Sled 5 Cut Help Please...

[Derek Arita]

I'm able to get the sled down to around .002 error, using the Charles Eng formula. Bouncing around that, I can't get to .001 or less. To add to it, every time I try to screw down the rest of the fence, even with one screw, it throws the fence off. I'm predrilling the screw holes and going in as straight as possible. What am I doing wrong?

[Johnny Barr]

Are you clamping the fence down before adding screws?

[Derek Arita]

Yes I’m clamping. Could the fence be moving as I make the cut? The center of the fence has no screws in it while finding the exact position. Could it be warping with the force of the workpiece against it? Wish I could clamp the whole fence down while making the test cuts.

[ChrisA Edwards]

My guess, and it is that, your screw is hitting a slightly harder area on one side of the fence that you are screwing it into, maybe glue, if it's a plywood fence or grain, if natural wood, or the point of your screw is not the dead center of the screw, so one side of the screw is kind of cam'img your fence.

[Michael Dean]

You say the error is .002". Over what distance? That amount of error over 24" or greater might not be worth chasing.

[Mel Fulks]

I don’t think you are going to get any more accuracy without going to platinum !

[Alan Schwabacher]

Are you sure it’s screwing down the fence that’s deflecting it, and not simply random error? If you get the deviation to a minimum, and then repeat the measurement without
making any changes, how similar are your results?

[Derek Arita]

Well, will wonders never cease. After many attempts, a few adjustments to the saw and remaking the sled, I finally was able to get a 5th cut variance of .001 consistently. That was an ordeal. I used a threaded insert on the adjusting end to make it easier to make the adjustments and to keep from running out of space for screws. Thanks for the help.

https://i.imgur.com/NU3XuIG.jpg

https://i.imgur.com/0QLGhlF.jpg