Download IPC 7801 - Reflow Oven Process Control Standard

IPC-7801 Reflow Oven Process Control Standard
Working Draft 2-27-2014
(Text highlighted in GREEN is for the use of typesetting only and will be removed before publication)
1 Scope This standard is provides process control for (conveyor/convection – traditional SMT?) solder reflow ovens
by baseline and periodic verifications of oven profiles using a standard methodology. Equipment calibration and
maintenance guidelines are provided.
For detailed information on development or verification of a product profile/recipe see IPC-7530.
1.2 Purpose Intended for users of reflow equipment to baseline performance and periodically verify and demonstrate
acceptable oven performance repeatability.
2 Applicable Documents
The following documents of the issue in effect on the invitation for bid form a part of this specification to the extent
specified herein.
2.1 Joint Industry Standards1
J-STD-033 Handling, Packing, Shipping and Use of Moisture/Reflow Sensitive Surface Mount Devices
J-STD-075 Classification of Non-IC Electronic Components for Assembly Processes
2.2 IPC2
IPC-1601 Printed Board Storage and Handling Guidelines
IPC-7530 Guidelines for Temperature Profiling for Mass Soldering Processes (Reflow & Wave)
IPC-9194 Implementation of Statistical Process Control (SPC) Applied to Printed Board Assembly Manufacture
Guideline
IPC-T-50 Terms and Definitions for Interconnecting and Packaging Electronic Circuits
2.3 ANSI3
ISA MC-96.1 – Temperature Measurement Thermocouples
(Footnotes)
1. www.ipc.org
2. www.ipc.org
3. www.ansi.org
3 Terms and Definitions Other than those terms listed below, the definitions of terms used in this standard are in
accordance with IPC-T-50.
Convection Reflow Soldering A solder reflow process where the primary means of heat transfer is by the
recirculating flow of heated air or nitrogen in a Reflow Oven.
Delta T Temperature variation across an assembly, specifically the difference between the highest and lowest
temperatures on an assembly when measured at Peak Temperature in the profile (Figure 3-1). A low Delta T is
desirable.
1
Figure 3-1 Identification of temperature delta at peak reflow temperature (NEW)
Delta T = Difference between highest measured temperature and lowest measured temperature on the
assembly.
Golden Board A reusable test vehicle with thermal characteristics similar to a production assembly, used to measure
Reflow Oven performance and repeatability, and establish reflow Recipes. Sometimes called a “witness board”,
“standard test vehicle”, “thermal profiling board”, or “reference board”.
Liquidus The temperature at which a solder alloy is completely melted.
Peak Temperature The maximum temperature experienced at any point on an assembly during reflow soldering. The
highest point on a Reflow Profile.
Preheat A preliminary phase of the soldering process during which the product is heated at a predetermined rate from
ambient temperature to a desired elevated temperature. Preheat reduces thermal shock and ensures uniform heating of
the assembly up to Liquidus.
Profiling System A data logger or measuring instrument for logging temperature and time data from thermocouples.
Also “Profiler”.
Ramp A controlled and uniform increase or decrease in temperature, represented as a constant slope on the Reflow
Profile. The magnitude of the slope is known as the Ramp Rate.
Recipe A defined set of process parameters programmed into a Reflow Oven (Figure 3-2). It includes a specific
conveyor speed and temperatures within each Reflow Zone, and possibly the flow rates of air or nitrogen. The recipe
typically varies with the thermal mass and other heat transfer characteristics of the assembly being soldered.
2
Figure 3-2 Example of reflow oven recipe set points (NEW)
SP = Set Point
PV = Present Value (Measured Value)
Reflow Oven A solder reflow system, typically using mostly convection heating in an air or nitrogen environment.
Conveyorized systems incorporate multiple reflow zones in series. Batch ovens offer lower throughput and are less
common.
Reflow Profile A graphical representation of temperature for a single or multiple locations on an assembly, plotted
against time, during the reflow process. It may be a recommended “baseline” or Target Profile, or reflect actual
measurements. Also known as "Oven Profile".(Figure 3-3 & 3-4)
Figure 3-3 Zones is a typical reflow profile (IPC07530-3-2)
3
Figure 3-4 Typical graph representation of a thermal profile (NEW)
Reflow Zone A local environment within the Reflow Oven where a specific temperature is maintained in accordance
with a Recipe. Heating zones are arranged in order of increasing temperature along the direction of conveyor travel.
Cooling zones control the rate of solder solidification and cooldown of the assembly.
Soak A portion of the Reflow Profile where the oven temperature is held nearly constant to allow temperatures to
stabilize or become uniform across the assembly being soldered, minimizing Delta T.
Spike A portion of the Reflow Profile where the oven temperature increases rapidly past the Liquidus point and then
cools off quickly. This permits wetting and flow of solder during reflow while minimizing time at Peak Temperature.
Thermocouple (T/C) A temperature sensor consisting of two dissimilar metals in physical contact. The electrical
potential (voltage) between the contact point and a reference point in the circuit is proportional to the temperature
difference between those two points. This small voltage is measured by a test instrument, which typically converts it to
a temperature.
User The individual responsible for controlling and performing reflow soldering.
4 Thermal Profiles – SnPb and Pb-Free
4.1 Recommended Reflow Profile Specifications The recommended baseline reflow profile is defined by the solder
paste supplier/manufacturer for the flux component and the alloy used and the component suppliers and board
specifications to ensure temperature limitations are not exceeded (MB,HP). The oven supplier does not define the
recommended reflow profile specification. The specification recommendation should be cross referenced with
component specifications and board specifications to ensure the lowest maximum temperature limitation on materials
used is not exceeded.
Meeting the requirements of an acceptable reflow profile will produce consistently acceptable solder joint quality as
well as undamaged components and PCBs for a typical printed board assembly. (MB,HP)
. Variations/deviations from this recommended baseline reflow profile may be required for printed board assemblies
which have a large variation of thermal mass (eg. very large thermal mass components and very small thermal mass
components on the same assembly).
The baseline recommended profile is typically available within the suppliers solder paste data sheet provided with the
solder paste. The data sheet is typically available on the solder paste supplier’s website.
For guidance on setting up a specific recipe, see IPC-7530.
4
4.2 Example of a Typical Tin-Lead Reflow Profile Specifications A general ramp/soak/spike profile for an SnPb
alloy may look like Figure 4-1.The elements of a ramp/soak/spike profile are:
 Preheat/soak time = 60-90 seconds
 Liquidus = 183˚C
 Recommended peak temperature = 210-225˚C
 Time above liquidus = 30-90 seconds
 Recommendations from solder paste manufacturer will vary based on formula and manufacturer. Consult
your solder paste manufacturer’s data sheet for specific time/temperature recommendations.
• Maximum package temp = J-STD-020 or Suppliers recommendation
• Maximum heating ramp rate = 3°C/s or as defined by component supplier
• Maximum cooling ramp rate = ≤ 3°C/s or as defined by component supplier
• Preheat/soak time = Refer to solder paste supplier guidance
• Liquidous temp = 183˚C
• Recommended maximum peak temperature = 230˚C
• Recommended minimum peak temperature = 210˚C
• Time above 183° C ( liquid) = 60-90 seconds (Minimum 60 seconds and maximum dependent on range of
solder paste and part supplier guidance)
• Maximum temperature difference (Delta T) across measured locations = 15°C
• Recommendations from solder paste manufacturer and component supplier will vary. Consult your solder
paste manufacturer and component supplier data sheets for specific time/temperature recommendations.
Figure 4-1 Typical SnPb ramp/soak/spike reflow profile (NEW)
4.3 Example of a Typical Pb-free Reflow Profile Specifications A general ramp/spike profile may look like Figure
4-2. The elements of a ramp/spike profile are:
 Preheat/soak time = 60-90 seconds
 Liquidus = 217-220˚C
 Recommended peak temperature = 235-255˚C
 Time above liquidus = 40-75 seconds
(sometimes 40-60 seconds)
 Recommendations from solder paste manufacturer will vary based on formula and manufacturer. Consult
your solder paste manufacturer’s data sheet for specific time/temperature recommendations.
• Maximum component/package body temperature = As defined by the component supplier
• Maximum temp. ramp up rate = 3 °C/s or as defined by component supplier
5
• Maximum temp. cool down ramp rate = ≤ 3 °C/s or as defined by component supplier
• Preheat/soak time = Refer to solder paste supplier guidance
• Liquidous temp = 217 ˚C
• Recommended minimum peak solder joint temperature = 230˚C
• Time above liquid = 60-120 seconds (Minimum 60 seconds and maximum dependent on range of solder
paste and part supplier guidance)
• Maximum temperature difference (Delta T) across measured locations = 15°C
• Recommendations from solder paste manufacturer and component supplier will vary. Consult your solder paste
manufacturer and component supplier data sheets for specific time/temperature recommendations.
Figure 4-2 Typical Pb-Free ramp/soak/spike reflow profile (NEW)
6
Users should monitor time above the component manufacturer’s maximum temperature threshold to insure that
sensitive components are not damaged. See J-STD-033 and J-STD-075 for thermal damage considerations for
moisture sensitive components, and IPC-1601 for printed boards.
4.4 Pb-free (SAC305) reflow profile temperatures Pb-free reflow profiles go to higher temperatures than those for
SnPb solder (Figure 4-3), and the user shall ensure all materials used in the printed board assembly (i.e. the
components and printed board) can withstand these temperatures.
Figure 4-3 Comparision of SnPb vs. Pb-free reflow profiles (NEW)
5 Golden Board Design for Process Verifications
5.1 Optimal Golden Board Using the actual product is always the best Golden board. The actual printed circuit
assembly gives feedback on the actual process and should provide more accurate data for troubleshooting or process
improvement. However, it is not always possible to use actual product; customers may not supply extra assemblies
and/or cost may be prohibitive. In these cases, a proxy for the Golden board may be useful.
5.2 Other Options for the Golden board The Golden Board design should be a board that can be produced easily and
quickly for low cost. After multiple uses, the board can break down or delaminate, components sustain damage, and
intermetallic compound (IMC) thickening changes thermal properties. The Golden Board can provide representative
data on the thermal excursion of product through the oven. When used consistently, the purpose of the Golden Board
is to show consistency and repeatability information to confirm that the way the process was operating yesterday is the
way it is still running today. The Golden Board can be any size or shape that best represents the product. Golden
Boards do not have to be complicated or complex. In fact, since their main function is to show repeatability, simpler is
often better because simpler is easier to make in quantity and maintain consistency.
5.3 Golden board Materials The laminate construction should be the same as that used in the product or approximate
the thermal properties.
Most printed board laminates will experience physical changes including delamination after repeated reflow cycles. To
extend use of the Golden Board, alternative materials can be considered.
5.3.1 T/C Location Thermocouples can be placed in a variety of configurations including cross belt (transverse to the
line of travel), along the line of travel, or a combination. (Figure 5-1)
7
Figure 5-1 Thermocouple placements (NEW)
Locating the T/Cs on the Golden Board is critical for acquiring accurate and useable data from any Profiling System.
Typical locations on the Golden Board include leading and trailing edges, left, center, and right.
It is common to use 3-5 measurement points (Figure 5-2). More connection points can be made depending on the
capability of the profiler being used.
8
Figure 5-2 Common thermocouple configurations (NEW)
5.3.2 T/C Attachment Methods – Bolt on T/C The bolt on t/c method is very simple to do and repeat using a screw
and a nut. This method is inexpensive, robust, and can survive many passes through the oven. Care shall be taken not
to touch the T/C wires to the outside of the screw head.
9
Some view the thermal mass of a screw head as not representative of an electronic component and as an “insulator”.
Still others consider the screw as repeatable and consistent from location to location, and using them on a Golden
Board can provide a repeatable thermal mass and rate of heat transfer.
A locking mechanism, thread locker, or adhesive should be used to hold the nut in place. See Figure 5-3.
Figure 5-3 Bolt on thermocouple method (NEW)
5.3.3 T/C Attachment Methods – Eyelet T/C The eyelet method is also simple and repeatable, using a screw and nut
to secure the eyelet. This method is a slightly more expensive but is extremely robust. See Figure 5-4.
Figure 5-4 Bolt on eyelet T/C attachment method (NEW)
5.3.4 Golden Board T/C Assembly methods UV and non-UV cure adhesives are relatively inexpensive and quick to
apply. Attachment may be less robust, but a small adhesive dot has less thermal mass and insulating properties than a
bolt or eyelet, and will give a more accurate temperature. The T/C should make intimate contact with the Golden
Board. See Figure 5-5.
10
Figure 5-5 Epoxy T/C attachment method (NEW)
6 Profiling Equipment Requirements
6.1 Types of Profiling Equipment
6.1.1 “Built-in” Profilers Profilers which are integrated as part of the oven system.
These systems typically use long, trailing T/C wires to acquire real time data (AKA the “fishing method”). They can
be included on most ovens at little additional cost, but often have limited features. See Figure 6-1.
Figure 6-1 “Built-in” thermocouple attachment sites on a reflow oven (NEW)
6.1.2 Remote Profilers Remote profilers offered by third party suppliers use short T/C wires with the profiler box
trailing the product as it passes through the oven. The box can be a singular item or can be an integral part of a larger
reflow pallet. These systems are more expensive but have many additional features and are generally easier to use.
Remote profilers can log data to be downloaded after the test run, or provide real-time RF data transmission. See
Figure 6-2.
11
Figure 6-2 Remote Profiling System
6.2 Minimum Data Acquisition The minimum data acquisition rate should be sufficient to capture the temperature
changes, but not so fast as to overload the user with data. A common rate is one reading per second.
6.3 Data Recording Unit The data unit shall be capable of withstanding the maximum temperature of the recipes in
use.
6.6 Number of T/C’s The minimum number of T/Cs should be three.
6.7 Measurement Accuracy The measurement tool (Profiling System) should be calibrated and should be accurate to
within ± 2°C.
The Profiling System is typically calibrated annually. The calibration process shall be performed carefully in a
thermally stable environment. Typically the instruments are returned to the manufacturer or an outside lab for
calibration. If your facility has a captive calibration/metrology lab, you should contact the manufacturer for their
recommended calibration procedure.
A calibration compares the instrument’s measurements to known standards. There are two common methods used to
make this comparison:
1. Place the instrument’s T/C in a temperature bath of known value, comparing the instruments
measurements to that value.
2. Using a thermocouple simulator standard, source an equivalent voltage to the instrument’s inputs,
comparing the instruments measurements to that value.
If the instrument’s measurements are within the manufacturer’s specification for accuracy, the instrument is in
calibration. Method 2 is most common since the T/C simulator can source most any temperature value within the range
of temperatures at which reflow ovens operate.
If the instrument is shown to be out of specification, it shall be taken out of service and recalibrated. The exact
procedure for this depends on the manufacturer of the instrument, and their procedure should be followed carefully, or
the instrument returned to the manufacturer.
12
7 THERMOCOUPLES
This section contains a brief discussion of T/Cs and common attachment methods. For a more in depth discussion on
T/Cs see IPC-7530.
7.1 “K” – Type T/C The most common T/C in use is the “K”-type. Type K T/Cs wires are typically made of chromel
(90% nickel and 10% chromium) and alumel (95% nickel, 2% manganese, 2% aluminum, and 1% silicon). A
temperature sensing range of -200°C to +1350°C [-330°F to +2460°F] is common.
7.2 Wire Length The wire should be long enough to allow the Profiling System to trail the board by at least one oven
zone. Short wires (61cm [24in]) can be used for remote data collection devices that ride through the oven (see 6.1.2).
Longer wires (which extend the full length of the oven plus approximately 1.5m [5ft] of service length) allow for builtin data collection systems (see 6.1.1).
7.3 Wire Gage The gage of the T/C wires can vary according to use (Figure 7-1). 20-24 gage wire is common when
using built-in data collection systems. Insulation is usually fiberglass, which can insulate up to approximately 480°C
[896°F]. This gage is recommended when durability is a concern or when higher temperatures will be used.
Smaller gage wires (30-40 gage) are common when remote Profilers are used or when the T/C will be attached to
BGAs or other small components. Fine gage wires are insulated with Teflon® which limits the temperature range to
approximately 260°C [500°F]. Other insulations such as Kapton®, glass, or hybrids may be able to withstand higher
temperatures.
Figure 7-1 Example of a heavy gage wire T/C vs. a fine gage T/C compared to a U.S. Dime (NEW)
7.4 Limit of Error The typical limit of error on most T/Cs is approximately 2.2°C [4.2°F]. It is important to note that
two T/Cs on the same board may vary by as much as 2.2°C [4.2°F] due to variability in the T/C itself. Multiple test
boards may also exhibit variations of up to 2.2°C [4.2°F] on the same test point.
If more precise data is required, Special Limits of Error (SLE) T/Cs are available which have a 1.1°C [2°F] nominal
variation. Cost for the SLE T/Cs are higher, but are an excellent choice when evaluating repeatability of a particular
oven profile.
ISA MC-96.1 defines the limits of error on T/Cs.
7.5 T/C Attachment This section addresses attaching to a production assembly where the T/C will be removede after
reflow. For attachment to a Golden Board, see sections 5.3.4, 5.3.3, and 5.3.4.
13
7.5.1 Polyimide Tape Polyimide tape (Figure 7-2) is an inexpensive and easy way to attach T/Cs to a board. The tape
can be removed and replaced at any time to move the T/Cs to any location on the surface of the PB or component. No
special tools are required to use polyimide tape, and it can easily be packed into a tool kit for travel if necessary.
There are drawbacks to the use of polyimide tape. Resulting data is often not repeatable, because adhesive softens with
heat, and tape can become dislodged or lift from the surface being evaluated. Lifted tape usually means a lifted T/C.
Over all, polyimide tape provides a relatively weak bond to the surface of the assembly. Also, it is not possible to use
polyimide tape to secure a T/C underneath a BGA or other component.
This method should be used when the assembly cannot be used as a Golden Board and will only be used for profiling
once.
Figure 7-2 Polyimide Tape (NEW)
7.5.2 Double-Sided Metal Tape Some Profile System kits include metal tape (figure 7-3), which attaches the T/C to
the surface easily, and like polyimide, can be easily placed or moved to various locations. Metal tape is a good choice
for attachment when the surface of the assembly shall not be damaged. The residue left by the tape can easily be
removed from the assembly.
Double-sided metal tapes are stronger then polyimide, but still have relatively weak bonding strength and are not very
durable. A small amount of force on the T/C can detach the metal tape, leading to erroneous data. Double-sided metal
tape is also impossible to use in confined areas or under BGAs and other components.
Figure 7-3 Double-Sided Metal Tape (NEW)
7.5.3 High-Temperature Solder High-temperature solder provides accurate results since it is highly thermally
conductive, similar to the assembly solder. High-temperature solders can have liquidus points in excess of 280°C
[536°F]. The higher melting temperature allows the solder to remain attached to the assembly well above the reflow
temperature for a typical electronic assembly.
14
The equipment needed to attach a T/C with high-temperature solder is the same as most other solders. A handsoldering unit shall have the capability of reaching the liquidus point of the solder being used.
High-temperature solders allow a solid connection to the assembly while still providing a good thermal transfer to the
T/C. Multiple reflow excursions can be made using the same assembly with T/Cs connected by high-temperature
solder, and it is unlikely for the T/C to become dislodged. Results from the assembly are usually reliable and
repeatable.
However, high-temperature solders can be difficult for an inexperienced technician to use, which can result in a poor
attachment. The solder usually takes longer to reflow, and the localized high heat can damage the assembly. In
addition, a large deposit of the solder may become aheat sink over the T/C, which will skew the resulting data.
7.5.4 U/V Cured/Thermally Conductive Adhesive U/V cured adhesive, like high-temperature solder, can provide
reliable and repeatable results. It is generally easy to use, being dispensed from a syringe or other dispenser system
(Figure 7-4). Many adhesives cure in 10-20 seconds with the use of a spot cure U/V system. The adhesive is fairly
thermally conductive, and does not act as an insulator for the T/C.
The resulting attachment is very strong and the T/C is unlikely to become dislodged. In addition, the adhesive is very
durable, which means the assembly can be used on multiple heat excursions allowing dozens of test runs.
One drawback for a U/V cured adhesive is that a U/V curing system is required. Also, the curing units can be
expensive and require an initial investment which is higher than other methods of T/C attachment.
7.5.5 Attaching without Damage In some cases, a limited number of assemblies are available for a
production run, making it essential to avoid damage to the tested assembly. Polyimide or double-sided
metal tape should be used to attach the T/Cs to the assembly. While the connections are not as durable as
the soldered or adhesive attached T/Cs, the tape can easily be removed, along with any residue.
(Line drawing place holder. Completed drawing will be placed here before publication.)
Figure 7-4 Adhesive Dispensing Unit (NEW)
7.7 Additional Suggestions The following are additional suggestions which can increase the life of the T/C and help
ensure reliable, repeatable data results.
7.7.1 Long Wires When using long trailing T/C wires, add a protective, insulating sleeve over the T/C wires to protect
them and increase the life of the test assembly (Figure 7-6). Wires can be taped together every 1 to 2 feet to prevent
tangling.
15
7-5 Added insulation to protect long run wires (NEW)
7.7.2 Stress Relief Wrap T/C wires through cutouts or holes in the product to provide stress relief. This will result in
less wear and tear on the T/C attachment, and the T/Cs will be less likely to detach (Figure 7-6).
7-6 Stress Relief Option (NEW)
7.7.3 Adhesive Support Using adhesive to attach the T/C wires to the trailing edge of the Golden Board (Figure 7-7)
will provide stress relief for the T/C attachment, and will increase the life of the test assembly.
16
7-7 Adhesive Support (NEW)
7.8 Recognizing Potential Issues
7.8.1 Variations in Attachments Variations in the attachment method can cause variation in the accuracy of the
collected data. More accurate results can be achieved by using uniform amounts of solder or adhesive on all
connections
7.8.2 Twisted Wires Avoid twisting bare T/C wires (Figure 7-8). Temperature is measured where the two wires
FIRST contact each other. Twisted wires will likely measure air temperature instead of board temperature.
Figure 7-8 Twisted T/C Wires (NEW)
A. Not Twisted (Correct)
B. Twisted (Incorrect)
C. Temperature will be measured at this point.
7.8.3 “Spikes” in Temperature Data Figure 7-9 shows a data result with spikes in temperature instead of a consistent
temperature ramp. These spikes are usually caused by a loose T/C which is measuring the ambient temperature instead
of the board temperature. The T/C should be re-attached and the test should be run again.
17
Figure 7-9 Spikes in Temperature Data (NEW)
7.8.4 Excessively High or Low Temperature An excessively high temperature line can be an indication of a broken
or “open” T/C. In these cases the T/C wire should be replaced.
An excessively low temperature (below 100°C [212°F]) can indicate a short in the T/C. Commonly in these cases, the
wires are connecting within the T/C connector. Open the connector and rework the wire terminations.
8 Verification Profiling Frequency The Golden Board should be profiled on a regular basis. The maximum time
between runs should be be no more than one month if the process has objective data that supports Cpk ≥1.33(see 9.1).
The user may choose a shorter or longer frequency between runs, based on the users internal control plan (see IPC9194) and as supported by historical data. If the oven is equipped with a continuous monitoring system, data from that
system may be used in lieu of the reference board check run to demonstrate Cpk values.
The Golden Board should be profiled after any maintenance procedure that affects the exhaust, heating/cooling, or
conveyor subsystems of the machine.
Data from the Golden Board should be recorded on a control chart per the user’s internal control plan. One or more of
the following process parameters should be selected for charting
a. Ramp Up Rate
b. Preheat Dwell Time between temperatures
c. Time above Liquidus
d. Peak or Maximum Temperature
e. Ramp Down Rate
Minimum Specification Limit: If any characteristic is outside established limits, the reflow system and the Golden
Board should be evaluated, and corrective action taken before proceeding with production.
9 Requirements for Oven Repeatability Calculations
9.1 Cpk - Process Capability and Cp - Oven Stability Process Capability is a measurable property of a process that
compares the product specification to the variability of the process. It is expressed as a Process Capability Index (Cpk)
or Capability Performance (Cp). These measurements use statistics to predict how many parts will be produced out of
specification.
Process capability can also be defined as the capability of a process to meet its purpose as required by an
organization's management. In most cases, when people ask about Cpk without specifying a product, they want
management information and are really asking for Cp.
In both cases the calculation requires:
18
1. A measure of the variability of a process (standard deviation), and
2. A specification or product tolerance.
The result is a statistical ratio of variability (standard deviation) to the process tolerance (specification).
9.2 SMT Oven Stability SMT Oven Stability uses a Golden Board with 4 to 6 T/Cs and usually evaluates
the peak temperature and Time Above Liquidus (TAL). It can also include the Flux Activation Time (FAT)
and the heating/cooling rates. It normally requires a minimum of 15 consecutive runs. For machine
capability (Cp), the specification target is equal to the average of the runs, and the control limits are set at
at industry standards for SMT solders” with:
Upper Control Limit (UCL) = 𝑋̿ + A2*𝑅̅
Lower Control Limit (LCL) = 𝑋̿ - A2*𝑅̅
Where:
You have at least two sets of 15 measurements
A2 = 0.223 for sample size of 15
𝑋̿ = the average of the averages of the values in the sample set
𝑅̅ = the average of the differences between the highest and lowest measurement in the sample set
See table “Factors for computing Control Chart Limits” in appendix for sample size less than 15
.
9.3 Process Capability Process Capability (Cpk) is done with actual production boards. The number of
T/Cs depends upon the number of critical components. Sometimes the cooling rate, hold time, heating
ramp, and time to peak are added to the peak temperature and TAL. The specification (target and range) is
usually from solder supplier’s documentation, but may vary in special cases. This usually requires runs to
be repeated over an extended time, such as once an hour for 12 hours, or every 12 hours for a week, but can
be done in a shorter time. When complete, Cpk can be calculated for that specific production board and
process. If either the board or process changes, the Cpk should be rerun.
The calculations are the same for oven stability and process capability.
1. Mean (average) and standard deviation for each criteria being measured. (Note: for Cp the target is the
calculated mean)
2. Tolerance range = Upper spec limit minus the lower spec limit
3. Cp = Tolerance range/6 times the standard deviation
4. Cpk = the lesser value of (USL – 𝑋)/(3 x StDev) or (𝑋 – LSL)/(3 x StDev)
Where:
USL = Upper Spec Limit
LSL = Lower Spec Limit
𝑋 = the average of the values in the set
StDev = the standard deviation of the values in the set
5. Cp should be > 1.0 and CpK should be >1.33 to demonstrate control. (Note: Some references say that
the machine capability (Cpm) should be > 2.0.)
Normally a Cpk question from a customer refers to the process capability study, because he wants to know
about the repeatability of the thermal process on his product. The customer may also be interested in
machine capability and repeatability. In a process study, the Cpk calculations can be done for all the T/Cs
or for each individual T/C depending upon the requirements.
10 Maintenance and Calibration Guidelines
10.1 Nitrogen Usage Many systems use a flow of gaseous nitrogen to reduce oxygen levels inside the system during
critical phases of reflow. If the process was established using nitrogen then ensure that nitrogen is on during the board
check run.
19
10.2 Reflow Oven Operation
10.2.1 General The system operator shall perform system startup and cool down/shutdown procedures as
recommended by the equipment manufacturer, and lubricate fittings at specified intervals as required. Monitor system
operation, and report any alarms or operating problems to the appropriate supervisor or maintenance personnel.
10.3 General Housekeeping and Daily Maintenance Daily maintenance will usually be performed by the
responsible system operator or process owner. Wipe down the machine daily. Clean any tools or fixtures and put them
away.
Conveyor chains, sprockets, and the machine chamber also require regular cleaning using an approved vacuum or
aqueous cleaning agent. Never use flammable solvents for cleaning. Recommended cleaning materials include lint-free
cloths, an ammonia based cleaner, scale stripper, and defoamer.
10.4 Calibration To assure proper control of the reflow profile, periodic calibration should be performed on conveyor
speed and the thermocouples in all zones.
A regular schedule of oven calibration should be in place, typically annually or as recommended by the oven
manufacture. This should include at least all temperature measurement and conveyor speed systems. Some may
recommend confirming proper convection air flow, N2 (O2) measurements systems, and other sensors. The oven
manufacture shall specify the process for performing these calibrations.
The manufacturer should recommend a calibration procedure. However, lacking any procedure from the manufacturer,
there are again two methods for temperature: 1) Place a T/C of known calibration at or on the oven’s T/Cs, or 2) Use a
T/C simulator to source thermocouple voltage into the oven’s measurement system, and then compare the
measurements to verify they are within the manufacturer’s specification. It is reasonable to expect these measurements
to be within ±2°C.
For conveyor speed, measure the length of the conveyor, measure the time to travel that length, and divide length by
time to determine if the conveyor speed is within the accuracy specified by the manufacturer.
10.5 Preventive Maintenance Tasks, procedures, and recommended intervals should be specified by the equipment
manufacturer in machine operation and service manuals. Table 10-1 is an example, and is included for reference only.
20
Table 10-1 Example of Maintenance Schedule (SAMPLE ONLY)
Appendix A
A-1 Factors for Computing Control Chart Limits
21
Samples (n)
A2
2
1.880
3
1.023
4
0.729
5
0.577
6
0.483
7
0.419
8
0.373
9
0.337
22
10
0.308
11
0.285
12
0.266
13
0.249
14
0.235
15
0.223