Premium masterclass
The Complete Closed-Loop Color Masterclass
The definitive course on building a closed-loop offset operation, from setup to scale.
Flagship masterclassCourse syllabus
- The closed-loop concept: sensor, decision, actuator
- Anatomy of the Rutherford system: console interface, ink-zone control, learning logic
- Integrating with your press brand: Heidelberg, KBA, Komori, Manroland workflows
- CIP3/CIP4 presetting: turning prepress data into ink-key opens
- Spectral targets and ΔE strategy in production
- Operator workflow: what changes day one vs week one vs month three
- Closed-loop on extended gamut (ECG, 7-color)
- Scaling across presses, shifts, sites
Course content
The full lesson, module by module
The video is the introduction. The complete written course is below, structured to match the syllabus. Read it in one sitting or come back module by module.
Every closed-loop system, whether in a chemical plant, an aircraft autopilot, or a printing press, has the same three components: a sensor that reads what is actually happening, a decision layer that compares the reading to a target and chooses an action, and an actuator that executes the action. Remove any one and the loop opens.
In a sheetfed offset closed-loop installation, the sensor is the spectral measurement device, typically an inline scanner like IntelliTrax2 reading every sheet (or every Nth sheet) as it leaves the delivery. The decision layer compares the measured ΔE 00 per ink zone against the target and decides which zones need more ink and which need less. The actuator is the ink-key servo that physically opens or closes the zone on the next pull.
Open-loop production is the alternative: the operator looks at the sheet, decides subjectively whether it is good, and adjusts the keys manually when something looks off. The operator is both sensor and decision layer in this model, and their judgment varies by shift, by mood, by the day.
The economic case for closing the loop comes from removing that variance. A measurement-driven decision is the same every time. Ink-key correction within 5 seconds of measurement is faster than any operator can do manually. Multiplied across a year of makereadies, this compounds into the ROI numbers the free courses opened with: 30 to 55 % less waste, 25 to 40 % less makeready time.
The Rutherford system installs alongside the press without replacing the press OEM's native console. The operator keeps their familiar interface. Rutherford sits on a second screen, typically a touchscreen mounted near the existing console, and overlays the existing controls with closed-loop logic.
The console interface shows the live ΔE 00 heatmap per ink zone, current density and density-target deltas, and recommended corrections. The operator can accept the system's decisions, override them, or run the system fully autonomous depending on shift confidence and job risk level.
Ink-zone control is the link between Rutherford's decisions and the press. On Heidelberg presses, Rutherford talks to the Prinect Press Center via INK-Net. On KBA presses, it integrates with LogoTronic via ECS XML. On Komori, the path is via PDC. On Manroland, through the InkDriver protocol. Each integration is one-time; once it is in place, the operator does not see the protocol, only the result.
Learning logic is the Rutherford-specific layer. The system records what worked on similar jobs in the past, same substrate, same ink set, same coverage profile, and pre-positions ink keys before the first sheet rolls. After 30 jobs on similar work, the cold-start drops dramatically. After 90 jobs, the system has a confident first-sheet correction. After a year, it predicts.
Heidelberg presses are the most common installation target in Europe. The Prinect ecosystem exposes a stable XML schema (INK-Net) for ink-key control. CIP3 preset import is straightforward via the PPF or JDF route. Rutherford has been integrated with Heidelberg consoles for over two decades; expect a clean install with minimal press downtime.
KBA (now Koenig & Bauer) machines use LogoTronic for press control. The integration surface is ECS XML, which exposes ink-key set points and reads back actual positions. Older Rapida-series presses required custom adapters; modern Rapida 75 / 105 / 145 generations have ECS as a standard interface.
Komori machines integrate via Komori PDC (Print Density Control) or KP-Connect on newer presses. The Rutherford layer reads PQS color-bar measurements and writes back ink-key corrections to the H-UV or G-Series consoles. Komori also provides their own closed-loop layer; Rutherford typically replaces or complements it depending on customer choice.
Manroland Sheetfed uses Pecom or InkDriver. Their proprietary control software requires a specific adapter; once installed, the protocol is reliable but the initial commissioning is slower than Heidelberg installations.
Across all four, the operator-facing experience is the same: Rutherford touchscreen on the right, native console on the left, closed-loop running behind both.
CIP3 is the industry consortium that defined the PPF (Print Production Format) file in the mid-1990s. CIP4 is its successor, defining JDF (Job Definition Format) and JMF (Job Messaging Format). Both encode the dot coverage of each color separation across the press sheet, divided into ink-zone columns.
The math is simple: ink demand for a zone is proportional to the total dot coverage of that zone summed across the sheet. A zone covered 80 % needs roughly twice the ink-key opening of a zone covered 40 %. The press OEM consoles have used this principle for thirty years to pre-position ink keys before the makeready starts.
Where Rutherford adds value is in calibrating the relationship. The raw coverage-to-opening formula varies by ink, by substrate, by press, by ambient temperature. Rutherford's learning logic adjusts the formula based on what actually happened on previous similar jobs, so the first sheet lands closer to target than a vanilla CIP3 preset would deliver.
In production, this is the difference between a 600-sheet makeready (vanilla CIP3) and a 250-sheet makeready (CIP3 + Rutherford learned offsets). The same prepress data, processed with site-specific learning, produces a meaningfully better starting point.
Brand-owner specifications usually arrive as a list of solids in CIELAB or spectral form, plus a ΔE 00 budget per channel. A typical packaging specification might read: C ΔE 00 < 2.0, M ΔE 00 < 2.0, Y ΔE 00 < 2.5, K ΔE 00 < 2.5, spot colors ΔE 00 < 1.5. Those are the contractual numbers.
Production targets need to sit tighter than the contractual numbers. If contract is ΔE 00 < 2.0, production should aim for ΔE 00 < 1.4. The buffer absorbs measurement uncertainty, instrument drift between your shop and the brand owner's lab, and the natural drift of an in-process run.
Closed-loop systems work best when tolerances are defined per channel and per region of the gamut. Solids tighter than midtones; grays tightest of all (gray balance is the most visible failure mode); spot colors negotiated case by case. Rutherford lets you express tolerances at this granularity and applies different correction aggressiveness based on which target is at risk.
Fail-fast triggers are the safety layer. If ΔE 00 exceeds the production target by more than twice for three consecutive sheets, the system should alert the operator rather than blindly continue correcting. The trigger catches scenarios where the issue is not "wrong ink-key position" but "wrong plate", "wrong ink", or "contaminated fountain", situations no amount of ink-key correction will fix.
Day one of a closed-loop install is hostile. Experienced press operators have spent years mastering ink-key adjustment by feel; they read the closed-loop screen as an attempt to deskill them. Expect frequent overrides, suspicion of the recommendations, and a tendency to "fight the system". Plan for it. Schedule day one for a low-pressure job, not a brand-owner audit.
Week one shifts the dynamic. Operators notice the system catching subtle drift that the eye missed, or holding stability through a substrate-batch change. Confidence builds slowly. The shift lead becomes critical here: if the most respected operator on the floor publicly endorses the system, adoption accelerates across the shift.
Month three is the inflection point. By this time the learning logic has accumulated enough job history to be visibly better than vanilla presetting on common work. Makereadies are shorter; rejects are rarer; the data trail is intact. Operators stop overriding except on genuinely unusual jobs. This is the milestone where ROI starts showing in monthly waste reports.
Year one is the new normal. Closed-loop is part of the workflow; new hires are trained on it from their first shift. The conversation shifts from "is this working" to "how do we extend this to the second press, second site, extended gamut work". Plan the rollout with this arc in mind; do not promise month-three results in week one.
Extended Color Gamut workflows print with seven inks instead of four: typically CMYK plus orange, green, and violet. The expanded gamut covers more brand spot colors directly in process, eliminating the need to mix custom inks for short-run jobs. The trade-off is operational complexity: seven channels of ink-key control, seven calibration curves, seven targets.
Closed-loop on ECG is harder than on 4-color in three ways. First, the measurement device must read all seven inks reliably; not all spectros handle high-chroma orange and violet well. IntelliTrax2 reads ECG inks without clipping; older inline systems struggle.
Second, trapping behavior in ECG is more complex. Three additional inks mean more overprint combinations, more sensitivity to ink sequence, and more places for trapping errors to cascade. The closed-loop system needs to recognize when a target deviation is caused by trapping rather than ink density and decline to over-correct.
Third, the prepress workflow has to feed the press a clean ECG separation. Equinox is the most common conversion technology; Esko Equinox-aware RIPs produce ECG separations from CMYK or RGB inputs. The closed-loop layer reads the resulting ECG color bars and corrects per channel.
Bottom line: ECG closed-loop works in production today on packaging presses, but it requires a pressroom that has already mastered 4-color closed-loop and a prepress team comfortable with extended gamut separation.
A single closed-loop installation is a project. A fleet deployment is a program. The two require different management. Single installations focus on machine commissioning, operator training, and first-month ROI tracking. Fleet deployments focus on standardization, governance, and central reporting.
Standardization across presses starts with the color bar. Same bar layout, same patch sequence, same dimensions, same position on the sheet across every machine. Without this, your data is not comparable; with it, you can benchmark machine A against machine B on identical metrics.
Shift standardization is procedural. Each shift uses the same tolerances, the same fail-fast triggers, the same approval workflow. The shift lead owns adherence; the plant manager owns the policy. Closed-loop systems make divergence visible: if night-shift ΔE 00 averages are consistently higher than day-shift, the data will say so.
Multi-site governance answers three questions: who owns the targets, who is allowed to change tolerances, and who issues the brand-owner report. The cleanest setup centralizes target management (one person in quality owns the master target library) and decentralizes execution (each plant runs its own production). MeasureColor Reports provides the aggregation layer for this model.
Ten years into a fleet deployment, the closed-loop system is invisible to the operator (it just works) and central to the business case (audit-ready data, defensible KPIs, faster new-customer onboarding). That is the destination this masterclass aims at.
Other masterclasses