Ongoing trial-and-error calibration when commissioning a high speed, 2.5 meter-per-second, 90000 cans-per-hour advanced sampling system.  “Advanced” referred to the capability to separate automatically for Quality Control purposes,  pre-selected rows of cans filled by specific Filler Valves or Seamer Heads, following time, or pulses, external or manual control by an Operator.   In evidence, a can fallen during the Sampling of the Filler and Seamer, machines both controlled by an Electronic Inspector configured with locating sensors and inspections in-the-Machine.  Video filmed at high frame rate (70 frames per second), the unique way to allow calibrations of the sampler and rejector timings.  It is the cans’ speed, an alias of kinetic energy, what makes critical these calibrations



      Flash animation, 17 MB, ZIP

             Click  index.html     


            17 pages, 3.4 MB

Sampling is a Quality Control recursive operation, devoted to systematic control of net contents for all filling valves, removal torques for all capping or seaming heads.  It allows to automatise an activity which if man-made requires long times and production stops.  

They exist three different kinds of Sampling: 

  • random, 
  • standard,
  • advanced.  

In its most efficient execution, it requires a special conveyor Layout, with an independent outfeed devoted to Sampling: its unique outfeeding conveyor, thought to host simultaneously rejected and sampled containers: a mix clearly incoherent.  All times it is really desired to sample frequently (i.e.: each 8 hours) all filling valves, it has to be preferred the most complex solution with separate conveyors.  

Sampling is the most critical and demanding type of classification, the most demanding type of rejection, because it implies action on long rows of containers, rather than rare contacts.  Rejection Table has to be specially designed for the task of sampling and also its lubrication has to be a special one.

Random sampling

Random sampling is the simplest version of sampling to let a preset amount of containers be separated by the production way in a row.

       Random sampling

“A simple way to quantify if all is correct lies in the observation of the Electronic Inspector’s False Triggers Counter.   

This has to be <0.5 %, thus assuring correctness to >99.5 % of the Locating and Sampling operations”

Control can be obtained in different ways:

  • manual control by external push-button;
  • manual control by soft-key in the Operator panel;
  • Inspector's restart;
  • Inspector counters’ reset.

With random sampling, no information exists about which valve filled, nor which head closed each container. The only real advantage with respect to human manual sampling is the fact that a precise permanent register of the operation is recorded in the Inspector’s memory.   If this is Ethernet-connected to the Factory Data Aquisition System (DAS), then the informations will be automatically transferred there and made available to the Production staff.


Standard sampling

Standard sampling the version yet allowing to sample specified rows of bottles, coming from filler and capper.  As an example, all filling valves and all capping heads, in the same sequence 1, 2, 3, …, N, as they are filled and closed.  

Its disadvantage is related to the fact that rarely Sampling Tables and rejectors are optimised to sample an entire consecutive row of all of the Filler valves and all of the Seamer or Closer heads. Then, a long manual operation is made necessary, to set manually the sampling of tens of shorter rows.  And it is because of this reason, that in the practice “Sampling” really and only means “Advanced Sampling”, object of the following section.

Advanced sampling


Advanced sampling is the most complex and professional way to sample.  Its complexity starts by the layout which is necessary to sample.   The figure at right side shows a configuration referred to a cans’ Filler-Seamer system. Similar configurations apply to the most common cases of glass or PET bottles.

Here, the containers:

  • infeeding the Filler Machine and the Electronic Inspector's Shifting-Register extended into Filler Machine, are marked with yellow colour (Positives + Negatives);
  • rejected (Positives) after the basic controls in the Electronic Inspector main cabinet (level and lid, cap or metal closure presence), are marked with red colour;
  • produced, because considered “correct”, (Negatives) are marked with green colour;
  • sampled, on a separate one-way conveyor, are marked as a long row of circles.

Tracking logic

The Electronic Inspector, in these kinds of applications may only be one with sensors in-the-Machines.  In the case at right side: Filler, Seamer and Conveyor.   Under the term sensors meant those digital necessary to provide to the Inspector trigger signals (references) of:

  • Filler synchronisation (what an angle is the Filler’s valve 1);
  • Filler Machine Cycle (how many valves since the Filler synchro);
  • Seamer Synchronisation (what an angle is the Seamer’s head 1);
  • Container Presence.

By the image on side it is possible to see that the Electronic Inspector's Shifting-Register is extended tens of meters in the Filler and Seamer.

  Sampling preconditions are complexity and pricing, starting by the Layout. The area the Electronic Inspector has to control is a huge one, several tens of meters extended in the Filler and Seamer system. In the figure it is possible to see the one-way conveyor destined to receive and accumulate a complete round-of-the-Filler-Machine sampling


Cans or bottles pass from a machine to the following in the sequence: 

                  Filler      Seamer /Closer /Capper     Conveyor 

Each one machine having its own peculiarities, like the:

  • outfeed of the Seamer/Capper, which being mechanically independent on the outfeeding Conveyor, lets containers slide on Filler Machine ramp-up or ramp-down phases;
  • pitch of the Seamer/Capper;
  • lateral guides of the Conveyors which shall be touched by the containers; 
  • Conveyor belts’ lubrication which cannot be constantly assured;
  • incremental wearing of the Conveyor belts;
  • incremental extension of the Conveyor belts, implying progressive dephasing of the Shifting-Register really existing, with respect to the programmed one;

and it is because of the superposition of the effects of these and other causes that, i.e., a single 40 meters long Shifting-Register should be surely unsuccessful.   

 (Click-to-enlarge)   The real Layout, corresponding to the drawing excerpted above, of the 90000 cans-per-hour in-the-Machine Full can Inspector Layout in DAMM® Brewery at El Prat, Barcelona, Spain.  In evidence the binary classification action of the orange coloured containers in green colour production and red colour Rejects.   Blue colour the classification of the Advanced Sampled containers, made possible by the existence of total 4 detectors in-the-Machines (here, Filler and Seamer).  Filler Synchro, Seamer or Closer Synchro, Container Presence (an inspection) and Machine Cycle.   The entire containers’ tracking system, here corresponding to 3 Shifting-Registers (2 of them in-the-Machines) plus a FIFO, has its origin or reference point in that rotation angle (Transfer Point) of the Seamer which let a container start to be released by its outfeed starwheel to the out feed conveyor.  All of the other phases are forced and dependent by this reference 

Or, successful in the only slowest thinkable Beverage Bottling Lines, i.e., < 8000 container-per-hour, < 0.3 m/s.  

Exactly those where Locating and Sampling functions are rarely or never applied simply because the containers are so slow that Operator has no doubt to identify the number of the: 

  • Seamer/Closer/Capper head not applying correctly positioned lids/closures/caps;
  • Filler valve not correctly filling the containers.

By this second category, obviously excluded the Can Filler Machines: whatever their speed, it is always extremely useful at least the Locating function of the Electronic Inspector.

Two or more Shifting-Registers

Because of these reasons and following the particular Layout, minimum two consecutive Shifting-Registers exist in the Electronic Inspector's memory:  

  1. one related to the Rotary Machines (Filler, Seamer, Capper, Closer), 
  2. the others to the outfeed Conveyors.

Why we’d need over two Shifting-Registers ?

The Shifting-Registers can be more than two as an example when it is necessary to gradually slow down the containers outfeeding the Machine (a Filler, Capper, Closer, Seamer or Labeller), before the main cabinet of the controlling Electronic Inspector and associated Rejector.   This is a case with containers' extreme speed at the Tranfer Point.  Transfer Point being the Machine's revolution angle which lets a container in the outfeed starwheel start to be released to the outfeed Conveyor. 

The Contractual relevance of the containers’ Sliding

In the special case depicted above, a very fast Canning Line, cans’ sliding at the Seamer out feed reach 500 mm.   Knowingly, the maximum sliding the Shifting-Register may compensate is limited to one-half of the container diameter.   In the case of cans, it’d mean always less than 32 mm.   But,  32 mm <<  500 mm.  

On practice:

  • at 90000 cans-per-hour production speed;
  • traditional 64 mm diameter cans;

  Advanced Sampling means an Inspector in-the-Machine configuration totalling 4 sensors.  3 of them Triggers devoted to synchronisation and the 4th an inspection of bottle or can presence. Here visible only 3 Triggers synchronizing events happened 10 seconds and 20 meters in the Past, with a Sampling action 1 second and 3 meters in the Future.  The entire system satisfies implicitly the canons of Shifting-Registers only in the Filler + Seamer / Closer / Capper area, because of a cardan mechanical connection between Filler and Seamer / Closer / Capper.  Containers cannot slide at all between adiacent cells, Filler's valves or Seamer / Closer / Capper's heads.  On the opposite, it does not satisfies at all the Shifting-Register's canons the Conveyor area after the Seamer/Closer/Capper's transfer point, where containers slide until 500 mm

~30 % of the cans’ identities first triggered in the Filler (Container Presence sensors) and synchronised with future events in the Seamer, Conveyor and Electronic Inspector, result lost after the Seamer’s discharge. 

Now, imagine what should happen with the lighter today growing productions of sleek and ultra-sleek cans (see figure below), featuring smaller diameter…..   On practice, ~70 % of the cans’ identities synchronised in the Filler and Seamer, lost after the discharge at the Seamer's transfer point.  

  The new “Sleek”® design, slimmer than the former existing “Slim”.  Slim & “Sleek”® cans present two factual advantages over the Standard can at the price of a gigantic negative. The Beverage can be cooled faster than requested by Standard cans and the ergonomic characteristics fit an human hand.   A downside is visible when trying to keep them synchronised at the Seamer Machine's out feed.   Because of the smaller relation between diameter and height, they are unstable.  Sliding > 300 mm at 90000 cph.  When sliding is > 1/2 of the container's diameter, the container identity shall be lost by the Shifting-Register of the Electronic Inspector which has to control Filler, Seamer Machines and outfeed Conveyor.   “Lost” meaning the detection of defective Filler Valves and Seamer is reduced from a truly correct information, one whose dispersion is spiked, to a dispersed statistical one (image credit Rexam®)

No “Statistically-significative” Sampling or Locator

If the: 

Easy to recognise the typical footprint of an erroneous or missing commissioning for the joint set of in-the-Machine functions underlying also the Advanced Sampling.   The location of the Filler's valve no. 1 and of the Seamer /Capper /Closer head no. 1 will not be correctly attributed.  Meaning that the reference system in the Electronic Inspector is rotating with respect to the real system in-the-Machines
  • Design, 
  • and/or Integration, 
  • and/or Commissioning, 

of a Sampling system are inadequate to the purpose, the location of the Filler's valve no. 1 and of the Seamer /Capper /Closer head no. 1 will not be correctly attributed.  

To understand the point, we can imagine a rotary reference system whose angular velocity ω is constantly changing in nearly random way.  

  Easy to recognise the typical footprint of an erroneous or missing commissioning for the joint set of in-the-Machine functions underlying also the Advanced Sampling.   The location of the Filler's valve no. 1 and of the Seamer /Capper /Closer head no. 1 will not be correctly attributed.  Meaning that the reference system in the Electronic Inspector is rotating with respect to the real system in-the-Machines

Be wary of arguments trying to convince you that in-the-Machine functions (like all those described before) inherent to:

  • Filler valves', Seamer or Closer heads' Locating;
  • Filler valves', Seamer or Closer heads' Sampling;

should be only statistically-guaranteed.   

Below an example of the difference between deep and poor knowledge of the status of a system defined in a bidimensional space.   X and Y are values measured for physical properties related to the system status.   In an ideal-world case, the sharp observed distribution of the measured values (right side graphics) is a vertical line of infinitesimal thickness, infinitely thinner than what here depicted.   These infinities correspond to 100 % of information about the true value of one of the physical properties characterising a system.  

In our practical cases, distributions can only be spread and then the key point becomes the answer to a single question: how-much spread ?  

What a variance to consider “acceptable” to say that we have enough knowledge about a system ?   

 3-D sharp (spiked) and unsharp distributions for the Probability of a status of an objects. Status derived by the informations arising by physical measurements of its properties X, Y.   Electronic Inspectors' physical measurements, kept apart extremely rare exceptions, are based on electromagnetic interactions between a sensor and the object (e.g., a container)

“...only if the design, integration, installation or commissioning are erroneous let the in-the-Machine functions become statistical, say generally wrong”

To say that the knowledge available at a given time for a physical property, as an example, the couple of angles of revolution of the Filler’s Valve no. 1 and of the Capper's head no. 1 is statistical, means an unsharp distribution like that one above, at left side.  In the reality, only if the Design and/or Integration and/or the Installation and/or or the Commissioning are erroneous, the in-the-Machine functions surely become statistical, say generally wrong.  

An excellent way to quantify if all is correct lies in the observation of the Electronic Inspector’s False Triggers Counter.   This has always to be < 0.5 %, thus assuring correctness to > 99.5 % of the Locating and Sampling operations.

“Equipments Audited typically because the Food and Beverage Bottling Company technical staff, with the time started to “feel that something basic is wrong or missing”

Is it fiducial the False Triggers Counter ?  

Can the False Triggers Counter: 

  • protect you completely ?  
  • be always evidence of the Truth about what really is operative front of you, in the Electronic Inspector ?  
  • prove that an Electronic Inspector really is wired and commissioned as the expensive in-the-Machine you paid ?

No.   They have also be Audited installations where a brown-coloured wire had been connected where someone else suggested, exactly and only to let the alarm red colour traffic light of the Electronic Inspector deprived of its in-the-Machine functions, be showing a calming green colour traffic light.   Thanks to the complex sabotage designed by someone else, definetely too complex for the incompetent Service Technicians sent to the Bottling Site, the False Triggers Counter menu, looked stellar with its 0.00 % of False Triggers.  

   By mean of this cable some Service Technicians bridged a constant GND (0 volt) to give to the Bottler just the sensation of a correct Container Presence Signal.  In the reality, before they disconnected the Container Presence Sensor and all others devoted to in-the-Machine Locating functions.  This sabotaging cable had been months in that Electronic Inspector.  Enough to let the Brewery's most skilled Electronic Engineers suppose that an Electronic Inspector is complete, corresponding to the Contract and correctly functioning.   Removing the sabotaging cable, the Truth immediately revealed itself front of many Witnesses and cameras.  The worldwide second Beer Group discovered that just < 35 % of the price paid for that Machine was commissioned and operative.  

Removing the sabotaging bridge, the Bottler discovered that:

  1. the in-the-Machine electronic inspector’s Bottle Burst inspection and rejection system was totally bypassed;  
  2. the expensive Inspector reduced to a standalone;
  3. since months glass splinters into the beer bottles were proceeding unrejected to the Market, exposing people to illnesses and the Brewery to claims

Disconnecting the brown-colour sabotage cable, immediately the Truth gots out with the opposite reality: 

  1. nearly 100 % of Falsely Triggered bottles;
  2. the in-the-Machine electronic inspector’s Bottle Burst inspection and rejection system was totally bypassed; 
  3. the expensive inspector reduced to a standalone;
  4. since months glass splinters into the beer bottles were proceeding unrejected to the Market, exposing the Brewery to claims related to Product Liability.

Readers will try to imagine the rational for such actions: “Why to behave this way ?”.  Reason is always and only the Principle of Maximum Profit.   A skilled Service Technician need over 10 years to be slowly built over stratified experiences.   More, the Technician has to be one who really studied and operatively understood Math, Electronics, Mechanics, etc. and not just someone with a piece of paper.   

Then, you understand that the maximum profit is reached when a someone, just camouflaged to look like a Service Technician, costing 50/day is resold to the Beverage Bottler for 750/day.   The Bottling Company's Maintenance electronic engineers specialization closes the circle.   Because they are highly specialized to assure Production, they are typically highly skilled in Machinery Automation.    But, …..but the Electronic Inspectors are not “Machinery Automation” !    And that’s why the sabotage briefly described above, passed undetected to several Electronic Engineers who tried to assess the operation of that Electronic Inspector. 

If who installed or started or commissioned or however guaranteed you a Sampling of Locating system, part of an Electronic Inspector, emits any statement sounding like “statistically guaranteed only” then he:

  1. has no idea of what is speaking;


  1. under-evaluates your know-how;


  1. knows that his/her Staff did not synchronised the system (it operates standalone and not in-the-Machine) and tries the easy way to get rid of the fact you have discovered a hole in the Commissioning whose huge economic value may had yet been paid by your Company.

With the time, you’ll discover that very strangely, there is something like a hole in the technical literature made public about this special subject.   A huge hole, when considering that if a standalone Electronic Inspector costs X, then its in-the-Machine variant shall be always paid by you:  (2 - 3) times X.     So much more because you’ll be paying a great difference in the hardware, installation and commissioning activities.    

The Writer of this notes Audited ~70 Electronic Inspectors in worldwide Food and Beverage Bottling Lines guaranteed in-the-Machine to the Bottling Companies.   Companies who paid the correspondingly higher price (over 100 % more expensive) than an Electronic Inspection equipment configured with the same inspections, installed and commissioned standalone.   Equipments Audited typically because the Food and Beverage Bottling Company technical staff, with the time started to “feel that something basic is wrong or missing”.   The Audits revealed an amount of missing functions yet paid.   


Conceiving the case of a relatively fast line (i.e., over 2.7 m/s), the Shifting-Registers’ tracking method should not be capable to preserve an univoquely defined identity for each one container.   Because of this reason, at the out feed of the Seamer (or, Closer), and precisely at the Seamers’s (or, Closer) Transfer Point, it acts a non-kinematical algorithm named FIFO (click here for for details).   The first of them starts before the Filler Synchronisation sensor.  

The second Shifting-Register starts after the sensor named FIFO TRIGGER in the drawing on right side.   You may briefly and correctly imagine a FIFO like a Shifting-Register without the Encoder.   A non-kinematic system, just counting containers.  The FIFOs allow the possibility to maintain correct attribution for the identity of each one container, until a container sliding equivalent to 255 containers’ diameters.   Then, a maximum sliding extended to several tens of meters.  

After the FIFO photosensor, containers’ motions results tracked by the triggers’ sequence: 



  FIFO is the unique configuration to choose when the containers’ speed exceeds 2.7 m/s

  1. TRIGGER 1 provides the Shifting-Register reference for the basic inspections in the Electronic Inspector's main cabinet (level and lid or closure presence) and for the Rejection/Sampling device;
  2. REJECT VERIFICATION verifies that defective or preset containers have been rejected or sampled.

FIFO is always the best solution for the Electronic Inspectors' in-the-Machine configuration when and where any of these conditions is met:

  • containers’ velocity exceeds 2.7 m/s;
  • light weight containers (i.e., < 200 grams) then particularly prone to sliding;
  • Conveyors’ layout is irregular and not including enough containers’ Tracking Triggers;
  • the inspections in-the-Machine (Filler, Seamer, Capper, Closer or Labeller) have not the character of Food and Beverage Safety.  Then, as an example, it cannot be adopted to control the sensitive beverages filling technologies.   Remember that the  Locating functions are not in-the-Machine’s inspections.   Then, Filler or Capper Monitoring and Location functions only report informations about positions equivalent to Rotary Machines’ angles, without any additional physical measurement say, an inspection.   As an example, Container Presence sensor in the Canning Filler Machines and Labeller Machines is an inspection.   However, if the beverage filled is not a sensitive one, FIFO still is the best option over the traditional kinematic tracking systems based on Shifting-Register.


The Advanced Sampling allows the possibility to define:

  • from which valve and head, until which valve and head, to sample;
  • along how many revolutions of the Filler and Capper;
  • how many revolutions, during other sampling operations.

Control can be one of the following:

  • manual pushbutton or Operator Panel;
  • time;
  • a number of bottles processed;
  • a number of revolutions of the filler and capper;
  • reset of the Inspector’s counters;
  • the Inspector's power on;
  • the stop of the Production condition.

  Advanced sampling parameterization, in a Coca-Cola® Bottling Line.  In the example, the Filler has 150 valves whose action shall be sampled along 25 revolutions 6 bottles each. A pause of 15 revolutions set to give time to Quality Control staff to crate the bottles.  Shown the initial 13 revolutions 

  FULL CONTAINER INSPECTION WITH POSITION SENSORS AND INSPECTIONS IN THE SEAMER AND FILLER machines.  Total spatial developement in the Seamer and Filler of this device including tens of CPUs and controlling 90000 cans-per-hour, is ~ 25 meters.   The second rejector is devoted to the advanced sampling function visible in the video above, the automatic separation for Quality Control purposes,  of pre-selected rows of cans filled by specific Filler Valves or Seamer Heads, following time, or pulses, external or manual control by an Operator.    In evidence also three additional trigger Laser photosensors, vital to track fast moving cans.   Image shot during the startup phase.   Visible by the orange coloured handles of the REJECT VERIFICATION trigger (right side), TRIGGER 2 (middle) and FIFO TRIGGER (left side). The entire system stabilized with tens of stainless steel bars glued with chemicals into the ground

How-to assure an immediately functional Sampling

Production speed defines the kinetic energy of the containers and, as derived amounts, the length of the Sampling or rejection table and the number of conveyor belts. 

From this fundamental amount, the containers’ maximum kinetic energy, it derives a Table of purely empiric origin, allowing to design an immediately successful Sampling Table:


  Production speed                       36000                  48000                  60000                72000             90000



  Rejection Table length [ m ]            2                         3                          3                        4                      5

  Conveyor belts                                4                         5                          6                        8                      9

  Polished belts                               1-3                      1-4                       1-4                    1-6                   1-7

  Syncronised belts                          all                 4+5: -20%                  all               7+8: -20%       8+9: -20%


  Polished overlubricated belts and table length are the preconditions for Sampling.   Rejection Table Length, is meant that part of the table truly active.  “Active” in terms of separation of the containers by the production flow and successive alignment in an individual belt.   A practical rule implies to measure the distance in between the central point of the Rejector, until that point of the outfeeding individual conveyor where the space is 1.5 times the container diameter.  Applying this to the Table, it means to add ~0.5 m of length

“...if the Projects' successes are counted in terms of Customers' acceptations for the unacceptable, with the time they'll create an artificial sense of correcteness of the technical solutions from whom the row of acceptations and payments derived”


  • fast moving belts have to be polished, to ease containers’ sliding;
  • until 48000 containers-per-hour, the only non-polished (then, cheaper) belt is the last one, the one far from the rejection (or, ejection) or sampling device;
  • over 48000 containers-per-hour, the last two belts (those far from the rejection device) can be of the cheaper non-polished kind;
  • Rejection Table Length, is meant that part of the table truly active in terms of separation of the containers by the production flow and successive alignment in an individual belt.   A practical rule implies to measure the distance in between the central point of the Rejector, until that point of the out feeding individual conveyor where the space is 1.5 times the container diameter.   Applying this to the Table above, it means to add around 0.5 m of length.   If the prescripted length and number of belts are not respected, however expensive may result a table so wide, bitter surprises wait for the Bottler’s Quality Control: jams and stops to production. 

Sampling made for speed lower than nominal of the format.   It’s a common trick of several Vendors of Bottling Lines: conveyor’s price is proportional to their weight due to the high price of the stainless steel.   The total price of the Line can be reduced simply reducing the total amount of tons of stainless steel.  This is true also for the dimension and number of belts of the Sampling Table requested by a Bottler.  If the majority of the Sampling attempts at nominal speed fails, becomes a jam and one more stop to production, Conveyors are hiding the root cause of the problem. 

Diagnosing non-functional sampling tables 

If during the Inspector's commissioning, it emerges that the Staff of the Inspector's Vendor insists objecting that:

  • table is not wide enough;
  • table is excessively inclined;
  • belts are not polished;
  • belts are not lubricated enough;
  • belts are not enough;

and, simultaneously: 

  • no self-evident error may be detected in the Rejector's installation;
  • containers do not fall immediately after Rejector’s action;

then the Line's Vendor or OEM is probably not compliant with the Order received.  To make an example, to sample 6 containers only in a Filler Machine hosting 168 valves, then wait 20 seconds to sample a following adiacent serie of 6, ….., and so on until 168, violates the very basics of Metrology.  The entire sampling of the filling levels and closing of the Filler-Seamer/Closer/Capper shall be referred this way to a period of time over 9 minutes.  Along 9 minutes they can happen lots of things...  

The Analysts of the Quality Control Dept. when objecting exactly this, frequently hears the answer:

“…it is not necessary the entire revolution of the Filler Machine, because these are electronic and not mechanical Valves…"

If the argument should be factual, to sample today or tomorrow should be equivalent, then nearly unuseful: “… because they are electronic and not mechanical valves…”.    But, unfortunately, Electronics can fail and fails and that’s why after the Electronic Valves were born the Advanced Sampling continued to be necessary as before.   More, they are sold today much more frequently than yesterday, when the electronic Valves were not existing, because their function is made vital by progressive increase of the Filler Machines.

How can this happen ?

Now we can ask ourselves how could it happen that, i.e., a Sampling Table is not adapt at all to let the Advanced Sampling be operative and successful. A straight answer is that if the Projects' successes are counted in terms of Customers' acceptations (and, payments) for the unacceptable, with the time they'll create an artificial sense of correcteness of the technical solutions from whom the row of acceptations and payments derived.   To understand how this can happen, there is to think that the Root Cause of the problem, lies in the fact that frequently OEMs and Vendors measure the Projects’ successes, attributing the highest weight to the fact that a Project (design included) provided full payments. 

The rationale is that if: 

  1. after long time, a Food and Beverage Packaging Line Project remains partially unpaid by the Customer, then its snagging-list is nearly surely including issues it was not possible, viable or economic to fulfill for the Vendor;
  2. the Customers accepted the Food and Beverage Packaging Line, it meant that that Design fulfilled the minimum requirements expressed as Technical Guarantees, integral part of the Contract.

If one of these this is your Case, remember that the extra-Contractual after-Commissioning agreements, trying to limit to ridicolously small amounts the number of bottles and cans to sample, have not to be accepted.   OEMs and Vendors know how to really solve the problem, adding the missing ton of stainless steel, separated lubrication circuit, its automation and expertise, honouring the Contract and reaching your final satisfaction.   Notoriously, Marketing studies showed since many years that a satisfied Customer has a probability four time higher to reacquire products of the compliant Vendor, than from new untested Third Parties.   By the cold analysis of facts, there is no doubt that many directives between OEMs and Vendors ignore systematically the Marketing Golden Rule.   

How-much lubrication requires a Sampling Table ?

On Bottler’s side there is one very important requirement: 

                  he has to provide a separate lubrication to the Sampling table.

Why “separate” ?  

 A rejection Table devoted to empty glass bottles' standstill rejection.  An application as critical as the Sampling of rows of full bottles, requesting similar amounts of lubricant.  What visible is a reference for the amount sufficient to assure correct sampling of 60000 bottles-per-hour (~17 bottles-per-second) on polished chains at a linear speed of ~1.3 m/s.  We are suggesting a separate lubrication system to have so high lubricant's levels, specific for Sampling, only around each Advanced Sampling's operation

Because it has to be very dense and constantly applied.  A hyperlubrication different than what is expected, designed and integrated for a simple Rejection application.  To reject roughly speaking means <10 consecutive containers.   But, Advanced Sampling may signifies the 173 consecutive Valves of a Filler Machine.  In other words, rarely or never the Conveyors’ Vendor has the know-how necessary to fulfill a correct high speed sampling.   Then, the Conveyors’ Vendor is the first who did not set a lubrication system adapt to assure high-speed sampling.    Not all Bottlers are so available to accept the unacceptable and the lubrication is, in the end, their only duty toward a correct Sampling in their own interest.  Separate lubrication means a hydraulic circuit providing extremely high amounts of lubricant originating by the same common circuit of the Bottling Line, but whose timing is controlled by a separate timer.  Timer changing the standard production duty-cycle to a heavy-duty, after an external command.  Command which in the simplest case, can be incoming by an Operator-controlled manual push-button.   Today, all Advanced Sampling operations can be made automatic directly in the Electronic Inspector.   They always exist opto-coupled potential-free digital inputs specific to start a Sampling operation, after an external command, as an example by the Programmable Logic Controller of the Filler Machine or as a timed output by the Lubrication Controller.

How much lubricant shall be applied in a sampling application, with respect to the lubrication it’d have been in case of simple rejection ?  

The amount depends on several factors, main being containers scalar speed and production. 

As an example, we consider the case of sampling for:

  • 500 ml;
  • PET bottles, 5-leaves petaloid bottom;
  • production 50000 bph;
  • linear speed 1570 mm/s.

Sampling table shall be:

  • 6-belts;
  • 3 m of active sampling area (table length > 3.5 m);
  • plastic polished belts;
  • optional guiding system for bottles;
  • lubricated > 4 times the amount recommended for a common rejection table.

Case Study  

Sampling 90000 cans-per-hour

First sight 

A Sampling Table was guaranteed allowing cans' Sampling at 90000 cph of 168 Filler valves (~ 2.5 m/s).   Like example for this Sampling Table, refer to the illustrative Layout drawing in the top of this page.    All “looked” correct: a separate reject in a bin before a sampling table, careful execution, shining stainless steel, etc.   But starting to look closely the Sampling Table and comparing its sizes and architecture with the Electronic Inspector Vendor's specifications for Sampling, they are discovered differences.   

  A Sampling Table had been guaranteed allowing cans' Sampling at 90000 cph of 168 Filler valves.  Visibly, the reference Table above does not applies.  At over 60000 container per hour, are necessary at least 8 belts and a rejection table length of > 4 m.  In the specific case of the video opening this web page,  the speed is so high (90000 cans-per-hour) that our own precedent experiences showed necessity for at least 9 belts and a rejection table length of at least 5 m.  But, as visible, there are only 4 belts.  Worse, the Rejection Table Length is < 3 m.  This way, also the presence of an excellent lubrication assured by the Bottler let us all understand the Sampling Table arrived 4 times smaller than due

Differences in the amount of the belts and in Rejection Table.  Enormous differences.   Yet after this comparison and before to start to run 2.5 m/s that system, it’d be possible to predict that cans shall fall, jam and block the Canning Filler out feed.  

This, after hitting the opposite guide close to the fourth belt, and rebound back toward the inclined one attached to the wedge.   And, what about the belts?   Definitely non-polished, when in the reality they also exists high quality (and, expensive) plastic belts whose friction coefficients are not poor with respect to those expected by polished stainless steel.   

An example of violation of rules, which shall limit the sampling of Filler and Seamer whose nominal production is 90000 cph, making unuseful at all the operation along next decade.  Comparing the Sampling Table arrived with the Table below, the situation gets out immediately until its fine-details. And this, well before to start the system.  

In brief:  

  1. over 60000 container per hour are necessary at least 8 belts and at least a rejection table length of 4 m.   In the specific case in the video opening this web page, the speed is so high (90000 cans-per-hour) that our own experience showed necessity for 9 belts. But, as visible in the figure here, there are only 4 belts;
  2. the expected length of the Rejection Table Length is at least 4 m. Our own experience shows necessity for a Rejection Table Length of 5 m.  But, visibly, the effective Rejection Table Length is limited to only 1.2 m.   

Making a few calculations:

9 belts / 4 belts  =  2.25              

5 m / 1.2 m  =  4.2                    2.25 x 4.2   =   9.

Say a Sampling Table whose surface is 9.4 times smaller than necessary to fulfill the Contract Guarantees.   Nearly one ton of stainlees steel paid by the Bottler and missing.

 Close-up on the Sampling Table over 9 times smaller than due.  It’d have had at least 9 belts, with the last two, the 8th and 9th running 20 % slower than all others.  But it has only 4 belts.  It’d have been 5.5 m long, so to have at least 5.0 m assured as Rejection Table Length.  But it has a Rejection Table Length limited to 1.2 m.   Why all this ?   The most common reason is to keep artificially low the final price of a commercial offer, making it more interesting for the prospected Buyer.   Who designs and fabricates Conveyors knows that Packagers are more sensible to the single amount represented by the total price, than to the fine-details about the Advanced Sampling.  Later, one time an Offer is accepted, the task passes to other Staff to obtain Bottler’s acceptation for something unacceptable.  The entire Bottling Line hosting such a Quality Control limitation, shall never deliver the Quality expected by who accepted that Offer


   Production speed                        90000



   Rejection Table length [ m ]             5

   Conveyor belts                                 9

   Polished belts                                 1-7

   Synchronised belts                      8+9: -20%



Filler Sampling

Implying the Sampling Table arrived over 4 times smaller than the Sampling Table recommended by the Vendor of the Electronic Inspector.   We reduced ourselves to ask to the Bottler to accept to sample only a few of the 168 valves, just no more than 6 consecutive.  


  • 168 consecutive cans, resulted in a 100 % of crashes (porcentage measured along 4 tests) and associated expensive and painful stops to production to free the entire zone by the hundredths of broken cans sprinkling beer.
  • 12 consecutive cans, resulted in a 70 % of crashes, porcentage measured along 6 tests.
  • 6 consecutive cans, resulted in a 10 % of crashes, porcentage measured along 10 tests.

Seamer Sampling

Now, imagine a Seamer born with 12 seaming heads: nor it was possible to sample correctly, say in a single row, at least its seaming heads' function.  


  • 12 consecutive cans, resulted in a 70 % of crashes’ occurrences, in the minuscule Sampling Table.
  • 6 consecutive cans, resulted in a 10 % of crashes’ occurrences, porcentage measured along 10 tests.


Sampling Table always jammed


The most typical problems manifest in a Food and Beverage Bottling Line derive by missing materials like motors, Frequency Converters, detectors, etc.   In the following we’ll briefly study a Case where the “material” was much more than necessary.   The same, it was not possible to let the Advanced Sampling of a big Canning Filler Machine and related Seamer be feasible.    

Close-up on the sampling Table and related Rejector

The figure below shows the Sampling Table designed with:

  • 10 belts;
  • 5 meters Rejection Table Length, which is the distance from the center of the Rejector to that point of the Table out feed “cone” corresponding to a space equal to 1.5 container's diameter (1 + 1/2 can's diameter equals 96 mm, in this Case).


   A great Sampling Table with 10 belts, whose Rejection Table Length is 5 meters.  Considering that the Filler Machine to control has a nominal speed of only 72000 cans-per-hour, this Sampling Table is over-dimensioned.  A design with 8 belts and 4 meters is cheaper and adequate to the task.  The out feed of the sampling area visible in the image enters in a single-belt conveyor (at left-side) necessary to prevent that cans arrived before exchange position with cans arrived later. 

Results of this over-sized Sampling Table resulted poor because:

  1. (Installation)  The side guide, indicated by the arrow, presents a profile inducing fallen cans due to cans’ anomalous fast deceleration.  Fallen cans becomes jams in the end of the Table, where it starts the single way belt devoted to accumulate the orderly row of Filler Machine valves and Seamer heads;
  2. (Design)  The belts chosen have high friction coefficients, favouring fallen cans and successive jams;
  3. (Design)  No separate lubrification system existing with the effect that front of many more belts than necessary, no lubricant at all is visible, favouring fallen cans and jams.

A Table of this size should have been excellent to sample can at a production speed:

  • Cans at > 90000 cans-per-hour, > 2.3 m/s linear speed; 
  • PET bottles with petaloid base, at > 66000 bottles-per-hour, > 1.8 m/s linear speed;
  • Glass bottles, with standard knurling marks under the base, at > 72000 bottles-per-hour, > 1.8 m/s.

But, when considering that the Can Filler Machine to control has a maximum nominal speed of 72000 cans-per-hour, this Sampling Table results definetely over-dimensioned. 

A design with 8 belts and 4 meters should have been adequate to the task and cheaper.   The out feed of the Sampling area visible in the image above, enters in a single-belt conveyor (visible in the figure below) necessary to prevent that cans arrived before exchange position with cans arrived later.   This single belt Conveyor lies in the left side of the image.  

3 Root Causes 

Results of this over-sized Sampling Table were poor because of three independent Root Causes (italics) related to:

  1. Installation:   the black colour side guide indicated by the arrow below, presents a profile inducing fallen cans due to cans’ anomalous fast deceleration.   Fallen cans becomes jams in the end of the Table, where it starts the single way belt devoted to accumulate the orderly row of Filler Machine valves and Seamer heads;
  2. Design:   the belts chosen are models with high friction coefficients, favouring fallen cans and successive jams;
  3. Design:   no separate lubrification system existing with the effect that front of many more belts than necessary, no lubricant at all is visible, favouring fallen cans and jams.

Incoherences like these are frequently encountered and the list above pin-points what to check first, when Sampling operations results in stopped Bottling or Canning Filler Machines.

  Cans jammed, blocking the entire Table, after attempting a Filler-, Capper-, Closer- or Seamer-Sampling operation.  Root Causes for the impossible Samplings are the superimposed issues numbered above 1., 2. and 3

Links to other pages:

This website has no affiliation with, endorsement, sponsorship, or support of Heuft Systemtechnik GmbH, MingJia Packaging Inspection Tech Co., Pressco Technology Inc., miho Inspektionsysteme GmbH, Krones AG, KHS GmbH, Bbull Technology, Industrial Dynamics Co., FT System srl, Cognex Co., ICS Inex Inspection Systems, Mettler-Toledo Inc., Logics & Controls srl, Symplex Vision Systems GmbH, Teledyne Dalsa Inc., Microscan Systems Inc., Andor Technology plc, Newton Research Labs Inc., Basler AG, Datalogic SpA, Sidel AG, Matrox Electronics Systems Ltd. 

                                                                                                                      .                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         .                   
TRUSTe Certified Privacy Seal

.                 .                                                                                         .                                                                                              .                                                                                                                                 66                                                                                 Webutation

                                                                                                                                                                                                                                                                                                                                                          .                                                                                                                         PRICE LIST       REFERENCES      SERVICES      CONTACT     MULTIMEDIA       TECHNOLOGIES       DEVELOPER       TERMS                                                                                                                                                                                                                                                                                                                         .                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           
TRUSTe Privacy Policy Privacy Policy
Site protected by 6Scan