Plane Blade Camber

[tnimble]

When taking too many through shavings the board will become convex indeed. This should not happen when taking only 1 or 2 through shavings.

 

[adrian]

When taking too many through shavings the board will become convex indeed. This should not happen when taking only 1 or 2 through shavings.

I've flattened several boards without the benefit of the methods described in the DVD and I definitely noticed the tendency of the board to become convex. (In my first effort, I started with inch thick lumber and by the time I declared it flat, several hours later, it was less than 1/2" in places.) In this case, though, I am talking about taking one shaving only (to the best of my ability). Maybe I'm overlapping in some places. The board went from being about 0.0015" concave to 0.0015" convex.

 

[tnimble]

Try it again on a edge of a board where you take full width shavings. This way you can verify if its your technique of overlapping the shavings of a set.

If it still happens, check the sole of the plane, espacially look at the ara close to the mouth. Use a good steel rule for this that is at least the length of the plane.

I don't know how much the bump at 4" from the beginning is. But this bump could be caused by either a flatness problem of the plane just behind the blade or can be caused by the plane offered to the wood with the toe slightly upward / heel slightly downward.

Also did the plane stop when making the stop shavings?

 

[MarcW]

Try it again on a edge of a board where you take full width shavings. This way you can verify if its your technique of overlapping the shavings of a set.

If it still happens, check the sole of the plane, espacially look at the ara close to the mouth. Use a good steel rule for this that is at least the length of the plane.

I don't know how much the bump at 4" from the beginning is. But this bump could be caused by either a flatness problem of the plane just behind the blade or can be caused by the plane offered to the wood with the toe slightly upward / heel slightly downward.

Also did the plane stop when making the stop shavings?

Hi all,

Another idea:

Often the tail vise is a bit lower than the bench surface. This could cause the board to depress under the weight of the plane and to jump up to a bump later on.

If you do not clamp in the tail vise, maybe it's a depression in the surface of the workbench...

Try against a planing stop on another spot of the bench or on a piece of MDF. If the problem maintains it probably is the plane or the technique.

David's technique is foolproof unless you make work with a plane with an uneven sole. I guess I'd check the plane's sole again with a confirmed straightedge.

I bet on the plane.

 

[adrian]

Often the tail vise is a bit lower than the bench surface. This could cause the board to depress under the weight of the plane and to jump up to a bump later on.

If you do not clamp in the tail vise, maybe it's a depression in the surface of the workbench...

Try against a planing stop on another spot of the bench or on a piece of MDF. If the problem maintains it probably is the plane or the technique.

I don't have a tail vise. The workpiece is secured in on the bench top with a bench dog at one end and the Veritas wonderdog at the other end. (The wonderdog is a clamp that fits into a dog hole.) To prevent the board from depressing under the weight of the plane, I have inserted a bunch of wooden shims. Before I deployed the shims I couldn't even get a concave surface while taking stop shavings.

David's technique is foolproof unless you make work with a plane with an uneven sole. I guess I'd check the plane's sole again with a confirmed straightedge.

I bet on the plane.

I checked the plane's sole against a Starrett straight edge. I found it to agree to within 0.0015". (The 0.001" but not the 0.0015" feeler would fit under the straight edge in various spots along the sole towards the middle, so the sole appears slightly concave.) Another observation is that there is a small movable piece right behind the blade which is 0.002" inset. (This has to do with mouth adjustment.) If this isn't good enough I'm not optimistic about my prospects. (The Clifton jointer, which I quit using, was worse, with a hollow of about 0.003".) What is the implication of having a slight hollow in the plane?

I guess I should also point out that having a 0.0015" bump is better than I ever achieved in the past. I generally considered the wood flat if I got it to better than 0.004" and I didn't worry if it was convex. I could maintain concavity by cutting 1/2" off each end of the board.

 

[MarcW]

Adrian,

A concave sole is no good. Your bump in the board is IMO a consequence of the concave sole. In order to take a consistent shaving, the area of the sole just ahead of the mouth has to make good contact to the board. A concave sole makes good contact at the start of the planing movement and at the end when the front of the sole isn't in touch with the wood anymore. This way you're taking off most wood at the ends. A bump is the result. Try to take shaving at the same thickness or less as is the hollow of the sole. The plane will shave only at both ends, not in the middle.

A slightly convexe sole is preferable to a hollow sole. So, don't hesitate to flatten the sole. It hasn't to be absolutely flat, just three areas, the tip, the spot before the mouth and the heel have to be aligned. The inset piece is of no importance as long as it doesn't protrude on the sole.

I added a picture of the hollow in the width, which I took a while back. The plane I used had a slight hollow to at the beginning, a 20 minutes flattening session improved a lot.

 

[adrian]

I have found it difficult (impossible?) to get shavings much below about 0.0015" in the past.

The one time I tried to flatten a plane I spent a few hours and made it less flat. (However, it started out convex, I believe.) I guess I'll have to reinvestigate the methods.

It does seem like trying to flatten the 25" long planes presents a problem of finding a reference surface substantially longer than 25". (Even the 15" long plane is a problem.)

 

[tnimble]

The longer your reference surface is the faster you can flatten the sole of the plane.

However a perfect flat surface is not needed as long as that surface has no twist / wind.

When your not sure about the surface and you have a convex plane use two sheet of sand paper with about 1 to 2 third the length of the plane between them.

Then take a number of passes and remeasure the hollow in the sole of the plane. Make a guesstimate on how long you should continue this.

Remeasre the plane. If the hollow in the plane gets less but moves to the toe or the heal of the plane move presure or remove one of the sheets for a while.

 

[Paul Kierstead]

Random thought of the day...
So, if one has 3 planes of the same length, they could in theory make them all perfectly flat without a reference surface, just like you can do with a straightedge

In other things, I have a #6 which is kind of ridiculously convex. It can still take a very fine shaving (actually leaves a very nice finish) and actually is pretty good at jointing too. Go figure. Concave would be a whole different ballgame though.

 

[David C]

I find that very slight convexity is not such a bad thing in a plane.

Infinitely preferable to any concavity, of either length or width.

In nice dense Sycamore, my thinnest shavings are just over 0.0005 inch. These are for curiosity only. One to 1.5 thou is sometimes useful for fine finishing of cranky timber.

Tried to do some sums on this and decided this was not far from 12 to 15 microns?

Choice of wood species is an issue and it seems the Japanese planing masters use a special slow grown timber from the north.

best wishes
David

 

[adrian]

The longer your reference surface is the faster you can flatten the sole of the plane.

However a perfect flat surface is not needed as long as that surface has no twist / wind.

When your not sure about the surface and you have a convex plane use two sheet of sand paper with about 1 to 2 third the length of the plane between them.

Then take a number of passes and remeasure the hollow in the sole of the plane. Make a guesstimate on how long you should continue this.

Remeasre the plane. If the hollow in the plane gets less but moves to the toe or the heal of the plane move presure or remove one of the sheets for a while.

Could you elaborate on this. The planes I've checked are concave, though I didn't check the side to side profile. (I no longer have the one that was greatly convex.) So if I have a 12" long reference and a 25" long plane I can use the reference to lap the plane? I didn't follow the description above with two sheets of sand paper.

If I'm trying to remove a 0.0015" hollow, what grit paper would I want to start with?

Another thing I wonder about is the tolerance of my references. I mean, I think the Starrett straight edge is flat to within 0.001", so if I observe an apparent concavity of 0.001" then perhaps it's really the straight edge that is wrong rather than the plane (?). I suppose if the goal is not to make the plane flat but to make it convex then this isn't such an issue, as long as I make it more convex than the tolerance of my reference.

 

[bugbear]

If I'm trying to remove a 0.0015" hollow, what grit paper would I want to start with?

I prefer seperating the metal removal process from the measurement process.

I have had success with scrapers, files, and abrasive (SiC and AlZi) on small blocks.

The nice thing is, you only need a reference as big as your plane, and the reference used under low loads, so doesn't need to be killer rigid.

(a 24" long sheet of glass has a LOT of flex)

More details here:

http://www.geocities.com/plybench/flatten.html

BugBear

 

[tnimble]

Could you elaborate on this. The planes I've checked are concave, though I didn't check the side to side profile. (I no longer have the one that was greatly convex.) So if I have a 12" long reference and a 25" long plane I can use the reference to lap the plane? I didn't follow the description above with two sheets of sand paper.

This is more or less the setup.

Use this only if the plane has a hollow!

Move the plane back and forth aover the two sheets. Due to the gap and the travel over it the middle of the sole is not sanded, the tips of the toe and heel have the most sanding.

If I'm trying to remove a 0.0015" hollow, what grit paper would I want to start with?

When one would look at he particle size (the maximum amount of metal that can be removed equals the particle size) between a P320 to P400 . However the plane's metal is rather hard and this would take a fair amount of time. There fore it would be sane to start with about P80 without applying much downwards pressure and move up the grids.

Another thing I wonder about is the tolerance of my references. I mean, I think the Starrett straight edge is flat to within 0.001", so if I observe an apparent concavity of 0.001" then perhaps it's really the straight edge that is wrong rather than the plane (?). I suppose if the goal is not to make the plane flat but to make it convex then this isn't such an issue, as long as I make it more convex than the tolerance of my reference.

When the plane is flat (but you measure a hollow) you can take very fine shavings. When the plane is slightly convex you can still take the same fine shavings. However there are some differences.

- The plane will stop making shavings when doing stop shavings a tad later
- The maximum amount of hollow is a bit more along the same amount of board lengh
- The plane tip of the toe and heal are a bit lifted off the board being planed, making it a bit easier to use on odd grain and not so flat or smooth boards.

Therefore it is better to go for a slight convexity than for a perfect flat.

 

[adrian]

The key picture isn't showing up.

 

[tnimble]

The key picture isn't showing up.

All fixed, somehow the browser does not support the PNG image format with the settings I used.

 

[adrian]

Could you elaborate on this. The planes I've checked are concave, though I didn't check the side to side profile. (I no longer have the one that was greatly convex.) So if I have a 12" long reference and a 25" long plane I can use the reference to lap the plane? I didn't follow the description above with two sheets of sand paper.

This is more or less the setup.

Use this only if the plane has a hollow!

Move the plane back and forth aover the two sheets. Due to the gap and the travel over it the middle of the sole is not sanded, the tips of the toe and heel have the most sanding.

This assumes implicitly a reference surface as long as the plane. If the plane is longer than the reference, it would be necessary to do one end at a time. Would that procedure be ok?

 

[adrian]

If I'm trying to remove a 0.0015" hollow, what grit paper would I want to start with?

I prefer seperating the metal removal process from the measurement process.

I have had success with scrapers, files, and abrasive (SiC and AlZi) on small blocks.

The nice thing is, you only need a reference as big as your plane, and the reference used under low loads, so doesn't need to be killer rigid.

(a 24" long sheet of glass has a LOT of flex)

More details here:

http://www.geocities.com/plybench/flatten.html

BugBear

I remember reading about this process in a set of materials by Jeff Gorman (which sadly seem to have vanished from the web).

I suppose I'd have to try this myself to get a feel for it. I also note that a flat reference the size of the plane is needed. (Are all sheets of glass guaranteed to be flat?)

But I guess another question is whether the goal is to be "flat" or to be convex. If concave is bad then specifically seeking convexity would be desirable because otherwise you're limited by the tolerances of your reference and you can't be sure that you're not actually still concave a bit. It's not clear how to use this procedure to make the sole convex.

 

[MarcW]

...This is more or less the setup.

Use this only if the plane has a hollow!
...
Move the plane back and forth aover the two sheets. Due to the gap and the travel over it the middle of the sole is not sanded, the tips of the toe and heel have the most sanding.
...

Hm,

This setup will produce a convexe sole if one doesn't stop at the right moment. And the convexity might be bigger than the former concavity. I wouldn't do it. I would lap the whole sole, i.e. put wet and dry under the whole length of the sole. The high spots are at both ends. These touch the wet and dry first. You can mark with a felt pen. The markings at both ends will disappear first. Only when the spot ahead of the mouth wears no felt marking anymore the sole should be flat enough.

I had a low angle jack plane with a convexe sole. It was very awkward to plane with it . It was not possible to take a consistent shaving over the whole length of a board. I tried to remedy and flatten the sole but I gave up. The belly was too big. What I want to say is, if the convexity is too bad, it will become a tedious job to put it right again. Pushing downwards at the same spot in the centre without rubbing the ends on the sheet is highly exhausting.

 

[bugbear]

http://www.geocities.com/plybench/flatten.html

I remember reading about this process in a set of materials by Jeff Gorman (which sadly seem to have vanished from the web).

I suppose I'd have to try this myself to get a feel for it. I also note that a flat reference the size of the plane is needed. (Are all sheets of glass guaranteed to be flat?)

Note that for the "traditional" so-called-lapping, your reference needs to be a good deal bigger than your plane (for good results), and rigid enough to stay accurate (as a reference) as you press down hard on the abrasive.

If the reference isn't rigid (under the working loads) accuracy it had under zero load means nothing.

But I guess another question is whether the goal is to be "flat" or to be convex. If concave is bad then specifically seeking convexity would be desirable because otherwise you're limited by the tolerances of your reference and you can't be sure that you're not actually still concave a bit. It's not clear how to use this procedure to make the sole convex.

Most people aim for slightly convex, because it's MUCH better than any-amount-of-concave, and they can't hit "flat". This amounts to erring on the side of caution.

This is how Lee Valley (used to?) specify their plane sole flatness; guaranteed flat within 3 thou, and never concave (so the tolerance was flat - 3 thou convex, NOT flat plus or minus 1 1/2 thou)

The approach I describe can get you to very flat indeed, and in a well controlled and predictable manner.

Attack the blue bits, re-mark, attack, repeat!

BugBear

 

[adrian]

This is how Lee Valley (used to?) specify their plane sole flatness; guaranteed flat within 3 thou, and never concave (so the tolerance was flat - 3 thou convex, NOT flat plus or minus 1 1/2 thou)

It appears that you have this backwards:

A slight correction to "concave sole" information - we make planes with flat soles, with our manufacturing tolerances only allowing concavity as a deviation from flat - as opposed a +- from flat. Then too, where you allow variance has the greatest bearing on quality. A concave variance in the sole 3" in front of the mouth will have no effect on performance - a variance at the edge of the mouth will...

Consider a corrugated sole plane - that has considerable concave variances...! Or - the back face of a japanese chisel...they still cut perfectly flat!

If the sole perimeter and a given locus around the mouth (fore and aft) is coplanar, than you'l get maximum performance. However - any amount of convexity (a bump) on sole WILL affect performance. We allow zero convexity - so what you get from us is flat to concave, and with good reason!

This is from http://www.woodworkforums.ubeaut.com.au/archive/index.php/t-14737.html


I was curious about the reference issue because tnimble wrote that it was possible to flatten a plane without a reference that is as long as the plane. Unless I can assume that any sheet of glass I stumble across is flat, I don't have a clear source of any flat reference that is as long as a jointer.