Jerry, .003 is too tiny to worry about. Humidity changes will alter this. Your problem, however; can be merely one sawtooth out of kilter. You must need to true your sawblade just a hair. Metal machinists work on the accuracies you say you are concerned with. Your optometrist must be proud, because none of the rest of us can see when something is 1/300th of an inch "off". Good luck.

 

Otis,
My vision is so bad that I have a restricted driver's license and I read a"nymore even with my best glasses, but I can sure see an error of .003" when I hold the joint up tothe light and look through where the two ends of joint are supposed to butte up together and be glued. Yes, wood can move, but when it is first cut it should fit and be able to be glued up before the wood shifts, or at least I think so, am I wrong?

 

Jerry, I will take photos of the triangles I use and post them later. .003" falls somewhere between a red human hair and a brown one. If you put glue in a joint and clamp it up this will disappear.

There could be several reasons for your cut being off if the miter gauge has been calibrated. How long has it been since your saw was run through a complete set up check? It could be as simple as a bolt that has come loose; it could be a bearing that is worn; you coul be getting vibration from using an old belt; your saw blade may have a defect. My point is the miter gauge will have the same amount of slop in any position... if you get a good cut at 45° you should get the same result at 30°.

 

Mike,
Maybe the error that I am talking about is greater than .003" since glue does not cause the gap to disappear. The way that I came to the measurement is by slipping a piece of typing paper into the gap, it slips into the gap with a slight bit of resistance. The typing paper measures .003" when measured with a dial caliper so the measurment is rough at best and only an approximation, but the actual measurement is not critical.

When I examine the fit of the joints, I take a picture of them and then download them photos into the computer and open the files in photoshop. Once that is done I zoom in on the joint and this allows me see the guality of the joint much better than just looking at it with my poor vision.

Today I cut six parts at 30 degrees, cutting one end of each piece on one side of the blade and then cutting the other end on the opposite of the blade and then matching the ends so that the joint is matched up with each end being cut from opposite sides of the blade, this helped alot but still did not pass muster. I then glued up two halves, three pieces each and these joints were tight of course. I then put the resulting two pieces on the sled, one half at a time, and just let the blade of the saw kiss the ends and this cleaned things up pretty good and the two halves are being glued up now and will be examined after the glue sets.

It seems to me that any error introduced to the cut due to one of the issues that you have listed would be off set by following the procedure listed above. Please comment.

Jerry



Jerry

 

Jerry, is there any side play as you move the miter gauge through a normal cut?

I included photo's of the Rockler heavy duty triangles I use for set up's and the Betterley Una-Gauge with the optional digital gauge. The sheet shows several applications for checking your saw table.

 

Jerry, I agree with you about initial fit immediately after your cut and I am also thinking that glue will easily bridge the gap. It still sounds to me like your saw blade is untrue and that possibly only one tooth is dragging. Can you make a cut on something that will enable you to accurately compare the side of the kerf for parallelism? If those two freshly cut faces are not exactly parallel - it is most likely your sawblade. It really takes less effort than one may realize to knock one sawtooth out-of-position.

 

No Mike, there is no side play in the Incra Express Sled that the miter gauge is attached to, all of the slop has been taken out with the adjustments that are part of the sled. The trianges that I just purchased are made of thin aluminum and I am interested in one like you are usning, I had not thought about the thickness of the triangle until I used it. Thanks for the sending the photos.

Jerry

 

I think that the test now would be a 90° cut on a board at least 4" wide, but the wider the better and see how accurately the two pieces fit together. It sure sounds like a saw problem and if so, probably the heal and toe a fraction out of alignment.

 

Harry has my vote. Just wondring if mitre fence is being calibrated to mitre slot or to saw blade

 

Jerry if you can't get the accuracy you need you might want to consider a Lion Trimmer
Pootatuck's Lion Miter Trimmer

 

Back to Otis and Mike in regard to the saw. I will check the alignment on the saw today and also will change blades, right now I have a Freud fusion combo blade on the saw, I will try my cuts after doing check on the alignment with Forester cross cut blade and see what happens.

Jerry

 

For accurate segment cuts I favor the cut-off sled. Jerry Cole makes an excellent one although expensive. You can build your own. I have watched his demo at wood shows and like his techniques. Having done some segmented bowl turning I found that if you glue up pieces in two or three pieces at a time you can have all kinds of problems. Jerry's method of sticking the outside edge of the pieces and then pulling the pieces into a circle distributes any errors evenly between parts. If your total error is .003 then it disappears between parts. I have a smaller version of a sled that I made form plywood that works fine.

 

I am a retired metal machinist but a novice woodworker. I agree,.003 is readily visible and can also be easily felt with the fingers.However, instead of fooling further with my machine set-up,I would finish the ends of the pieces with a handplane and shooting board if I needed an exact fit. Just my 2 cents.

 

I thought that was why they made sandpaper, to clean up saw inaccuracies. I don't have a table saw but I do have a cutting jig for my circular saw that is perfectly calibrated to 45 degrees. Even at that it is rare to get a perfect 45 degree cut and I blame that on the heat of the blade causing the wood to change configuration when making the cut. Making an angled cut with a router will result in the same problem, particularly an end grain cut. If I can see light between the cut and a straight edge I whip out my sanding block and sand until the light goes out. That solution is much more simple than aging oneself because a perfect cut can't be done with a saw 100% of the time.

 

Having worked in a drafting situation where accuracy is EVERYTHING, I can tell you that I never trust drafting triangles! We had to true every one of them up before use. Having built a waterwheel (6 segments), I ran into the same problem as you. In my case, I glued three segments together to form two halves. These showed about >010 to .012 gap. I flat sanded both segments until there was no gap, and then glued them together.

To true up the triangles for a good 90 deg, you need to clamp a good straightedge to a table with a white sheet of paper or other suitable material under it. Using a very sharp pencil, draw a line down the triangle edge with it securely against the straightedge. Then flip the triangle over and see how well it matches the previous line. Using fine sandpaper flat on the table, run the triangle edge over it, taking off material in the required direction, then repeat the line business with a new line until you get an exact match. This can take a lot of time, but the results are worth it.

 

If you want dead on miters make and use a shooting board. All of my miters are spot on after I shoot them with a low angle plane. I do not know how I got along with out it and a shooting board is cheaper than most other tools.

 

A couple of thoughts. Are you sure your wood is absolutely squared up? The smallest difference in width will throw off the mitre. Second thought is when cutting a mitre the wood really wants to move. Is your wood clamped solidly to prevent this?

 

Your problem could be from simple run out of your saw blade or saw or both. You have to remember it doesn't take much run out to give that small amount of error after so many cuts. My guess is its in the saw or blade. Just my two cents. Jimmy