CX-259
`
`UNITED STATES
`CONSUMER PRODUCT SAFETY COMMISSION
`WASHINGTON, OC 20207
`
`Memorandum
`
`Date:
`
`July 19, 2001
`
`TO
`
`Ronald L. Medford, Assistant Executive Director
`Office of Hazard Identification and Reduction
`
`THROUGH:
`
`FROM
`
`Hugh M. McLaurin, Associate Executive Director
`Directorate for Engineering Sciences
`Caroleene Paul, Division of Mechanical Engineering
`Roy W. Deppa, Division of Mechanical Engineering
`Dean LaRue, Division of Electrical Engineering
`
`SUBJECT: Evaluation of Prototype Tablesaw Safety Device
`
`INTRODUCTION: The Directorate for Engineering Sciences received a sample of a
`prototype tablesaw safety device, as well as a detailed demonstration from its inventor,
`on July 11,2001 to evaluate its potential to address injury. The inventor also provided
`an information package that combines the extensive technical information of the 26
`different patents obtained in designing the safety system. The device consists of a
`modified commercial consumer-grade tablesaw, including an electrical blade contact
`detection circuit, logic circuit, and electromechanical device that stops blade rotation
`and lowers the blade below the table surface upon contact with a human body part.
`This system is under development and was demonstrated by SawStop LLC. of
`Wilsonville, OR.
`
`BACKGROUND: Tablesaws account for approximately 30,000 injuries to the hand or
`finger per year, with approximately 10% of these injuries involving amputation.
`Tablesaw blades are typically 10 in. in diameter and rotate at about 4,000 rpm. A
`typical 40-tooth blade’s teeth cut at a rate of about 2,700 cuts per second; these saw
`teeth are travelling at about 120 mph. Resulting injuries are usually severe.
`
`Review of In-Depth Investigations shows that typical incident scenarios involve
`inadvertent contact with the blade. The operator allows his hand to contact the blade
`while sawing due to inattention, or the workpiece slips or moves suddenly and the
`operator reaches, falls, or slips and contacts the blade from the top or rear of the blade.
`In some cases the work piece is kicked back by the blade and draws the operator’s
`hand into the blade.
`
`CPSC Hotline: 1-800-638-CPSC(2772) # CPSC’s Web Site: http://www.cpsc.gov
`
`SawS top IT C-0094070
`
`CX-259.0001
`
`SD3 Exhibit 2014 – Page 1
`
`

`
`Safety engineering on a systematic basis takes a tiered approach to address
`hazards:
`
`1. The most effective measures are those that design the hazard out of the product.
`This has not been possible with tablesaws; the operational requirements of tablesaws
`seem to preclude the possibility of removing the hazard.
`2. The second most effective measures are those that guard or shield against the
`hazard. This is the approach that has been applied to tablesaws, but it has not been
`effective because the guards are optional and they do not work very well.
`3. When design and shielding approaches do not work, the next most effective method
`is to introduce an intervention strategy in the development of the hazard. That is, allow
`the events that lead to injury to begin, but introduce some element that stops or diverts
`the process before the injury occurs, or at least before the injury becomes very severe.
`This is the philosophical basis for the SawStop. The operator’s hand actually contacts
`the spinning blade, but the device senses this contact and stops the blade and moves it
`before severe injury occurs. This approach is sophisticated and potentially vulnerable.
`Timing is everything; the blade begins to cut into the operator’s finger before the system
`can work, and it must work reliably and very quickly to limit the injury.
`
`The SawStop demonstration model is a prototype, therefore issues of reliability
`and robustness over the life of the product cannot be evaluated. These issues will be
`dependent upon choices made in the development and manufacture of production
`products, and they are likely to differ significantly between manufacturers. While the
`ability of the product to function properly under different conditions, or incident
`scenarios, can be addressed with a prototype, these factors may differ depending upon
`the manufacturing design. Consideration of details that are dependent upon design
`and manufacturing must be evaluated on production products, or may be considered in
`establishing standards of performance.
`
`What can be evaluated at the prototype stage is whether the basic concept of the
`device addresses the known hazard pattern in an effective way, and thus can establish
`whether the device demonstrates the feasibility of eliminating or reducing the hazard.
`The basic concept of the SawStop is to electrically sense when contact with a body part
`has been made, and to mechanically remove the cutting hazard before severe injury
`can occur. In the next three sections, the electrical operation, the mechanical operation,
`and the testing will be discussed.
`
`Electrical Operation
`
`The theory of operation is based upon the electrical capacitive nature of the
`human body, or the ability of the human body to store electrical charge. A small
`electrical field is placed by the SawStop circuitry onto the saw blade by a supply
`electrode, and a sensing electrode senses the electrical field coming from the saw
`blade. If a person touches the saw blade, some of the electrical field is redirected into
`the person’s body rather than into the sensing electrode. When the field measured by
`
`SawStoplTC-0094071
`
`CX-259.0002
`
`SD3 Exhibit 2014 – Page 2
`
`

`
`the sensing electrode drops below a pre-determined percentage of the normal value,
`the stopping mechanism is activated.
`
`The electrodes are not actually in contact with the saw blade. They are a small
`distance away from the blade. This is what is called capacitive coupling. Two conductive
`materials with a dielectric material (in this case air) between them creates a capacitor.
`Essentially, there is a capacitor created by the supply electrode to the saw blade in
`series with another capacitor created by the saw blade to the sensing electrode.
`Electrical energy can therefore flow between the supply electrode to the saw blade and
`from the saw blade to the sensing electrode.
`
`In the prototype received for evaluation, the supply and sensing electrodes are
`capacitively coupled to the arbor shaft. In most cases, this is not a problem because the
`saw blade is electrically connected to the arbor shaft. However, a few saw blades used
`in the evaluation had plastic hubs. The safety mechanism will not work with these
`blades because the plastic hub insulates the metal part of the blade from the arbor
`shaft. In this particular implementation, there is no means to determine whether an
`appropriate saw blade is attached to the arbor shaft or that the blade is actually coupled
`to the circuit. This is an issue that will need to be addressed in the development of
`manufactured products, to ensure that a user knows when they are protected. This
`does not affect the evaluation of the basic safety mechanisms and principles of this
`device.
`
`The remainder of the circuitry is designed to detect and react to a person
`touching the saw blade. The circuitry is controlled by a microcontroller. The
`m icrocontroller reads various inputs and makes a decision to activate the saw brake or
`to allow the saw to keep running. Using the example of the 40-tooth blade operating at
`4,000 rpm, one tooth goes by a point every 370 #s. The circuit samples the status and
`makes a decision every 18 ps, which is more than 20 times per tooth. The circuitry
`reacts quickly enough to minimize the damage to a person’s hand should it come in
`contact with the saw blade.
`
`The microcontroller is programmed to react quickly to a person touching the saw
`blade while adjusting itself for scenarios involving wood that may be slightly conductive.
`Wet green wood or wet pressure treated wood can be conductive and could make the
`saw brake trip without any danger. Logic has been built into the program to monitor not
`only the magnitude of the signal but also the rate at which it changes. Conductive wood
`would cause a slow change in the signal magnitude where a person would generate a
`quick change in the signal magnitude. If the controller detects a slow change in the
`signal magnitude, it changes the supply voltage to maintain a relatively constant
`sensing voltage. However, it is designed so that it cannot change the supply voltage fast
`enough to miss an actual human event. This is designed to reduce nuisance trips
`without reducing the protection to people.
`
`S awS top IT C-0094072
`
`CX-259.0003
`
`SD3 Exhibit 2014 – Page 3
`
`

`
`There are several self-tests designed into the circuitry to ensure that the safety
`mechanism will work if needed. If any of these self-tests fail, the saw will either stop if
`running or will not start if not running. The self-tests are:
`
`1. Watchdog error- this is monitoring the status of the microcontroller.
`2. Saw brake triggered or trigger circuit open - this will sense if the saw brake has
`already been spent or if the electrical connection to the saw brake is missing.
`3. Supplies out of regulation - this senses the voltage on the power supply to ensure
`that it is adequate to operate the circuitry.
`4. Capacitor over voltage - this senses the voltage on the capacitor to ensure that the
`capacitor is working properly.
`5. Hall sensor defective - the Hall sensor detects motion of the saw blade. This is used
`to allow protection during a shut down of the saw. The electronics is capable of
`activating the saw brake as long as the saw blade is rotating, even after the saw is
`turned off.
`6. Capacitor not charging - senses to see if the capacitor is charging to prevent a
`misfire.
`7. Capacitor under value (discharges too fast) - the system is measuring the time
`constant during operation to ensure that the capacitor is properly charged.
`8. Sense calibration circuit error- the microcontroller monitors the sensing portion of
`the circuit to verify adequate signal.
`9. Sense circuit error- the microcontroller monitors the sensing portion of the circuit to
`verify it is receiving the signal it expects.
`
`Mechanical Operation
`
`The mechanical theory of operation uses the potential energy stored in a spring
`to force a plastic brake into the teeth of the rotating saw blade, and the angular
`momentum of the rotating blade to retract the blade below the surface of the table saw.
`A brake cartridge consisting of a spring loaded plastic pawl and controller circuitry is
`positioned on a shaft directly behind the blade arbor. Once a saw blade has been
`installed, care must be taken to adjust the pawl side of the brake cartridge as closely to
`the blade as possible without interfering with the blade’s free movement. An electrical
`lead from the m icrocontroller attaches to the brake cartridge. When the m icrocontroller
`determines that a person has touched the saw blade, it sends a signal to discharge a
`capacitor in the brake cartridge. The capacitor is discharged through a thin wire whose
`function is to suppress a 100 Ib spring against the plastic pawl. When the current from
`the capacitor goes through the wire, the wire melts and releases the spring. The plastic
`pawl is then forced into the teeth of the saw blade. The plastic pawl begins to stop the
`saw blade rotation within milliseconds of when the detection circuitry senses human
`contact.
`
`The saw blade is raised and lowered by way of a worm screw, keyed to a shaft
`that is manually rotated by the operator. The worm screw slides freely on this shaft until
`a U-pin on the worm screw locks into a groove on the shaft. When the worm screw is
`locked into place, rotation of the worm screw drives the saw blade up and down. The
`
`SawS top IT C-0094073
`
`CX-259.0004
`
`SD3 Exhibit 2014 – Page 4
`
`

`
`sudden braking of a rotating blade creates so much momentum that the worm screw is
`knocked loose from its locked position on the shaft. With the worm screw now free to
`slide on the shaft, the angular momentum of the blade carries the blade straight down
`below the table saw surface. As with the blade braking, the blade retraction occurs in
`the time frame of milliseconds.
`
`SawStop Prototype Testing
`
`A table saw is among the most diverse of power tools. A variety of blades can be
`installed to make straight thru cuts, angled bevel and mitre cuts, or non-thru dado and
`rabet cuts. The SawStop was tested using a variety of blades to make common cuts.
`Contact between the saw blade and a finger was simulated using a hot dog in lieu of a
`finger. The signal change (detected by the SawStop circuitry) caused by contact with a
`human finger is comparable to the signal change caused by contact with a hot dog that
`is in contact with a human body. The inventor verified this similarity in signal changes
`by measuring the signal of a human finger as it was cut on a saw blade and measuring
`the same on a hot dog as it was cut on a saw blade. The following table summarizes the
`testing performed on the SawStop.
`
`Trial
`
`1
`2
`3
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10"
`
`11"
`
`Blade
`Type
`10" carbide
`10" carbide
`10" plywood
`10" rip
`
`Teeth
`4O
`4O
`25O
`12
`
`12
`
`40
`
`40
`
`40
`
`40
`
`24
`
`10" rip
`
`10" carbide
`
`10" carbide
`
`10" carbide
`
`10" carbide
`
`7" dado with
`plastic hub
`7" dado with
`plastic hub
`
`Type Cut
`
`straight cut
`straight cut
`straight cut
`straight cut
`
`straight cut
`
`35 deg bevel
`60 deg mitre
`kick back into
`rear of blade
`contact during
`coast to stop
`
`straight cut
`with glove
`NA
`
`NA
`
`Blade
`Stop
`6 ms
`4 ms
`24 ms
`
`Hot Dog
`Damage
`no
`no
`no
`no
`
`no
`
`no
`
`no
`
`no
`
`no
`
`NA
`
`NA
`
`4 ms
`
`4 ms
`
`1 ms
`
`4 ms
`
`NA
`
`NA
`
`Comments
`
`slow feed, hot dog on wood piece
`fast feed, hot dog on wood piece
`blade retract prevented injury
`blade retracted before stop and
`prevented injury
`blade retracted before stop and
`prevented injury
`average feed, hot dog on wood
`piece
`contact to rear of blade simulated
`kick back
`blade stopped immediately
`contact made approximately 4
`seconds after shut off
`cut thru glove, activation upon hot
`dog contact
`no reaction, blade insulated from
`arbor
`no reaction, blade insulated from
`arbor
`
`* These tests were performed with the drive belt removed from the blade and a specialized test box in
`place of the brake cartridge. The test box simulates braking by cutting power to the motor.
`
`The reaction time of the SawStop system is too fast for the human eye to detect.
`Each test trial was recorded using a high speed camera at 1000 frames per second.
`The slowest replay of events possible is 1 frame per second. A typical SawStop
`reaction to contact with a hot dog resulted in almost immediate retraction of the blade
`and cessation of the blade rotation within 4 milliseconds. Time for the blade to retract
`
`SawS top IT C-0094074
`
`CX-259.0005
`
`SD3 Exhibit 2014 – Page 5
`
`

`
`below the surface of the table saw depends on the blade height set for the cut. An
`important factor is the fact that however long it takes for the blade to stop rotating, the
`hazardous cutting edge of the blade is already moving away from the contact point.
`
`A 40 tooth, 10" carbide blade stopped in approximately 4 milliseconds. This was
`true whether it was contact made during a straight cut, during a compound cut, from the
`rear of the blade, or through a glove. A straight cut made with a 250 tooth, 10" plywood
`blade resulted in a longer blade stop time of 24 milliseconds. However, despite the
`longer blade stop reaction time, minimal damage to the hot dog occurred because the
`blade still retracted from the point of contact almost immediately. Similarly, cuts made
`with a 12 tooth, 10" rip blade resulted in a blade stop time of approximately 35
`milliseconds (the blade retracted below the table saw surface before blade stop), but
`minimal damage to the hot dog occurred because of the immediate blade retraction.
`
`As stated before, because the prototype design capacitively couples the arbor,
`conductivity between the blade and the arbor is necessary in order for the system to
`react to contact between the blade and a body part. Two different blades with plastic
`hubs were tested and resulted in operation of the table saw in an unsafe condition -- if
`contact were made, the system would not have worked. The blades were specialized
`dado blades; however, their use is not uncommon among serious woodworkers.
`
`The limited amount of time allotted for evaluation did not allow for electrical
`interference testing. Electrical interference transmitted through the electrical supply line
`or the air could potentially cause nuisance tripping or possibly prevent the circuitry from
`detecting someone touching the saw blade. If any of these types of interference should
`cause problems with the circuitry, the problems could likely be remedied by minor
`changes to the circuitry or how they are shielded from outside interference. Testing for
`the effects of electrical interference should be conducted in future evaluations of this
`product.
`
`CONCLUSION
`
`Based upon the evaluation reported here, it appears that the SawStop concept
`is valid and the prototype impressively demonstrates its feasibility. The electrical and
`mechanical components operated without failure in a time frame that would greatly
`reduce blade contact injury. The design concept is very flexible and can be modified to
`address foreseeable areas of concern.
`
`The device that was evaluated is a prototype, with handmade, non-production
`components. Production products will include modifications due to design and
`manufacturing decisions that may result in different performance. In addition, the
`robustness and life-cycle details of production units will be different from those of the
`the prototype. The evaluations that were performed therefore concentrated on the
`validity of the concept and the performance of the components used in the prototype
`system. A significant amount of further development work may be required before this
`device could be incorporated into production saws, both because of the need to adapt
`
`SawS top IT C-0094075
`
`CX-259.0006
`
`SD3 Exhibit 2014 – Page 6
`
`

`
`the concept to mass production, and to address some issues that still require
`refinement.
`
`Of highest concern are those areas where the SawStop may not perform, and
`more importantly, may not indicate to the user that it will not perform. As discussed
`earlier, the device is dependent upon electrical conductivity from the hand through the
`blade to the saw arbor and thence to the circuitry. There are tablesaw blades that have
`plastic or other non-conductive hubs or centers, and even a painted or coated metal
`blade may not make electrical contact with the arbor. In this event, the saw may be
`operated, but the SawStop will not work as presently configured. This failure may likely
`be addressed through further design refinement.
`
`Of secondary concern are those areas where the SawStop system may be
`perceived as a nuisance and therefore a candidate for bypassing by the user. The
`prototype SawStop uses a brake cartridge that may only be used with a 10 inch blade.
`The cartridge location does not accommodate smaller diameter blades or thicker
`specialty blades. In addition, specialized blades such as molding sets, which only have
`one to three teeth, may not work with the current brake configuration. As stated before,
`these areas of concern would need to be addressed during production design of each
`specific table saw.
`
`SawS top IT C-0094076
`
`CX-259.0007
`
`SD3 Exhibit 2014 – Page 7