guide rail reference edge

[OzarkNerd]

I watched a video from Sedge where he talks about connecting two rails together and he specifically says putting a straightedge on the outside of the rails is wrong.  He says the inner edge is the reference edge as that's what the saw rides on (makes sense).  So how then, do these various (TSO, woodpeckers, etc) rail squares guarantee squareness if they are butting against the non-reference edge?

Is it simply the case that the offsets between the inner and outer edges are not guaranteed to be the same across different rails but in all cases those edges are guaranteed to be parallel?

 

[Dr. P. Venkman]

The inner edge is the splinter guard strip and is made by your saw blade.  If your saw is calibrated to ride properly in the track, it'll be parallel to the track on which your saw is moving (give or take a tiny amount of slop in the saw/track).

I don't believe the outer edge is "guaranteed" to be anything.  I assume it tends to be pretty much parallel to the track in which the saw rides (and thus the splinter guard) because there are any number of cutting setups using dogs, etc. that reference the outer edge and do just fine. But the track on which the saw rides and the splinter guard edge are what have a relationship to one another, so far as I know.

I don't really know what is guaranteed or what tolerances the outer edge is subject to.  But it sounds like Sedge and others suggest you shouldn't count on it being the same distance from the splinter guard on different rails (maybe it is always parallel).

 

[OzarkNerd]

Thanks for the response and apologies for my unclear language.  When I said "inner edge" I meant the side of the hump in the center (ish) of the rail that the saw rides on and tightens against with the knobs ("2" in my image below).  This is what dictates the actual line the saw travels along.  If this inner surface is not guaranteed to be parallel to the outer rail surface ("1" in my image), then I'm not sure how any of these rail squares can guarantee a square cut since they are referencing off of 1 and not 2.

[attachimg=1]

 

[jeffinsgf]

I'm always surprised at these discussions. My guide rails are really old. So old they only have one rib and no green stripes on the top. When I join them, it is simply a matter of butting the ends together and locking them down. The ends are perfectly square to each other and I have never had anything but great results from joining as many as three sections of guide rail.

 

[Dr. P. Venkman]

Gotcha - my mistake.

As you mention, there are a bunch of products and methods that rely on that outer edge being parallel to the cut line, so I would imagine that's reasonable to expect/count on.

My best guess: It could be an issue of distance, rather than parallel, but I'll bet it's just a matter of good practice. He has you using the primary interface with the saw as the reference edge for joining the rails (even if the outer edge should theoretically be sufficient, why not go ahead and use the surface the saw actually contacts?).

Beyond that, I don't know.  I've always taken for granted that the outer edge is parallel enough to the cut line, and the terrific results I get cutting with the outer edge of the rail up against dogs support that.

 

[Packard]

TSO's system assumes that all the shapes on the extrusion run parallel.  I think that is a reasonable assumption.  They guide off the area of the extrusion where the connectors fit in.  They specifically state that a spacer be placed between the butt ends of the two adjoining tracks.  They used a 3" x 5" index card for that purpose (0.007" thick).

This video shows how it works.

 

[OzarkNerd]

I'm always surprised at these discussions. My guide rails are really old. So old they only have one rib and no green stripes on the top. When I join them, it is simply a matter of butting the ends together and locking them down. The ends are perfectly square to each other and I have never had anything but great results from joining as many as three sections of guide rail.

I would suggest you are getting lucky and that your results cannot be assumed as a given.  I say this because Festool themselves indicate the rail ends are not guaranteed to be square and further that if you are joining 2 sections of rail you should actually leave a gap between them.  Butting them directly can pull them out of alignment due to the lack of end squareness guarantee.  Honestly, I'm not super impressed with any of that given the prices of these things, but knowing all of these little tidbits does increase the likelihood of success and repeatability. 

At the same time I've read many posts about rails bought at different times having slightly different sizes so again, if you are connecting 2 rails that aren't identical in width, the exact surface used for aligning them will produce different results.  This is where the TSO rail connectors are apparently not the best choice since they will force a single repeatable alignment each time, even though that may not be the alignment you need.

 

[jeffinsgf]

That would be about 21 years of "lucky". I think I'll keep rolling the dice instead of changing what I'm doing.

 

[OzarkNerd]

That would be about 21 years of "lucky". I think I'll keep rolling the dice instead of changing what I'm doing.

Unless you have the T-1000 liquid metal rails yours are either square or they aren't, right?  You got "lucky" once - the day you bought them and they happened to be square.  It's not something that would change cut by cut unless you have the saw rotated 90 degrees.  But your results also have no bearing on what anybody else would experience given Festool says they aren't guaranteed to be square.  Hence your procedure is considered a "common mistake"...
=98

 

[rubber_ducky]

Just because the ends aren’t guaranteed to be square doesn’t mean that they aren’t. Or even that one is lucky to get a unit with square ends. It’s more likely that because they don’t need to be square (to make a straight cut), festool won’t replace a unit that isn’t.

Sent from my iPhone using Tapatalk

 

[mino]

This is bit more complicated as one needs to understand how the physics of extrusion production work as that is what defines what will be accurate and where will be a variance. This is a universal comment, irrespective of the rail make.

Some basics:
- it is extremely difficult /read: impossible/ to ensure the "size" (think width and height, not length) is consistent BETWEEN manufacturing runs
- it is easy to ensure ONE surface of your extrusion has almost no bends and so is mostly straight, class 1 engineer's straight edge straight, but the other surfaces will vary based on the varying width/height of the extrusion in relation to that "reference" surface
- it is relatively easy - within a single extrusion production - to ensure the width WITHIN this extrusion is pretty consistent, this is because the biggest problem is not that specific piece, but the variance in the aluminum material composition as it comes from the mill

What this translates to in laymen's terms:
- manufacturer can choose a surface which he will setup the extruder to produce as straight as possible /at the expense of the other surfaces/
- manufacturer can STILL ensure that surfaces parallel with the reference surface will be "very much parallel" and very close to the straightness of the reference surface, just even so slightly worse than the reference surface precision
- manufacturer CANNOT ensure all produced rail - even those produces on the same machine during the same day - will be the same width /I have observed close to 0.2 mm (0.01") difference between new rails inner and outer rib "1" and "2" surfaces distance/

What this translates to in practical use:
- for max precision when joining, always use the reference surface
- using the "far/outer" surface for squaring is very much fine /hence GRS practical accuracy/
- using the "far/outer" surface for distance/width reference is NOT gonna produce consistent results when using joined rails as those can have slightly different widths
- using the "ribs" relative position to each other for aligning the rails will produce hit/miss results depending how lucky one was in the "widths lottery" when buying his rails

Hope helps.

P.S.:
When I say this post is universal, I mean that the relative comparisons - where is accuracy and where is inaccuracy potential - will be so with any extrusion /rail/ maker. One would have to go with a milled rail to change this. There are and will be differences in how precise each maker makes his rails.
But what is the same is that the biggest "issue" /besides alignment/ when joining extruded rails is always the varying width. And the most important property indicating "quality/accuracy" of a rail is always the "saw-guiding" rib straightness as that is what guides the saw. Heh.
The rib-outer to rib-outer width within a rail is something worth checking with a caliper when buying, but will generally be in the "good enough" territory so squares using the outer edge are fine. Actually best possible, as that is the straightest "accessible" surface given the reference rib has a saw riding on it ...

 

[Steve1]

It's true that it is good practice to align to the same edge as the one that the saw rides on.

But we may be overthinking this.  Its all one piece of extruded aluminum that comes out of one die.  Its not like one edge gets precision ground after extrusion.  I would expect the parallelism error between surfaces is measured in microns.

 

[mino]

It's true that it is good practice to align to the same edge as the one that the saw rides on.

But we may be overthinking this.  Its all one piece of extruded aluminum that comes out of one die.  Its not like one edge gets precision ground after extrusion.  I would expect the parallelism error between surfaces is measured in microns.

Correct, tens of microns though.

The "bigger" differences are seen only between different rails unless one hits a bad rail, so can affect the joining and PGS calibration but not angle accessories like squares etc.

 

[Packard]

A small ding on the end of the rail will also be able to cause a problem.  Aluminum is fluid.  That is, if you make a dent in one area, the material has to flow to another area. 

Even if you are careful, it is conceivable that the chopped end of the extrusion can get dingged. 

Also, note that there are two ways to cut extrusions to length. 

The primary way is called a "flying cutoff", and the extrusion continues to travel through the extrusion process as it is sawn to length. 

These are never quite precise. 

It is not clear that Festool is machining those cut edges.  Typically extrusions are made (in the USA) in 12 foot or 20 foot lengths.  So the rails might have only one machined end, and the other end might be a factory cutoff.  You would not be able to see the difference, though sometimes if the cutoff is not setup correctly the cut will be at a slight angle.

 

[mino]

Per my understanding the Festool rails are made in-house on custom tooling. Not by general extrusion shops.

So other than being "extrusions" one cannot infer how exactly they process, post process or what lengths are the "initial" result from the industry defaults. For one I would certainly expect Festool to use the same German-made metric tooling like they use on their German line and not a local US make.

 

[Packard]

It would be highly unusual for a manufacturer to have its own aluminum extrusion line.  It's been many years since I visited an extrusion company, but the one I visited probably needed 50,000 square feet of space.  It ran 24 hours per day.  The line needed its own smelter and it is not cost-effective to let it cool off and re-heat it each day. 

At best, Festool's aluminum extrusion requirements would be "modest" or likely "small" by industry standards.  Aluminum window and door manufacturers would probably be larger.  And they typically buy extrusions, and don't manufacture them.

I am seeing from this video that the cut to length lines are done differently than I had seen it, and these cuts are probably more accurate.  But not the stuff of a machine shop.

The 8 or 10 inch billet rods are the raw material that gets fed into the smelter.

 

[mino]

Yeah, I did not mean Festool "made one" but that they would have a tested/validated solution where the tool will be designed as optimized for their use case which is quite specific - they have only one high-volume product in the FS/2 rails to begin.

When you have such a requirement - a single specific size product with a specific /above standard/ accuracy requirement, it is common you will have a customized tool made for your use case, instead of buying a "general" machine which would be versatile.

This means one cannot infer from "generalist/flexible" extruders capabilities what are the parameters of those Festool uses for FS/2.

When making a specific product machine like I believe they use, one has a LOT of design dials that can be tuned where you trade machine flexibility /able to make different extrusions/ for improving specific parameters for that one mass product you are using it for.

E.g. I have measured across about 5 rails sample, all bought independently, that the "bending" of the reference surface is about 2x worse on the Makita rails AND that the Makita rails straightness is similar to the accuracy of general extrusions I get to buy from my Aluminum supplier.

Festool having such a significantly higher accuracy than is "normal" for this type of a product is a clear indication to me they must have their production optimized in some way beyond what a "standard" extruder provides. I have also checked and they do not have a milling step, they seem to get their accuracy directly out of the extrusion process.

 

[TSO_Products]

Good morning everyone -
this is an interesting thread which deserves comment and clarification from TSO. We will prepare a thoughtful response and post it to this thread.
Stay tuned . . .

Hans

 

[fshanno]

Seems to me that a good tool for aligning rails would be a straight edge with a daddo in it that would fit snug over the guiding channels on the rails.  It would be parallel and straight, assuming the guide channels on the rails are the same width.  It would have to be wide enough to be very stiff.  I would say 8 inches or so.  And it would have to be pretty long.  3' maybe.

Could such a device be made of wood or plastic or some other material that a wood worker could mill?  UHMW plastic maybe?  The dado on the straight edge would be as true as the fences on our table saws or router tables could make it.  And we count on that straightness all the time for many many things.

 

[thender]

I think there are more than one truecway to skin this cat.

The only thing that matters is if you are getting straight cuts using the rail alignment method of choice. 

If the cuts you are getting are straight — your aligning methods are fine.  If not….  Consider other techniques.

The alignment is just a means toward this end.

Any method that gets straight cuts from the joined rails is a good method.  Use whatever method that works for you.