An Improved Table Saw Fence With Threaded Rod

Back in the bad old days, table saw fences were terrible. You would have to measure the top and bottom of the fence before each cut, just to make sure the fence was square to the blade. In the 1970s, [Bill Biesemeyer] invented a better table saw fence, one that was always square, and included a measuring tape, right on the table saw.

[Jer] wanted an upgrade for his table saw and came up with what might be the next evolution of the table saw fence. It will always produce a square cut, but unlike the 1970s version, this fence has repeatability. If you rip a board to 1″, move the fence, come back to it after a month, and try to rip another board to 1″, those two boards will be exactly the same width.

The secret to this repeatability is a threaded rod. On the front of the fence is a big, beefy piece of threaded rod with 16 threads per inch. On the fence itself is two nuts, cut in half, welded to the guide, with a lever and cam to lock them in place.

When the lever is up and the nuts are disengaged from the threaded rod, the fence easily moves from one side of the table to the other. When the fence is locked down, it locks to the nearest 16th of an inch, and only the nearest 16th of an inch. While that may seem a little large for a relatively expensive tool, this is wood we’re talking about here. There’s not much reason to make the resolution of this fence any smaller; wait until the humidity changes and you’ll have a piece of wood that’s the desired dimension.


68 thoughts on “An Improved Table Saw Fence With Threaded Rod

  1. “There’s not much reason to make the resolution of this fence any smaller…” Be amazed at how quickly that attitude changes once a piece of furniture full of 3/64″ gaps enters your home.

    1. Ikea?

      Seriuosly though, standards have dropped very low with the majority. I grew up woth furniture built to last generations (yeah… antiques). Now with stores like Ikea, Wal*Mart, K-mart to neme a few selling, what is essentially, disposable furniture.

      Knock that price low enough and the buyer won’t care about 3/64″ gaps.

      1. I don’t think quality has dropped, we have an old cabinet made from shipping crate wood and it ain’t pretty. What happens is that people throw away crappy or ugly furniture, and only the nice stuff gets passed down to become antiques.

        1. Quality has most certainly dropped. You won’t find a single dovetail or mortice and tenon in any Ikea or what-have-you furniture store. It’s not that we can’t do it, it’s that people don’t want to pay the money for the quality.

          1. Actually the decline ff dovetails and mortise and tenon is because of the increase in strength and durability of modern adhesives. You no longer need the wood to provide the majority of strength to a joint because the adhesive can take it.

      2. Yes ikea has stuff meant to not last generations, but they also have stuff made from slightly better materials and it’s up to the consumer to pick which he buys. But sure it’s still not likely to going to become antiques. But if you have stuff made of real wood or sturdy welded metal it should last a bit.
        But the question is: Is it quality dropping? Do you think the furniture the poor masses bought in the past were all of such quality that they ended up antiques? Or is it that you only know the stuff from the past that lasted and the most common cheap stuff you never saw because it’s long been burned as fire wood?
        Personally I think the truth lies in the middle and the quality was probably generally slightly better but at the same time they had stuff that was not meant to become antiques in the past too.

        1. 1/16″ is 0.0625. 3/64″ is three quarters of 2/32″, and 2/32″ is 1/16″, so about 0.045″ or something.

          Welcome to Hackaday, where people call programmers noobs if they can’t count in octal, but refuse to if a tape measure is involved.

          1. Yes, I can figure it out easily enough, but in developed nations we use a better system which eliminates any conversion steps and the errors they can lead to.

        2. He’s using threaded rod that has about 629,921 threads per kilometer. I think the nuts take a .1429 decimeter wrench. .005″ is 1,270,000 angstroms, but 1/16″ is 1587.5 microns, so, duh, that one’s pretty obvious. As for 1/16 vs. 3/64, come on.

        3. I’m so glad the metric system is here! We can now do away with arbitrary units of measure such as the foot, degree F, and Inch of Mercury.

          The new ones meter, degree C, and Pascal are based on universal measures that would be quickly understood by anyone from a civilization on a distant planet.

          Also our measure of time (the second), our division of time into units of 60 and 24 (and 365), our use of base 10 instead of base 2, our division of a circle into 360 degrees, and our reliance on a lump of metal in Paris for mass. Which lump is, incidentally, getting progressively lighter over time.

          We’ve got candles, moles, inductance, reluctance, resistance, remanance, reluctivity(!).

          Quickly now, what’s the time zone difference between Phoenix, AZ and San Diego, CA?

          (And what was it last month?)

          Not to mention that the prime advocates of said system still use quaint and local customary units such as angstrom, electron volt, light year, parsec, and torr. There’s Planck length, Planck mass, Plank time, Planck temperature, Planck charge, and happy hour.

          Quickly now – this vacuum pump pulls to 50 microns. What is that in Pascals?

          There’s Ohm’s law, Young’s modulus, Archimedes principle, Euler’s constant, Avogadro’s number, Peano’s postulates, LeChatelier’s principle, the Kollsman window, Galois theory, Simpson’s inequality, Heisenberg uncertainty, Pascal’s wager, the Goldilocks zone, and Einstein’s famous equation. Darmok and Jalad at Tanagra!

          This all makes *perfect* sense to the beginning student. Feynman once quipped that about 1/3 of all physics is converting between units of measure.

          Be like Feynman. Deal with it.

          1. Surely you’re trolling.

            “Quickly now – this vacuum pump pulls to 50 microns. What is that in Pascals?”

            They don’t measure the same thing. Microns is another word for micrometers, Pascals is Newtons per square meter. Most of the others are either constants, derivative units, convenience units, or have nothing to do with units of measurement or the metric system.

            Euler’s constant for example comes from mathematics.

      1. Checked another video and zoomed in and saw that the Dial Indicator [Jer] used was accurate to +/- 0.001″ instead of 0.005″.
        Either way, I love the design.
        One modification I was thinking of was to use two threaded rods instead of the half nuts.

        1. In fact, I feel a nut embedded in the fence and moved with turn of a know fixed to one of the threaded rod would move the fence by 1/16″ and half a turn of knob by 1/32″ etc. For non-micro adjustment of fence, just disengage nut and move the fence, for which only a disengaging mechanism should be thought of. If something flashes in my head, i would share it here. Meanwhile, someone could think on this way and share any Eureka!

      2. as an apprentice I was taught to file ,to within 1/10th of a thou, the average human hair is about .005 thick I try to work on furniture to 1 /10 th of a thou. no you cannot patent split dies or nuts have been around on lathes cenurys., for patents systems must be unique for purpose

  2. Add an indexed wheel to the threaded rod that allows you to rotate the rod a fixed amount and you’d be able to subdivide the 1/16 inch resolution down to whatever you’d like. A wheel with 4 index positions would allow for 64 (4×16) unique positions for the fence per inch. You could even go crazy and have 8 or 16 positions for extra accuracy.

    1. I conversed with the dude who made it a little bit and got some clarifications – the detents are only there so that you can get to 1/32 or 1/64th without looking at the wheel. You can set the wheel in any other position and it’s stable, and so it’s not hard to dial in a few thou. He’s considering an indicator on the wheel itself (instead of more notches).

      1. Taps in drills work fine. Try it someday.

        To put some math on it: Let’s say that you want your tap to turn at 30 ft/min surface speed (not very fast for HSS cutting cutting any non super high strength steel) and are using a 1/4″ tap. RPM = surface speed * 3.82 / diameter. So 30 * 3.82 / 0.25 = 458 RPM. While you could go faster than that with a drill, it isn’t very hard to keep it down to that.

      1. I tend to agree. I find that I need to work to 0.5 (this is closer to 1). But an earlier comment makes tons (tonnes if you insist) of sense – that sacrificing some resolution is OK if it means gaining good repeatability and accuracy. In other words, I would rather measure in increments of 1 if I know that I’m getting within that to an accuracy of around ±0.1. That would be really impressive and incredibly useful.

      2. You don’t seem to understand the system. It effectively has locking positions every 1/16″ which serve as a common repeatable quick placement. Rotation of the rod allows you to rezero this to the blade so that positional accuracy is the accuracy of the rod and nut mating and should be in the range of .001″ or about .025mm. So if you need to make a cut in the shop that is 3′ 5 1/4″, you could do that almost exactly and quickly with this setup.

        It is effectively the exact same system that my Southbend lathe uses to drive the tool post, just made stationary to hold the fence at an exact distance.

        1. So it’s possible to adjust the fence less than 1/16″ by using the “quick placement” then rotating the rod and mating nut? Repeatability is great, but it’s the one-off cuts for relative dimensioning of tight-fitting pieces where you end up taking several tiny passes.

          1. I conversed with the dude who made it a little bit and got some clarifications – the detents on the wheel are only there so that you can get to 1/32 or 1/64th without even looking at the wheel (so yes, up to 1/64th, repeatably and fast). You can set the wheel in any other position, however, so it’s not hard to dial in a few thou. He’s considering an indicator on the wheel itself (instead of more notches).

      3. all you have to do is add 2 more detents, that will give you 1/64″ add 6 more and you will be at 1/128″.

        Also I would love to see examples of your woodworking that requires greater that 1/32″ precision, and I would love to see you setup that gets you there.

        1. 0.8 mm is still quite a lot, even in wood. Nearly a millimeter gap is huge, and it’s going to show.

          Of course you can design all your stuff in 1/32″ increments and get no gaps, but that’s not always going to work. You can easily get to 0.5 mm precision with a ruler and eyeballs, so one would expect the fence to be adjustable at least to that.

      1. On the contrary. A patent is granted to the person who can prove original authorship. Posting it for everyone to see helps you, because it trumps anyone who later tries to claim the same invention as their own.

        The point of a patent is to publicize the invention in exchange of a limited monopoly to it, to prevent people from inventing things and then refusing to share their findings.

      2. You guys are both wrong, almost every country now uses a first to file system, the US recently moved to it with passing of this law:

        So it doesn’t matter when it was invented it is the first inventor to actually file for a patent that will be first in line to be granted the intellectual property. In Europe i think if you have disclosed your invention in the public domain (hackaday for instance) you are ineligible to apply for a patent at all. In the US this is not the case. I’m not an attorney, but I have read up a little bit on this.

        1. Wrong. First to file, yes, but prior art still trumps your patent application.

          “If an invention has been described in the prior art, a patent on that invention is not valid.”

          In the EU, you need to explicitly make your invention public domain before it becomes unpatentable. Simply showing it around doesn’t make it public domain.

          1. It’s all so confusing. You are right prior art trumps any patent app, but what is prior art and what is not? In the US public disclosure does not constitute prior art until a grace period of 1 year has passed. So if you post your idea on the internet and don’t have at least a provisional application in, I think you will be in trouble if someone else tries to file a patent for the idea. Of course in this case it is a moot point because Incra has already made a jig exactly like the one in the article.

    1. I’m sure if this was commercially available by him, he would be sued by Incra. I know their fence design is different, but the way he is achieving repeatability is a direct copy of their system. I’m sure HE would have to prove his design is significantly different from theirs first to not get sued before applying for a patent.

  3. Digital positioning without a ‘duino, wow!
    Still I’d rather have those screws linked together and turning, and might as well use a stepper and ‘duino.
    Nothing wrong with a hand crank though. Easier to count to do rabbits etc. Home changes when you change a blade though.
    I was taught to get it close then use a piece of scrap to get it on the spec. Then you can repeat success.
    The table saw is a place where the work slides and wiggles and scoots every which way like cattle along a fence.
    Fasten the work down on a grid and move the saw in a machine way, straight. Now that’s progress! The grid can lean against a wall taking up much less space than 8-by-10 feet needed for a table saw for support and incoming and outgoing areas to move a sheet in. And don’t forget the rollers on stands, or jerky helping hands. These kind of saws are so much safer too!
    Building speaker cabinets is one example of the need for near zero slop or deviation. Add mitered joints, even more so.

  4. I built a router table with leadscrews and a dial to position the fence, accurate to 0.001″. I was a beginner woodworker with some experience in machining. I never once used those damn leadscrews. I quickly realized for woodworking, you don’t need much more than feel and sight, and maybe some setup blocks. I used that router table to make a very nice black walnut machinists tool chest with very tight tolerances and never once had to rely on calipers or dials.

    1. Pretty much.

      You take the measurement off the part and cut to the line, and that’s close enough. When two pieces need to be the exact same size, you cut them together.

      It’s only when you make larger pieces, or things that go into assemblies where errors can stack, that you need precise measurements. I’ve done trimmings on wooden doors and window frames, and when you fit them in place you need to make the various pieces quite precisely the right length, just long enough so they’re tight but not too tight or you can’t get them in, and in the right relative size or the corners will jut out. If you’re off by a millimeter then you need to start shimming things with cardboard or sanding the ends down, and that’s tedious.

      1. I mean, you can fit the thing once, but when you’re doing a hundred, there’s just no time to fiddle about. The trim needs to be cut the right size or you’ll be fighting it all the way.

  5. I think that this is an absolutely stellar tool mod, and is definitely something I see as having a lot of general utility in the woodshop…


    While I would not call myself a wood worker, I have done enough to raise a couple of questions. I think that all of this discussion that boils down to resolution, accuracy, and repeatability is dumb. If you are doing all your own work (i.e.: not using pre-cut materials from somewhere else), measured dimensions don’t matter so much. Getting a good, repeatable dimension (whether in inches, millimeters, cubits, or chains; it doesn’t matter) is more important than ‘Is this part 48 1/4″, or 48 3/64″ long?’.

    Also, here are a couple of things that will kill your accuracy even with this setup.
    What happens when you change saw blades? I assume that direct metal-to-metal contact (the saw blade to the spindle flange) will be less than 0.001″, which is better than most people can measure with a tape, but what about the blade kerf? Tolerances are reasonable for most wood work even with the cheapest blades, but do vary.

    Also, what about the dimensional accuracy of your sacrificial fence guard? That can easily kill 1/16″ accuracy, even if it is the same guard. MDF is a common sacrificial guard material, and that stuff can vary considerably with changes in humidity.

    As an aside, MDF is the reason you don’t see dovetails in IKEA furniture. Fiberboard (of any density) will not tolerate any type of torsion, twisting, or wracking that a dovetail would be subjected to. Neither would plywood. Fiberboard and plywood do well with static stresses, but can fail with dynamic stresses. Meaning, your MDF bookcase might survive several years holding up your books, but an MDF toolbox wouldn’t last a week of normal handling.

    1. The fence guard is made of aluminium.

      The leadscrew can be adjusted to zero the fence with a new blade exactly how he zeroes it in the first time. The indicator scale will be offset then, but I’m sure if that becomes a problem he can make a slot and insert a ruler in it.

  6. Are threaded rods really accurate enough over 24+” to do something like this? Repeatability is great but if your table says 24″ but is really 24.25″ you’ve got a problem…

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.