Question about being squared

[Mini Me]

This discussion reinforces my dislike of track saws for spot on accurate work as required by the OP, for carpentry work on site where near enough is good enough they are an ideal tool. I vaguely recall seeing somewhere a tape measure with an end that made measuring diagonals accurately very easy but that only works with a very accurate tape measure and I haven't seen any mention of that being the case in this thread.

 

[woodferret]

Rockler has a stupid square check jig that provides a 1" offset hook point for corners.  But honestly, any stable hook should do it.  I prefer my Talmeter for it because it comes down to a point that I can at least ref place on the corner while dragging to the left/right cant. 

The Rockler won't give an accurate reading because you're not pivoting on the corner, but rather it's a 45deg flat, which makes it hard to get the opposite corner read on the markings (which you're pulling at a slight non-45 angle).

 

[waho6o9]

View attachment 1

The Talmeter is good at measuring diagonals. 

 

[makpacman]

33" should be plenty to align the tracks but sight down them while holding with a slight concave bow to see how they look. I often use a 2 foot level to align and then an 8 foot to verify the cut and it's almost always money.

By holding a concave bow, I assume you mean I should hold one end of the track up while the other end rests on the work surface?  What does that bow do for me and what would I be looking for?

When slightly lifting the track you can see and then line up one of the groves or a ridge from one end to the other, ensuring both edges and center are in alignment. I check this every couple of cuts if not every because small bumps or even dragging the tracks from the end can misalign things.

This holds true for 2700 and 3000 mm tracks also because it's easier than you'd think to pull them into a slight curve inadvertantly. So I'll set the track where I want, lift, check for straightness, then lower trying to set right back on my mark with as little side to side movement as possible. This technique has improved my long rips, which is often the reference edge for a new sheet of plywood. That pays off throughout working with that sheet of ply.

I also have an 8 foot stabila to double check all this every once in a while, especially for very expensive or hard to get ply. The thing is, that's large and cumbersome so at times does not make it to the site.

 

[Mini Me]

The tape measure issue is key to the accuracy problem IMHO. Doing the layout in Fusion or similar is a great idea to obtain the required numbers but then you must have a very good measuring system that matches those numbers and we all know how accurate the majority of tapes are. I don't use the tape tab instead I measure from the 100mm mark on the tape but that can be a bit of a fiddle without a helper and I use a Talmeter or similar with a specified accuracy. In the end do you want to match the mathematical number or simply have the diagonals read the same on the tape? 

 

[Crazyraceguy]

The accuracy of the tape measure only matters if the measurement you are getting matters. In the case of checking diagonals, it does not matter at all. The actual number is irrelevant, is just whether they are the same or not.

That whole Rockler thing of offsetting the hook is just plain stupid.

 

[woodferret]

I should have said repeatable, not accurate.  My bad.

 

[Crazyraceguy]

This discussion reinforces my dislike of track saws for spot on accurate work as required by the OP, for carpentry work on site where near enough is good enough they are an ideal tool. I vaguely recall seeing somewhere a tape measure with an end that made measuring diagonals accurately very easy but that only works with a very accurate tape measure and I haven't seen any mention of that being the case in this thread.

That is unfounded. I routinely cut Corian to seem-able accuracy with my TS55. That was the reason the company bought the Makita unit, so others could do the same. It didn't work out like that though. The track had a bow. If you placed to cut edges together, they would bump bellies in the middle.
The exact same cuts with my TS55 was totally gapless.
Corian is not forgiving, you cannot "clamp it out".
They returned the Makita and bought a TSC55. I'm not really sure why, but it wasn't my money. If you already have batteries, fine, otherwise I don't see it.

 

[sawdust-samurai]

This morning I was able to check the straightness of my guide rail using the edge tracing method.  There was less than a 1mm of divergence along the entire length with the exception of about 125mm on one end.  That end is where the splinter guard tapers back towards the metal.  I've order a new splinter strip to replace the current one to see if that accounts for the divergence. 

 

[woodferret]

Semi-unrelated, you can shift the splinter guard by carefully peeling, offsetting, resticking, and recutting.  That said, it sounds straight-ish (given a pencil test).  IIRC, the saw cams aren't on the rail 125mm so anything there wouldn't be reliable anyways.  Thanks for checking.

 

[Michael Kellough]

This morning I was able to check the straightness of my guide rail using the edge tracing method.  There was less than a 1mm of divergence along the entire length with the exception of about 125mm on one end.  That end is where the splinter guard tapers back towards the metal.  I've order a new splinter strip to replace the current one to see if that accounts for the divergence.

You can’t rely on a worn rubber splinter guard to determine if the aluminum is straight. Mark along the opposite of the rail.

 

[Packard]

The main cause of extrusions warping is that the manufacturer speeds up the process so that the aluminum is still soft when exiting the extrusion die. 

But when you have a wide, basically flat, piece like the track, if the part was going to warp, it would be that it would not lie flat on a flat surface.  So, basically, if your track lies flat on a known flat surface, you can be reasonably certain that the other plane is going to be straight and true also.

If the part were to warp, it would likely be the flatness, not the straightness.

I have two tracks.  Before I posted this, I pushed two pieces together so that aluminum edge pressed against aluminum edge.  Mine were perfectly straight.

Also, speeding up the extrusion process affects dimensions.  So saw to track fit would also show up as being “off”. 

The same issue (but far more noticeable) occurs with plastic extrusions.  Injection molded pieces can exhibit excessive shrinkage when the molder shortens the cycle times.

As stated in Mike’s post above, checking on the splinter guard end is probably not ideal.

 

[sawdust-samurai]

I rechecked (pencil method) both tracks individually on the back sides and they both straight.  I also assembled them with the TSO GRC-12 self-aligning guide rail connectors and after using the 33" WP story stick pro to get them straight with each other, the full setup passed the pencil test.

I guess all that is left is to try and re-cut this top slightly smaller and get it square this time.  Any suggestions on which of my current four sides I should use as the reference edge?

 

[Michael Kellough]

Use the green side as reference edge.
Place the splinter guard/Post-It about a mm in from the existing corner.
Hold the TSO rail square tight to the panel and lift the far end of the rail to make sure the foam isn’t throwing the rail off, then let it down easy. 

It’s best to clamp both ends of the rail but at least clamp the far end.

Lay a scrap of equal thickness stock alongside to keep dust from escaping and make the cut.

Flip the panel over so you can still use the TSO on the green side and do the above.

That will leave you with a panel 1624mm wide with two pretty square corners.

If you dare to make the long sides truly parallel turn the panel upside down and make the black (1194) side the reference edge and again set the splinter guard a little in from the corner. (Actually since you NOT depending on the TSO rail square for this cut you don’t have to flip the stock over but you do have to be more careful to insure that the splinter guard is a good reference for the kerf. Or just commit to cutting off more of the stock)

Before you clamp, take a roughly 4 foot stick, lay it along the reference edge and butt it against the guide rail right in front of the TSO. (You’re not really depending in the TSO for this cut but no need to remove it) Clamp the stick and then clamp a small stop (block of wood) to the overhanging stick while snug against the green edge.

Park the saw on the rail so that it overhangs the beginning of the cut.
This is simply to help keep the rail in place while you move the stick to the far end of the stock.
Gently push the stick along the magenta edge until the stop is snug to the green edge and clamp it.
Now pull the guide rail snug to the end of the stick and clamp the guide rail.

Lay some dust barrier stock alongside the edge and make the cut.

Now you should have a pretty square rectangular panel, just a little bit smaller than the trapezoid was.

 

[sawdust-samurai]

I cut one short side using the green edge as the reference.  When I flip the sheet and held the TSO square guide tight against the green edge, the cut line of the track is 3mm in (to the right) at the far end.  I've checked this on two separate pencil lines, both 122mm in from the edges.  Granted, this pencil check is tracing the splinter guard.

Edit: The 2nd track shows off by 2mm in the same direction.

That makes me think there is something wrong with the rail or the TSO square.

I've only made the additional cut as of now.

 

[sawdust-samurai]

Lol.  I just moved over to the cyan side with the same track setup.  I was expecting it to be 2-3mm off at the far end again. Nope, it was still off 2mm on the near side (same side as when checked from the opposite side).

I have no clue what is going on.

 

[Michael Kellough]

Assuming you've been scrupulous about making sure there is no debris between the square's fence and the mdc panel.

First figure out how to confirm that the cut edge you're holding the rail square against is actually dead straight.

If it is, figure out if the rail square is actually square to the rail.

Draw a large 90* triangle on the mdf panel. Use the simple 3-4-5 formula.
Short leg 3, long leg 4, hypotenuse 5. Make it 30 inches x 40 inches. Short leg and hypotenuse to the left.

Make a trammel if you need to because the triangle must be made precisely.
Mark with an .5mm (or .3) mechanical pencil.
Tape the rule (can be a tape measure) to a stable surface so you can concentrate on setting the trammel precisely.

Move the rail square and rail over to the long side of the drawn triangle and see if the aluminum lies exactly along the line. Forget about the rubber strip.

If the rail doesn't match the line then move the rail square a few inches (to another spot on the rail) and test again. Sometimes there is a bad spot on the rail and just moving the square makes the difference.

If you had an MFT I'd say do the 4/5 cut process to make a piece of plywood with a very accurate 90* corner and use that for testing. It's good to have a big accurate square. If you can't make one consider buying the big Woodpeckers square or the large TSO triangle. There was a place in California that made and sold large 90* triangles from phenolic coated plywood for a reasonable price. Don't know who they are or if they're still doing it.

 

[Mini Me]

The accuracy of the tape measure only matters if the measurement you are getting matters. In the case of checking diagonals, it does not matter at all. The actual number is irrelevant, is just whether they are the same or not.

That whole Rockler thing of offsetting the hook is just plain stupid.

I totally agree so theoretical numbers are not a good guide and using a tape measure to try and match those numbers is mission impossible.

 

[sawdust-samurai]

Assuming you've been scrupulous about making sure there is no debris between the square's fence and the mdc panel.

First figure out how to confirm that the cut edge you're holding the rail square against is actually dead straight.

If it is, figure out if the rail square is actually square to the rail.

Draw a large 90* triangle on the mdf panel. Use the simple 3-4-5 formula.
Short leg 3, long leg 4, hypotenuse 5. Make it 30 inches x 40 inches. Short leg and hypotenuse to the left.

Make a trammel if you need to because the triangle must be made precisely.
Mark with an .5mm (or .3) mechanical pencil.
Tape the rule (can be a tape measure) to a stable surface so you can concentrate on setting the trammel precisely.

Move the rail square and rail over to the long side of the drawn triangle and see if the aluminum lies exactly along the line. Forget about the rubber strip.

If the rail doesn't match the line then move the rail square a few inches (to another spot on the rail) and test again. Sometimes there is a bad spot on the rail and just moving the square makes the difference.

If you had an MFT I'd say do the 4/5 cut process to make a piece of plywood with a very accurate 90* corner and use that for testing. It's good to have a big accurate square. If you can't make one consider buying the big Woodpeckers square or the large TSO triangle. There was a place in California that made and sold large 90* triangles from phenolic coated plywood for a reasonable price. Don't know who they are or if they're still doing it.

Thanks for the additional steps to checkout.  Last night I ordered the 18" TSO triangle and since this bench is for my CNC machine, I went ahead and continued the build.  I got the top (a single layer) and BB plywood edge banding applied over the weekend.  Hopefully tonight I will be able to get the black laminate installed.  Once the CNC is setup and running, I'll use it to make an MTF style bench and a large square to help troubleshoot this further.  I'm not giving up on figuring this issue out, I just need to get the CNC running to make sure everything is working with it before my return/support window runs out.

I'll post more as I sort this out.