Table Saw--Problems ripping straight

[Grant Lasson]

This one may seem a little ticky but it's driving me crazy. My 10" Jet left tilt cuts in a slight wedge shape--say +0.004" thicker at the end than the beginning (over 2 feet). The phenomenon is quite repeatable.

My blade is out no more than 1.5 thou from front to back. The fence is very slightly toed out--although I've tried no toe to much more toe. That doesn't seem to make much difference. I'm not noticing any fence deflection and I'm not pushing against the fence too hard. Just enough to keep the board flush against it.

Normally, I haven't worried about it but right now I'm getting ready to final cut the pieces of 16 panels for case work (width and length). It would be very helpful if they were all exactly square and parallel. I'm going to assemble the cases with HB DTs.

BTW, some sharpies may suggest technique is my problem. I'm open to that. So far, the problem exists with and without feather board pressure (gentle to firm pressure ending 1/2" before the front of the blade). I'm getting this result using plywood so case hardening or tension release shouldn't be an issue. The edge against the fence should be dead flat. I've checked with 1.5 thou feeler gauge.

Any recommendations before I plunge ahead?

[Peter Quinn]

If your panels are truly out of parallel by .004" I'd have to say who gives a rats behind, and frankly how are you measuring a .004" deviation over 2' to begin with. That's one sheet of paper, and also one more sheet of paper than my tired eyes can see on a ruler. Or call it two swipes with a block plane. The reason I prefer woodworking to say machining parts for formula one engines is that wood is far more forgiving than stainless or aluminum. Forgive yourself and move on.

I worked in a shop where the lead molder had been a machinist and actually had a pair of calipers (mitiyomo?) with 24" jaws.

That said try getting that .0015" out of parallel out of the the blade (and good luck with that), try a different blade (could be slightly tweaked), run the fence parallel to the blade not toed out, and try clamping a block to the table on the opposite side of the fence at the outboard side.

[Ben Cadotte]

My guess would be the fence bars are not fully perpendicular to the blade. May be good when close to the blade and measuring, but once the fence moves right along the rails that are not totally perpendicular. You will pick up more of an error.

.004 doesn't seem that bad to me though. Thats only 1/256th of an inch.

How old is the saw? Problem since you have had it? Did you clean the front rail mount good before installing the rail? Do you have a right side table? Is it aligned propperly with the front of the saw?

Take a thin piece of wood and cut it with just the front of the blade. Use the miter gauge in the right slot (keep it square). Use the fence for width of cut and lock it down. After the cut, shut down. Move the cut to the back of the blade. Running against the miter gauge make sure the wood is against the fence still. Now see if you can get a feeler gauge between the cut and the blade. If you can and its more than with the fence close to the blade (original toe out). Its the front rail (guessing its a T-square type fence).

[Dan Lee]

Not sure I'd worry to much about 0.004".
But anyway did you measure intermediate points to confirm a wedge. If you checked only the 2 ends it could be as the piece exits the cut the stock is shifting ever so slightly into the blade.

[Bob Genovesi]

This one may seem a little ticky but it's driving me crazy. My 10" Jet left tilt cuts in a slight wedge shape--say +0.004" thicker at the end than the beginning (over 2 feet). The phenomenon is quite repeatable.

Any recommendations before I plunge ahead?

.004 over 2 FEET!!!! http://img291.imageshack.us/img291/6236/shockedsmileyzn3.gif http://img291.imageshack.us/img291/6236/shockedsmileyzn3.gif http://img291.imageshack.us/img291/6236/shockedsmileyzn3.gif

I'd consider taking the saw back and buying something with a bit more precision like a BridgePort milling machine.......

[M Toupin]

I'd consider taking the saw back and buying something with a bit more precision like a BridgePort milling machine.......

What Bob said...

We're talking wood here, it's not perfect to begin with, put down the micrometers, verniers and eye lopes and take a deep breadth.

Mike

[Bruce Wrenn]

First thing I would do is get blade and left miter slot aligned, not close, but exact. Then check your fence over it's entire length, to the left miter slot. Kinda sounds like the fence may have a slight bow in it, to me.

[Grant Lasson]

Dan: I did measurements all along the board--it's a wedge.

Ben: Thanks for the rail idea. I hadn't considered that and I'll check tomorrow.

M & Bob: I know. It's small. I've been getting by just fine. On the other hand, it's systematic--not random. It happens every time. I'll let it go if I can't get a handle on it but since it is systematic, it seems solveable. And then my panels will be the same top to bottom and side to side. That way I won't have to keep track of how I put each panel through the saw. Since I'm matching panels, magnification of the error makes it quite noticeable.

Bruce: I think you're on to something here. The HDPE face on the fence is absolutely bowed. At either end of the fence, the dial indicator reads the same but in the middle it bows somewhat toward the blade. Coming into the cut, the board must be slightly pushed into the blade. On the far side of the bow, it feeds away from the blade. Hence my "wedge" cut. I'll confirm tomorrow.

Thanks for all the insightful comments. I'll post whether I get it solved or whether I have to move on.

[Steven Wilson]

You should spend time and make the HDPE fence face straight. You can do this with shims and by varying the tightness of the nuts that hold the fence face on. You should take the HDPE fence off and smooth out the metal subfence. Often times the main fence has too much paint that leads to a dip or bow to show through the fence face. A .004" in 2' is too much. I have my slider set up to hold .002" in 8-9'. Being able to make parallel and square cuts makes glue ups and fitting much easier.

[Simon Dupay]

Why are you worried about .004 difference between panels? THIS IS TOO SMALL A DIFFERENCE FOR ANYONE TO NOTICE!!! Don't worry about it.

[Peter Quadarella]

I don't have a table saw, but for a difference that small, couldn't you just tape a piece of paper to the side of the fence where the wood is wider?

[Scott Loven]

Isn't this a problem of fence deflection?
Scott

[John Thompson]

You will quite often get more deflection than .004 from just tension release when the wood is severed. No offense to anyone, just a general statement but... IMO the biggest con and obstacle to advancement of skill I have seen in WW over the last 36 years is:

Taking a manual or digital micro-meter out of the hands of a machinist who needs one and putting it in the hands of WW's. Much ado over nothing in the majority of cases. And I'm not saying it cannot be useful as it can but.. simply it is over-hyped in it's true value to a WW.

I have seen too many times someone spend countless hours tuning machines to perfection.. but can't build a box without a set of plans. Where have all the flowers gone?

BTW.. if you want and need your ripped stock precise.. rip it 1/8" over-size and take it to the jointer to take off 1/16" on both edges.

Sarge..

[Jon Crowley]

Do you have a riving knife or splitter on your saw? If it is not aligned with the blade, it will push to one side and can cause this and other problems. I know from experience. :o

[Lee Schierer]

Isn't this a problem of fence deflection?
Scott

This was my thought exactly. I am not familiar with your fence, but getting deflection when ripping would not be hard to do. I your fence is aluminum, aluminum fences are not all that ridgid and .004" deflection would not take much side force.

[Josiah Bartlett]

If its not your fence face bowing, I would try changing blades. Your blade could be dished or duller on one side than the other.

[Grant Lasson]

Steven: your suggestions were gold! I took the HDPE face off and worked the steel tube until it was flat. There was too much paint in some spots just as you mentioned there might be. In addition, the screws raised an edge around the holes in the tube. Filed those down.

Even with the flat tube surface, the HDPE didn't sit flat. Unfortunately, it's also very flexible. I shimmed in front and behind each screw. (The fence surface was depressed around each screw.) I was able to get it pretty flat but not absolutely flat. (The surface resembles a very, very gentle sine wave.) I'd never thought about it before but the more rigid melamine-style faces may be easier to work with.

Now, the ripped piece has no measurable difference from end to end.

Thanks for the help! I was not thinking about the fence as the source of the trouble. It really didn't take too long to fix once you got me on the right track.:D

[Grant Lasson]

There's a sizeable group that expressed the sentiment (essentially) of "Quit being a Dork!" (My paraphrase:o.) I appreciate the sentiment. I've had the saw 8 years and haven't done anything about this before. But, with your permission, I'd like to challenge the assertion that it doesn't matter. I'll agree that in most cases, it doesn't matter. But it does depend on the situation.

Here's mine: I've got two daughters for whom I'm trying to build chest-on-chest dressers. That makes four cases. Eight corners each case. Thirty two corners in total. Let's say they're about 4 ft long. My width error is then doubled to 0.008 to 0.01". If I don't keep track of how I put every panel through the saw, then I'll inevitably mix wider edges with narrower edges. The difference in width will be very noticeable and I'll have to manually adjust 32 corners so they feel right. Not a tragedy but in this case I thought it worth while to invest some time up front.

BTW, John. I think your insight is absolutely correct. The precision required to work to close tolerances with machines wastes a lot of creative time. The width difference would cause trouble with any other precision tool involved in the process. In my case, that's a Woodrat. I'm planning on hogging out most of the DT waste with it before getting busy with the chisels.

My shop has evolved to a weird hybrid of hand tools (with which I prefer to work) and tools on which I've installed digital readouts (e.g., Woodrat, Ornamental mill, table saw, etc.). I'm not sure where this is going. Can I make peace between the two worlds or will I lose my mind?

[Burt Waddell]

Grant,

I don't mean to be disagreeable but I am with those who say that small amount is not enough to worry about. A board can expand or contract more that in one day. I also prefer precision but over years learned that isn't always the wisest approach.

Burt

[John Thompson]

I can't answer that Grant.. only you can. And I can't comment on how accurate stock must be fed to a precision machine as Wood-rat as I don't have any. I just finished sharpening two sets of chisels and 3 hand-plane irons hand holding without a jig. They have gone through a bed and 6 drawer chest the last 6 months and did an excellent job with guessed angles.

Furniture has been made for around 5000 years by craftsman. Until just recently I doubt anyone spent more time tuning machines and fiddling with angles than they did building pieces. I have seen entry levels that were afraid to start a project until every tool in their shop was precise. Someone out there has done an excellent job of marketing.

Wood tension will move wood.. moisture will move wood.. clamping will compress wood. What I suggest to each and every one is to puchase a cheap feeler guage at Auto Zone.. etc. Take a look at what .004 .008 .012 is.. and what you will discover is you can't force a knat's hinny between any of them.

Then relate that to tension release.. contraction and expansion and clamp pressure then make up your own mind to whether every machine has to function as a NASA space capsule. That's my challenge.

Regards and good luck with the chest...

Sarge..

[Joe Chritz]

From a distributor of Starrett.

"Note: When comparing Starrett straight edges to others on the market, it is important to note that they are precisely milled on both the edges and flat sides to a straightness of + / - .0002 in. per foot. - They are also parallel to: + / - .0004 in per foot."

You are cutting parallel to 1/5th of a precision straight edge by ramming a sheet of plywood through a blade.

Don't sweat it.

That could be technique causing fence deflection, rails, the wind or whatever.

Joe