Thread: Saw sharpening files 101

The formula I worked out was
25.4mm / tpi * 2.2 so in this case
25.4 / 3.5 * 2.2 = 16
(25.4mm to the inch, and 2.2 to give an unused gap in the middle of the file - i.e. technically it could be * 2 so that as you rotate to a new face you are using all new file face).

Actually when I read off 14mm before that was for 4 tpi and 3.5 needs a 15-16mm face, but I'm sure a 14 will be just fine.

Ahhh, was that a clumsy midnight explanation or what?

Originally Posted by FenceFurniture

Orrite den, my little chart says for 3.5 tpi you need a file with a 14mm face, and for 4.5 a 12.5mm face. Although you may get away with the same 14mm file for the 4.5 teeth if the gullet can have the same radius.

Are you sure about "tpi" vs "ppi" ?

I used a Nicholson 8" slim taper for a 4-1/2ppi (3-1/2tpi) saw ... it was 11.5mm on a face.

Despite this whole thread I still think that if you are having a first go, with a handsaw under say 7ppi, look for a file from the local and try it - if they have any.

Of the modern stock I've used a blatantly chinese "Supertool" file and three Nicholsons (on closer inspection, the 5" says made in Brazil but the 4" and 8" both say Mexico). All have been ok for me. YMMV.

Touching up a ripsaw that is basically the right shape can be just like honing an aftermarket plane iron ... ok maybe NOS.
Straightening a hollow toothline adds challenge ... or irregular teeth ... same as a pitted chisel or a wonky iron.

For a ripsaw at least, sharpening should become seen as just as learnable as grinding and honing.


Paul

Paul, I agree largely with what you said there. I'm never going to talk ppi because I think it obfuscates the most meaningful number (tpi), as it is always tpi + 1. That's not a big deal in fine tooth saws (20 or 21 - meh), but in rips saws "1" can make an enormous difference ( 3 verses 4 is a 33% difference). Actually, the most significant, and least confusing measure of the tpi/ppi is what you might call the wavelength or the distance from point to point. It is actually that measurement that defines what file should be used (i.e. 25.4 / tpi OR 25.4 / (ppi -1).

You'd certainly get away with a narrower faced file as long as the gullet radius was ok - it's just that the middle smidgen of the file would get used twice. So, as you say, for a first go get your hands on whatever you can - lots of luck finding a single cut file (apart from a Brazilian Milled File POS) in any Australian Hardware Store.

I'll put up some diagrams tomorrow to illustrate how I arrived at that formula (which may need some work, but as an approximation it's pretty close, disregarding rake).

Originally Posted by pmcgee

He he he ... I was just thinking that, but I heard it as a turkey-cooking story.

Kid comes home from school ... thanksgiving coming up ... and says "We cooked a turkey at school."
"Mom" is cooking a turkey too, so says "You can help me."
They go through it all until the point where mum cuts the wings off the turkey ...
Kid says "We didn't do that at school ... why do you do it?"

Mom ... "I don't really remember, that's how Mom taught me."
So they phone up Grandma down the road to ask why we cut the wings off the turkey.
She can't remember either.

So they get in the car, pick up Grandma, and go to see Great-Grandma in the nursing home.
"Great Grandma ... why did you teach us to cut the wings off the turkey when we cook it?"

... "Well, when I was growing up, we only had this narrow pan ..."


Always loved that "joke".

Similar but with a leg of lamb with the shank cut and bent back to fit the roasting tin

great grandma "it was the only way to fit it in the roasting pan I had"

The interesting thing is that the local butcher sells all legs of lamb that way, I have a little chuckle to meself everytime I buy one.


and so this is relevant I find it easier to file a saw after having a roast lamb and mustard pickle sandwich for lunch

Originally Posted by FenceFurniture

Paul, I agree largely with what you said there. I'm never going to talk ppi because I think it obfuscates the most meaningful number (tpi), as it is always tpi + 1. That's not a big deal in fine tooth saws (20 or 21 - meh), but in rips saws "1" can make an enormous difference ( 3 verses 4 is a 33% difference). Actually, the most significant, and least confusing measure of the tpi/ppi is what you might call the wavelength or the distance from point to point. It is actually that measurement that defines what file should be used (i.e. 25.4 / tpi OR 25.4 / (ppi -1).
).

Yes the Germans and there annoying decimal logic do it that way: ''zahnteilung' tooth-spacing is done in nice round numbers, 2mm, 3mm, 5mm etc.

Nice.

Hell will freeze over before before that crazy system is ever adopted over here.

Firstly: Matt, if you mean that the Germans measure the distance between two teeth points then to me that is perfectly acceptable, in fact desirable, as it takes out any confusion between tpi and ppi.

Funnily enough, for once the Imperial system gives a more user friendly method for measuring tooth pitch. In large tooth saw such as 3.5 tpi the metric system would be able to cope reasonably well (3tpi is 8.5mm point to point), but in smaller tooth saws it falls down to cumbersome numbers in metric: 14 tpi is 1.8mm point to point, and 19 tpi is 1.34mm point to point.



Ok, then. The difference in tpi verses ppi can be shown as follows:



I'm not in favour of ppi because one of teeth is actually counted twice, if you know what I mean. When measuring the frequency of soundwaves it is done in the same way as tpi.


Remembering that for Western saws, the files are always an equilateral triangle in profile, and the bottom of the gullet is always 60°, we can see that a file fits in the gullet like so (ignoring the fact that Excel doesn't allow me to show radiused gullets):



These teeth are perfectly symmetrical, with the back and the front of each tooth being the same length.

So, to get maximum use out of each file face, the face has to be at least double the face of the tooth, otherwise when the file is rotated to use a fresh corner the tip of the saw tooth will be passing over a part of the file that has already been used (or used up - blunt), and so the tooth won't be filed evenly.

However, this is not a typical line of teeth in respect to the rake, or angle that the teeth are set at. This next diagram might be a little closer to the mark:



What we can see here is that unlike the previous diagram, these teeth have quite a distinct difference between the length of the back and the front, and this can also be seen in Tony's saw:



That means that we can still use the same file that was a perfect fit in the symmetrical teeth:
because we use a wider part of the file face on the back of the tooth, we use an exactly proportionately smaller width of the file face on the front of the tooth. In other words, if we add together the lengths of the back and front of the tooth the total will always be the same regardless of the rake, or angle of the tooth.

The only time a wider face file would be needed to avoid a double wear spot in the file face middle, is when filing teeth if the same TPI on two different saws with different rakes. If this is the case then a double wear spot may become problematic, and perhaps two identical files may be required (say one for the Rip Saw, and one for the Crosscut). For a professional saw doctor it would be a different story - they'll never know what they are going to get.

So, I arrived at the formula for calculating the file width required as follows (restated from a previous post):
File width = 25.4 / tpi X 2 (but preferably 2.1)

For a 3.5 tpi we would need a minimum file face width of 14.5mm, say 15mm
For a 12 tpi we would need a minimum file face width of 4.25mm
For a 19 tpi we would need a minimum file face width of 2.65mm, say 2.75mm

You can use a file of a slightly wider face as long as the corner radius is still ok, and yields the correct shaped gullets.

You're so far down the rabbit hole I'm worried you won't be sane again when you've finally got to grips with saw files.

...backs away slowly...

One small issue with the above Brett, I think if you changed from a file with a small radius gullet and went to a wider radius the file would drop down in the gullet, this might cause you to start pushing out teeth or pushing them along. Could get messy. I may be wrong but it would be worth checking before you started. And I have a crosscut and a rip with the same tpi (just to make a point) 7 actually

Originally Posted by FenceFurniture

That means that if we can still use the same file that was a perfect fit in the symmetrical teeth:
because we use a wider part of the file face on the back of the tooth, we use an exactly proportionately smaller width of the file face on the front of the tooth. In other words, if we add together the lengths of the back and front of the tooth the total will always be the same regardless of the rake, or angle of the tooth.

Yes, that's right Matt. My previous mumblings about "it may be different when rake is introduced" were off the mark.
I'll edit that post.

Yep, I modified my comment.

Originally Posted by Berlin

You're so far down the rabbit hole I'm worried you won't be sane again when you've finally got to grips with saw files.

...backs away slowly...

Where does the "again" come from?

Originally Posted by Claw Hama

...if you changed from a file with a small radius gullet and went to a wider radius the file would drop down in the gullet...

Are you quite sure about that Claw.....

Originally Posted by Claw Hama

And I have a crosscut and a rip with the same tpi (just to make a point) 7 actually

Yeah, and you don't exactly fit into the casual group

Originally Posted by FenceFurniture

Where does the "again" come from?

You said that

Originally Posted by pmcgee

Are you sure about "tpi" vs "ppi" ?
I used a Nicholson 8" slim taper for a 4-1/2ppi (3-1/2tpi) saw ... it was 11.5mm on a face.
l

Sorry - I didn't really complete that train of thought.
What I had meant to write was "... and I thought it was still a little large."

I was filing teeth with zero rake, and the vertical tooth height was less than half that of the file.
I was thinking that something was out of whack, cos 14mm or more seemed like "Woah - that'd be a big file"

At that point I started doodling and thence drifted off from there ... and actually drifted off.

I drew a diagram and realised (as you have said) that the front of the gullet in that instance is greater than the vertical.
I was then wondering whether I cared about that or not.

teeth.JPG
By the way FF ... are you wanting to call the tooth-point distance (red) "λ" (lambda) ???

At zero rake, with the bottom of the gullet 60* ...
- front of gullet (blue) = 1.15 x red = 2 x vertical
- vertical height = 0.58 x red

So the blue section + vertical height = red x 1.73

[Edit: ... alzheimers ... so for 3.5tpi at "λ" = 7.3mm ---> 12.5mm which seems more reasonable to me.]

Changing the rake on my saw, a lot of the filing was into the front of the teeth. But I can see that for a touch-up to a saw in regular condition, where the front of the tooth is to be kept in position as much as possible, then the back of the tooth will be the main focus of the filing.**

Paul
**Edit again: That was a stupid thing to write. It only applies to zero rake teeth.