Who needs precision, I do!!!!

[Adam9453]

I often read posts on here saying you don't need to be that precise etc it's woodwork after all not engineering.
Well it's just not bl@@dy true, I've been cursing for the last two hours trying to get the diagonal measurements on a carcass equal, despite much effort I've got a difference of 4mm!
After much checking I've concluded that my saw is cross cutting approx 0.75mm out of square over 600mm length.
So tomorrow I'll be squaring up the cross cuts with the track saw instead of the tablesaw and then painstakingly adjusting the joinery to get everything to fit properly and most importantly square!
So next time someone says you don't need to be that accurate, feel free to slap them with a wet fish for me :twisted:

 

[Jacob]

Classic case of not understanding the difference between precision and accuracy.
If your saw can consistently cut "approx 0.75mm out of square over 600mm length" then it's very precise, but also, unfortunately, inaccurate.
Your problem is about making accurate cuts. This is down to machining skill, working methods, machine set up etc, not precision.
Splitting hairs I know - but people often think that a more precise bit of kit (e.g. vernier callipers, or an expensive straight edge) will improve accuracy but it may not help at all.

Getting the same error means "systematic" error which is often easier to remedy than a really inaccurate bit of kit where the error is more random.

In the end it's hands on - how do you get the result you want. It may be by fine adjustment to your very precise machine, but it might be to work around it in another way, somehow.

So the answer to "Who needs precision, I do?" is no, you've already got precision, what you need is accuracy.

 

[skipdiver]

My cross cut jig is slightly out too. I just put a 1mm packer behind the piece i'm cutting against the fence. Feel free to slap me with a wet fish.

 

[thetyreman]

from my recent workbench build, it has made me realise accuracy is everything, for example you are a 64th of an inch off your 0.3mm thick knifeline on 4 joints, when all 4 joints come together that's a 16th on an inch out...which is actually noticeable, it's how the slight inaccuracies add up when put together as one object. It really matters!

 

[Jacob]

from my recent workbench build, it has made me realise accuracy is everything, for example you are a 64th of an inch off your 0.3mm thick knifeline on 4 joints, when all 4 joints come together that's a 16th on an inch out...which is actually noticeable, it's how the slight inaccuracies add up when put together as one object. It really matters!

Use a pencil. Marking with a knife is a mistake - you can't rub it out. They are called "marking knives" but they aren't really suitable for marking - unless you actually want an indelible mark

Good example really. Knife marking is very precise but it doesn't guarantee accuracy if you've put it in the wrong place. Pencil less precise but more accurate in that you can rub it out and do it again if necessary.

 

[thetyreman]

from my recent workbench build, it has made me realise accuracy is everything, for example you are a 64th of an inch off your 0.3mm thick knifeline on 4 joints, when all 4 joints come together that's a 16th on an inch out...which is actually noticeable, it's how the slight inaccuracies add up when put together as one object. It really matters!

Use a pencil. Marking with a knife is a mistake - you can't rub it out. They are called "marking knives" but they aren't really suitable for marking - unless you actually want an indelible mark

Good example really. Knife marking is very precise but it doesn't guarantee accuracy if you've put it in the wrong place. Pencil less precise but more accurate in that you can rub it out and do it again if necessary.

isn't that just common sense though?

I use a pencil line first, then the marking knife, but it's the knife that gives me the reference line.

 

[Beau]

from my recent workbench build, it has made me realise accuracy is everything, for example you are a 64th of an inch off your 0.3mm thick knifeline on 4 joints, when all 4 joints come together that's a 16th on an inch out...which is actually noticeable, it's how the slight inaccuracies add up when put together as one object. It really matters!

Use a pencil. Marking with a knife is a mistake - you can't rub it out. They are called "marking knives" but they aren't really suitable for marking - unless you actually want an indelible mark

Good example really. Knife marking is very precise but it doesn't guarantee accuracy if you've put it in the wrong place. Pencil less precise but more accurate in that you can rub it out and do it again if necessary.

Don't put it in the wrong place?

I have used a marking knife for many years and could count on one hand the amount of times I wanted to remove one. Great for precision when you can locate a chisel in the groove for the final cut. Pencil is way to vague IMO for a lot of marking out.

 

[Adam9453]

Not quite so sure if the pedantic argument is fool proof Jacob
It takes precise measuring and adjustments to get the machine to cut accurately which is the fundamental cause of my current woes.
The machine cuts consistently providing the user operates it consistently. I don't have any issue with consistency as all the cuts are off by precisely the same amount lol
However I have also discovered that my large L square (1350mm) long is unfortunately not as accurate as I thought so I will be resetting the table saw based on the smaller but more accurate 500mm square.
The cutting inaccuracy I'm trying to get rid of is circa 0.5mm difference on the diagonal measurements of a panel sized 2200x574mm.
The reason I'm so frustrated by this is because it's my fault for designing the piece to require absolute accuracy (within a very fine tolerance) and that it's the second go I've had at setting the saw to cut bang on square (again within a very fine tolerance).
Suffice to say I won't be building a carcass using this design again!!

 

[screwpainting]

The cutting inaccuracy I'm trying to get rid of is circa 0.5mm difference on the diagonal measurements of a panel sized 2200x574mm.

If you don't mind me asking, what are you making that require such accuracy?

 

[RossJarvis]

I'd have to disagree that woodworking needs to be as accurate as precision engineering. In precision engineering you may be looking at a difference of 0.1mm over a 2m diagonal, I can't see that this would be necessary in any woodworking project. 4mm difference over the diagonal may well be too much in the woodworking sense so a greater level of accuracy may be required. I personally wouldn't think a square with a 1350mm length would be accurate to any great degree so a better marking out technique would be more appropriate.

Having taught engineering I am well aware that accuracy is important and it is one of the hardest things to do to teach people to mark out accurately. Unfortunately people place too much reliance in the marking out tools and their accuracy when in fact they need to understand the principles behind marking out better. The basic principle is to never ever assume the tool itself is accurate, once you understand that it is much easier to mark out accurately.

 

[MattRoberts]

I'm not sure what benefit you would have using a marking knife over a (thin) pencil line anyway - unless you are cutting across the grain.

I always took marking knives to be used to sever the grain and avoid tear out or provide a cut line to pare back to

 

[skipdiver]

The cutting inaccuracy I'm trying to get rid of is circa 0.5mm difference on the diagonal measurements of a panel sized 2200x574mm.

If you don't mind me asking, what are you making that require such accuracy?

I was wondering that too. Seems so minor as to be non existent.

 

[Jacob]

The cutting inaccuracy I'm trying to get rid of is circa 0.5mm difference on the diagonal measurements of a panel sized 2200x574mm.

If you don't mind me asking, what are you making that require such accuracy?

I was wondering that too. Seems so minor as to be non existent.

Way out beyond normal level of precision for woodwork. Thermal/humidity movement would be greater than that.
Can't imagine anything which would need that degree of precision. Don't see how you could measure it that precisely either, without some sort of lab equipment - lasers etc.
So what are you making? :lol:

 

[Beau]

I'm not sure what benefit you would have using a marking knife over a (thin) pencil line anyway - unless you are cutting across the grain.

I always took marking knives to be used to sever the grain and avoid tear out or provide a cut line to pare back to

Agreed but most joints are marked out across the grain and benefit from protection from tare out and a line to pare back to. Not saying every line should be done with a knife and not a pencil just were it's advantageous.


To be fair it does depend on the work at hand. If I was building a wooden shed or some built in cupboards cant say I would be reaching for the marking knife but when it comes to quality fitted mortice and tenons and dovetails etc the marking knife lines across the grain are invaluable.

 

[PAC1]

Adam what makes you think the track saw is going to get you this sort of accuracy. The level of accuracy you are looking for is way beyond most woodworking machines without getting into your "Not Square" or wondering if the slider on your table saw is parallel to the blade etc. etc.
You might be able to get this sort of accuracy with a CNC machine or in the engineering world but not ordinary woodworking equipment.

 

[Adam9453]

Ok to answer a few questions:
I'm making a large carcass which forms part of a large wardrobe
The issue is that when I assemble the carcass it is not square (hence the 4mm difference in the diagonal measurements).
After much investigation the root cause can only be (as far as I can tell) that the sides are not actually cut completely square. The compound of the inaccuracy of the cutting of the sides is what I believe is throwing the diagonal measurement of the assembled carcass out.
I don't usually have to work to this kind of precision in woodwork but unfortunately this design apparently calls for it.
The solution I've worked out should work but today will tell
The reason I can achieve greater accuracy with the tracksaw is I can set the cutting track in position using the accurate 500mm square and measure it multiple times before cutting.
Using the sliding table on the tablesaw relies on the table saw being setup accurately in the first place which I where I've fallen foul so far (it's close but not close enough for this design apparently).
Thanks for the comments chaps, it was more of a rant from me anyway lol

 

[RobinBHM]

I must admit Im struggling to see where the accuracy of the panels could make your carcase 4mm out.

I can only visualise that the overall squareness of a wardrobe is determined by the squareness of the back. The sides and top and bottom will only have a slight impact.

The biggest issue usually with wardrobes is that if they arent sitting on a dead flat surface a twist is introduced into the carcase that makes the doors sit badly.

Issues with accuracy are often the result of not working from a consistent datum edge. A classic example of this is CNC panel drilling, often seen in kitchen carcasses. If the panels are cut slightly out of square, dimension when put on the point to point, the holes may not end up the correct distance from the nearest edge, it because the cnc is not referencing from that edge only from an x y coordinate.

For markung out, a 2h pencil is ideal it gives a nice crisp line to work to.

When marking atound a section of timber, the ljnes often dont meet on the last edge, this can be avoided by always using the square against the face or edge, that should be datum marked.

 

[Adam9453]

I agree robin, I wouldn't have thought the sides would be the cause of the problem but it's the only thing I can find that's out. The back panel is square (diagonals are the same and my 500mm square agrees it's square).
The carcass is square where the sides attach into the back panel but it's out at the front face and the only reason I can think of for this is the sides are throwing it out, if I shift the sides (one towards top and one towards bottom) I can get the diagonal measurement very close at the front face so I think one side is skewed towards the bottom and one side skewed towards the top. I'm confident the sides are parallel just the ends are skewed cut.
I'm about to starting trying to fix the issue so we'll see if my theory is right or whether there's another problem adding to or causing the issue.
One way or another I know I will fix it, it's just a question of systematic problem solving.
Well that or remake the bloody thing :shock:

 

[Jacob]

The other answer is to use the various trad joiners' tricks such as taking off the arris and making an edge look straight against the next one, even though it isn't - the details being in the shadow.
Or even simpler - make the door tight or oversize and trim to fit the hole. This is common procedure in a small workshop - the assumption is that nothing is ever straight or exactly square however hard you try. You avoid having to fit as much as possible but use it as a last resort.
There are all sorts of ways of pulling things in (with screws etc,) or pushing them out with wedges and packing.
A lot of things go in nice and square but the following morning have moved about.
A very common universal design solution is to have things set-in rather than flush, then any mismatch of the gaps is less visible.

 

[Adam9453]

Normally I could use those tricks but as this has a face frame with overlay doors it'll show the discrepancy.
I've bought the doors in so there's nothing I can do to the doors to disguise it.
The reason I realised the issue was there is because the face frame joints were skewed so I started measuring and found the carcass is out.
I'm confident with a little careful trimming and tweaking the joints I can resolve it.
I'll report back later today either happy or more bemused by it