Paul Industries designs, installs and commissions cleaning systems across Minnesota. The state runs on contract manufacturing, and that creates a cleaning problem with no equivalent in a single-product plant: shared equipment carrying one customer product after another, where you frequently cannot calculate a carryover limit because you do not have the next customer formulation and are not entitled to it. The usual method assumes knowledge you do not possess, so the evidence has to be built a different way.

Request a quote or call 201-450-8280

The contract manufacturing problem You cannot derive a limit from a formulation you are not given
What replaces it Cleaning to a demonstrated non-detect, on methods each customer accepts
The second decision Which parts are shared and which are dedicated per customer
Who audits you Your customers, their notified bodies, and FDA
Industrial power 9.15 cents/kWh, 1.13x the US average of 8.13 (EIA, 2024)
Reviewed September 2026, against current state code and EIA 2024 energy data

You cannot calculate a limit for a product you have never been shown

Minnesota has an unusually deep contract manufacturing ecosystem sitting underneath the primes, and this is the cleaning question that ecosystem generates.

The textbook approach to a cleaning limit works forward from the next product. You establish what exposure is acceptable, carry it through batch size and dosing, and arrive at a permitted surface residue. Every step depends on knowing what the next product is.

A contract manufacturer frequently does not. The next run may belong to a different customer whose formulation, adhesive chemistry or coating is proprietary and deliberately withheld. Sometimes the customer does not know what the following customer will bring either. The chain of reasoning breaks at the first step, and a limit calculated on assumptions about an unknown product is not a limit, it is a guess with arithmetic attached.

The workable answer has three parts and it is quite different from single-product practice.

First, clean to a demonstrated non-detect rather than to a calculated threshold. Rather than arguing that residue below a computed figure is acceptable, demonstrate that the cleaning process removes the residue to the limit of detection of methods that are sensitive and general: total organic carbon on rinse water, gravimetric residue, particle counts and structured visual inspection under magnification. An argument that nothing measurable remains does not require knowledge of the next product.

Second, decide deliberately what is shared and what is dedicated. Shared equipment needs the cleaning argument above. Dedicated parts, tooling, fixtures and sometimes whole lines remove the question entirely for the customers that justify it. That is a capital and floor-space decision, and making it explicitly at design stage is much cheaper than discovering during an audit that a fixture everybody assumed was dedicated has been moving between programs.

Third, design so the argument can be made. A cleaning process defended by demonstrated removal has to be reproducible, and reproducibility is an engineering property: verified coverage, controlled and recorded chemistry, temperature and time, drainable equipment without traps, and change control on anything that affects the cycle. A cycle that varies with the operator cannot carry this kind of evidence.

Single-product and contract manufacturing compared

How the cleaning argument changes when equipment is shared between customers
 Single-product plantContract manufacturer
Next productKnownFrequently unknown or proprietary
Limit basisCalculated carryover into a known productDemonstrated removal to limit of detection
AnalyticsCan be compound-specificMust be general: TOC, gravimetric, particulate, visual
Worst caseHardest-to-clean locationHardest location plus the most tenacious residue seen
Equipment strategyShared throughoutExplicit shared and dedicated split
AudienceYour own regulatorEvery customer quality organization, plus regulators
Cleaning cycle electricity at Minnesota’s 9.15 cents/kWh
Cycle profilePer cycle400 cycles a year
30 kW for 45 minutes$2.06$824
60 kW for 90 minutes$8.24$3,294
120 kW for 2 hours$21.96$8,784

The number that matters to a contract manufacturer is not on this table. Changeover frequency is the constraint, because a contract operation changes over far more often than a single-product plant, and every changeover is capacity that is not being sold. A cleaning cycle shortened by twenty minutes returns that time on every changeover, which in a high-mix operation is a materially larger number than the entire annual electricity cost of the cleaning system.

What a customer audit will actually ask

Customer quality organizations auditing a Minnesota contract manufacturer tend to ask a predictable set of questions, and they are worth designing for rather than preparing for.

They will want to know what else runs on this equipment and how you know the previous product is gone. They will ask which parts are dedicated to them and how that is enforced physically rather than procedurally. They will ask how the cycle is controlled and recorded, and what happens when a parameter falls outside its range. They will ask what your worst-case residue is and why you believe it is the worst case. And they will ask to see the data, not the procedure. Equipment that can be cleaned reproducibly and recorded automatically answers all of those far more comfortably than equipment that depends on an experienced operator doing the right thing.

Frequently asked questions

Do you build cleaning systems for Minnesota manufacturers?

Yes, across the Twin Cities metro and statewide: ultrasonic and parts-cleaning lines, CIP skids and distribution where vessels are involved, rinse systems, chemical dosing, instrumentation and controls. On contract manufacturing work we design for recorded, reproducible cycles because that is what a customer audit actually examines.

How do you set a cleaning limit without the next formulation?

You change the form of the argument. Instead of calculating an acceptable carryover into a product you have not been shown, demonstrate that the cleaning process removes residue to the limit of detection using sensitive general methods: total organic carbon on rinse water, gravimetric residue, particle counts and structured visual inspection. Showing nothing measurable remains does not require knowing what comes next.

Should we dedicate equipment per customer?

For some customers and some parts, yes, and the decision should be explicit rather than inherited. Dedicating fixtures, tooling or a line removes the shared-equipment question entirely for that program, at a cost in capital and floor space. The failure to avoid is an assumption of dedication that is not physically enforced, because that is discovered during an audit rather than before one.

What analytics suit a shared-equipment argument?

General and sensitive ones, because a compound-specific assay presupposes knowing the compound. Total organic carbon on rinse water, gravimetric residue determination, particulate counts and structured visual inspection under magnification between them cover organic, inorganic and particulate residues without requiring a formulation. Each customer should confirm the methods are acceptable to them.

How do we choose a worst case with a changing product mix?

By the residue rather than the product. Identify the most tenacious material you actually handle, the geometry hardest to clean on the equipment, and validate against that combination, then reassess whenever a new program introduces something more difficult. That makes the worst case a living determination with a trigger for review rather than a one-off statement in a report.

What do customer auditors ask about cleaning?

What else runs on this equipment and how you know the previous product is gone; which parts are dedicated and how that is physically enforced; how the cycle is controlled, recorded and handled when a parameter goes out of range; what your worst-case residue is and why. And they ask for the data rather than the procedure, which favors equipment that records automatically over equipment that depends on operator diligence.

Does this apply to pharmaceutical contract work too?

The structure does, though pharmaceutical contract manufacturers usually have more information because a customer supplying an active ingredient generally has to disclose enough toxicological data for a limit to be derived. The device and component side is where the information gap is widest, and it is the reason the demonstrated-removal argument does more work in Minnesota than in a pharmaceutical state.

Can we shorten a validated cleaning cycle?

Only with data, and in a high-mix operation it is usually worth generating. Shortening requires development work against deliberately soiled worst-case parts, recovery studies, and a change assessment. What makes it pay here is changeover frequency: twenty minutes saved returns on every changeover, and a contract operation changes over far more often than a single-product plant.

Does Minnesota energy cost affect cleaning economics?

Hardly at all relative to capacity. At 9.15 cents per kilowatt-hour against a national average of 8.13 (EIA, 2024), even a demanding cycle run four hundred times a year is a few thousand dollars of electricity. The number that matters is the changeover time the cycle consumes, because in a contract operation that is capacity you are not selling.

How do I get a quote for a Minnesota cleaning system?

Use the form on this page or call 201-450-8280. Useful inputs are how many customers and product types share the equipment, which residues are hardest to remove, what is dedicated today, your changeover frequency and current cycle time, and whether any customer imposes cleaning requirements beyond your own. If a customer audit has raised a cleaning observation, send it, because it usually identifies the gap directly.

Are generic default limits defensible?

Only with a rationale behind them. A commonly used figure adopted because it is commonly used is not a rationale. The defensible version demonstrates that the chosen limit is more conservative than a health-based calculation for the plausible range of next products, and states the assumption explicitly so an auditor can test it.

Should cleaning limits be written into customer contracts?

It is worth doing, because otherwise the site carries an obligation whose boundaries are undefined. Agreeing what the customer requires, what the site will demonstrate, and who owns the assessment when a new product is introduced removes the most common source of dispute during a customer audit.

What analytical method suits a shared-equipment argument?

One matched to the residue and sensitive enough to reach the limit. Non-specific methods measuring total organic carbon are efficient for an argument about general cleanliness across a changing product mix, while a specific method is required where a particular compound drives the limit. Most contract sites use both for different purposes.

What are the limits of a non-specific method?

It tells you how much carbon-bearing residue is present and not what it is, so it cannot distinguish product from cleaning agent, and it cannot demonstrate that a specific toxic compound is below its own limit. It is well suited to routine monitoring of a validated state and poorly suited to being the sole evidence.

How is a worst case chosen when the product mix changes?

Through a documented rating of products by the properties that matter, solubility, cleaning difficulty and toxicological limit, with the highest-rated product validated and others bracketed beneath it. The structure has to be maintained as products come and go, and it is what allows a new product to be assessed in days rather than months.

What happens when a customer product is withdrawn?

The grouping is reviewed, because the withdrawn product may have been the worst case supporting the whole validation. Sites frequently notice a new worst case arriving and fail to notice the old one leaving, which leaves a validation resting on a product no longer made and an unexamined new maximum.

How is swab recovery established?

Per surface material and per residue, because recovery differs considerably between a polished stainless surface, an elastomer and a plastic, and between residues. Without a recovery study the swab result understates what is present by an unknown factor, which undermines every limit calculation that rests on it.

Can a validated cleaning cycle be shortened?

Yes, on evidence and through change control. Many cycles carry time added for reassurance rather than derived from data, and on a contract site cycle time is capacity. Demonstrating the cycle reaches its endpoint earlier, at worst case, converts directly into more changeovers per week.

Should change parts be dedicated?

Frequently yes, because change parts are the components with the most difficult geometry, the most manual cleaning and the highest carryover risk, and they are relatively cheap to duplicate. Dedicating change parts while sharing the major equipment is often the most efficient point on the dedication spectrum.

Does automation help on a shared-equipment site?

Considerably, because the variable that undermines shared-equipment arguments is operator technique. An automated cycle with recorded parameters produces the same result regardless of who ran it, which is exactly the claim a customer auditor is testing. It also removes cycle time variability, which is capacity.

How is the validated state maintained?

Through routine monitoring against the limits, trending rather than pass-fail review, and a change control route that catches equipment modifications. The most common failure on contract sites is maintenance modifying equipment without reference to the cleaning validation, so the physical system drifts away from the one that was demonstrated.

Does cleaning agent residue need its own limit?

It does, and it is the obligation most often forgotten because the detergent is the cleaning agent rather than the contaminant. Residual detergent on a device or in a product is a chemical the patient receives, and the rinse verification should demonstrate removal of both the soil and the chemistry used to remove it.

What is the commonest gap on contract manufacturing sites?

A cleaning validation that was sound when it was written and has not been maintained as the product mix changed. New products were added without reassessment, the original worst case left, equipment was modified, and the documentation still describes the plant as it was three years ago.

Should cleaning be verified between every changeover?

Verification between changeovers is normal practice on shared equipment, using a faster indicator once its relationship to the validated method is established. Full analytical verification every time is rarely practical and rarely necessary; what matters is that the routine check is capable of detecting a genuine failure.

What happens if a customer’s product is more toxic than assumed?

The default limit no longer protects it and the grouping has to be revisited, which is why a new product assessment must include the toxicological limit rather than only the cleaning difficulty. A product that cleans easily and carries a very low permissible residue can become the worst case on limit alone.

Planning a cleaning system in Minnesota?

Tell us how many customers share the equipment and what is hardest to remove. Call 201-450-8280 or use the form below.

Service needed *