UV Ink Spitting: FQC Team Isolates Variables, IDs Factors Increasing Likelihood of Phenomenon

Flexo Quality Consortium FQC LogoUV ink spitting, nemesis for many a narrow web printer, has been an elusive quarry for quite some time. It is a rare but certainly troublesome aspect of printing flexo. The process of inking an anilox roll, shearing off excess ink via the use of a doctor blade and transferring a consistent ink film is the basis of today’s flexographic printing. Unfortunately, there are times when this process does not behave as one would expect. A recurring challenge for flexographic printers is the UV ink spitting phenomenon.

Spitting can be commonly described as the escape of the ink from the confines of the blade/anilox contact point. Once past the blade, the ink builds up on the opposite side of the blade. Accumulation of ink releases and creates print flaws, typically in the form of a randomly placed teardrop shape. The frequency of UV ink spitting is often so small for a converter that the incidents tend to be treated as isolated occurrences. This method tends to find individual cures but no overall understanding of the cause.

Combating the spitting phenomenon by a converter typically consists of trying any number of solutions, including changing inks, blades, anilox rolls and even manipulating any or all of the three, moving to different print stations, etc. Unfortunately this does not allow for the isolation of the causes, and every day brings yet another solution to make the problem go away. The enigma of spitting persists in the endless need for so many solutions.

The Flexo Quality Consortium (FQC) assembled a team of experts to explore the problem in hopes of coming up with a better approach. Consisting of representatives from the fields of converting, doctor blade, ink, anilox, measuring equipment, press manufacturers, statistics and educational institutions, the team was very well-balanced and contributed mightily to the conclusions.

By charter, the goal was to determine possible preventative setup measures and isolate variables to the process. FQC analysts do not think there is anything inherently wrong with anilox rolls, inks, blades or press equipment, so the question the teams set out to answer was how to get the four to work together without fault in the flexo process.

Initially, the group met via conference call to determine how it wanted to approach the testing and analysis phases. In order to get everyone involved, the FQC team spread the “wealth” of testing as much as possible. Initial testing was done at Clemson University, with follow-up trials at locations in Plymouth, MN and Dayton OH.

Evaluation Criteria

Ink Spitting
Focus & Findings

  • FQC team chartered to explore ink-spitting phenomenon, isolate variables and determine possible preventative setup measures
  • There is no universal cure for UV spitting
  • Discovery of causes linked to measurements and
  • Use of a matrix evaluation isolates spitting to more identifiable root causes
  • A fingerprint is the prescribed method to understand the nuances of any particular situation

The first and most critical step was to identify what to use as the measurements of success. Each “operator” graded the set of print samples from a trial and scaled it on the following:

  • 0: No spitting on the sample (2 print repeats)
  • 1: 1-5 incidences
  • 2: 6-10 incidences
  • 3: More than 10 incidences

Based on scoring evaluations, the FQC team was able to then have a statistician determine significance.

Key Variable IDs

It was agreed to take a matrix approach so we would not get lost, but rather stay organized in the mountain of potential data. Our matrix consisted of categories that the team felt were very important to the discovery of the causes of UV spitting. It decided to measure or control the following variables (examples):

  1. Anilox (500 cpi vs. 800 cpi)
  2. Run speed (Ex. 300 fpm vs. 500 fpm)
  3. Blade pressure (light, constant)
  4. Blade width (28-mm., 30-mm., 32-mm.)
  5. Blade tip (standard 15 degree, Lamella, long bevel 4 degree)
  6. Blade material (carbon steel, tool steel, plastic)
  7. Viscosity (low, medium, high, very high, measured in centipoise)
  8. Operator (how results were judged)

The test matrix was designed to allow for adjustment based on what FQC team members wanted to hold constant and what they wanted to evaluate. Headings contained the test parameters and the amount of variation were designated accordingly.


Many of the trials conducted found no spitting whatsoever to evaluate. Even what was considered to be a verification trial at Clemson, yielded no repeat in the spitting results experienced before. FQC was left to ponder the meaning of this information. It put a lot of effort in making sure there were no outside variables, so the conditions of the blade holders were paramount to the process. Each was meticulously cleaned so interference would not be an issue. Anilox rolls were prepared and verified so the cell condition was at maximum volume. Inks were verified beforehand for viscosity. In the end, the focus on preparation mattered to the outcomes.

FQC turned to the statistician, who took the original data from the initial success and evaluated it. Examining the data, the statistician segregated ratings within the parameters (variables FQC chose to examine) if there was a great difference, he used standard deviation of the difference between predicted and actual ratings and finally corrected the ratings, taking the most significant into account. He examined each parameter for the largest effect, one at a time. The most significant was omitted each time, significance meaning the most reduction in variation. The cycle was repeated until the remaining parameters were minimized. In one case, the statistician separated out differences within a variable. The significant difference within operator was classified as “harsh” and “not harsh” based on the subjective evaluation difference of the same print samples by the participants.

For this particular trial, the results by significance came in this descending order:

  1. Anilox cell
  2. Blade material
  3. Blade width
  4. Operator
  5. Speed

The statistician prepared a chart to explain the significance further by placing what he deemed an “up charge” based on the known variables. Essentially, this means that depending on what components you chose within the variables, you could increase or decrease the likelihood of experiencing UV spitting. If a variable was assigned an “up charge” greater than the “up charge” of the current variable employed, the likelihood of spitting increased. In contrast, switching to a lower value for the same variable decreased the chance of spitting.

For this particular trial, the chart accompanying this story outlines the “up charge” cost:

Variables that Influence UV Ink Spitting

For the values 0.0, there is of course no additional chance of spitting. It also said that by increasing viscosity, there would also be a scalable increase in risk. Going with a blade that was wider (32-mm.) than the OEM spec for the blade holder (30-mm.) increased risk.

In this setup, the use of a lower line, higher volume roll would also increase the likelihood, and so on. We learned in later trials that the differences found in this particular trial had no similar bearing in other tests, which speaks to the fact that there are things still unmeasured and unaccounted for in our analysis.

We had an occasion to change viscosities from 1,000 centipoise to 3,000 centipoise, also increasing the speeds to the press maximum of 400 fpm and nary a spit was identified in any of the samples generated. This again points to the need of “fingerprinting” the UV spitting situation for that particular press, ink, anilox roll, blade and blade holder combination in order to determine cause.

Explaining Unknowns

FQC UV Spitting Project Team

Tom Allison – Allison Systems Corp.
Mike Buystedt – Flint Group Narrow Web
Dat Tony Chu – Mark Andy Inc.
Steven Cicchese – Brookfield Engineering Laboratories
Kern Cox – Clemson University
Bobby Furr – Max Daetwyler Corp.
John Hailey – Esterlam International
Arthur Hazleton – Creative Labels Inc.
Mike Huey – Harper Corporation of America
Jean Jackson – Praxair Surface Technologies, Inc.
John Kilbo – Zeller+Gmelin
David Larson – Brookfield Engineering Laboratories
Daniel Muthig – Fox Valley Technical College
Bobbi Olp – All Printing Resources, Inc.
Dean Radford – Environmental Inks & Coatings
Richard Selvidio – Actega WIT
John Seymour – QuadTech World
Perry Stacks – Environmental Inks & Coatings
Johnny Stamey – Max Daetwyler Corp.
John Valla – Nazdar Inc.
Steve Webendorfer – Standard Register

The statistician also determined that the variables that were controlled accounted for 30 percent of the variation, leaving 70 percent unknown, which means the experiments were deemed “out of control” in that the variables “controlled” weren’t all the key things controlling output.

FQC came up with a list of possible variables that were not measured, because they were not incurred, making up the 70 percent:

  1. Eliminating blade flutter from weak or worn blade clamp assembly. When any blade wiggle/flutter is allowed mechanically, it is magnitudes above any control you have with an ink, anilox or blade itself
  2. Adding variability to the installation process by having dirty blade holders, causing blades to not lay flat in the holder
  3. Ink that had not been disturbed/mixed prior to used.
  4. Unknown, unmeasured properties of the inks, blades and anilox rolls
  5. Anilox rolls were not clean, bearing failure, TIR (total indicated runout) issues

You can look at the lack of spitting in many of the trials as a failure, but it can also be said that to repeatedly avoid the sequence is a testament to what FQC researchers were doing correct. Meticulous preparation went into each trial to avoid such unknown, unmeasured occurrences. Team members believe that addressing spitting with the preparation points listed above, as well as evaluating the spitting with the use of a matrix, helped isolate spitting to more identifiable causes for the converter.


FQC research has determined that while there is no universal cure for UV spitting, a comprehensive effort in the form of a fingerprint is the prescribed method to understand the nuances of any particular situation. It is similar to the challenges of uncontrolled process printing. Once you fingerprint and define variables, then you can progress to a workflow solution.

FQC developed an effective method by using the matrix analysis that it hopes to see applied to similar situations, so the root cause can be better defined. The matrix method has been applied in the field and has netted positive results for those converters.

FTA’s FQC team will remain in place and the customizable matrix will be available for any converter who chooses to run an analysis of the situation to help improve this process to eliminate a majority of spitting issues. If you need help with your UV spitting, contact FTA for assistance.

About the Author: Sean Teufler proudly serves the central United States as a technical graphics advisor for Harper Graphic Solutions, a division of Harper Corporation of America. He is currently leading the Flexo Quality Consortium (FQC) project on Narrow Web UV Ink Spitting and will serve as co-chair for FTA’s 2012 Fall Conference, “Efficiency in Flexography: Connecting the Dots.” The event is set for Milwaukee, WI, Oct. 15-17.