Making a precise bevel rip

[MarkSHancock]

I need to make some 8" x 16" box colums out of 3/4" MDF for my basement project. For this I need to be able to rip four pieces (two pairs) for each column with a trapezoidal cross-section that have opposing 45 deg bevels along each side and a max width of 8" for two and 16" for the other two.[8D]

This is my first time to do a bevel rip so I went about practicing the cut on my BT3100 (according to the manual) and have been unable to figure how to set the zero properly for the rip fence for a bevel rip. First, why do the instructions say to rip with the fence on the left for a bevel rip? This forces you to use side of the fence with less surface area which makes lining up an already beveled piece along the fance harder. Second, when I set the zero with the saw at 90 deg, set the fence to 4", rotate the blade to 45 deg, then make a rip, I don't end up with a 4" wide beveled board. I have tried it several ways and I can't find anything that is 4" wide.[V]

Is my problem with my technique, my saw, or my expectation?

Thank,
Mark

[stoli]

The rip fence should go on the left of the blade for a right tilting blade. With this setup, the offcut is free to fall. Imagine the setup the other way around. The piece between the blade and the fence is trapped -- can't move down because of the table, can't move up because of the blade, can't move right because of the fence. Any slight motion of this piece will cause it to bind up on the blade, and you will get some serious kickback.

You are correct about the scale not lining up. SInce the blade pivots about a point that is well below the table, any scale that was correct before the tilt will be incorrect after the tilt. However, most people don't really use the scale for exact results anyway. It is always better to measure for cuts with the one measuring device you are using for your project (you are only using one measuring device, right?).

To get around the limitation of the side of the rip fence, you could always make a secondary fence that attaches to the rip fence out of MDF and some of Jeff's T-nuts (or a pair of clamps).

-=[doug]=-

[RodKirby]

Perfect answer from Doug (Stoli)!

[BobSch]

quote:Originally posted by RodKirby

Perfect answer from Doug (Stoli)!

I'm going to second Rod on this one and add to Doug's comment on using one measuring device.

Mark, Doug isn't kidding about always using the same rule/tape/etc. throughout a project. I found out the hard way that the two tapes and the 18" rule I had in the shop couldn't agree on just how long a foot was. Ruined some nice lumber figuring out why 9" didn't equal 9" [:0] Sounds silly, but it's the truth.

[boblon]

Yeah, I've learned that lesson on the 'one measuring device' the hard way myself. Also, if your using a tape measure try to remember to make critical measurements using the same edge of the tape. DAMHIKT.

BobL.

[MarkSHancock]

Thanks for all you info.
Regarding the side the fence needs to be on, you safety concern really makes sense. There is a lot I need to learn so I stay safe. Table Saws have a lot of potential for harm to a newbie like me.

Regarding a single measuring device, generally I use the same one for a project but from what I have seen they all seem to match to within the accuracy/resolution of the measure. Certainly I don't try use a tape measure to measure something I need better than 1/16" (I would use my calipers for that) and I don't try to use calipers to measure 48" either. But, if I measured something within the range of both I expect them to match within 1/16th. I agree you can beat this some when repeatability is more important than absolute accuracy by using the same rule for everything. Is that what you mean? ( Can you tell I'm an engineer )

Regarding the scale not lining up because of the saw pivot; I thought the saw pivot point was supposed to be the point where the surface of the blade intersects the surface of the table so that this is not an issue. I have a friend at work that has a Delta Commercial table saw and so I asked him and he said that he sees no variation with blade angle; the scale is true no matter the angle. He mentioned that I could make tool/jig to account for the difference but I had already figured that as well. Also, when you read the directions in the owner's manual for making a bevel rip, they mention setting the fence zero before setting the blade angle then setting the fence for the cut distance after. To me that implies that they expect it to be true based on the scale. What I am seeing is a 9/32 variation of the material length from perpendicular to 45 deg. This seems quite excessive. I was expecting no more than a 1/16th variation. For this project, repeatability for each identical cut (parallel sides of the box) is what is key and I would like cut accuracy to be within 1/8 (but that is not as important); so, I should be fine setting up with a jig but that is inconventient. I would prefer if the saw worked as my friends does. I wonder if the problem is just something is out of allignment?

Thanks,
Mark

[LCHIEN]

I don't think there's an easy way to make the bevel pivot point at the surface of the table.

Everyone would like the indicators to be perfect for all settings and bevels but its not that easy to do.

I always make test cuts on scraps of wood to verify and or adjust dimensions for final cuts.

[stoli]

For most woodworking, precision is much more important that accuracy (one engineer to another). However, I think 1/16 is way, way to coarse. For most furniture, you need 1/64 or better -- the results of 1/16 error from precision on your column will be noticable in that you won't be able to get good joints or a square column. My approach is to use 1 ruler, but cut long, then sneak up on my fits using the micro-positioner, which should easily be able to move the fence by 1/128 inch. Another precision hint is to use the same setup whenever possible. For example in your case I'd probably rip the column sides about 2" oversized, but square. Then I'd bevel rip one side of all your columns. Finally, I'd sneak up on the desired width of one column (using featherboards if possible), and then rip the other 3. This guarentees that all 4 sides will be the same, to within operator experience. If you try to bevel rip one side of a piece, adjust the fence to the correct position, bevel rip the other side of the first piece, then repeat this set 3 more times, I guarentee that you will get variation in the final widths, no matter how finely you think you can reposition the fence (incra fences exempted).

Regarding the offset when you bevel rip. I thnk that must be the case for all saws. First off, the only way to get a design where the position of the blade stays constant on the table through all bevel rips would be for the axis of the blade change to be on the table. Any change in blade height would likely also move the tilt mechanism, so you could only get this "feature" at one blade height. But, the pivot point probably also needs to always be under the table to protect the mechanism. All of this is guess work on my part -- can anyone confirm these assumptions?

-=[doug]=-

[MarkSHancock]

Certainly you need something more complicated that a single point pivot (like a hinge) to pivot at a point that keeps the fence true for a bevel cut, but isn't that what makes a table saw complicated. I was/am not expecting 1/128" precision Doug (stoli) is expecting; but, come on, I can do better than 9/32 with my hand saw (maybe even with my chain saw ).
Regarding the affect of blade height, when the saw rotates the blade height mechanism has to rotate with it so that the blade height doesn't affect the cut location (the Ryobi does this well, blade height does not seem to affect cut location).
I am not an ME (thought I work with several world-class ones) but it seems having the motor travel along two circular races would create a virtual pivot point that could be alligned to the intersection of the table and blade making the fence accurate. I am sure there are other (probably better) ways as well. I tried to figure out how Ryobi does it from their exploded diagrams but it wan't clear to me. Has anybody dissassembled their saw and noted how they do it?

Doug (stoli),
I have my cuts planned to rip a rough square-cut the 49x97 sheet in half down the length to make it managable (probably with my circular saw). After that I would rip the pieces all with bevel cuts. That way I can get the full column out of a single piece of MDF. I will cut opposite sides without changing the saw setup changes so that opposite pieces are precisely the same.

Mark

[MarkSHancock]

<center>Here is a research poll for everybody</center>

How much does the cut location of your saw vary with the bevel angle?

Procedure:

• Square up a couple of pieces of wood
• Set the fence to a fixed location and the blade to the vertical position
• Mark a cut
• Move the bevel the 45 deg and repeat the cut (without moving the fence)
• Compare the length of the side that faced the table of the bevel-cut piece (should be the short side) to the length of the square-cut piece. If the fence stays true, they should be the same.

Here is what I found for my saw:
Ryobi BT3100 - Bevel piece was longer by 0.32 in (5/16")

[LCHIEN]

Basically you cannot make the saw blade INTERSECT the top of the table at any bevel angle. Which is what is required to keep the cut width the same as the bevel angle changes.

Look at this illustration:


In order for the blade to rip fence distance to be constant, the axis of rotation (for the bevel) of the blade, arbor and motor assembly will have to be positioned at the blue dot. That would put the axis flush with the top of the table.
It would be virtually impossible to buld a compact pivot flush with the table top since a pivot would require a finite thickness and any such thickness would interfere with the path of the wood above the table.

Mark, I think your expectations are too great. I for one will not waste my time measuring the change in distance from the rip fence - it'll be different for every bevel angle. I simply pretend there's no scale (many people don't even use it, ever) and simply measure using a steel rule. And then (here's the complicated part...) I make a test cut and confirm/adjust/confirm, etc until its right. With my rip fence microadjuster it usually takes only one adjustment. I use the scale but I know its only good for 90° cuts.

I don't think even your friend's commercial Delta saw for $3000 or whatever can keep the fence-blade distance constant at different bevel angles. Basically its a geometric impossibility unless the physical pivots are placed way outside the normal saw body dimensions in which case they'd have to be way stout to not deflect, etc. and make the saw ungainly and difficult to use.

(Everyone, I apologize for showing the tilt opposite of the way the BT3000 operates. Oh, maybe I intended for you to view this from the rear. Yeah, that's the ticket! It's viewed from the rear.)

[MBG]

Informative post!

I had a project in the past that required a bevel cut on both sides of the board. The first bevel is easy because there is a nice square edge to ride against the fence. For the second bevel I carefully pushed the beveled edge against an auxilary fence - was not to happy making the cut this way.

I feel dumb now - never thought to leave a little extra stock on and flip the board over so the other square side could ride against the fence.


Thanks for the post,

Mike

[gonesailing]

Boy, amazing how a good diagram helps one to think. First of all, the tilting mechanism is not at the surface of the table saw, but well below it at the arbor. From what I can see, the machanism causes the arbor to travel in a circular arc that keeps the blade perpendicular to the arc. This is nicely shown in Lorring's diagram. This motion to the right, (looking from behind) if done with precision, does establish a pivot point that is fixed in space. Since the blade only travels from verticle to 45 degrees if should be possible, if not simple, to build such a mechanism where the virtual pivot point lies in the plane of the table.

Now comes the difficult part, getting a zero thickness blade. If the pivot point is established at the table surface, but the center of the blade, the scale will still be off by an amount that depends on the thickness of the blade and the angle of the cut. Still, the pivot pooint could by established at the edge of the blade, BUT if you changed blades and the carbide teeth were a little wider or narrower, you would be back to the same problem. The maximum error should be at 45 dregrees and be about .7 x the width of the kerf, for the BT3k about 1/16 of an inch.

I think this is just an engineering problem, and not a difficult one at that. That doesn't say it would not be expensive, and probably is not worth it for most applications that I will ever be involved in.

Mike

[LCHIEN]

quote:Originally posted by gonesailing

Boy, amazing how a good diagram helps one to think. First of all, the tilting mechanism is not at the surface of the table saw, but well below it at the arbor. From what I can see, the machanism causes the arbor to travel in a circular arc that keeps the blade perpendicular to the arc. This is nicely shown in Lorring's diagram. This motion to the right, (looking from behind) if done with precision, does establish a pivot point that is fixed in space. Since the blade only travels from verticle to 45 degrees if should be possible, if not simple, to build such a mechanism where the virtual pivot point lies in the plane of the table.

Now comes the difficult part, getting a zero thickness blade. If the pivot point is established at the table surface, but the center of the blade, the scale will still be off by an amount that depends on the thickness of the blade and the angle of the cut. Still, the pivot pooint could by established at the edge of the blade, BUT if you changed blades and the carbide teeth were a little wider or narrower, you would be back to the same problem. The maximum error should be at 45 dregrees and be about .7 x the width of the kerf, for the BT3k about 1/16 of an inch.

I think this is just an engineering problem, and not a difficult one at that. That doesn't say it would not be expensive, and probably is not worth it for most applications that I will ever be involved in.

Mike

Actually you are correct that the axis should not only be at the surface of the table but at the edge of the blade. I thought of mentioning that but it seemed only to complicate it more.

The axis of bevel can be above or below the arbor depending uponthe elevation dialed in; the arbor actually moves up an down but hte mecahism tilts, that's why the blade stays in the same plane when you elevate. If the elevation boved the bevel instead, each elevation at any non 90° bevel would result in a parallel plane and the bevel cut would move all over the map.

I think that the geometric location of the pivot point is problably located below the table surface the exact same amount as the translation of the blade at the table surface when doing a 45° bevel.

If Mike found the cut became .32" wider then the bevel mechanism is pivoted .32" below the top of the saw. It should also be obvious that the bevel pivot is located at the center of the arc made by the bevel crank handle slot in the front of the saw.

Finally I'll disagree that its "not a difficult engineering problem" to place the axis of the bevel at the surface of the saw. I gave the reasons in the post with the diagram.

Given your correct observation that the apparent thickness of the blade changes with bevel, you'd also have to decide which side of the blade you wanted to make right.

I ran out of time but maybe later I'll make a diagram of the way the saw is actually set up.

Note I do now see there is useable pivot point location configuration wherein the rip-fence to blade edge distance can be maintained the same but it would only work for 45° and not any other angle.

[RainShadow]

This post is way beyond my little pea brain...............

Good job guys, Randy from Oceanside, CA