Center Joint gap

[David Burgess]

36 minutes ago, Nestorvass said:

You probably mean the back plate the top plate had no gaps. I will upload the photos again hopefully they will be visible 

I would cut or soak that one apart, let it dry thoroughly, and then re-do it. Not everything goes perfectly on the first try, even for the most experienced in our profession.

[Guest]

4 minutes ago, Bill Yacey said:

For maple, I don't use a rub joint. I use 3 pipe clamps, one about 1 inch in from each end, and one in the center. I apply less pressure to the center clamp, and more pressure on the ends.

Why not? Is it because of strength? If that is the reason I can guarantee you the rub joint is extremely strong. I tried to take it apart today as I said above I put a lot of hot water in the joint and I heated both with hair dryer and heat gun. It wouldn't come apart not even move. It was rock solid despite all the heat I applied.  I had to use a bandsaw to take it apart.

[David Burgess]

8 minutes ago, Nestorvass said:

Why not? Is it because of strength? If that is the reason I can guarantee you the rub joint is extremely strong. I tried to take it apart today as I said above I put a lot of hot water in the joint and I heated both with hair dryer and heat gun. It wouldn't come apart not even move. It was rock solid despite all the heat I applied.  I had to use a bandsaw to take it apart.

You'd probably need to keep the joint moist for at least a couple of days to turn the glue back into a gel all the way through, and get it to release.

Heat doesn't seem to do much for releasing hot hide glue joints. That's used more for Titebond and epoxy joints.

 

[Guest]

13 minutes ago, David Burgess said:

I would cut or soak that one apart, let it dry thoroughly, and then re-do it. Not everything goes perfectly on the first try, even for the most experienced in our profession.

Thank you for the advice. I decided to cut the joint using the bandsaw. Then used a no. 5 to remove the saw marks and then the wooden jointer upside down to make the two joints. I used glue to size the joint and I will leave it until tomorrow. There is still a hairline gap at the end but I will fix it when the glue dries. I still have to decide whether I need to take another shaving in the jointer or use a finely set block plane to remove only a tiny amount of wood as well as the excess glue used in sizing. I also need to take a shaving or two in the middle to make the joint slightly concave.

[Guest]

2 minutes ago, David Burgess said:

You'd probably need to keep the joint moist for at least a couple of days to turn the glue back into a gel all the way through, and get it to release.

Heat doesn't seem to do much for releasing hot hide glue joints. That's used more for Titebond and epoxy joints.

 

Yes you are right. It just wouldn't come apart I was practically leaning on the joint. Only the outside part started to seperate a bit. The rest was practically untouched. So I had to use the saw. I probably should have gone for the saw in the first place as I am not a huge fan of making a perfectly dried wood wet. But oh well we live and learn  it is my first instrument and I expect to make many more mistakes in the process from which I will learn.

[Bill Yacey]

58 minutes ago, Nestorvass said:

Why not? Is it because of strength? If that is the reason I can guarantee you the rub joint is extremely strong. I tried to take it apart today as I said above I put a lot of hot water in the joint and I heated both with hair dryer and heat gun. It wouldn't come apart not even move. It was rock solid despite all the heat I applied.  I had to use a bandsaw to take it apart.

Maple seems to swell more in the center causing the ends to open up, or maybe the glue shrinks faster on the ends than in the center; I'm not sure exactly what the cause is, but it has happened enough that I just clamp everything up and not leave anything to chance.

[Don Noon]

3 hours ago, Nestorvass said:

... i followed your advice and I sized the joint with glue. I will let it dry until tomorrow and then, I will do a pass on the jointer plane to take the glue off and flatten it again. Then I will take my block plane and take a few thin shavings in the middle to make it slightly concave. I have my block plane set up with a 45 degree bevel on the iron and it takes shavings which are about 10 to 20  microns thin.

If you plane completely through the glue line, then the effect of the glue sizing is gone.  Only take off what will get it to flat, and no more.  I would use a scraper, and see how the joint looks. I do violas this way, without the concave, and haven't had the ends open up.

1 hour ago, Bill Yacey said:

Maple seems to swell more in the center causing the ends to open up, or maybe the glue shrinks faster on the ends than in the center; I'm not sure exactly what the cause is, but it has happened enough that I just clamp everything up and not leave anything to chance.

I don't think it swells in the center... the damp edge expands in all directions, including longitudinally.  The rest of the plate doesn't expand, so it has to bend slightly, making the joint line convex.  Spruce is far more stable along the grain with humidity and moisture, so you don't get it as badly if at all.  I imagine that highly figured, flatsawn maple (where the flame squiggles laterally) would be the worst, on several accounts.

[HoGo]

3 hours ago, Nestorvass said:

I don't see how clamping could improve such a strong joint.

You're not looking in the right direction. The joint must be not only strong but also closed all along the plate and that's what the clamps are for... You CAN do the rubbed joint without clamps but that requires lots of experience. There are many things that can go wrong even holding the one half in vice can deform it enough so the joint won't close perfectly. Experienced worker has feel for these details and will get good results. The clamps are there not to squeeze the s#!t out of the joint but to hold it closed while the glue dries. You must be aware of all the forces you are inducing. I would suggest cutting "steps" in the plates at the very ends some 3 cm ffrom joint so you can use simple C clamps at he ends and not clamp on thin edges of plate. These will ensure the ends are well closed. You can add one light clamp in the center (and you can cut out wood in the bout for that as well). I did this for bass top I was joining years ago and often do it for mandolin plates as well.

[Bodacious Cowboy]

Many roads lead to Rome. 

[Guest]

14 minutes ago, HoGo said:

You're not looking in the right direction. The joint must be not only strong but also closed all along the plate and that's what the clamps are for... You CAN do the rubbed joint without clamps but that requires lots of experience. There are many things that can go wrong even holding the one half in vice can deform it enough so the joint won't close perfectly. Experienced worker has feel for these details and will get good results. The clamps are there not to squeeze the s#!t out of the joint but to hold it closed while the glue dries. You must be aware of all the forces you are inducing. I would suggest cutting "steps" in the plates at the very ends some 3 cm ffrom joint so you can use simple C clamps at he ends and not clamp on thin edges of plate. These will ensure the ends are well closed. You can add one light clamp in the center (and you can cut out wood in the bout for that as well). I did this for bass top I was joining years ago and often do it for mandolin plates as well.

There are people who believe that a rub joint isn't sufficient enough for a strong joint. I just pointed out that they are wrong for reasons I explained above. A clamp on the ends is not a bad idea however if a joint is carefully made I don't see a need for clamps, which if not carefully placed could open a joint instead of closing it. For the spruce the rub joint worked perfect I have absolutely no gaps they are barely visible under my stereoscopic microscope. The reason why I got gaps in the back plate is probably because I didn't size the joint and because I used a block plane at the end. That could have caused a small gap. Anyway I will redo the joint, but I am convinced a rub joint is the way to go. At least for me.

[Davide Sora]

51 minutes ago, Nestorvass said:

There are people who believe that a rub joint isn't sufficient enough for a strong joint. I just pointed out that they are wrong for reasons I explained above. A clamp on the ends is not a bad idea however if a joint is carefully made I don't see a need for clamps, which if not carefully placed could open a joint instead of closing it. For the spruce the rub joint worked perfect I have absolutely no gaps they are barely visible under my stereoscopic microscope. The reason why I got gaps in the back plate is probably because I didn't size the joint and because I used a block plane at the end. That could have caused a small gap. Anyway I will redo the joint, but I am convinced a rub joint is the way to go. At least for me.

Now I can see the photos, thanks.

I think rub joint can work for maple too, but I don't think it can work with a perfectly straight joint and no light passing through at all. The result will probably be 9 times out of 10 what happened to you, caused by the deformation (bending) of the pieces due to the moisture absorbed. Things will probably go better with the glue sizing, but in my opinion it may not be enough. I think there must be a very very slight concavity in the two pieces, with very very little visible light passing through, so that it becomes straight with moisture absorption without causing the ends to rise. It is impossible to quantify these things, but with a little practice it could be done.

In any case, I prefer to use clamps, because if you can compensate for the longitudinal deformation you can not do anything for a possible helical deformation (twist), which not too rarely occur with very flamed woods, and I don't like surprises.

PS All this assuming a perfectly planed joint, if there are slight bumps just before the ends (or in other places) it is sure they will open and even the clamps could not solve the problem. To be sure that there are no bumps, you need to check with a short straightedge (10 cm or less) the entire length of both pieces.

[Guest]

9 minutes ago, Davide Sora said:

Now I can see the photos, thanks.

I think rub joint can work for maple too, but I don't think it can work with a perfectly straight joint and no light passing through at all. The result will probably be 9 times out of 10 what happened to you, caused by the deformation (bending) of the pieces due to the moisture absorbed. Things will probably go better with the glue sizing, but in my opinion it may not be enough. I think there must be a very very slight concavity in the two pieces, with very very little visible light passing through, so that it becomes straight with moisture absorption without causing the ends to rise. It is impossible to quantify these things, but with a little practice it could be done.

In any case, I prefer to use clamps, because if you can compensate for the longitudinal deformation you can not do anything for a possible helical deformation (twist), which not too rarely occur with very flamed woods, and I don't like surprises.

Thank you for the advice Mr. Sora I've been following your videos to make my  violin. I have the art of violin making book but to be honest with you I didn't find a need to use it. Your videos cover all I need, so thank you for those as well. I would use clamps but the ones i have are very cheap like 5 euros a piece. They are too flexible and dont apply the force perpendicularly to the joint (they are large f clamps not parallel jaw clamps). So if i where to use them i believe they would create even more gaps. Thus rub joint is my only choice. If it turns out bad i will cut it open again and redo it. Hopefully I will succeed within the first ten tries  If not there will be no material left.

A sidenote:

I think i need to invest in a new set of clamps. I saw the video where you glued the ribs to the blocks and you made it look very easy. With the clamps i have and the lack of experience as well it took me 30 minutes just to glue the upper bout ribs. They would slide and leave gaps between them and the mold, as i tightened the clamps they moved out of position. Though i believe its mainly the clamps fault. They are too cheap and flexible

[Don Noon]

1 hour ago, Davide Sora said:

I think rub joint can work for maple too, but I don't think it can work with a perfectly straight joint and no light passing through at all. The result will probably be 9 times out of 10 what happened to you, caused by the deformation (bending) of the pieces due to the moisture absorbed. Things will probably go better with the glue sizing, but in my opinion it may not be enough.

This voice of authority scared be a bit, so I checked all 21 joined maple plates that I have in stock.  No gaps.  All of mine were planed dead-flat, no light passing through, and glue sized.  However, all of my wood is torrefied, and not as prone to expansion with moisture... so that would help.

It's probably just a theoretical concern I have about planing a concave, but when the glue totally dries out, there's a residual stress attempting to pull the joint apart in the middle.  Admittedly, this would be reduced a lot after the plates are carved.

[Dennis J]

The obvious cause of the end gaps is that one or both boards were not planed flat. And checking for gaps by holding two planed edges together in front of a bright light will not necessarily show that gap, especially if you are holding the boards off centre. It is difficult to check both ends at once.

To plane such a long narrow edge flat is, not surprisingly, problematic. In my experience the only way to get a perfect fit without using clamps is to use a straight edge to check each planed edge. I've found that planing each board slightly hollow using a smaller plane and then finishing with one long stroke is the best way to go.

And I agree that a rubbed joint should be all that is necessary.

 

 

[Guest]

4 hours ago, Dennis J said:

The obvious cause of the end gaps is that one or both boards were not planed flat. And checking for gaps by holding two planed edges together in front of a bright light will not necessarily show that gap, especially if you are holding the boards off centre. It is difficult to check both ends at once.

To plane such a long narrow edge flat is, not surprisingly, problematic. In my experience the only way to get a perfect fit without using clamps is to use a straight edge to check each planed edge. I've found that planing each board slightly hollow using a smaller plane and then finishing with one long stroke is the best way to go.

And I agree that a rubbed joint should be all that is necessary.

 

 

I used a straight edge as well. I have a machinists straight edge . In my opinion the probably was a twist on the board leaving low spots on the opposite ends of both the front and the back side of the plate.  For that reason a set of winding sticks could be more useful than straight edge in my opinion, as these are the tools that are traditionally used to check for twists. My only concern, about whether they would work is that they are typically used to check for twists on the faces of boards and not the edges since the reference surface would be too small at the edges. The reason I believe there was a twist is because the gaps didn't extend all the way to other face of the plates.

[Davide Sora]

8 hours ago, Don Noon said:

This voice of authority scared be a bit, so I checked all 21 joined maple plates that I have in stock.  No gaps.  All of mine were planed dead-flat, no light passing through, and glue sized.  However, all of my wood is torrefied, and not as prone to expansion with moisture... so that would help.

It's probably just a theoretical concern I have about planing a concave, but when the glue totally dries out, there's a residual stress attempting to pull the joint apart in the middle.  Admittedly, this would be reduced a lot after the plates are carved.

 

1 hour ago, Nestorvass said:

I used a straight edge as well. I have a machinists straight edge . In my opinion the probably was a twist on the board leaving low spots on the opposite ends of both the front and the back side of the plate.  For that reason a set of winding sticks could be more useful than straight edge in my opinion, as these are the tools that are traditionally used to check for twists. My only concern, about whether they would work is that they are typically used to check for twists on the faces of boards and not the edges since the reference surface would be too small at the edges. The reason I believe there was a twist is because the gaps didn't extend all the way to other face of the plates.

The problem is that defining the amount of concavity or perfect flatness is rather difficult because it is not measurable (not with the tools available to me and being above all wood), so what I define as very light concavity could also be interpreted as almost dead flat by many. The problem of residual stress is only there if you leave a consistent gap (but how do you interpret consistency?) and this is the risk when clamps are used putting too much pressure to close a too consistent gap. This risk does not exist with rub joints, but the risk caused by the twist remains. To check for twists you could use the long straightedge diagonally on opposite corners, but I think the best and most accurate way is to check the two pieces against each other and try to perceive if they "rock". It is necessary to develop a certain sensitivity to perceive even the smallest rocking motion, and also to correctly perceive the sensation of "firmness" of the two pieces when the twist is absent.

[Dennis J]

1 hour ago, Nestorvass said:

I used a straight edge as well. I have a machinists straight edge . In my opinion the probably was a twist on the board leaving low spots on the opposite ends of both the front and the back side of the plate.  For that reason a set of winding sticks could be more useful than straight edge in my opinion, as these are the tools that are traditionally used to check for twists. My only concern, about whether they would work is that they are typically used to check for twists on the faces of boards and not the edges since the reference surface would be too small at the edges. The reason I believe there was a twist is because the gaps didn't extend all the way to other face of the plates.

I doubt that winding sticks would be of much use with such a small gap. I've come to the conclusion that a number 4 low angle block plane is about the best plane to use as long as it has been sharpened at a high enough angle to avoid tearout. Its main advantage is its low centre of gravity. I've found that planes tend to cut more deeply at the last few centimetres of the pass. I don't know why that is but I've found that to be the case by routinely using a straight edge when checking for flatness.

 

[Guest]

1 hour ago, Dennis J said:

I doubt that winding sticks would be of much use with such a small gap. I've come to the conclusion that a number 4 low angle block plane is about the best plane to use as long as it has been sharpened at a high enough angle to avoid tearout. Its main advantage is its low centre of gravity. I've found that planes tend to cut more deeply at the last few centimetres of the pass. I don't know why that is but I've found that to be the case by routinely using a straight edge when checking for flatness.

 

Yes planes have a tendency to make surfaces slightly convex. So to counteract that more pressure should be added in the middle. The way Christopher Schwarz describes it who is not a luthier but an authority in fine woodworking, is imagine the wood is icecream you short of push the plane down as you go into the middle like you would a spoon if the wood was  ice cream. Silly metaphor but  quite intuitive. Also what is a number 4 block plane? Theres a number 4 bench plane and a block plane which is another plane held with the palm. Also what is the point of using a low angle block plane and put a high angle microbevel when you can use a normal angle block plane with more acute bevel?  Unless you have another blade with a more accute angle to use for trimming end grain such as the blocks, i dont see a reason to own a low angle plane at all. On the contrary most of the planing work on violin is on maple and ebony which are wood species very prone to tear out, which benefit for a very high angle block plane

[HoGo]

How did you hold the plate during rubbing? I prefer rather small vice with 10cm wide soft rubber lined jaws that don't tend to cause twist in the half held but I always check the fit with one half in the vice. No rocking or any sign of gap. I glue much like Mr. Sora and add clamps. I like to do a very light sweep of sharp cabinet scraper along maple joint before gluing and apply tiny bit more pressure in the center. This cleans the surface for succesful joint and perhaps adds some of the concavity mentioned above. I use standard (~190 strength) hide glue that doesn't contain as much water as the higher grades when prepared. I test offcuts from ends for joint strength - hit them with hammer and see if the glue separates or it just breaks through wood.

[Guest]

7 minutes ago, HoGo said:

How did you hold the plate during rubbing? I prefer rather small vice with 10cm wide soft rubber lined jaws that don't tend to cause twist in the half held but I always check the fit with one half in the vice. No rocking or any sign of gap. I glue much like Mr. Sora and add clamps. I like to do a very light sweep of sharp cabinet scraper along maple joint before gluing and apply tiny bit more pressure in the center. This cleans the surface for succesful joint and perhaps adds some of the concavity mentioned above. I use standard (~190 strength) hide glue that doesn't contain as much water as the higher grades when prepared. I test offcuts from ends for joint strength - hit them with hammer and see if the glue separates or it just breaks through wood.

I have a european workbench i put the one side of the plate between the jaws of my tail vise as well as some rubbery material to prevent it from slipping. I did that because i barely tightened the vise, in order to prevent deformation. Also I put a stool under the wood held in the vise, because the vise barely held the piece. 

[Peter K-G]

Curly high dens maple can be quite strong and the glue joint breaks first even if strong enough.

[Mark Norfleet]

14 hours ago, Nestorvass said:

There are people who believe that a rub joint isn't sufficient enough for a strong joint. I just pointed out that they are wrong for reasons I explained above. A clamp on the ends is not a bad idea however if a joint is carefully made I don't see a need for clamps, which if not carefully placed could open a joint instead of closing it. For the spruce the rub joint worked perfect I have absolutely no gaps they are barely visible under my stereoscopic microscope. The reason why I got gaps in the back plate is probably because I didn't size the joint and because I used a block plane at the end. That could have caused a small gap. Anyway I will redo the joint, but I am convinced a rub joint is the way to go. At least for me.

Just for fun...  When you're redoing the joint and after you've gotten it perfect, set it aside for a few hours and check it again.  You may well find that the joint is no longer perfect, even without the effect that glue will have on the joint surfaces.   Producing a joint with no visible gaps doesn't mean that there aren't any and that all is well once things are glued.  I like to have the insurance of thoughtfully applied clamps.  Experience is valuable.
I find that it's a good idea to try to prove myself wrong once in a while...

 

[David Burgess]

8 hours ago, Nestorvass said:

I have a european workbench i put the one side of the plate between the jaws of my tail vise as well as some rubbery material to prevent it from slipping. I did that because i barely tightened the vise, in order to prevent deformation. Also I put a stool under the wood held in the vise, because the vise barely held the piece. 

That's an excellent method for supporting each plate half when planing the joint too, since it's a method of clamping and supporting which doesn't distort the plate.

[Bodacious Cowboy]

2 hours ago, Nestorvass said:

 Also what is the point of using a low angle block plane and put a high angle microbevel when you can use a normal angle block plane with more acute bevel?  Unless you have another blade with a more accute angle to use for trimming end grain such as the blocks, i dont see a reason to own a low angle plane at all. 

You seem to answer your own question. A low angle block plane is nice (although not absolutely essential) for trimming endgrain, and you can instantly turn it into a high angle plane for ebony etc simply by fitting a spare blade with a high bevel angle.

It's not so straightforward to go the other way (low bevel angle blade in a high angle plane), because the edge of the blade usually becomes too weak at bevel angles much less than 25 degrees. (you can actually get round this by also putting a bevel on the back side of the blade, but that's too much of a pain in the ass for me).

[Don Noon]

5 hours ago, Nestorvass said:

I used a straight edge as well. I have a machinists straight edge . In my opinion the probably was a twist on the board leaving low spots on the opposite ends of both the front and the back side of the plate.  For that reason a set of winding sticks could be more useful than straight edge in my opinion, as these are the tools that are traditionally used to check for twists. 

Granite surface plates are the tool for the job.  Check against the plate with chalk, and it will show you everything you want to know... twist, ripple, concave.  With a feeler gage, you can even measure how much concave you have, if that's what you want.  Surface plates aren't terribly expensive (although shipping can be), and they are astonishingly flat.  Every shop should have one.