Page 2 of 4 FirstFirst 1234 LastLast
Results 16 to 30 of 60

Thread: table saw tuning......

  1. #16
    Join Date
    Aug 2013
    Location
    Princeton, NJ
    Posts
    7,254
    Blog Entries
    7
    A very basic setup can measure .003”. Take the angle out of the indicator by eye.

    Breathing on wood doesn’t change it .003”.

    Wood moves in a relative fashion in many cases, your mortise and tenon fit does not change even as the parts move throughout the seasons.
    Bumbling forward into the unknown.

  2. #17
    Join Date
    Jul 2007
    Location
    Inkerman, Ontario, Canada
    Posts
    1,389
    Quote Originally Posted by Jacob Reverb View Post
    It also takes some pretty sophisticated setups to even measure .003" accurately. For example, if the rod in your dial indicator isn't perfectlly perpendicular to your fence/blade/whatever, you're gonna get another .006" or so of error.

    Then there's the issue of the straightness of the materials used to construct your fence. On my Powermatic's Accufence for example, the factory BB panels that cover the rectangular tubing of the fence vary by probably five times that – .015" – in and out, like a sine wave, all along the length of the fence, due to the clamping pressure of the t-nuts used to fasten the BB to the tubing. But even if you stripped off the BB panels, unless you scrape and polish the steel tubing of the fence, a speck of rust is gonna quickly throw in another .010" or so.

    If you add to that the fact that your wood is going to move maybe five or ten times that .003" every time you so much as breathe on it, you quickly realize that three mils has as much meaning in this discussion as three Angstroms.

    I'd be interested to see the woodwork done by those who think half the thickness of a human hair is "a lot." This kind of thing gets silly after a while.
    01-dr winsor .JPG06-051.JPG07-053.JPG11-Profil17.JPG09-Profil93.JPG10-Profil97.JPG03-SAM_1349.JPG04-rsz_1-woodmortise4.JPG

  3. #18
    Join Date
    Jul 2007
    Location
    Inkerman, Ontario, Canada
    Posts
    1,389

  4. #19
    Join Date
    Jul 2007
    Location
    Inkerman, Ontario, Canada
    Posts
    1,389
    these business card holders are made from one piece, tolerances are pretty tight.


    01-SAM_1329.JPG02-SAM_1334.jpg05-SAM_1338.JPG

    The ones below are even tighter tolerances.
    0.003" out on these card holders and they are trash, tops either wont fit or will fall off.
    Joints wont line up, they just wont work.

    06-SAM_1339.jpg08-SAM_1342.jpg09-SAM_1343.JPG

    I guess that you have never set up a supersurfacer; 13.250" long blade and back-knife on the small ones, Taking shavings down to 0.001"
    To some people 0.003" is a lot.

    12-rsz_shavings1a (2).JPG14-rsz_1sam_1812.JPG

  5. #20
    Join Date
    Apr 2008
    Location
    East Virginia
    Posts
    830
    I wonder how much glue fits into a gap of 15 ten-thousandths ... assuming you can get it assembled, even dry, without a sledgehammer.

  6. #21
    Join Date
    Jul 2007
    Location
    Inkerman, Ontario, Canada
    Posts
    1,389
    Jacob, In some instances, for some setups 0.003" is a lot, other situations its not.
    Depends what you are doing.
    When you are doing very complicated items with multiple parts and setups, it may be necessary to keep high tolerances to avoid accumulated problems down the road.
    Having your machines setup the best you can doesn't hurt anything.

    Each to his own of course, but don't dismiss others for working to whatever standards they see fit.

    The whole point of working to high precision is you don't need a sledgehammer.

  7. #22
    Join Date
    Aug 2013
    Location
    Princeton, NJ
    Posts
    7,254
    Blog Entries
    7
    Mark, those table bases are exceptional, really nice work.

    Jacob,

    Tolerance, or error, in machine setup depends on the machine and the operation. I like to minimize it to as near to nothing as possible because it allows me to find confidence in my machines.

    People often associate looser tolerances with a faster pace of work, but the opposite is often true. Maintaining tight tolerances throughout a project can greatly reduce the amount of time spent making joints look tight or marking out by hand, etc. I did a set of frames recently that used to take me days to do, and completed the cutout in a few hours because I have much tighter cutting tolerances now to the point where I don't need to trim to fit.
    Bumbling forward into the unknown.

  8. #23
    Join Date
    Jul 2007
    Location
    Inkerman, Ontario, Canada
    Posts
    1,389
    Thanks Brian, I had an awesome time doing them and hope to get back to doing some more of that stuff soon, its been way too long.
    The is no downside to running tight tolerances and having a good system to reference, check and verify each step; its just good project management, what makes stuff fit, without any problems or worry.

  9. #24
    Join Date
    Feb 2008
    Location
    E TN, near Knoxville
    Posts
    12,298
    Quote Originally Posted by George Makra View Post
    Forest blades use to have a real nice write up on setting up your saw. You can call them and ask if they could email it to you.
    There is also an article by Tom Burroughs in American Woodworker, Feb '95, on "Supertune Your Tablesaw". Excellent article and great diagrams. On the following page is a review by Ellis Walentine on a couple of tools that could help. Hmmm, also 4 pages...

    JKJ

  10. #25
    Join Date
    May 2015
    Location
    Ingleside, IL
    Posts
    1,417
    Quote Originally Posted by Chris Fournier View Post
    If you're using a tape measure then you're not really in the spirit of the post.
    Made me laugh out loud. Thanks. Needed that.
    Stand for something, or you'll fall for anything.

  11. #26
    Join Date
    Apr 2008
    Location
    East Virginia
    Posts
    830
    Quote Originally Posted by John K Jordan View Post
    There is also an article by Tom Burroughs in American Woodworker, Feb '95, on "Supertune Your Tablesaw". Excellent article and great diagrams. On the following page is a review by Ellis Walentine on a couple of tools that could help. Hmmm, also 4 pages...
    FWIW, if you have a contractor-style saw, the PALS gizmo makes what is ordinarily a nightmare – making the trunnion and miter slots parallel – much faster and easier.
    After seeing how it works, I considered making something to accomplish the same thing, but for $20 it wasn't worth the time and trouble.

    https://www.in-lineindustries.com/pr...ctor-saw-pals/

  12. #27
    Join Date
    Aug 2013
    Location
    Princeton, NJ
    Posts
    7,254
    Blog Entries
    7
    Pretty sure you're off by a decimal point on the above, everything you say makes perfect sense at one decimal point further (IE it's difficult to measure .0003", a spec of rust might be .001") I've seen BB ply panels that varied by a few thousandths but not by 15 thousandths.

    Most dial indicators read thousandths (.001"). 'Test' indicators often read .0005" or .0001".
    Last edited by Brian Holcombe; 03-26-2019 at 12:08 PM.
    Bumbling forward into the unknown.

  13. #28
    Join Date
    Apr 2008
    Location
    East Virginia
    Posts
    830
    Quote Originally Posted by Brian Holcombe View Post
    Pretty sure you're off by a decimal point on the above, everything you say makes perfect sense at one decimal point further (IE it's difficult to measure .0003", a spec of rust might be .001") I've seen BB ply panels that varied by a few thousandths but not by 15 thousandths.

    Most dial indicators read thousandths (.001"). 'Test' indicators often read .0005" or .0001".
    Thanks, Brian, but I'm aware of the difference between a mil and a tenth. (I was going by memory when I said my Accufence varied by .015" ... since my memory dims year by year, I just checked, and it varies by about .007" trough-to-crest, so we're talking ± .003" or .004" )

    For grins, I measured a rusty spot on my Freud blade, and it was raised by about .006" from the surrounding plate.

    I suspect there are also large variances in the straightness of the miter slot. Then there's the runout in the arbor, which I suspect is .002" or more, yet within the specked tolerance.

    Since my miter slot is 6-1/4" to the left of the blade, there's also a lot of rod between the dial indicator and the blade to "take the angle out of the indicator by eye," as you phrased it. But here's the thing: A small angle off from perpendicular makes for a large linear difference when we're talking about mils. If my trig is right, an error of just 2° off perpendicular there (88° or 92° instead of 90°) will cause an error of almost .004" in one's measurement of the distance between the miter slot and the blade/tooth.

    It's easy to say, "Well, gee, just take the smallest measurement there," but it ain't that easy, at least on my Freud LU84 blade, because it makes a large difference (like ± .005") where, exactly, on the tooth the tip of the dial indicator's probe bears, since the tooth is shaped like a trunked "V" and the tip of the probe is shaped like a ball-point pen tip. So if you slide the dial indicator forward and aft in the miter slot and try to get the minimum measurement (assuming that this will indicate that the probe is square), you'll also move where on the tooth the tip of the probe bears, and since the tooth is wedge-shaped laterally, the linear distance registered by the dial indicator changes. Lots of moving parts, lots of things to think about, and not enough hands to hold a square, and the blade, and the dial indicator, and the probe, and the mirror, and the magnifying glass, and the leash on Schroedinger's cat.

    All this and more makes it quite difficult (for me, at least) to measure to a .003" level of precision. Maybe you're better at it than me. I'm not a machinist. But I believe that even for a machinist (and I believe a machinist registered his views above), getting into the realm of .003" precision on many of these machines is a fantasy because it simply isn't in the wood, because the machines themselves simply aren't made to that level of precision. Quite simply, you're getting into such small numbers that the noise overpowers the signal. (And Mark's UberDooberSurfacer's ability to take shavings of half an Angstrom, while impressive, isn't really relevant to a discussion of .003" misalignment of a TS, or if it is, I've missed the relevance, though it wouldn't be the first time.)

    For a fairly new woodworker seeking advice on table saw tuning, this can be overwhelming and disconcerting. "Golly, my table saw has a whopping 10 mils of slop. I'll never be able to do quality work on that." And so they go down the rabbit hole chasing mils that don't matter when they would be far better served by using what they've got, building their birdhouses and workbenches and bookcases, and learning along the way. And this, not the striving for accuracy, is what I oppose: The notion that one has to have a $10k machine to do good work, or even that a $10k machine will necesarily produce better work than the $100 Craftsman 113 from 1972. Or that .003" is "a lot" – except in the very rarest of cases, and never in ordinary cabinetmaking for mortal men.

    You see the same thing in competitive shooting, where the guys with the $10k rifles sneer at the beginners with their $2k Savages, and try to discourage them, suggesting that there's no way they can be competitive unless they spend $X. It's baloney, and often enough, the beginners show them it's baloney.

    Anyway, the indian is far more important than the arrow. And the way to improve the indian isn't by chasing mils, but by milling wood. (I also believe design is more important than execution – so long as the execution is clean – but that's another argument for another thread.)

    Finally, as said before, TS blades are tensioned, so what you measure on the blade at rest isn't even necessarily what you have when the blade is running ... and measuring it while running isn't for the faint of heart. Better to measure what the blade does to the wood IMHO, since that's where the rubber meets the road, and all that matters is what comes out of the downstream end of the sausage factory, to mulch yet another metaphor.

    YMMV.
    Last edited by Jacob Reverb; 03-26-2019 at 8:50 PM.

  14. #29
    Join Date
    Feb 2011
    Location
    Central WI
    Posts
    5,666
    Biesemeyer and clone fence plates were notoriously out of flat so you need to know that if measuring. Measuring the kerf tells you the run out but not the cause. Different blades yield different results but when I swap the same blade from saw to saw, I get the same results, but only after I've trued the flanges. At the end of the day, you get a pretty good feel when a machine is operating well. When you get into the Mark and Brian world, you find that you are only happy with really good machines that you can calibrate and that stay that way. Dave

  15. #30
    Join Date
    Aug 2013
    Location
    Princeton, NJ
    Posts
    7,254
    Blog Entries
    7
    OK, These are my experiences and opinions:

    Machine setup is not relative to wood movement, it is independent of it and so it can be looked at independently. IE wood may bow or cup after I joint it, but I wouldn't want my jointer to produce a bow or cup intentionally, I want it to make flat boards. It's up to my ability to chose wood properly at that point to get something that remains reasonably flat. Same goes for any other operation, I want the operation to produce a proper result irrespective of what the wood may do at some point down the road. Wood movement is something I need to plan around as a designer, I plan my joinery with respect to wood movement. I wouldn't work around looser tolerances than can be had because wood moves.

    Tolerances can be tightened until the point at which they no longer have effect on the work that is noticeable. For fine furniture making my personal feeling is that it should be a few thousandths if the machine will allow for it (many of them will, even my Felder can hold that tolerance). My goal is to minimize time spent 'fitting', fitting by hand for a zero gap is extremely time consuming and I'd rather spend time up front prepping a machine to the best of my ability so that I can minimize fitting time. Fitting by hand offers no benefit to the end consumer if the same result can be had by machine quickly.

    I worked for a very careful machinist, if you mentioned to him 'a few thousandths' that would raise a suspicious eyebrow. I suppose I like working with that ideal in mind and find comfort in minimizing machine errors so that my results improve. I'm far from perfect, I find myself combing over my own work looking for spots to improve each time I have the opportunity to do so.

    My goal is to make a very good product and manage my time effectively. Tight tolerances help that for me.

    Specifics:

    Dial indicator. The distance from which you are measuring is not meaningful, a dial indicator is used to measure a relative distance. So, in respect to your saw blade you would be measuring low to high. The distance your indicator travels in that range is what will be effected by angle. I won't argue that angle has no effect, but your math is incorrect in that it's based on the assumption that the distance between the indicator's mounting point on the machine and the saw blade has an effect, it does not. It would be that angle measurement over the length of travel of the indicator. So if the indicator is traveling .003" at a perfect 90 degrees, it would theoretically travel .0045" if it were pointing at 45 degrees toward the same position (realistically the indicator probably wouldn't travel very nicely). So a few degrees is not going to have a meaningful effect in this case.

    I agree with you that a setup is only as good as it's result, I like to work out the errors I can find up front, and then check my work with material to see if my work is accurate. If I missed the mark, the wood will tell me.

    Assumptions:

    - No where does Karl mention that he is a beginner. No reason that an advanced user can't wonder what tolerance range is good.
    Last edited by Brian Holcombe; 03-26-2019 at 11:13 PM.
    Bumbling forward into the unknown.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •