>Just thought I'd post again because Ray brought me one of the drawer fronts. A picture is below. It was almost the exact width of my cabinet saw's table, which I know is flat. You can see some light coming under the middle of the panel. I pulled out some feeler gauges and found the middle to be .012" higher than the ends (3/250th).
>When you look at the edge, do you see any of the tooth marks from the saw cut, or did you sand it all away? If there is some still there especially near the middle, there would be something telling. You could rule out sanding for certain.
Did you say that you cut this on a TS sled? If so , I am suspecting that you may have used a thin kerf blade a little on the dull side. Sometimes if they are starting to dull, they can start to resonate, which makes the cut wider kerf than when really sharp, especially if you started into the cut slowly, then sped up a little, then slowed down to prevent chipping as the blade neared the exit point.
>My guess is the same as Keith's. The first sign that my tablesaw blade is getting dull is when it vibrates in the middle of a crosscut, leaving saw marks just in the middle but not on the ends. When it is sharp, it doesn't leave saw amrks anywhere along the cut. I use a Forrest WWII. If you only sand away the saw marks but not the entire length of the cut, it will end up just like what you have.
>The glue-up has absolutly nothing to do with it. I believe the ROS is the culprit. When sanding an edge like that, the sander is toward the center of the piece most of the time which would cause the slight "bow". You were probobly being careful as to not round over the corners. That looks like 1x material, right?
In order to fix the problem you should use a long sanding block, preferably longer than the width of the piece, and very carefully (220 grit) straighten the edge out. If you use anything other than elbow grease you may cause more problems than you already have.
>I sand that sort of thing one of two ways - assuming I want a pefectly straight edge. I use 3M spray adhesive to glue some sandpapaer to a LONG piece of straight, stiff scrap. That way you cannot dish out the middle. Or, since I have a lot of sandpaper glued to glass for sharpening tools, I just use that.
>These were crosscut on either the RAS using a Freud LU91 full kerf blade, or on the tablesaw on a sled using a Ridge Carbide Hi-ATB full kerf blade (very sharp).
Sanding could be the culprit I guess. These two blades really leave a very smooth crosscut in the end grain requiring little, if any, sanding. I used the ROS to just kiss the wood wherever I could see a blade swirl, but that still seems to be the only plausible reason.
>...retrace your steps with a piece of plywood the same width, using the same blade and sled arrangement. Check the straightness of the cut as it comes off the blade, and then check again when you have done your ROS routine. With the same regime, the errors should be repeatable.