Checking diagonals for table flatness?

[keith ouellette]

Yes. I am boaring everyone with yet another table flatness question in my attempt to make a better router table for myself.

What would it mean if a table is nearly flat along its length and width in all places checked

and is only out about .008 along one diagonal

but gently slopes down out of flat so that a section of the center reaches just over 1/32" low on the other diagonal?

(I hope I worded that right. Pretty sure I did.)

[Chip Lindley]

Keith, IMO if you have a 1/32" DIP in the center of your table close to the bit opening, that is a BIT much! .008" is Nothing to warrant increasing your Prozac dosage! Personally, I had rather have a slight crown at center than a depression. IF a 12" straight-edge lays flat across the hole, side to side, and front to back, with no obvious gap, I believe you are good to do all the precision routing you will ever need to do.

[keith ouellette]

thanks for the reply Chip but what i am really looking for is

Why would it be out of flat so much on one diagonal and not on the other or along the length or width?

I fixed the problem along the length but if I raise the dip in the center to fix the diagonal that is so out of wack then i ruin the flatness of the width, the length and the diagonal thats not so bad.

Its got me a little confused.

[Jeff VandeZande]

Yes I am totally bored but couldn't stop thinking about it.;)

If the top checks flat around the perimeter but has a high spot or gap in the center it may be twisted.
Try to get it mounted somewhere so you can clamp the four corners.
Then you can shim and clamp the corners to work out the error.
What you mount it on does not have to be flat as you will checking the top. Check across the corners only - then work on the actual flatness. (If it is twisted the error will move and effect the flatness.)
Hope this helps.

[keith ouellette]

I tried what you suggested.

It is fine in the center along its length but from corner to corner it is low in the center. When I change that problem so it isn't low in the center of the diagonal then it isn't flat any more across the center of its width or its length any more.

[keith ouellette]

I'm just trying one more time to get some help.

[David DeCristoforo]

I would not accept a flatness of less than one light band (11.6 millionths of an inch). Any more than that and you will never be able to mill wood accurately. You need to purchase a 6 inch thick cast iron lapping palate, some lapping fluid and some lapping compound. Or you could just get a diamond lapping plate and just use water for the lapping fluid. In either case you will need an overhead lift and track to move the lapping plate around the shop. After you have used this to flatten your router table top to a tolerance of one light band, you can use it on the rest of your equipment. In fact, you should do all of your other tools first because with everything out of whack, you will never be able to build a router table top that is flat enough to lap to begin with. So get started and post back when you are done. Anything less than this would be unworthy of anyone wanting to call themselves a woodworker.

Oh and BTW, it will probably take you several months to lap out a 1/32" error. But stick with it. Don't wimp out on us here... do it right.

[keith ouellette]

I would not accept a flatness of less than one light band (11.6 millionths of an inch). Any more than that and you will never be able to mill wood accurately. You need to purchase a 6 inch thick cast iron lapping palate, some lapping fluid and some lapping compound. Or you could just get a diamond lapping plate and just use water for the lapping fluid. In either case you will need an overhead lift and track to move the lapping plate around the shop. After you have used this to flatten your router table top to a tolerance of one light band, you can use it on the rest of your equipment. In fact, you should do all of your other tools first because with everything out of whack, you will never be able to build a router table top that is flat enough to lap to begin with. So get started and post back when you are done. Anything less than this would be unworthy of anyone wanting to call themselves a woodworker.

Oh and BTW, it will probably take you several months to lap out a 1/32" error. But stick with it. Don't wimp out on us here... do it right.

That was actually very funny. I do wish you would have actually read the question though. You may have given me an answer because the question did not relate the fine tolerances .

Very funny though. Thanks for the help. Sorry to have aggravated you so much.

[David DeCristoforo]

" Sorry to have aggravated you so much...."

Who's aggravated?

[Mats Bengtsson]

I tried what you suggested.

It is fine in the center along its length but from corner to corner it is low in the center. When I change that problem so it isn't low in the center of the diagonal then it isn't flat any more across the center of its width or its length any more.

Sounds to me you are facing a twisted surface with the "balloon problem". Almost all material contain some tension. If the material has been straightened, or bent in such a way that the material has become compressed/extended on a part (for example the konvex surface that was pressed down, then the compression has created a tension. That tension within the material is a force, stribing for equilibrium (causing a tendency to bending/twisting in such a way that the compressed material fits in the length it can achieve after bending/twisting or at least reduces the force caused by compression to be lower than stiffness/thickness).

If your materials stiffness and thickness, combine to make it strong enough to resist the tension, then it stays flat. If the stiffness is too low, or the thickness is too low, then the bending/twisting force wins.

The easiest way is to make such a material flat on only one side, sanding it. That sanding takes away some material, and also reduces the thickness, so it will cause some reactions where bending/twisting changes, but the sanding process naturally adopts to that, and in the end you can have one flat surface.

However, if you try to achieve two flat surfaces (top and bottom), or mount the bent/twisted surface on something which forces it to adjust its shape, then the job become much more iterative. Easiest way out is then to permanently mount the material you have on a flat bottom support which has a thickness and stiffness which combine to be strong enough to keep your material unmoving independent on the tension it contains, and independent to the changes to the tension your create when sanding the not mounted side to flattness, or mount it to supports.

--- Mats ----

[M Toupin]

That was actually very funny. I do wish you would have actually read the question though. You may have given me an answer because the question did not relate the fine tolerances .


Or you could just sell off all those sloppy tolerance woodworking machines and but yourself a nice metalworking lathe and mill and be done with it...

Seriously, the bottom line is does it effect your woodworking in a real world setting? Cut some wood and find out before you start to obsess.

Some darn fine craftsman have been making some darn good furniture for a heck of a long time without feeler gages and dial indicators. At a certain point it's about the craftsman, not the tool.

Mike

[Ben Cadotte]

I lean the same way as Mike. If one piece matches up with another piece from the router. Then its just fine.

I would think the most important line to have flat would the the line you use most when running stock through the router. And if your measuring it in thousands of an ich. I would think its just fine.

To me the most important part to have flat is the router plate (or center area around the bit). I use 4 feather boards. One on each side pushing towards the fence, and one on either side of the bit pushing down. I purposly over built my fence to allow a tall square surface to allow feather boards. I have been doing alot of trim and lots more to do. They come out as good or better than if I had gone to purchase it.

[Jeff VandeZande]

Couldn't get back here till this morning.


I tried what you suggested.
It is fine in the center along its length but from corner to corner it is low in the center. When I change that problem so it isn't low in the center of the diagonal then it isn't flat any more across the center of its width or its length any more.

I thought that is what would happen.
(If it is twisted the error will move and effect the flatness.)
So now if you can build your base fairly stiff so you can again shim the corners -and- around the opening for the router you should be good to go. The top of your base would look like the router top and you would slide the shims in between as needed and secure the top in place.
Or you could just throw the whole thing away and buy the cast iron one.
$$$$$$$$$$$$$$$ ouch!
I made mine out of and old piece of counter top and had to do this same thing, it sits on an old adapted angle iron frame, ain't nothin fancy but it does a nice job.

[Wilbur Pan]

My bet is that you have a bit of twist in your surface, and that the 1/32" that you are seeing is the result of cumulative error.

I know it seems like 0.008" may seem pretty flat, and 1/32" is a big gap, but remember that 1/32" = 0.03125". If you have an error of 0.008, and that happens four times across your surface, that will accumulate to 0.032" out of flat, and there's your 1/32".

I'm going to refrain from comments about whether this is acceptable in the building of a router table, as I don't actually have or use a router. But I will point out that if you use hand tools instead, a lot of these issues go away. ;)

[David DeCristoforo]

"...if you use hand tools...these issues go away..."

Right. Then all you need to worry about is the sole of your jointer plane being .000003 "out of flat". Or if the bevel on your paring chisel is ground at 25 degrees or 25.375 degrees or how many millionths of an inch of clearance you should have between the faces of your dovetails. Hand tool junkies never get into obsessing over tolerances or waste time hair splitting!

[keith ouellette]

"...if you use hand tools...these issues go away..."

Right. Then all you need to worry about is the sole of your jointer plane being .000003 "out of flat". Or if the bevel on your paring chisel is ground at 25 degrees or 25.375 degrees or how many millionths of an inch of clearance you should have between the faces of your dovetails. Hand tool junkies never get into obsessing over tolerances or waste time hair splitting!

I don't know David. I haven't seen anyone talking about anything to the 6th decimal or 1000's of a degree. I'm sure there isn't a feeler gage that thin but I know I have one at .002.

There aren't manufacturers who make money selling things with micro adjustments on fences or flatly ground machine tables.

You've taken a question about thousandths that can be seen and made it as absurd as possible talking about light widths and things to the millionth place.

So if it went the opposite way:

Did you buy a jointer with humps and depressions on the bed? I bet its off by 1/64 or maybe a 1/4 inch.

Did you hear the story of the guy who stubbed his toe and the guy next to him said "you might as well of killed yourself".

whats the purpose of making up extremes no one has mentioned.

[David DeCristoforo]

"Did you buy a jointer with humps and depressions on the bed? I bet its off by 1/64 or maybe a 1/4 inch."

You know, it's funny you should mention that. I am using a jointer I bought 25 years ago (a Delta 8 inch long bed) and I have probably run tens of thousands of feet of material over it. And never once has it ever occurred to me to check the tables with a precision straight edge and feeler gauges. I'm sorry if you think I am being insulting but I just think this is "blowing smoke". Woodwork is simply not that critical.

[Chip Lindley]

Keith, my *too-personal* post was deleted by our moderator. Not a problem! In all honesty and in the spirit of trying to further good woodworking everywhere I say:

You continue to try to tweak this first-time router table made of two thicknesses of 15/32 ply! (unless something has changed that I am not aware) That, IMO, is your problem. The stuff will NOT remain flat! When one area is shimmed, another area moves!

My best advice is: Go back to the drawing board! My old RT extension was 3/4 *particleboard*! I used it for years with good results, but never *lasered* it for flatness! Now there is MDF, which is more dense and lays pretty flat if it starts out flat! OR GO FOR BROKE and save up for a CI top! Then, if you have an issue, you can take it up with the manufacturer!

Take Heart Keith! Thomas Edison didn't get the Incandescent Light Bulb right until after 10,000 tries! You are way ahead of the game at this point!