Controlling Coater Gap Uniformity

Published: November 01, 2002, By Robert Foster and Bryan Manning, Capacitec

The common thread between end-users and various coater suppliers is the utilization of slot extruder dies to apply a variety of very thin chemical, adhesive, and photographic coatings to a variety of media.

Specific examples of applications include:

• adhesive coatings onto labels

• chemical coatings onto films and tapes

• photographic coatings onto films

• manufacture of plastic tapes.

In a typical application, the coater die slot gap sizes range from 0.006-0.024 in. (150-600 microns) with a typical set of slots being 0.006, 0.008, 0.010, 0.012, 0.014 in., etc. The length of the slot gap is typically 3-6 ft (1-2 m) wide.

Since there is a direct relationship between setting the width of the slot gap and the thickness of the coating material, it is critical for manufacturers to set a very uniform gap along the full length of the coater die. The uniformity must be held at the microinch level or 0.01 microns.

Traditional Measurement Methods
In earlier days, manufacturers using old measurement methods were forced to live with gap variations of more than 50 microinches (1.25 microns) over the length of the coater dies. They would set the gaps in a variety of time-consuming methods using traditional standard metrology tools.

One method was the use of feeler gauges. This procedure could not provide the required accuracy, and repeatability was poor due to the subjective variation between users. In addition, the feeler gauge method suffered from several problems:

• It could damage highly polished surfaces such as the mouth of a coater.

• Once the gap was set, it was very difficult to recheck the actual dimension.

• Feeler gauges could not accurately measure “inboard” gaps.

Another method was to split the extruder die in two and put each half on a granite metrology table. Flatness measurements then would be taken with the use of displacement sensors measuring from above. The next step was to polish each side to obtain an ideal matched set and bolt the two sides back together.

Gap uniformity variations are caused by two factors. The first is a result of variations in the planarity in each half of the die as measured on the table. Additional variability could be seen during re-assembly of the die as a result of variations in the torque applied to the mounting bolts. The combined impact of these factors resulted in a gap uniformity variability of 10-40 microinches (0.25-1.0 microns).

Another challenge with the granite table method was to measure not only the gap along the length of the coater mouth but also selected profile points within the coater to identify “choking points” inboard in the die.

The only way to check the results of this measurement method was to remount the die in the coater and make a test run of material. The coated material then was measured after the fact to determine relative uniformity of the coating deposited on the media.

This “metrology table” gap setting method presented a variety of problems and challenges, including the following:

1. The process was time-consuming, requiring hours from skilled technicians.

2. There were combined flatness variations of at least 10-40 microinches (0.25-1.0 microns).

3. Successful measurement required performing several test runs, wasting time and dollars, especially when exotic material was involved.

4. Since nominal gaps are set based upon minimum dimensions, material often was wasted to normalize coating thickness across the remainder of the product.

5. It was difficult to measure variation or “choke points” inbound in the coater.

Over many years of using traditional measurement methods, manufacturers concluded the most critical cause of these problems was the difficulty in maintaining coater die gap uniformity to higher levels of accuracy. Several set a benchmarking goal to attain a level of less than 10 microinches (0.25 microns) gap uniformity across the full length and width of the slot.

The ultimate advantage of controlling the constancy of the slot gap will be that manufacturers can move from post-process control (where measurement is done after the coating takes place) to pre-process control (where measurement and adjustments are done before the run is started).

Measurement Solution
The benchmark level of less than 10 microinches gap uniformity of slot die gaps has been accomplished recently. In fact, in some applications it has been exceeded, with uniformity being maintained at levels down to 5 microinches. The goal was met after several years' development in close participation with customers. This section will describe the steps taken on the road to success.

The first step was the selection of capacitive technology as the basis for sensor design. The choice of capacitive was based on several of its inherent advantages including:

• noncontact measurement method

• ultra-thin composite sensors

• linear analog output

• excellent repeatability

• high-temperature capability

• good value solution.

Principle of Operation
Coater gaps are measured with two capacitive displacement sensors mounted back-to-back at the end of a flat wand. Each sensor has a central sensing element with a typical diameter of between 0.079 and 0.197 in. (2 to 5 mm). A ring layer called a "guard," which is approximately twice the diameter of the sensor, surrounds the sensor. The guard serves to focus the capacitive change field, and both parts are connected separately to a 100% shielded coaxial cable.

When positioned parallel to an earth-grounded or conductive target, the sensor/guard combination measures a capacitance proportional to the air gap. When the signal is input to a specialized signal conditioner amplifier, the output can be spread linearly between 0-10.000 Vdc as related to a gap spacing from almost touch to some full-scale dimension. This operation allows the ability to show 1 part in 10,000 resolution. For example, a full scale range of 0.010 in. divided by 10,000 yields a resolution of 1 microinch/mVdc (0.25 mm divided by 10,000 yields 250 nanometers resolution).

Sensor Selection
The capacitive sensors are attached back-to-back on a sensor wand. The configuration, thickness, and material of the sensor wand depend on the application at hand. Sensor wands come in two groups (Kapton or Composite) depending on thickness. When very thin gaps (from 0.009 to 0.10 in. [0.23 to 2.5 mm] are measured, the sensor wands typically are made of Kapton in thickness ranging from 0.009 to 0.040 (0.23 to 1.0 mm). Standard Kapton wand length is 7.8 in. or 200 mm. In applications in which the gaps are wider (from 0.40 in. [1.0 mm] and up), the two sensors are mounted on composite stainless steel laminated wands with plastic protective covers.

Maximizing Accuracy
In the science of capacitive measurement technology, there is a relationship between range and accuracy. The smaller the range, the higher the accuracy and linearity. The selection of the sensor wand thickness therefore is made in relationship with the gap size.

Wand thickness is selected to create an overall range of 0.010 in. (250 microns). This produces an accuracy of ±0.2% full scale (FS). Even higher accuracy can be attained by selecting the wand thickness to be a maximum of 0.004 in. (100 microns) below the targeted slot gap. These two wand-section criteria combine to give an overall accuracy of better than ±0.1% FS (e.g., 4 microinches/0.01 micron accuracy).

Custom Fixtures
An additional discovery uncovered during the design process was the importance of wand positioning when taking gap measurements. The best measurements were attained when the sensor wand was held stable in a parallel position relative to the two halves of the coater die. When the wand was allowed to twist or rock out of this position, accuracy and repeatability would deteriorate.

In order to assure best-case parallelism between the sensor wand and the die slot, a special custom fixture was designed.

This fixture offers the following improvements to the measurement process:

• Allows for easy handling and positioning of the wand into coater die slots

• Prevents twisting of wand when positioning wand for measurement

• Prevents wand from rocking out of the ideal measurement area.

Instrumentation
The dual-sensor probes are combined with a matching electronics package that consists of an electronics rack, power supply, cables, and amplifiers. Usually they are matched with two channels of amplifiers, one amplifier for each of the two sensors on the probe.

Maximizing System Value
In traditional systems each gap sensing probe is matched to a dedicated set of two amplifiers. Since most coaters have a range of different die slot sizes (e.g., 0.008, 0.010, 0.012 in.), users would need a large number of dedicated amplifiers for each of the gap sizes. This requirement could push up the cost of a system significantly.

Concerned about total system costs, users asked for ways to reduce the overall expense of a multi-gap measurement system. In response to this need, a specialized software program was developed.

The software reduces cost through a method that allows two amplifiers, normally dedicated to only one dual-sensor gap wand, to be used on several sets of wands without additional cost. This is accomplished by using the software to create “virtual” amplifiers from two actual amplifiers used for one of the dual-sensor probes in the system's multi-probe sets. A pair of virtual amplifiers then is used for each of the additional dual-sensor probes in the set.

The technique for creating “virtual” amplifiers with this program is linked to the sensor channel-to-amplifier calibration process required for each channel (sensor) of the dual-sensor probes as part of the measurement system implementation. Each of the two channels of the first dual-sensor probe is calibrated to a matching amplifier. The calibration sensitivity is 0 to 0.10 in. (0 to 250 microns) = 0 to 10 Vdc.

The analog output voltage from the calibration process is fed into the program, which has a real-time calibration module that takes the analog output voltage from the amplifiers and turns it into linear engineering units (either English or metric) using polynomial interpolation. Then the next dual-sensor probe is matched to the same pair of amplifiers with the polynomials adjusted to calibrate the second wand. This process is repeated to create a set of virtual amplifiers for each additional dual-sensor probe used with the system.

The calibrations created for each dual-sensor probe are verified with a primary standard that has the precise dimensions of the extrusion slot being measured.

The linearization feature of the program also helps reduce cost. This feature allows the use of lower-priced amplifiers while still offering repeatability of ±0.01% of FS and an accuracy of ±0.1% of FS or better.

The program was developed under National Instruments' LabView program and operates under Windows 95/98/M.E. and 2000 on standard PCs (the program is typically used in conjunction with PCMCIA data acquisition cards widely available for use with portable computers).

Robert L. Foster is president/technical director at Capacitec, Ayer, MA. He has a BA in physics from St. Michael's College and spent the past 25 years developing noncontact capacitive and strain technology for use in the coating, nuclear quality control, aerospace, and manufacturing engineering sectors. His patents include one for a coaxial contact connector used in conjunction with super-thin gap measurement sensors. Contact him at 978/772-6033.

Bryan Manning is president of B&D Intl. He has an MBA from American Graduate School of Intl. Management (Thunderbird). He spent the past 25 years in senior sales and marketing positions for companies such as Invensys, Esterline, and Texas Instruments. B&D Intl. is a marketing and sales consulting company specializing in helping technical companies increase sales and profits in the US and Europe. Contact him at 978/772-6033.

The views and opinions expressed in Technical Reports are those of the author(s), not those of the editors of PFFC. Please address comments to author(s).