How Square Does "Square" Have To Be???

[onocoffee]

I've been making a crosscut sled for my little SawStop CTS and following the guidance of YouTubers like Ron Paulk (for the basic design of a small sled for the CTS), Greg Paolini from his SawStop tutorial on Making A Crosscut Sled For Your SawStop, and The 5-Cut Method Made Easy from Bike City Woodworks.

Overall, I think it's gone well but how square do I really need to be?

Using the 5 Cut Method and the Bike City Calculator, I've gone from being .45mm off to .41mm (both were with Imperial to Metric conversions), now after measuring in Metric, the correction amount is 0.79161227mm pivoting towards the blade.

The fifth cut measures at the Top 9.92mm, Bottom 9.15mm, Length 91.92mm. The Pivot from the screw is 378mm.

I originally measured it in Imperial: Top .389", Bottom .365", Length 3.62", Pivot 15"

On the last adjustment, it said to move the fence by .45mm, so I used a business card at .41mm to adjust the fence. I don't know if I'm doing it wrong, measuring wrong or what, but as I'm writing it, a nearly 1mm variance seems like a lot.

It's starting to drive me a bit mad. Thoughts? Thanks.

BTW, for measurements, I'm using my old Mitutoyo CD6B digital calipers from my machining days with the wood having slight measurement variances depending where I put the calipers on the same part of the piece.

 

[smorgasbord]

First, you're cutting a pretty small piece - less than 4" square. Ideally you should be cutting something at least a foot square.

The amount to move seems large because the distance from the pivot is much larger (4X) than the thing you're cutting. So, since the thing you're cutting is 0.77mm off, you divide that by 4 (because of the 4 cuts) but then have to multiply it by almost 4 due to the distance, so it's about what you measured.

So, use a much large piece of stock.
Also, check your fence to be sure it's really flat. A back fence that's not flat will give different results for different sized workpieces.

 

[onocoffee]

[member=77266]smorgasbord[/member] - Thanks. I was cutting smaller pieces because that's what I remember seeing on YouTube (or I could be wrong). The sled has a max depth shy of 14" but I did find a piece of 12mm ply that I cut to 12" square for the next 5 Cut - and I can see how off it is!

Top: 16.53mm
Bottom: 14.35mm
Length: 291mm
Pivot: 378

The Bike City calculator tells me I need to pivot the fence towards the blade by .70793814mm

Now, I'm gonna give it a try.

 

[onocoffee]

Alright, that last cut said I was off by .707mm and I had to pivot the fence forward. I did check the flatness of the fence by putting a straightedge to it with a light behind and there was no light leak along the edge, so i think we're good there.

I shimmed the fence with two business cards that measured .71mm and then redid the 5Cut for these results:
Top: 18.64mm
Bottom: 19.07mm
Length: 277mm
Pivot: 342mm

Calculator says that I need to adjust backwards now by 0.13272563mm

At what point are you just chasing your tail and driving yourself crazy?

The calculator says that the fence is square when the results are:

 

[smorgasbord]

At what point are you just chasing your tail and driving yourself crazy?

For me, as long as each iteration actually gets me closer, I keep going. As soon as the results go backwards I figure something else isn't quite right (fence not flat, slot in miter bar/slot/sliding mechansim, etc.) and either try to go back or just stop.

Note that I do cut some different size samples to be sure my results will work universally.

 

[MikeGE]

One of the best videos I've watched that explains and demonstrates the 5-cut method is William Ng's video from 2012.


The beginning of the video explains the 5-cut process, and then William builds a crosscut sled.  At about 28:13, he performs the five-cut method.

 

[Lincoln]

Now that you're close, set a stop on your sled and cut two identical panels, making sure to mark a front edge. Then, put them face-to-face, with the front edges aligned - it's a good 'real world' way of checking how square you are.

 

[smorgasbord]

Note that the 5-cut method is more sensitive than almost any other method of checking square, sort of 2-cut, flipping, checking against a known square, measuring diagonals, etc. should show perfect results if the 5-cut was done properly.

On the 5-cut method, be sure you're putting the previously cut edge against the fence. There is a YouTube video of someone rotating the piece the other way and struggling, lol.

 

[guybo]

I bought one of these made by a long time fogger  [smile]

 

[ChuckM]

When I built my crosscut sled for my SawStop PCS, I used the same method that I use for checking the squareness of my projects (cabinets, drawers, etc.): measuring and checking against the diagonals. I figured that the diagonal method was accurate enough since I never used a pair of calipers when I checked my projects for square.

So, does my current cross-cut sled give me a tolerance of 0.004 or 0.007 or something else? I wouldn't know (or care). I just know that I've been very happy with how my projects came out using the cross-cut sled that was tested with the diagonal method and some scraps.

As an exercise, I did use the multipl- cut method when I checked and adjusted the fence for the new Kapex after it arrived (as I didn't really quite trust the QC for any mass-produced miter saws).

 

[waho6o9]

if the diagonals are the same we're good to go. 

Big respect to those that get it down to the .0000s

All the Best

 

[woodferret]

It only needs to be as square as your edgebanding   But seriously, what do you want to build?  Cabinets?  With what?  Because ply is going to be bendy and your 'precision' is less important than your design around it and experience at fixing stupid material/design issues.

For me, that just means I'm happy with < 1mm over 36".  Most weirdness is absorbed by the edgeband treatments and the backside.  It also lets me use the rails that are ever so slightly out of line without too much worry.

addendum: cut angles do matter greatly though as you're very likely going to notice that more.  Human vision thing and all that.

 

[Stone Message]

Hi Everyone,

The 5 cut test is anexcellent way to check a machine or process for square. I created a new tracksaw cutting station for my workshop using the Parf Guide System (you can ask Ron Paulk about this) and I then tested it to see how accurate the array of 20 mm holes were.

I made a video, some of which is shot in one take, that shows the 5 cut test in detail and also how to calcuate the angular error in degrees:


Peter

 

[mino]

As is usual in life, it depends.

Case in point:
Using a TSO GRS 16 (FS/WA 90) for a squared cut I can reliably get within about 0.02" (0.5 mm) error per meter of cut length. That error is mostly from the GRS 16 contact with the stock not being perfect, stock not being perfect etc., not the tooling itself.

Good (enough) for furniture-grade 4' cuts? Nope.

Good for 1' cuts? Absolutely!

---
Moral of the story:

In my experience it is not about how precise/accurate a tool/method is.

It about understanding / knowing how precise/accurate a method/tool is. And using it appropriately.

 

[bobtskutter]

Check for dust on the inside of the fence, it doesn't take much to prevent the work piece sitting flush.  Check the videos carefully, the formula depends on which side of the blade you hold the work piece and which side the pivot is on.  When you make a cut you need to rotate the work piece so the fresh cut is against the fence.  I've been caught out by all of the above.
Bob

 

[onocoffee]

Check for dust on the inside of the fence, it doesn't take much to prevent the work piece sitting flush.  Check the videos carefully, the formula depends on which side of the blade you hold the work piece and which side the pivot is on.  When you make a cut you need to rotate the work piece so the fresh cut is against the fence.  I've been caught out by all of the above.
Bob

Does this mean I have to make the cuts from the same side of the blade as the pivot? If I'm standing at the saw facing the blade, the pivot point is to my right on the fence. Should the cuts all be from the right side of the blade? I'm pretty sure I've been doing that but for no other reason that I just did, but I can't be sure if all of them were done from the right!

 

[bobtskutter]

I'm sorry I can't remember.  Watch the YouTube videos closely, you'll see how they set it up.
I like this one (about 10min in)

Bob

 

[Michael Kellough]

Check for dust on the inside of the fence, it doesn't take much to prevent the work piece sitting flush.  Check the videos carefully, the formula depends on which side of the blade you hold the work piece and which side the pivot is on.  When you make a cut you need to rotate the work piece so the fresh cut is against the fence.  I've been caught out by all of the above.
Bob

Does this mean I have to make the cuts from the same side of the blade as the pivot? If I'm standing at the saw facing the blade, the pivot point is to my right on the fence. Should the cuts all be from the right side of the blade? I'm pretty sure I've been doing that but for no other reason that I just did, but I can't be sure if all of them were done from the right!

The side that is freshly cut should rotate to the fence. If testing a miter gauge on a table saw (typically the miter gauge is in the slot to the left of the blade) then the piece being cut is rotated clockwise as the cuts progress.

It’s the opposite rotation when testing the MFT fence.

As has been said, you want to start with the biggest piece of wood that can be managed on the saw. To put mino’s comments another way, the smaller the stock the more significant the decimal places to the right are.

And on each cut both sides of the blade should be engaged with the wood. But you only want to cutoff a few mm’s so you don’t reduce the size of the piece too quickly.

 

[smorgasbord]

And on each cut both sides of the blade should be engaged with the wood.

I've heard this repeated a number of times from various sources, but is it real? I could see where it might be real on a sliding miter saw where the whole head assembly has lots of opportunities for play, but on a tablesaw with a normal kerf blade (blade stiffeners) does it really matter? Has anyone done a test? Even if the blade deflects, wouldn't it deflect consistently through the cut as long as it didn't vary between cutting off a complete kerf and a partial kerf?

There are lots of time where I sneak up on a cut, and that last cut may indeed be partial. Do I align things for a partial cut or always for a full cut, or, again, does it really matter? Someone should test this conventional wisdom.

 

[Michael Kellough]

And on each cut both sides of the blade should be engaged with the wood.

I've heard this repeated a number of times from various sources, but is it real? I could see where it might be real on a sliding miter saw where the whole head assembly has lots of opportunities for play, but on a tablesaw with a normal kerf blade (blade stiffeners) does it really matter? Has anyone done a test? Even if the blade deflects, wouldn't it deflect consistently through the cut as long as it didn't vary between cutting off a complete kerf and a partial kerf?

There are lots of time where I sneak up on a cut, and that last cut may indeed be partial. Do I align things for a partial cut or always for a full cut, or, again, does it really matter? Someone should test this conventional wisdom.

It might not deflect, depending on the sharpness and stiffness of the blade, but when you’re worrying about tiny fractions of mm’s you want to remove variables.