Amongst the methods for checking the accuracy of a crosscut
fence on a saw, (meaning, a fence which is positioned perpendicular to the blade) the Five Sided Cut Method is one of the most conclusive. This method is
commonly used with sliding table saws, and is a useful method for truing up aftermarket saw accessories, and home made cross cut sleds.
For our purposes, read this as if a crosscut sled were being calbrated, but keep in mind this process is adaptable to any circular saw device with a fence perpendicular to the blade, and that this method can be juxtaposed to either side of the blade, as long as the freshest cut made is being positioned against the fence prior to the next cut made. In other words, either clockwise or counter clockwise will work, depending on the side of the blade you choose to test from, as long as that rule is followed.
The benefit of this method is that it is "error additive", meaning the error gets easier to resolve as the test progresses, easy for anyone to do, and gives the tester the ability to tune the tool to very high accuracy. The process can also be used on other machines which cut with a similar process.
1. Select a piece of sheet goods to use for the test, something
24" square is a good size that will give the test the resolution for higher accuracy. If 24" square
is too large for your application, simply use the largest size square you can suitably fit.
Error is increased by scale, so the higher you can make accuracy possible in the large and long,
the higher accuracy will be when things are average or small. The sheet material should be stable and flat for the best results.
1/2 to 3/4 inch MDF is a best choice, because it is inexpensive, dense, solid throughout with no flaws or blows, generally very flat and leaves a fine edge for measuring tools to register from.
You will need a caliper to measure some fairly small differences in the off cuts made by the test cuts, the best way to do this is with dial calipers.
You may use vernier calipers or a micrometer as alternatives, if that is what you have and can read them accurately.
2. Assuming a properly set up table saw, where the blade has been properly calibrated to be parallel to one of the miter slots,
set the crosscut fence that you wish to calibrate, perpendicular to the blade. (ninety degrees to the blade) Using the largest, highest quality square you have
would be the best practice. You are only establishing the beginning. This test is intended to increase accuracy beyond this initial setting.
Once you have established this angle, the fence should be securely fixtured in place for further testing.
3. Ensure that the crosscut sled and fence are completely clean and re-cleaned of all dust and debris between each cut performed
during this test. Failure to vacuum or sweep everything thoroughly can adversely affect accuracy. How you regard the cleanliness of your sled during
normal use is up to you, but for testing and adjusting the sleds accuracy, it is very important.
4. Place the sheet goods intended for use in this test against the crosscut fence
and cut off a thickness of material where the offcut will measure in the proximity of 3/4 inch wide, but not more than 1 inch wide.
This first cut is the establishing cut which creates the fresh base from which the subsequent four cuts (totalling five) will cumulatively resolve fence error.
The reasons for this approximate test cut width are:
A, The material is presented to the blade as a stable load for the motor, the cut
should be consistent and equal on both sides of the kerf, as far as power to the blade is concerned.
B, The offcut is a width which is thick enough to be stable and not affected by machine vibration and proximity to the blade.
C, All the off cuts will generally be in the same range of measurement on the instrument reading them, so calibration will be consistent in that range, and the user
will not become confused by reading errors going back and forth over the 1 inch zero point.
5. Remember to re-perform step 3 and proceed. Rotate the stock 90 degrees counter clockwise. Place the fresh cut edge
against the crosscut fence and make another cut. Repeat this procedure until you have cut all four sides and rotated
the stock back to the original position.
6. In this position, take one last cut, (this is the fifth cut) and
save this off cut.
7. Take this fifth off cut and mark the end that was oriented furthest away from the fence "F" for forward, (this end
of the off cut was cut first by the blade) and "A" for aft for the end closest to the fence, in the correct positions.
8. Measure the thicknesses of each end of the fifth offcut, ends "F" and "A". Write these measurements down.
Compare the difference in width between ends "F" and "A". The difference between the two is the accuracy of the fence to blade angle, multiplied by four.
As an example, say the difference between each end is 1/8th inch, and the goal is to be equal, or parallel on this final test cut, then the fence is out of square by an
amount of about 1/8th inch over the sum of the sides that were cut.
If a 24 inch square test piece was used, the fence is out of square by
1/8 inch over approximately 90 linear inches of test cut length. One would think 4 x 24 = 96, but this test resolves cuts to shorter lengths as the test board rotates.
Cut 1 = 24 inches long, but is not counted in the test length as it is used to establish the base cut for the test.
Cuts 2 and 3 = 23 inches in length, and cuts 4 and 5 = 22 inches respectively. This will total approximately 90 linear inches of test cut if the method previously outlined was carefully followed. Remember, the first cut made is not additive to the test's linear length,
but rather is used to establish the trued base for the last four cuts so that each side of the test pieces to be measured has a freshly cut surface..
This method reveals an additive quantity of error with each of the four
rotations and cuts of the test piece.
Resolving the error to the .001 level of accuracy is often easiest when the fractions are converted to a decimal. If decimal inch reading instruments are used to measure, fractions can be avoided altogether. It also makes the error easier to calculate. In this case, 1/8th inch is equal to .125 inch.
Next, this number must be divided by four to resolve it to the correction needed for the approximate capacity of the sled. So .125 / 4 = .031 inches or 1/32nd in 22-24 inches.
The squareness this test is helping attain is important and is a basis for future project accuracy. The test has revealed the necessary correction required.
Adjusting the fence is the next step. Assuming the hinge point of the fence being corrected is chosen to be on the left side of the sled, if the "F" end of test offcut five is wider than the "A" end, then the right hand side of the fence needs to swing .031 or fractionally expressed, 1/32nd towards the aft end of the sled to make the appropriate correction.
Keep in mind that the accuracy is scalar. This article speaks of a 24" test piece,
however, not all devices will have the same capacity, so compensate, and use the largest test piece possible, and figure accuracy as
derived from the final side of the size that can be tested.
Once this correction has been made, the complete test should be re-performed to determine current accuracy and establish further correction. Since wood can move in the tangential and radial directions, seasonal tests for fence flatness and fence to blade
squareness can be performed, reaccomplished and compensated for.
If the fence isn't glued to the sled base, it can be removed for maintenance, repair,
or replacement should the need arise. It is worth noting that if glue is employed to fixture the fence,
future calibration will not be possible, and the working time available to calibrate before the glue kicks will be short.
A lead screw can be employed, so the sled can be tested and calibrated seasonally.
If a 10-32 threaded rod is used as a lead screw to establish
final fence to blade squareness accuracy, the screw will dial .032 per full turn. That is .016 per half and .008 per quarter turn.
With care and comparison, a very high quality level of accuracy is temporarily possible and re-acquirable with repeatability, even with wood as a tooling material.
Ultimately, when the fifth off cut measures the same on both ends, your sled is cutting square. and that can be as close as .001 if you work at it.
Without the lead screw and decimal dial caliper accuracy, one can only get close to squareness, and adjustment may only be reliably possible to the 1/64th to 1/128th range.