’s where most plants come up short. The Control Plan lives in a binder or on a shared drive. The control system was programmed by an integrator who never saw it. The result is familiar:

• Alarm limits in the PLC don’t match Control Plan limits, or don’t exist.
• Oven temperature is monitored, but nothing links cure to conveyor speed or stoppages.
• Reaction plans exist only on paper. Nothing holds or flags suspect parts.
• Data isn’t historized, so when the customer asks, you can’t prove the part was in control.
• The highest-risk PFMEA items have no automated detection at all.

This is the part insiders don’t say out loud: in a coating line, most of the prevention and detection controls a credible PFMEA calls for can only realistically be delivered by the control system. A person with a clipboard every two hours is not a prevention control for a variable that moves in minutes.

So why hasn’t it been done? Honestly, because instrumenting the process has been painful. Every analog point meant shielded cable home-run to a panel, an input card, scaling, calibration, and commissioning time from someone who understood all of it. When budgets got tight, instrumentation got cut, and the clipboard filled the gap.

What IO-Link changes

For years, the pain of instrumenting the process was the reason we didn’t. That pain is mostly gone.

IO-Link (IEC 61131-9) is a point-to-point digital connection between a sensor and a master module. In practice, that means:

• Simple wiring. Standard unshielded three-wire M12 cordsets run to an IP67 master mounted at the equipment, not back to a distant panel.
• No analog headaches. Values arrive digitally, with no scaling errors, noise pickup, or input-card drift.
• Painless replacement. When a sensor fails, the master reloads its configuration onto the new one. Second shift swaps it and keeps running.
• Built-in diagnostics. Devices report their own health, so you know whether to trust the reading.
• Easy integration. Masters sit on PROFINET, EtherNet/IP, or similar networks, so data flows straight into the PLC and historian.

In a finishing system, that puts a long list within easy reach: pretreatment stage temperatures, conductivity, and spray header pressure to catch plugged nozzles; rinse flow; booth temperature, humidity, and filter differential pressure; oven zone temperatures through IO-Link temperature converters; conveyor speed from an encoder; powder hopper level and compressed air dew point.

IO-Link isn’t the answer for everything. Burner management stays on safety-rated systems (but some flame safety systems manufacturers are now offering communications gateways for monitoring and rest function), and some chemistry analyzers and oven profiling belong on their own technology. But for most of the characteristics on a typical Control Plan, the cost and effort barrier has dropped dramatically.

Make the control system enforce the Control Plan

With the excuse gone, the principle is simple: every critical characteristic on the Control Plan should map one-to-one to a point in the control system that is measured, limited, alarmed, and logged. Where the risk justifies it, the reaction plan should live in the logic (a hold, an interlock, a flagged load), not in a binder.

This is where the PFMEA starts paying for itself. When the control system catches an out-of-window condition and contains the product, you’ve honestly earned a better Detection rating. When it prevents the condition outright, you’ve lowered Occurrence. Controls investment becomes documented risk reduction, in the language your customer’s supplier quality engineer speaks.

Go back to that Monday morning call. With conveyor speed on an encoder, time at temperature calculated for each load, a limit taken straight from the Control Plan, and a reaction plan coded as “flag and hold any load outside the window,” second shift’s speed change gets caught at the oven exit, not three weeks later in the field.

A quick audit for your own line:

1. Pull your Control Plan and your PLC alarm list. Do the limits match?
2. Do the step numbers on your flow diagram, PFMEA, Control Plan, and PLC tag list line up?
3. For each high-risk PFMEA item, what detects it automatically? Is the Detection rating honest?
4. If conveyor speed changes or the line stops, what happens to cure, and who knows?
5. Could you prove, with data, that a part shipped three weeks ago was in control?
6. When an escape happens, does it update the PFMEA, the Control Plan, and the logic?

That last question closes the loop. Escapes feed the PFMEA, the PFMEA updates the Control Plan, and the Control Plan updates the controls.

Spec it in, don’t bolt it on

The cheapest time to get this right is during a capital project, when the controls specification is being written. APQP already provides the moment: the controls spec is a process design output, and it should be written from the PFMEA and Control Plan, not the other way around. Hand your integrator the Control Plan with the P&IDs. Retrofitting is possible, and IO-Link makes it far easier than it used to be, but it’s never cheaper than doing it once.

One of the biggest benefits of IO-Link is how much easier it can make maintenance. Instead of spending time tracking down a sensor issue or manually reconfiguring devices, maintenance teams can access diagnostic information and replace devices with configurations restored automatically. That means less time troubleshooting and more time keeping equipment running.

New to these tools? AIAG’s Quality Core Tools page is a good place to start. Old friends? Take the audit above to your own line.

The tools to take the PAIN out of PAINT exist. The only question is whether you spec them in before the line is built or after the first escape.

Remember, start at the beginning, not at the end.

Sources and further reading

• AIAG. Advanced Product Quality Planning (APQP), 3rd ed., and Control Plan, 1st ed. Automotive Industry Action Group, 2024. go.aiag.org/apqp-cp
• AIAG. “Our presale for the new APQP 3rd Edition and standalone Control Plan is finally here.” AIAG Blog, March 2024. blog.aiag.org
• AIAG & VDA. FMEA Handbook, 1st ed. Automotive Industry Action Group and Verband der Automobilindustrie, 2019. aiag.org
• AIAG. CQI-12: Special Process: Coating System Assessment, 3rd ed. Automotive Industry Action Group, 2020. go.aiag.org/special-process-assessments
• AIAG. “Quality Core Tools (APQP, CP, PPAP, FMEA, MSA, SPC).” aiag.org
• IATF. IATF 16949:2016, Quality Management System Requirements for Automotive Production and Relevant Service Parts Organizations. International Automotive Task Force, 2016.
• IEC. IEC 61131-9: Programmable controllers, Part 9: Single-drop digital communication interface for small sensors and actuators (SDCI). International Electrotechnical Commission.
• RAMCO Innovation “IO-Link advantages” RAMCOI.com
• IO-Link Community. “IO-Link: The First Globally Standardized IO Technology.” io-link.com

Smilor Consulting LLC provides vendor-neutral paint process engineering and capital project support for liquid and powder coating operations in the U.S. and Mexico.