Final rinse conductivity in cataphoresis on anodized aluminum — one haze and the part is cosmetic scrap.

The problem

The rinse that decides the cosmetics of the profile.

On a cataphoresis line running anodized aluminum, the final rinse is not a formality: it is the cut-off between what goes into the oven clean and what carries dissolved paint with it. Rinse conductivity is the direct indicator of that drag-out. It rises when too much product comes in and the demineralised water make-up does not compensate. And the symptom does not show up at the rinse: it shows up after the oven, as haze, staining or uneven gloss on profiles that already cost time and energy.

1. What comes into the rinse — the previous rack, the paint drag-out from the cataphoresis bath, the thermal mass of the profile. Data that lives between the tunnel PLC and the shift leader's head.
2. What happens in the rinse — conductivity, DI flow rate, temperature. Usually on the chemical supplier's panel, sometimes never leaving it.
3. What comes out of the rinse — the curing oven and cosmetic quality control, which pick up the cost of everything above. And architectural glazing does not forgive hazes.

The paint shop manager knows this — but manual sampling once a shift does not cover long profiles with frequent SKU changes. What is critical is the slow drift of the rinse, the one you notice when five racks have already been cured with haze.

How it fits the IRIS system

iLEAN does not add another panel — it seals the cracks between chemical control, the line and the customer.

The cataphoresis rinse problem is not a lack of instrumentation: it is that the rinse data, the paint bath data and the cosmetics measured by quality control live on islands, and the technician finds out too late. iLEAN acts as the filler that closes those gaps, without asking you to change the tunnel or the chemical supplier.

Edge sees rinse conductivity every second. Connect reads the flow rate and the times wherever they sit. The agent cross-checks against the production order and, if the drift points to cosmetic haze, holds the rack before the oven. The person signs off — never the other way round.

The three iLEAN pieces applied to the final rinse of cataphoresis on anodized aluminum:

• Edge — terminal with vision + probe integration in the final rinse tank. It reads conductivity in seconds, cross-checks it against the PLC reading and, if the drift crosses the threshold, triggers a traffic light or a rack hold before the oven. It works with no network.
• Connect — captures the DI flow rate and the immersion times from the tunnel control, even if they live on an old panel belonging to the chemical supplier. And it captures from outside: a customer email with a change of shade or a tighter finish requirement, a message from the anodizing lead about a change in the prior sealing step.
• Agent — cross-checks conductivity, flow rate, the thermal mass of the rack and the cosmetic history of the SKU. If the drift points to haze, it proposes increasing the DI make-up, purging the tank or diverting the rack before the oven. The technician validates and signs off; the oven does not resume on its own.

See the full IRIS architecture →

Before and after

Manual rinse sampling vs. in-line tracking with iLEAN

Aspect	Manual sampling + final quality control	With iLEAN Edge + Connect + Agent
Rinse reading	Per shift or by exception	Continuous, second by second, in line
Reaction to drift	When the haze appears after curing	Traffic light and rack hold before the oven
DI make-up	Fixed recipe per SKU	Adjustment suggested to the technician from real drift
Tank purge	Planned by time or by appearance	Planned by real accumulated load
Traceability per rack	Whatever the MES plus the shift report holds	Per-rack file with conductivity curve and make-up
Operation with no network	n/a	Edge keeps reading and holding on the cabinet's power

Impact estimate

Impact estimate for your plant — to be validated with your numbers.

The block below is an estimate to be validated with the specific data of your line. We lay it out so the committee has an order of magnitude; we refine it during the diagnostic.

• Cataphoresis line on anodized aluminum profiles for architectural glazing, multi-SKU with frequent colour changes.
• Edge pilot on the final rinse tank (conductivity probe + integration with the tunnel PLC + rack hold actuator). First value expected within a few weeks.
• Reduction of cosmetic rejects from haze / staining of 30% or more on the recurring rinse incident (the other failure modes remain a matter for the process).
• Indicative payback between 4 and 9 months, depending on the frequency of cosmetic rejects and the average cost of rework (stripping + repainting) or scrap.
• The hard lever: a single rack avoided pays for a good part of the pilot.

And the CAIO's reasonable doubt

"What if the AI decides to hold a rack that was fine?" — hallucination is a problem of free generation, not of anchored tasks. In tasks where the AI merely compares a physical reading against a threshold trained on the line's own history, the best models pushed the error below 1.5% ^[1]. And even so, nothing critical is decided alone: iLEAN holds and the person signs off. The three safety rings exist for exactly this.

^[1] OpenAI paper "Why Language Models Hallucinate", 2025 — on the reliability of AI in anchored tasks.