Squaring advice...

[Paul alan]

...I have made myself a table saw workbench.
It’s ( Bosch gets 10j) sat in the bench at around 98% (ish) square to the surface.
I now need to align it as square as is possible for the fence build.
I have a 12” starret square which I trust but is way too short for this situation.
I have a 600mm Axminster square which I had to square up myself by sanding the internal parts, I think this pretty good but can’t be sure.
I have a framing square which is carp, I have tried to square up but was in a rush and didn’t do a good job (poor).
The front of the table is flat to a decent degree, I have no precision straight reference edge to test though .
I did take my time with all this and the 6ft by 4ft table top from corner to corner is around 1/2 mm out.
I kept checking for level with a “level” and planed any high spots out.
I need a reliable way to square up for a fence with out spend lots of £££ on a massive square.
What to do???
Please help me not pull my hair( what’s left) out tomorrow.
I’ve taken some time from work to deal with this and am under pressure to get done.
Thanks

 

[Ttrees]

Hello
What fence are you going to make?
The biesemeyer fence is well regarded, and adjustable.
I think Steve Maskery of this forum may have a thread on that.

You could try this method to get it perfect if you wanted, but you may like the fence to toe out instead,

I suppose it depends if you wish to cut tenons on the saw for instance, and you wanted to make sliding carriages for the fence.

Hopefully I have understood your query?
Tom

 

[MusicMan]

The squareness you are measuring is not the way to go. The reference you will use is the mitre gauge slot, and this needs to be parallel to the blade. Measure the distance from the blade (just at the bottom of the gullet) to the edge of the channel at the front of the blade and at the back of the blade (ideally mark the tooth and rotate the blade on its spindle so that the same tooth is measured at back and front). Normally one makes this adjustment simply by slackening the bolts holding the saw mechanism to the table top and using the free movement to twist the table. To measure it you could set up a dial gauge, but a combination square is ok, just use it as a depth gauge with the base part pushed against the mitre slot edge and the end of the ruler just touching the blade. Then tighten down and check again. If there isn't enough movement then make the mounting holes slightly larger till there is.

Then you set your fence parallel to the mitre gauge, when it will also be parallel to the blade and all your cuts will be happy ones. The bosch unit should be set up this way anyway (blade parallel to slot), so all you may have to do is align the fence to the mitre slot.

It looks as if you are going to have a fence running from the front to the back of the table. This is fine for sheet goods (but best to have a 'toe-out' by a mm or so from true square, to resist jamming. But then you should have a short fence mounted on this, ending no further than at the mid point of the saw. This prevents jamming or kickback on ripping timber if it warps outward slightly. Here's mine on a Wadkin (guard removed to show the short fence (I never run it like this. Not once.).

As long as your mitre slot, blade and fence are parallel (apart from a very slight toe-out on the fence) you don't really care what the front of the table does, as long as the fence runs straight and smooth.

 

[Paul alan]

Hello
What fence are you going to make?
The biesemeyer fence is well regarded, and adjustable.
I think Steve Maskery of this forum may have a thread on that.

You could try this method to get it perfect if you wanted, but you may like the fence to toe out instead,

I suppose it depends if you wish to cut tenons on the saw for instance, and you wanted to make sliding carriages for the fence.

Hopefully I have understood your query?
Tom

Thanks, I really like the look of the fence you reference to, bu that would require me adding the rail for the fence to slide along to the front of my table which gives me 2 problems.

Firstly is getting it very straight on a budget, the front of the table is straight as I spent a lot of time trying my best with pine and cheap shuttering ply, but I wouldn't be confident that it will not put the fence out of alignment to a small degree here and there. I want something I can rely on.

Secondly is having the rail protruding from the front, the shop is 12x10 and the table is 6x4. So it has castors on and will serve other functions. I will add a router, and It will be an assembly table also. Its a bit of a squeeze but I wanted something I can assemble upon. I think adding a protrusion on the front would not be a good idea, it'll take knocks and I'll also get bruises from it.

Its my first "proper" (decent) table saw and I want the most functions I can get from it.

So, would I toe out the fence or would it be better to toe in the back of the blade?

Many thanks

 

[Ttrees]

Musicmans sound advice about alligning the blade to the miter slots I should have mentioned.
I have seen the technique referred as a pick a tooth method


Then you can adjust your fence however you like, but never never "toe in"
as you would get burning, and cause a kickback.
I presume you mean parallel with the blade

I suggest you set it toe out and see how you get on, unless you want to make something like this


I should mention the one part of the saw that you shouldn't trust is the edge of the table, the mitre slots are the bassline for setup.
I wouldn't think you could damage the biesemeyer fence as the rail is fixed
and its the fence that has adjustment.
Most I've seen are made from square tubular steel.

The only other fence I've seen that sits on the table instead of attached to the rail is the Incra fence, but havn't seen a shop made one...
I presume it would need good tolerances to work well or it would either be flimsy
or get stuck.
I'm sure something else has the same system if you could find it, and adapt.

Just for the sake of it, a good video for viewing some safe practices


Tom

 

[cammy9r]

So, would I toe out the fence or would it be better to toe in the back of the blade?

Bad idea. The aim of toe out at the back of the fence is to stop binding of sheets. The distance between the rear of the blade and fence is ever so slightly greater than at the front be it 0.5mm or 1mm. Toe in blade at the rear will bind and cause dangerous situations for you to get yourself out off.
I hate to be one of the safety police, i really do, but try and get some practical knowledge from someone in the UK before messing with a table saw.

 

[bp122]

I am also in the design phase for a larger surface for my table saw and a fence upgrade.

After passing on an incra ls system due to lack of space, I have settled on an incremental position version of the biesmeyer style fence, DIY of course.
I think the biggest advantage of a biesmeyer style system is that the fence adjustment is very easy and in all three axes.

Anyway, regards to the squaring issue, Perhaps you may have already thought about it but the easiest way to check if a square is accurate is to do the flip over trick. Put the square in place, draw a line on the inner or the outer vertical (whichever you are trying to determine is square or not), flip the square over (mirror its position) and use the same reference as before and draw another line right on top of the first line.

If the lines don't converge or divulge, the square is good. If they don't overlap on top of one another and you see a wedge, your square is off.

You can determine the error by how many degrees etc mathematically, but it is a black or white matter - either it is square or its not.

 

[Paul alan]

Bad idea. The aim of toe out at the back of the fence is to stop binding of sheets. The distance between the rear of the blade and fence is ever so slightly greater than at the front be it 0.5mm or 1mm. Toe in blade at the rear will bind and cause dangerous situations for you to get yourself out off.
I hate to be one of the safety police, i really do, but try and get some practical knowledge from someone in the UK before messing with a table saw.

Yeah, Ive no idea why I wrote that the opposite to what I meant.
Toe in as it is very slightly now, but very soon to be corrected is obviously a very bad idea and safety is very important to me as I have two dependents.
Thanks for pointing that out though.

 

[MusicMan]

Anyway, regards to the squaring issue, Perhaps you may have already thought about it but the easiest way to check if a square is accurate is to do the flip over trick. Put the square in place, draw a line on the inner or the outer vertical (whichever you are trying to determine is square or not), flip the square over (mirror its position) and use the same reference as before and draw another line right on top of the first line.

If the lines don't converge or divulge, the square is good. If they don't overlap on top of one another and you see a wedge, your square is off.

You can determine the error by how many degrees etc mathematically, but it is a black or white matter - either it is square or its not.

I'm going to be pedantic and say that it is never black and white. The best you can do is say that it is square within a certain tolerance, be that ever so small.

Yes the flip over trick is useful, but it requires a bit more than you say. The edge that you are setting it against must be straight, not bowed; in fact it is the bow of this edge that is the main tolerance in the method. The other tolerance is the thickness of the line that you draw, and its straightness. Did you hold the pencil at just the same angle all the way?

The five-cut method referred to already is a very good test since it magnifies the error. So does my square setting jig, (search old threads for this). A quality straight edge is also useful for setting up machines.

 

[Paul alan]

Musicmans sound advice about alligning the blade to the miter slots I should have mentioned.
I have seen the technique referred as a pick a tooth method


Then you can adjust your fence however you like, but never never "toe in"
as you would get burning, and cause a kickback.
I presume you mean parallel with the blade

I suggest you set it toe out and see how you get on, unless you want to make something like this


I should mention the one part of the saw that you shouldn't trust is the edge of the table, the mitre slots are the bassline for setup.
I wouldn't think you could damage the biesemeyer fence as the rail is fixed
and its the fence that has adjustment.
Most I've seen are made from square tubular steel.

The only other fence I've seen that sits on the table instead of attached to the rail is the Incra fence, but havn't seen a shop made one...
I presume it would need good tolerances to work well or it would either be flimsy
or get stuck.
I'm sure something else has the same system if you could find it, and adapt.

Just for the sake of it, a good video for viewing some safe practices


Tom

Thanks Tom,
I don't really work with metal so would have made the fence you talk of out of wood, meaning it might not of been as sturdy. To be fair it would just get in the way, especially if I was trying to clamp onto the table when using it as an assembly.

I do plan to make a sled, and a tenon jig etc I need the saw to bas versatile as possible as I dont have room for much other equipment, would it then be a bad idea to toe out??

 

[Paul alan]

I am also in the design phase for a larger surface for my table saw and a fence upgrade.

After passing on an incra ls system due to lack of space, I have settled on an incremental position version of the biesmeyer style fence, DIY of course.
I think the biggest advantage of a biesmeyer style system is that the fence adjustment is very easy and in all three axes.

Anyway, regards to the squaring issue, Perhaps you may have already thought about it but the easiest way to check if a square is accurate is to do the flip over trick. Put the square in place, draw a line on the inner or the outer vertical (whichever you are trying to determine is square or not), flip the square over (mirror its position) and use the same reference as before and draw another line right on top of the first line.

If the lines don't converge or divulge, the square is good. If they don't overlap on top of one another and you see a wedge, your square is off.

You can determine the error by how many degrees etc mathematically, but it is a black or white matter - either it is square or its not.

Great tip thanks

Can you tell me more about the fence system you plan to make please?

 

[bp122]

Great tip thanks

Can you tell me more about the fence system you plan to make please?

It is a variation of the system shown by Jeremy Schmidt (watch video)
Quite an elegant solution and doesn't break the bank. Apart from the threaded rod (my substitute would be a m10x1.0 stud from eBay, cost about £15 a metre) and using an indexing wheel with either 10 indents to get to 0.1mm increments or 4 indents giving 0.25mm increments. The latter is more likely to be my option as it is simpler to make and I don't anticipate needing increments of 0.1mm in my projects any time soon. But can be swapped at a later stage if needed.



If you aren't that bothered about the incremental feature, perhaps the attached pdf is a better option as it is even simpler than the first one and inexpensive and most of all, can be adapted for a no-welding design with bolts.

In your case, you can adapt these designs to your table. And it is very easy to square the fence to the mitre slots/ blade with Biesmeyer style fence even if the fence rail is not strictly perpendicular to the blade. Although every effort to get this right will make the system that much more reliable.

My third option for a fence is to use incra shop stop system and incorporate that into my main fence by making a t square with the locking feature to utilize incra's incremental system. Like I said, I'm still designing to see which one suits my needs based on time, cost and complexity

Hope this helps.

 

[bp122]

I'm going to be pedantic and say that it is never black and white. The best you can do is say that it is square within a certain tolerance, be that ever so small.

Yes the flip over trick is useful, but it requires a bit more than you say. The edge that you are setting it against must be straight, not bowed; in fact it is the bow of this edge that is the main tolerance in the method. The other tolerance is the thickness of the line that you draw, and its straightness. Did you hold the pencil at just the same angle all the way?

The five-cut method referred to already is a very good test since it magnifies the error. So does my square setting jig, (search old threads for this). A quality straight edge is also useful for setting up machines.

Regarding the black and white matter, I'm afraid it is still black and white. As soon as the angle is 90.001 degrees, it is by definition not square. But like all things in life, we stop the pursuit of perfection when it meets our needs and reaches what we determine as a point of diminishing returns. Also I didn't want to give the op a metrology 101 and finding the error to the nth degree (I know, because it is my day job!)
And I assumed the edge the op is using to bolt the fence assembly to is straight and not bowed.

I strongly agree with the 5 cut method. I have never tried it myself but practically that is the best solution, you are right.

However the flip over method can be used to quickly determine whether the squares op has at his disposal can be relied upon or not.





I

 

[MusicMan]

It used to be my day job too. I agree that the flip method is a useful quick check, but I do think that one should check the straightness of the edge being used, to use it reliably. It's the main source of error. And if three straight edges fit one another they are all straight, so that can easily be realised in a home workshop.

Mathematically, yes 90 degrees can be defined by one quarter of a full circle, which is a natural standard. Unfortunately there is no way to realise this except to some degree of tolerance, however small. We do it, as you say, to a tolerance that meets our needs.

I once provided NPL with an instrument that measured length to an absolute accuracy of 1 ppm and a tolerance of 10 picometres. Bit over the top for woodwork though!

 

[bp122]

I once provided NPL with an instrument that measured length to an absolute accuracy of 1 ppm and a tolerance of 10 picometres. Bit over the top for woodwork though!

That sounds fascinating. How did you achieve that? Is that using spectroscopy?

And I agree about the straightness of the edge. That's why when I see people making saw fences out of wood, it concerns me. Good quality ply or if you are in a pinch, MDF, are good choices, but I'd still struggle to trust the system when compared to a metal fence system built and mounted correctly.

I do see a lot of people building the fence rails out of 4040 or 4080 aluminium extrusions, but they are so expensive if you want anything longer than 1m. Even most low and mid range table saw fences are made of custom aluminium extrusions.

I do get the feeling that Americans are blessed to get raw materials at a lower cost than we do, even tools.

 

[MusicMan]

That sounds fascinating. How did you achieve that? Is that using spectroscopy?

And I agree about the straightness of the edge. That's why when I see people making saw fences out of wood, it concerns me. Good quality ply or if you are in a pinch, MDF, are good choices, but I'd still struggle to trust the system when compared to a metal fence system built and mounted correctly.

I do see a lot of people building the fence rails out of 4040 or 4080 aluminium extrusions, but they are so expensive if you want anything longer than 1m. Even most low and mid range table saw fences are made of custom aluminium extrusions.

I do get the feeling that Americans are blessed to get raw materials at a lower cost than we do, even tools.

You are right about that. And Chinese even cheaper, which is partly why they make everything in the world, it seems. I once knew a CEO whose company made refrigerators. He could buy complete refrigerators from Chine at lower cost than he could buy the steel to make them. Also from China.

The calibration instrument used X-ray interferometry. You can split an X-ray beam by diffracting it through a blade of essentially perfect crystal, semiconductor grade silicon. Then you recombine it with another blade, and analyse with a third. The third blade moves on a parallelogram spring mechanism. The second picture shows the device, which is all cut from one single crystal of silicon. The fine motion was achieved essentially with voice coils - stick a magnet on the moving part and surround it with a solenoid coil. It was a fun project!

 

[MorrisWoodman12]

@MusicMan XRay interferometry: very interesting. I used to design the electronics for HeNe laser interferometry. I suspect your equipment didn't measure distance as such but how far a reference piece had moved just like laser interferometer. Please tell how if I'm wrong. The standard HeNe LI used the wavelength of the laser and measured to 80nm though I worked out how to get 20nm easily electronically. Unfortunately the company was closed before we could get to market.
I find it really interesting the fantastic mix of people here on, an ostensibly woodworking forum.
Martin

 

[MusicMan]

@MusicMan XRay interferometry: very interesting. I used to design the electronics for HeNe laser interferometry. I suspect your equipment didn't measure distance as such but how far a reference piece had moved just like laser interferometer. Please tell how if I'm wrong. The standard HeNe LI used the wavelength of the laser and measured to 80nm though I worked out how to get 20nm easily electronically. Unfortunately the company was closed before we could get to market.
I find it really interesting the fantastic mix of people here on, an ostensibly woodworking forum.
Martin

Martin, yes you are right, it measured displacement. It was designed to calibrate surface roughness machines such as the Talysurf.

Also right about the forum. Never assume ignorance in others here! And many good companies never got to market.

 

[pe2dave]

Back to your 'right angle'. Other woodwork sources suggest the 3,4,5 unit method?
Cut paper / thin wood whatever to make strips, 3,4,5 units long to make the 'unit' appropriate to your needs.
Lay (pin) down the 3 unit piece. Attach the 4 and five first markers to the ends of the 3 unit piece.
Where the 4 and 5 unit ends meet pin that down. You've marked out a perfect right angle.
Tip. Use the same 'unit' to measure out each one. Dividers / single piece of scrap wood.
Tip. Don't use the end of the piece of paper, make the first point one 'unit' in (makes it easier).
explains it.

 

[Paul alan]

The squareness you are measuring is not the way to go. The reference you will use is the mitre gauge slot, and this needs to be parallel to the blade. Measure the distance from the blade (just at the bottom of the gullet) to the edge of the channel at the front of the blade and at the back of the blade (ideally mark the tooth and rotate the blade on its spindle so that the same tooth is measured at back and front). Normally one makes this adjustment simply by slackening the bolts holding the saw mechanism to the table top and using the free movement to twist the table. To measure it you could set up a dial gauge, but a combination square is ok, just use it as a depth gauge with the base part pushed against the mitre slot edge and the end of the ruler just touching the blade. Then tighten down and check again. If there isn't enough movement then make the mounting holes slightly larger till there is.

Then you set your fence parallel to the mitre gauge, when it will also be parallel to the blade and all your cuts will be happy ones. The bosch unit should be set up this way anyway (blade parallel to slot), so all you may have to do is align the fence to the mitre slot.

It looks as if you are going to have a fence running from the front to the back of the table. This is fine for sheet goods (but best to have a 'toe-out' by a mm or so from true square, to resist jamming. But then you should have a short fence mounted on this, ending no further than at the mid point of the saw. This prevents jamming or kickback on ripping timber if it warps outward slightly. Here's mine on a Wadkin (guard removed to show the short fence (I never run it like this. Not once.).
View attachment 93030

As long as your mitre slot, blade and fence are parallel (apart from a very slight toe-out on the fence) you don't really care what the front of the table does, as long as the fence runs straight and smooth.

The squareness you are measuring is not the way to go. The reference you will use is the mitre gauge slot, and this needs to be parallel to the blade. Measure the distance from the blade (just at the bottom of the gullet) to the edge of the channel at the front of the blade and at the back of the blade (ideally mark the tooth and rotate the blade on its spindle so that the same tooth is measured at back and front). Normally one makes this adjustment simply by slackening the bolts holding the saw mechanism to the table top and using the free movement to twist the table. To measure it you could set up a dial gauge, but a combination square is ok, just use it as a depth gauge with the base part pushed against the mitre slot edge and the end of the ruler just touching the blade. Then tighten down and check again. If there isn't enough movement then make the mounting holes slightly larger till there is.

Then you set your fence parallel to the mitre gauge, when it will also be parallel to the blade and all your cuts will be happy ones. The bosch unit should be set up this way anyway (blade parallel to slot), so all you may have to do is align the fence to the mitre slot.

It looks as if you are going to have a fence running from the front to the back of the table. This is fine for sheet goods (but best to have a 'toe-out' by a mm or so from true square, to resist jamming. But then you should have a short fence mounted on this, ending no further than at the mid point of the saw. This prevents jamming or kickback on ripping timber if it warps outward slightly. Here's mine on a Wadkin (guard removed to show the short fence (I never run it like this. Not once.).
View attachment 93030

As long as your mitre slot, blade and fence are parallel (apart from a very slight toe-out on the fence) you don't really care what the front of the table does, as long as the fence runs straight and smooth.

Hello
What fence are you going to make?
The biesemeyer fence is well regarded, and adjustable.
I think Steve Maskery of this forum may have a thread on that.

You could try this method to get it perfect if you wanted, but you may like the fence to toe out instead,

I suppose it depends if you wish to cut tenons on the saw for instance, and you wanted to make sliding carriages for the fence.

Hopefully I have understood your query?
Tom

Back to your 'right angle'. Other woodwork sources suggest the 3,4,5 unit method?
Cut paper / thin wood whatever to make strips, 3,4,5 units long to make the 'unit' appropriate to your needs.
Lay (pin) down the 3 unit piece. Attach the 4 and five first markers to the ends of the 3 unit piece.
Where the 4 and 5 unit ends meet pin that down. You've marked out a perfect right angle.
Tip. Use the same 'unit' to measure out each one. Dividers / single piece of scrap wood.
Tip. Don't use the end of the piece of paper, make the first point one 'unit' in (makes it easier).
explains it.

More like it mate thanks, starting to think I was on the wrong forum there for a minute. Damn inferiority complex growing...