Question about being squared

[sawdust-samurai]

Okay, I re-measured everything this evening.  I switched to my metric tape as it is in better shape and has more resolution in the markings (every 1mm as opposed to every 1/16").  Here is what I get for each measurement being as accurate as I can:

Thoughts?

 

[bobtskutter]

Right...I'll put my engineers hat on
Using this tool:https://www.calculator.net/triangle-calculator.html
I drew out 4 triangles starting in the top left and going clockwise.
Green-red-pink = 90'
black-blue-green=90.1'
LightBlue-red-black=89.85'
pink-blue-lightblue=89.9'

Did you use the green edge as the reference edge for the TSO square when you cut the pink and black lines?
It looks to me like it's the length of the cut that's catching you out.  i.e. 1 corner is 90', the other 3 are ALMOST 90'.

There are other combinations of triangles to look at, e.g.
pink-lightblue-red
pink-green-blue
green-black-red
black-lightblue-blue
I don't have time to work out that angles right now (I need to do my day job)

Regards
Bob

 

[sawdust-samurai]

Right...I'll put my engineers hat on
Using this tool:
https://www.calculator.net/triangle-calculator.html
I drew out 4 triangles starting in the top left and going clockwise.
Green-red-pink = 90'
black-blue-green=90.1'
LightBlue-red-black=89.85'
pink-blue-lightblue=89.9'

Did you use the green edge as the reference edge for the TSO square when you cut the pink and black lines?
It looks to me like it's the length of the cut that's catching you out.  i.e. 1 corner is 90', the other 3 are ALMOST 90'.

There are other combinations of triangles to look at, e.g.
pink-lightblue-red
pink-green-blue
green-black-red
black-lightblue-blue
I don't have time to work out that angles right now (I need to do my day job)

Regards
Bob

Thanks Bob.  I'll run the other triangles just to see what they show. 

As to my actual cut order, it was Cyan/Lt Blue followed by Green.  I then cut both pink and black with the TSO square referenced off the Cyan line.

 

[Crazyraceguy]

As to my actual cut order, it was Cyan/Lt Blue followed by Green.  I then cut both pink and black with the TSO square referenced off the Cyan line.

Ok, this may be part of it. It took me a second to wrap my head around, since my track square only cuts one direction, that was my thinking. However, since you have a TSO, it could be done that way.
Plus, we are still not sure of the straightness of this rail? Moving/repositioning the square on the rail and cutting in opposite directions, introduces a lot of variables.

Assuming your rail and square can make a 90 degree cut, there is no reason that you couldn't reference the opposite sides, since they are "known" to be parallel.
Meaning, if the cyan/black cut is square, you take the same track/rail square around and cut the green/purple. 
At that point, the worst you can have is a parallelogram, every side will be equal, only the diagonals will be off. Then you know where you are....you are not cutting square. Easier to diagnose.

 

[waho6o9]

a + B = c  squared

1195 + 1630 = 2019 squared does not compute

If I'm reading the numbers correctly.

 

[bobtskutter]

a + B = c  squared

1195 + 1630 = 2019 squared does not compute

If I'm reading the numbers correctly.

You're thinking about Pythagoras Theorem, which only applies to right angle triangles.  You can't use Pythagoras maths on this problem because the triangles aren't right angle triangles.  That's the bit that messes with peoples heads.  Have a look at https://www.calculator.net/triangle-calculator.html it explains the maths involved.

However, if it was a right angle triangle the diagonal (hypotenuse) would be 2021mm vs the 2019mm measured.  2mm is a 0.1% error in the total distance, i.e. very small.

Regards
Bob

 

[sawdust-samurai]

As to my actual cut order, it was Cyan/Lt Blue followed by Green.  I then cut both pink and black with the TSO square referenced off the Cyan line.

Ok, this may be part of it. It took me a second to wrap my head around, since my track square only cuts one direction, that was my thinking. However, since you have a TSO, it could be done that way.
Plus, we are still not sure of the straightness of this rail? Moving/repositioning the square on the rail and cutting in opposite directions, introduces a lot of variables.

Assuming your rail and square can make a 90 degree cut, there is no reason that you couldn't reference the opposite sides, since they are "known" to be parallel.
Meaning, if the cyan/black cut is square, you take the same track/rail square around and cut the green/purple. 
At that point, the worst you can have is a parallelogram, every side will be equal, only the diagonals will be off. Then you know where you are....you are not cutting square. Easier to diagnose.

I think my method is still a little unclear because I've not give enough info to you guys.  I have the TSO GRS-16 and it only cuts from one side of the track.  I did not remove the square from the track between cutting the black line and the pink line.  I cut the black line first, using the TSO square to make that cut square to the cyan line.  I then measured from the black edge down the length of the MDF sheet and made a mark for where the pink cut should be made.  I slid my track with square still attached down so that that the track's splinter guard lined up on the pink line mark.  The track and square was resting on/against the remaining portion of the MDF sheet (I'm only using about 5.5' of the 8' sheet).  So I made both the black and pink cuts from the same direction and with the exact same square to track placement.

TSO does make a track square that will allow you to cut from either direction, it is the GRS-16 PE.  Had I realized this at the time, I would have ordered it instead.  [sad]

I still have not had a chance to double check straightness of my track.  Should I be checking that on the backside (where it is thicker) or along the splinter guard?  If it is the latter, that will be kind of hard as the splinter guard is not the same width down the length of the track.  It is thinner towards the edges of the track then it is in the middle.

 

[woodferret]

Okay, I re-measured everything this evening.  I switched to my metric tape as it is in better shape and has more resolution in the markings (every 1mm as opposed to every 1/16").  Here is what I get for each measurement being as accurate as I can:

Thoughts?

Thrown into Fusion 360 since it has a nice constraint / layout engine...

[attachimg=1]

A couple possibilities

A) You weren't fully seated on the TSO when you cut the left side.  You were off by 0.2deg.  For this to hold though, your measuring skills on diagonals leave a lot to be desired.

B) There's a kink in your rail at the far end, where your GRS-16PE is placed away from you.  When cutting ref cyan line for the L edge, you got 89.8deg.  Still doesn't explain the divergence in diagonals.

C) Your miswrote the diagonals, with the shorter 2018 actually being the blue line, and the 2019 is just a misread.  Still A,B holds for the weird left side. 

edit: I see you don't have the PE.  Then it's A and a combo of C or the rail not being straight over the 1600+ mm.

edit2: I think to get better insight into whether the rail is straight, I know you don't have a 1600mm+ trusted straightedge, but it doesn't matter.  Just put it on MDF, draw a pencil line on the splinter guard.  Flip the rail 180deg and see if that line still matches up.

 

[Michael Kellough]

I drew it out in cad too.

Slightly different results.

I found the diagonals a little too short.
Given how easily MDF corners crumble it isn't unexpected.

Adjusting the trapezoid to get the diagonals equally short of the corners (~.7mm) I get this,

[attachimg=2]

Since both TSO aided cross cuts were referenced from the same baseline (cyan) I would expect the resulting
angles to be nearly identical instead of having the opposite bias. Unless there was debris under the TSO on the first cut maybe the rail is curved.

The rail square being only on the left side of the cut line can't push the rail to the right so the first second crosscut is the one that is problematic to me.

 

[woodferret]

Ah shoot.  Trust Michael's result more.  I see a right angle constraint that snuck into mine.

 

[sawdust-samurai]

Okay, I re-measured everything this evening.  I switched to my metric tape as it is in better shape and has more resolution in the markings (every 1mm as opposed to every 1/16").  Here is what I get for each measurement being as accurate as I can:

Thoughts?

Thrown into Fusion 360 since it has a nice constraint / layout engine...

A couple possibilities

A) You weren't fully seated on the TSO when you cut the left side.  You were off by 0.2deg.  For this to hold though, your measuring skills on diagonals leave a lot to be desired.

B) There's a kink in your rail at the far end, where your GRS-16PE is placed away from you.  When cutting ref cyan line for the L edge, you got 89.8deg.  Still doesn't explain the divergence in diagonals.

C) Your miswrote the diagonals, with the shorter 2018 actually being the blue line, and the 2019 is just a misread.  Still A,B holds for the weird left side. 

edit: I see you don't have the PE.  Then it's A and a combo of C or the rail not being straight over the 1600+ mm.

edit2: I think to get better insight into whether the rail is straight, I know you don't have a 1600mm+ trusted straightedge, but it doesn't matter.  Just put it on MDF, draw a pencil line on the splinter guard.  Flip the rail 180deg and see if that line still matches up.

Thanks for the in-depth analysis and good thinking about throwing it into Fusion 360 (I'm still learning that program). 

Both A and C points are possible.  I'll try remeasuring the diagonals again and see what I get.  I will also try  your method of checking the rail for straightness.  That should be a quick and easy way to check it.

I drew it out in cad too.

Slightly different results.

I found the diagonals a little too short.
Given how easily MDF corners crumble it isn't unexpected.

Adjusting the trapezoid to get the diagonals equally short of the corners (~.7mm) I get this,

Since both TSO aided cross cuts were referenced from the same baseline (cyan) I would expect the resulting
angles to be nearly identical instead of having the opposite bias. Unless there was debris under the TSO on the first cut maybe the rail is curved.

The rail square being only on the left side of the cut line can't push the rail to the right so the first second crosscut is the one that is problematic to me.

Thanks for your analysis as well!

 

[Michael Kellough]

A cut of less than 90* on crosscut one is what I would expect but you can do better than .1+ degrees off.

Hold the rail up off the work (the minimum amount) so you can slide the TSO/rail to the mark and then lower the rest of the rail. Nearly four feet of foam can throttle rail off. The mark should be close to the beginning of the cut but if you read ant to be precise use geometry and make marks for both ends of the rail. Four feet is a little too much for the rail square.

You measured from the first crosscut.

To do that I would make marks on the work along the back of the guide rail at both ends. Then make marks for the new rail position, adding the kerf width. The splinter guard strip is usually unreliable if it’s been in service a while.

The PE versify the rail square would allow you to cut from the other side (the green side) while still referencing the cyan edge. But to be successful you should have confirmed that you could get square cuts with the rail square at that position on the rail. Even Festool rails can be problematic and the longer the cut…

 

[sawdust-samurai]

The rail square being only on the left side of the cut line can't push the rail to the right so the first second crosscut is the one that is problematic to me.

I just saw that you edited our post to call out the second crosscut (pink).  I also think this is where my issue was induced.  I was making these cuts on my current 4x8 workbench which is also an outfeed table for my table saw.... on the pink side.  So the other cuts I could easily walk the saw down the track.  For the pink cut, I had to lean across the full 4' width of the table.  I could see that effecting my cut.

 

[Crazyraceguy]

Since both TSO aided cross cuts were referenced from the same baseline (cyan) I would expect the resulting angles to be nearly identical instead of having the opposite bias. Unless there was debris under the TSO on the first cut maybe the rail is curved.

The rail square being only on the left side of the cut line can't push the rail to the right so the first second crosscut is the one that is problematic to me.

Agreed on the first part, that's what I would expect too.

The second part however is not so definite. If there is more than the slightest amount of toe to the saw, it can indeed pull. "How much" depends on the amount of the mis-adjustment. The more the blade is exposed (depth of cut) the more this is exaggerated. I very small amount can pull the blade, cutting further into the splinter strip. More can actually move the track.

I don't recall the foam grip strips being as vigorous as Festool? Maybe another factor?
Either way, this kind of seems like a technique issue.....or the track is not straight?

 

[waho6o9]

Thank you  bobtskutter

 

[makpacman]

Psst.... your rails aren't straight.

That could very well be the case.  I do however always check them with a Woodpeckers 33" straight edge after putting them together.  Perhaps I need to check them with a longer straight edge?

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.

I use Sedges trick and keep a biz card gap between the tracks during the alignment process or else the straight edge doesn't really have proper affect because the mating surfaces of the tracks (and any imperfections they contain) become the determining factor for straightness. Also, the tiny gap actually helps the saw make the transition from one track to another smoothly.

It's been insinuated at times in this thread that Makita tracks aren't straight but that's not been my findings at all. I have several of them, as well as various types and sizes of festool tracks and there is no dicernable difference in cuts made with any of them or even the generic no name tracks I have.

The profiles are ever so slightly different so sticking to one brand is advisable, but as long as your saw is adjusted to fit them and the tracks are not defective, cut quality is equal. Same can be said about using Bosch tracks in place of Mafell.

If you're breaking down full sheets regularly then the longer TSO parallel guides may be helpful but really there's no substitute for plain good craftsmanship. You're on the right track trying to figure out how and why this happened as opposed to making a purchase believing it will remedy the situation.

 

[sawdust-samurai]

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?

I use Sedges trick and keep a biz card gap between the tracks during the alignment process or else the straight edge doesn't really have proper affect because the mating surfaces of the tracks (and any imperfections they contain) become the determining factor for straightness. Also, the tiny gap actually helps the saw make the transition from one track to another smoothly.

I saw that tip someplace else and have been using it for a while now.

 

[woodferret]

It's been insinuated at times in this thread that Makita tracks aren't straight but that's not been my findings at all. I have several of them, as well as various types and sizes of festool tracks and there is no dicernable difference in cuts made with any of them or even the generic no name tracks I have.

I've had Festool tracks that were bowed.... at 800mm no less.  So it's not just a Makita thing, but it's a warranted (low-effort) first step anyone should verify if they're not getting square, no?

It's the basic unplug/replug power cord thing we do for computers.  It's amazing how much people fight this.

 

[Michael Kellough]

I’ve never used Makita tracks. Just echoing what I’ve read. That’s what the internet is really good at…

 

[Allano]

The diagonals being almost equal is easy to explain:

In a symmetric trapezoid, the diagonals are equal, even though the top and bottom edges have different length.

I am not going to try to explain all corners being square, though.