Sanded glue joint not strong?

[Jim Underwood]

I was just skimming through Tage Frid's book 2; Shaping Veneering, Finishing and ran across this statement on page 6:



"Before bending, it's a good idea to remove planer marks by sanding. An uneven surface is often what gets fibers runnin in the first place. Don't sand when you're going to laminate, because a sanded joint is not a good glue joint."


(I'm assuming when he says "bending" he's talking about steam or heat bending.)

Why is a sanded glue joint not "good"?

[Jamie Buxton]

Those Frid books were my bible when I was learning about woodworking, but I can't make any sense of that quotation. Sanded wood can be glued, just like planed wood, and just like sawn wood. Heck, it is possible to glue wood chips together to get OSB, and it stays together. The best explanation I could make is that sanding has a tendency to round the edges of the sanded face. If you've cut the laminates exactly to size, and then you round the edges just a bit before you glue the laminates together, the glue line right at the edge of the bent lamination will be a little more obvious that it could be. But I don't recommend cutting laminates exactly to size before lamination. Laminates squooge around in the clamping process. Make the laminates over-size, and trim the glued-up piece.

[Kent Chasson]

There's a lot of disagreement about that topic among some folks and the scientific answers can get very involved. I think the practical answer depends on a lot of factors like what kind of glue, what kind of joint, what grit of sandpaper, and possibly what kind of wood. For most purposes, once a joint is strong enough so that any failure is 90% wood failure, then the glue joint is fine. From what I've seen, any joint prepped with 80 grit of finer will meet that criteria. And most joint failures are caused by a long list of problems other than the smoothness of the surface.

[Chris Fournier]

It is a tough quote to decipher but I think that the author is referring to two situations, one where he suggests sanding prior to steam bending and another where he does not recommend sanding prior to bent laminating.

In the first case Tage Frid is trying to save you the trouble of removing machine marks once you have a curved surface - he recommends getting rid of the machine marks while the piece to be bent is still flat with two parallel faces and square edges. Time saver with a better outcome.

In the second instance I think that for many of us our methods of work have changed since Tage wrote his book and had he written the book today he would have said: "Don't hand sand either manually or with an ROS prior to gluing up your bent lamination because the gluing surface will no longer be truly flat and a poor glue joint will result. If however you have a drum or wide belt machine sander in your shop you can sand prior to laminating and still maintain flat surfaces suitable for gluing."

I have pretty much put words in one of the recognized "Master's" mouth and I hope that I am not soon to be struck dead by a bolt of lightning but I think that I'm on track with this one.

[Chris Friesen]

I suspect the idea is that sanding will result in "loose" fibers--they're the ones that cause fuzziness with waterborne finishes. If you have a lot of loose fibers, it's going to give a reduced-strength bond with yellow or white glues.

If you want to even out the surface before laminating you can always use a hand plane.

[John Coloccia]

I've been ignoring the this wisdom for many years. I really don't personally think it matters as long as the joint is flat and dust free. Maybe all my stuff will fall apart in 50 years. LOL....I hope not!
:D

[Lee Schierer]

I've been ignoring the this wisdom for many years. I really don't personally think it matters as long as the joint is flat and dust free. Maybe all my stuff will fall apart in 50 years. LOL....I hope not!
:D

I think John has defined the essence of the question. If you can keep the surface flat while sanding then it shouldn't make any difference to glue joint strength, But if you sand in curved surfaces then the joint gap will be larger and the joint strength will be lower.

[Kent Chasson]

When he says, "a sanded joint is not a good glue joint", I'm assuming he's talking about surface adhesion of the glue. Maybe I'm reading my own biases into it though because that is often a hot topic in my world of guitar building. Ask that question on a guitar forum if you want a long-winded, heated debate!

The debate usually centers around gluing on bridges because they are seriously stressed and fail with some regularity. The old school wisdom is that you actually need a "toothed" surface to give the glue something to mechanically grab onto. They even make "toothing irons" to groove the nice flat surface of a bridge before gluing.

Many people disagree with this line of thinking based on the fact that most glues bond chemically, not mechanically, and any toothing (or sanding scratches) make for a thicker and weaker glue line.

Again, in my experience, it's a moot point because a well executed joint will be stronger than the wood regardless of a sanded vs planed surface. And a poorly executed joint will likely fail regardless of surface prep.

[Jim Underwood]

I pretty much agree with Chris Fournier's interpretation of what Frid says.

I'd only add that he's probably recommending that you sand to remove the knife marks from the stock before hot or steam bending, because those knife marks cut through the fibers and that uneven surface is where the tearing breaks in the bend start (or "running" as he puts it).

When bending by lamination, he's recommending not to sand because it doesn't make a good joint. I could only surmise essentially what Chris Friesen says is correct- the fuzzy loose fibers in the joint cause weakness because the glue doesn't bond with the main structure of the wood.

I had thought I might buy a card scraper and touch it up. But I have no experience with this...

The other variable in my scenario is the glue. I'm going to be using Weldwood's plastic resin glue. This supposedly reduced "creep" and "springback". Perhaps it will alleviate any weakness from sanded fibers too?

[Montgomery Scott]

Do a google search on The Encyclopedia of Wood, published by the USDA. Read Page 9-3, 9-4 for the reason why planed surfaces are better than sanded surfaces.

[Chris Fournier]

Great choice!

[Jim Underwood]

When he says, "a sanded joint is not a good glue joint", I'm assuming he's talking about surface adhesion of the glue. Maybe I'm reading my own biases into it though because that is often a hot topic in my world of guitar building. Ask that question on a guitar forum if you want a long-winded, heated debate!

The debate usually centers around gluing on bridges because they are seriously stressed and fail with some regularity. The old school wisdom is that you actually need a "toothed" surface to give the glue something to mechanically grab onto. They even make "toothing irons" to groove the nice flat surface of a bridge before gluing.

Many people disagree with this line of thinking based on the fact that most glues bond chemically, not mechanically, and any toothing (or sanding scratches) make for a thicker and weaker glue line.

Again, in my experience, it's a moot point because a well executed joint will be stronger than the wood regardless of a sanded vs planed surface. And a poorly executed joint will likely fail regardless of surface prep.

Huh. Now if I could just remember at what grit I sanded those laminations? I did use the widebelt sander, but I can't remember if I was using 80 or 100 grit. Perhaps I should run them back through at 150 or 180.

I figure I'm going to be clamping these in a preformed jig with lots of clamps and glue. And hopefully enough heat. (The temp dropped before I could get the glue shipped. And no one carries this stuff locally anymore. :mad:)

So I'm hoping the sanding won't affect the bent drawer front laminations...:confused:

[Jim Underwood]

Do a google search on The Encyclopedia of Wood, published by the USDA. Read Page 9-3, 9-4 for the reason why planed surfaces are better than sanded surfaces.

I'm looking, but I can't find anything that I don't have to spend money on....

[David Prince]

My interpretation is that a lamination requires critical joints that need to be strong to maintain the integrity of the newly formed board. Care needs to be taken when gluing these joints because they are critical. A sanded joint is not a good glue joint. Glue needs surface to adhere to. Sand it smooth on both sides and you have a smooth surface. Some topography on the surface makes a better joint. Try glueing two pieces of glass together! It can be done, but probably not the greatest joint because the glue is just adhering between the surfaces. Glue two flat 1 x 6 oak boards face to face and you will get a stronger joint because the glue goes into the small crevices and "locks" itself into the topography of the board. If it is too flat and too smooth, there isn't this "locking" process.
Find a way to split the glass and you will have a sheer glaze of glue that would be smooth. Split the wood joint and it is taking wood fibers with it and would be rough.
IMHO and IMHE.

[Kent Chasson]

My interpretation is that a lamination requires critical joints that need to be strong to maintain the integrity of the newly formed board. Care needs to be taken when gluing these joints because they are critical. A sanded joint is not a good glue joint. Glue needs surface to adhere to. Sand it smooth on both sides and you have a smooth surface. Some topography on the surface makes a better joint. Try glueing two pieces of glass together! It can be done, but probably not the greatest joint because the glue is just adhering between the surfaces. Glue two flat 1 x 6 oak boards face to face and you will get a stronger joint because the glue goes into the small crevices and "locks" itself into the topography of the board. If it is too flat and too smooth, there isn't this "locking" process.
Find a way to split the glass and you will have a sheer glaze of glue that would be smooth. Split the wood joint and it is taking wood fibers with it and would be rough.
IMHO and IMHE.

Wood and glass are very different materials and don't necessarily make a good analogy. This link should get you to the pages referred to by Montgomery Scott.

http://books.google.com/books?id=mUGSaiTsBAIC&pg=SA9-PA2&dq=encyclopedia+of+wood+usda&source=gbs_toc_r&cad=4#v=onepage&q=encyclopedia%20of%20wood%20usda&f=false

If not, do a "book" search in google and you can read the whole thing.

It gives a better explanation of the chemical bonding process and why smooth tends to be better where wood is concerned. Again, the basic idea is that furniture glues bond chemically, not mechanically, and the best way to get a good joint is to have relatively smooth, well mated, freshly surfaced areas. In my experience, 80 grit or finer meets that criteria.

[Jim Underwood]

Ok.... Thanks for the link.. I think. :confused: Can you translate that for me? Wow. That's a lot more technical than I want to get into at the moment. Anyway, I sorta got the gist of it. They recommend a well cut joint as opposed to a sanded joint, because the glue contacts the wood better and bonds at the molecular level...

[Kent Chasson]

..... Can you translate that for me?.....

Official translation: "Don't sweat it". A well mated joint with fresh surfaces and fresh glue sanded to 80 grit or finer, scraped, planed, or even sawn with the right blade and technique will be fine for just about anything.

[John Coloccia]

Yeah, it seemed like they were talking about "abrasive planing" and they hinted that finer sanding may or may not have the same effect. It would be nice to see some data about how much weaker the joint is because it kind of looks like they just looked at the joint and assumed it should be weaker.

Maybe I'll just test it myself tonight and see what I end up with. I'm really interested in knowing the answer to this, now.

[Chris Friesen]

Try the "Wood Handbook, Wood as an Engineering Material" by the Forest Products Laboratory of the USDA. Specifically, "Chapter 10: Adhesives with Wood Materials- Bond Formation and Performance" at http://www.fpl.fs.fed.us/documnts/fplgtr/fplgtr190/chapter_10.pdf

Some selections:

"Experience and testing have proven that a smooth, knife-cut surface is best for bonding."

"Dull cutting edges of planer or jointer knives crush and burnish the cells on the wood surface. Not only are these cells weaker, they also inhibit adhesive wetting and penetration."

"Abrasive planing with grit sizes from 24 to 60 causes surface and subsurface crushing of wood cells. The adhesive industry typically recommends 60–80-grit sanding as acceptable for wood bonding as this equates to 24 to 30 knife marks per inch when planing. Generally, anything above 200 grit fuzzes the wood surface and is not recommended."

"If abrasive planing is to be used before bonding, belts must be kept clean and sharp, and sanding dust must be removed completely from the surface. However, abrasive planing is not recommended for structural joints that will be subjected to high swelling and shrinkage stresses from water soaking and drying."

[Chris Padilla]

"Experience and testing have proven that a smooth, knife-cut surface is best for bonding."

"Dull cutting edges of planer or jointer knives crush and burnish the cells on the wood surface. Not only are these cells weaker, they also inhibit adhesive wetting and penetration."

This jives perfectly with my intuition and how I think about the surface of wood at the microscopic level.

Besides, don't most of us use freshly jointed edges to glue up small boards into larger table tops? That has been going on for a long, long time.

[Pat Barry]

What he is saying is that glue itself is not very strong. You are relying on the bond line formed with the wood itself. Take away some bond line by sanding in deeper grooves that you fill with a less strong material (glue), and you have made a weaker joint. I think this makes perfect sense - the thinnest glue line is the best. It on ly makes sense that the planed, smooth surface will be better. Would you take a freshly jointed pair of edges and sand them before gluing them into a panel?? I think not...