grain match in glue up?

[engineerone]

so if you are a cheap skate and you want to finish a project by gluing up some offies to make a wider panel how do you match them???

the pieces i have are different widths, and i have six to get the width i want. they will be finished off to 18mm deep after glue up.

but the question is do you match widths, or the patterns, since you are actually going to alternate the grain??

paul

[jfc]

How can you match widths if they are all different

[jasonb]

Match for grain/looks. You want to get it to look like one piece of wood unless you are painting it

Jason

[paulchapman]

Arrange the boards as best you can so that it "looks right" while ensuring that the grain alternates. Also try to get it so that the grain goes in the same direction along the boards so that you don't get too much tear-out when finish-planing the boards - although you could always use your scraper plane if the best visual effect results in the grain going up-hill on one board and down-hill on the next.

Cheers

Paul

[engineerone]

trust you to ask the difficult questions ;D

what i meant was, do you put two thin bits together, and then two fat bits, or do you like up the grains and not worry??

thing is a know i am going to get glue lines but will the grain
minimise or excerbate???

paul

[mrgrimsdale]

The only important detail is to put the widest boards on the outside edges. It kinda looks more 'purposeful' or 'settled' in some mysterious way. If you have a thin board at the edge it looks stuck on as an afterthought.
Otherwise I'd just turn em over trial and error to see what looks neatest. Doubt anything will match in a meaningful way, perhaps better to do the opposite and have alternate pieces with grain in opposite directions, as Paul suggests.

cheers
Jacob

[blacksheep]

I just arrange them so that it looks right - agree with Mr Grimsdale on not putting thin pieces at the edge.

At the risk of blaspheming, I don't worry too much about 'alternating the grain'. If the boards are properly acclimatised before planing it is unlikley that they will cup any more. For me, the aesthetics of the glued up board are more important than slavishly following the 'alternating' rule.

[wizer]

Paul, I think your talking about thickness here, aren't you? Not board width?

[craigmarshall]

I agree with the comment about not worrying too much about alternating the growth rings. Sam Maloof doesn't bother, he goes for best visual effect, and he says he's never had sigificant problems with boards moving in that way. I think I would do the same - go for the best look, whatever looks "right" as someone else mentioned. However, you have to make sure you're using well seasoned stuff, not the damp 3" thick stuff you cut down last month...

It also depends what you're doing with the board. If it's going to be dovetailed into another board as a drawer side or something, then the joint is going to help hold the board straight. Same with a panel inside a frame and panel door. If it's a single board cupboard door or a small table top only held on with buttons, you're more likely to have a problem.

Cheers,
Craig

[engineerone]

interesting comments and mainly in line with what i had thought.

however i will fess up the reason for asking is i am using cherry, and do not want to stain it too dark, so am looking for a sensible solution to whether or not the cherry will discolour in the same way>

it will be part of a table top, but this jointed piece will only be about 6 inches wide.

now i have to go back and make my scraper plane work properly ;D

i do agree with jacob the thinner pieces will look better on the inside.

paul

[engineerone]

ok next part of the process.

do you get the pieces down to almost the finished thickness before you do the glue up. or do you reduce it after glue up

paul

[jfc]

I go for finished thickness and then a belt sander .

[paulchapman]

Apr 16, 2008 22:35:34 GMT engineerone said:

do you get the pieces down to almost the finished thickness before you do the glue up. or do you reduce it after glue up

Depends how you join the boards. If you use biscuits or loose tongues (I always do), and can be sure that the boards will line up nicely, you can go close to finished size. However, if you are just going to glue the boards, it depends how confident you are that you will get them to line up accurately and they won't slide about when you tighten up the cramps.

Incidentally, if you use shaped cramping blocks it helps to direct the cramping pressure to the centre of the edge and counteracts the effect of the cramps bending slightly, like this



Either way, the top isn't that big so why not leave a bit extra to be on the safe side

Cheers

Paul

[craigmarshall]

That looks like a big bit of sapwood in the middle of your glue-up there paul? Or do my eyes deceive me?

Craig

[blacksheep]

I use Plano clamps which enable me to use boards of pretty well the finished thickness. A couple of passes though the drum sander and the job's done.

[paulchapman]

Apr 17, 2008 6:33:47 GMT craigmarshall said:

That looks like a big bit of sapwood in the middle of your glue-up there paul?

Yes, but it was OK

Cheers

Paul

[Sgian Dubh]

Apr 15, 2008 17:52:45 GMT engineerone said:

if you are a cheap skate and you want ... a wider panel how do you match them???

... do you match widths, or the patterns ...

since you are actually going to alternate the grain??
paul

Alternating the grain Paul? Not necessarily so. See below. Slainte.

7.6 Built Up Laminations.

Making furniture throws up all sorts of interesting construction quandaries for the furniture maker. For instance the dimensional changes and distortions that are sometimes barely noticeable or important in a single piece of wood can lead to unsatisfactory results when two or more pieces of wood are joined. Solid wood panels made of several pieces occasionally develop ridges at the glue lines. There is more than one cause of this, one of which is glue ‘creep’*. Although creep can be the cause another significant explanation is mis-matched end grain patterns in the arrangement of the boards- see figure 32 on the next page.

* Creep. Some glues remain slightly elastic even after their full cure. Polyvinyl acetate (PVA) glue is one of the classic ‘creepers’. Under stress the glue line gives and parts move out of alignment. For example, bent wood laminations glued with PVA tend to straighten out long after their removal from the bending form. Where the bend is slight and the stress on the glue line also similarly small the springback and long term creeping is often insignificant. Severe bends where stress occurs in both the glue line and laminations themselves show greater tendency to ‘springback’ or straighten. Another form of creep involves moisture cycling, again with PVA glue. In this second version both the wood and the glue in a joint swell as they take on moisture. The wood later shrinks as it dries but the glue doesn’t and sits above the joint line. Where this happens to a solid wood glued up table top the result is a line of small pimples along the glue line that is just perceptible to the finger tips although, quite often, the pimples are also visible. I’ve occasionally seen the ‘row of pimples’ fault where rather wet planks, 13% MC or so, were edge joined to make a panel which later shrank as the wood adapted to internal house conditions by drying to roughly 8% MC. Glues that never really cure hard and always remain a little elastic allow a little movement in the joint. PVA glues, being water based and affected by humidity sometimes allow adjacent planks to shift a little as the seasons change.



Figure 32, above. At 1 three boards make up a panel. From left to right the three boards are radially cut at left, a mixture of rift cut and radially cut in the centre board, and tangentially cut at right. At 2 arrows indicate the main direction of the end grain movement along with the resultant size and end sections of the individual boards. The sketch rather over-emphasises the movement and shrinkage but the indicated steps between each plank do occur. Sometimes the steps barely register to the touch; at others they are clearly visible.



Figure 33, above. Radially cut planks arranged in different ways along with the end result caused by differential shrinkage. At top the heart of one plank joins the bark edge of the next resulting in steps at the joint line. Below, alternating heart-to-heart and bark side to bark side edges reduces or eliminates the steps.

7.7 Panel Glue Ups

At some point almost every woodworker faces the need to make up wide panels out of narrow timber. There are several good reasons for edge joining narrow boards to make wide slabs.


• The selected timber is only available in narrow planks.

• The machinery available can only cope with modest sizes, assuming of course that stock is prepared with machinery rather than hand tools.

• It may be that cutting wide pieces into smaller sections and rearranging the order uses the grain more attractively.

• Edge joining narrow planks sometimes provides a means for controlling, predicting, and to some extent reducing the effects of warping.

• Preparing a long and/or wide board to a specified thickness might be impossible, ie, getting a warped piece flat where both sides are parallel leaves the material too thin for the job.


How you should arrange the planks is often a subject for heated debate. I use all the layouts illustrated in figures, 39, 40 and 41, and it depends on the circumstance. For instance, the most attractive grain might all be on the heart side of tangentially cut boards, and if I’m making a table top I will use the material this way, ie, figure 40. For a table top arranged thus I usually incorporate a heavy slot screwed bearer on the underside. Similarly, if it’s a cabinet side that is housed across the grain (dadoed in the US) to accept a shelf or similar I often choose the same arrangement because then, at worst ,I only have a wide cupped piece to flatten prior to running over it with the router. This is often preferable to dealing with a series of ripples. I tend to use the layout in Figure 39 most commonly when the strips are narrow, say 50 mm or less, or in cabinets where opposite cabinet sides are linked with tenoned or dovetailed rails.

It’s also a useful option to try and arrange all the grain so that it rises towards one end when viewing the panel from the narrow edge, although this is often not possible because the grain orientation varies so much. If successfully achieved, this results in reduced tear-out when planing the final glue up, whether by hand or machine, ie, you can plane in one direction without tearout. Sometimes sacrificing this consideration over appearance is a valid choice.

There is always the thorny issue of what to do with very wide planks. Should they be ripped into narrower strips and re-glued, or left as one whole wide piece? I’ve never found any hard and fast rules that suits every case. The nature of the job determines the decision. Take, for example, a very wide plank, say 20” (500 mm) wide with striking grain pattern and other figuring, perhaps kiln dried to 7% or 8%. If it’s remained stable and flat during storage since then in RH varying between about 40% and 60% there is no particular reason to think it will become unstable as a furniture part and put into service in a typical house. With suitable cross bracing and means to hold it flat in the structure a viable choice might be to keep it whole with no ripping into narrower planks and re-gluing. Without large machinery to do the flattening and squaring this decision could result in a lot of hand work with planes, but hand work is sometimes very much worth the effort.

Where I have to join up narrow planks to make up wider panels I tend to use random widths. I willingly glue up stuff 175 mm (7”) wide and wider if I am able to incorporate suitable bracing. However, if I decide that I want the boards to be around, for example, 80 mm (3-1/8”) wide each, I try to keep them all about 15 to 20 mm (5/8” to 7/8”) either side of this figure, i.e., 60-100 mm (2-3/8” to 3-7/8”.) This tactic prevents just one board standing out as markedly different from the rest, and I happen to like a little irregularity anyway.

There are times when I machine stuff all to exactly the same width. This applies particularly when making a cabinet in which the top does not overhang the side to disguise any offset join lines. Offset join lines that run around the side to the top, and back down the other side in a cabinet with mitred corners are usually visually disconcerting.


Figure 39, above. Joining tangentially, or flat sawn planks with the end grain arranged as at 1 results in a washboard effect as at 2 which is an exaggerated approximation.


Figure 40, above. Cupping of panel in one direction due to arranging the bark side of the growth rings all towards one face.


Figure 41, above. Joining quarter sawn planks to make up a panel. Typical movement in quarter sawn stock where the material merely become a little thinner because of the even shrinkage experienced.

7.8 Edge Joinery.

When joining narrow boards to make up wider slabs the traditional approach is the plain edge or butt joint. This is a good place to start because it helps with understanding other methods. It isn’t really an easy joint to prepare, and like every joint, proper preparation means reduced risk of failure. The idea is to prepare two edges that abut perfectly all along the length so that no pressure is required to close a gap. In practice, this is difficult to achieve and most workers aim for a ‘sprung’ joint.

The sprung joint has an evenly graduating gap produced at the centre point of the joint line of the two boards. After preparation very gentle hand pressure should be enough to close up two adjacent boards. When cramping up several well prepared boards in one go it is sometimes possible to close up all the gaps with one centrally located cramp.

The sprung joint helps counteract the comparatively volatile width expansion and contraction of the last 50- 100 mm (2- 4”) of a plank’s length. The end grain of timber absorbs and gives up moisture the most rapidly. During glue up of the sprung joint, the ends are closed just a little tighter than the middle, and this helps to compensate for the disparate movement.

In addition to this, custom and practise show that where the two joined edges are convex at glue-up the joint usually fails fairly quickly. In this situation bringing the edges together requires firm clamping pressure at both ends to close the gap. The joint will close, but both ends of the joint are under tensile stress and want to open up. As soon as a slight opening begins at the very extremity of the joint it’s my experience that serious failure soon follows-- a gap anything up to a few inches long extending towards the middle of the length develops in just days. The worst case is complete failure of the joint.


Figure 42, above. A sprung joint. The gap is somewhat exaggerated here for clarity. When joining boards up to about 800 mm (32”) long the ideal gap at the mid-point should be about the same thickness as perhaps a cigarette paper. Longer edge joints tolerate a slightly wider gap. The ideal gap is one that will close with little or no more pressure than you’d need to hold something like a downward facing telephone directory with fingers and thumb either side of the spine.

The sprung joint on the other hand is much more durable. When formed correctly both ends of the joint should remain tight. Even with the greater volatility of expansion and contraction at the ends of a plank a gap doesn’t form. I use a zip as an analogy to explain what goes on. Close a zip and lock the tab at the end and the zip stays closed. Pull up the zip and leave the tag unlocked and the zip gradually opens.

The basic sprung edge joint provides the largest possible matching long grain glue surfaces. If we accept the premise that the strongest glue joint between two pieces of timber is long grain to long grain, then anything that reduces the maximisation of such a joint in table tops, panels, etc. weakens it.

Adjoining boards in a table top can, and do, drift out of line in the thickness and form a step. Even the weight of a TV or computer monitor will do it with a foot placed next to a glue line and left in place for long enough. This is the case with PVA glue that creeps, see earlier discussion.

For the reason just described, and in many other situations, many woodworkers like to include an alignment aid or reinforcement of some kind into the basic sprung joint, eg, biscuits, dowels, loose tongue, finger joint, etc. They all reduce the strength of the plain edge joint, but compensate for this loss through increased glue line length and mechanical locking. The alignment aids and machined edge joints work best when the joined edges are as carefully and meticulously prepared as if doing a basic sprung edge joint.

There is a further wood movement factor to consider in edge joinery. Water based glues such as PVA, hide glue, urea formaldehyde, etc, all introduce water to the glue line. This water is very localised, but it does cause wood to swell creating a slight ridge at the line itself, perhaps no more than 6- 10 mm (1/4”- 3/8”) wide. If you glue up a panel with PVA glue at 9 am, take it out of the clamps at, say, 12 pm and immediately get to work preparing the panel for polishing, and actually follow up with polish you’re likely to encounter a problem. A week or two after polishing you’ll notice a slight furrow tracking every glue line.

If you’ve used a series of glued in biscuits to reinforce the joints you’ll see a slight furrow, and evenly spaced roughly elliptical shaped depressions that match the position of each biscuit. The chances are that you’ll get away with this error on vertical surfaces because there is often little or no light glancing off the surface. But you won’t with table tops, and the glossier the polish the more the furrows and depressions stand out. Biscuits are manufactured of compressed beech wood, and they work in part by swelling with the water introduced during the gluing up procedure. This makes them fit tightly in the slot and can exacerbate any swelling caused by water alone.

Sufficient time is needed between introducing moisture to the wood causing it to swell, and preparation of the surface for polishing. The wood around the joint line needs two or three days to lose the added moisture and shrink down to match the level of the rest of the surface. I like to leave edge joined panels at least three days to settle, and a week is better still.

I have only discussed biscuits in this section, but similar considerations are required when using other edge joint reinforcement or alignment aids, eg, dowels or loose tongues. Non water based glues such as epoxy resin or polyurethane types don’t cause the same problems, but these types aren’t always appropriate for the job.

[engineerone]

bloody hell, that was what i wanted and more, thanks richard it
makes it all clearer.

i will play around with the material i have and see how it can properly fit together.

paul

[engineerone]

well having re-read it richard i am really impressed.

it does cause me to consider whether some of the ideas that we accept since they have been handed down need to be reevaluated.

whatever happens with global warming, the fact is that accomodation whether it is offices or homes will become more sealed, and there will be less air movement within the rooms.

so adding to the fact is the wood is of course now not really old growth, and with kiln drying being more prevelant, due to economic considerations, we should really look more carefully at the long held beliefs.

also there is the chemical nature of many of the modern glues, plus the nature of many varnishes since they are basically plastic products, they do not breath in the same way as shellac does.
one also wonders whether they actually cause moisture to be retained within the wood, or through the glue line.

in the case of what i am hoping to complete, i have a table top which will be a maximum of 460mm square. with the two outer panels being solid boards about 140 by 460 long. and then the middle will be a total of six bits which will total the difference of 180. since it is as usual with me a relatively slender item, i am wondering how much i should consider the striking nature of the grain, and how much the actual colour, since i actually do not want to stain the cherry too dark. in fact i would prefer it to be clear varnished, and then weather itself.

i would guess from what i have seen, it will take a couple of years, and then i may have to strip the varnish and refinish the top to take account of the change of colours in the different pieces. i would obviously prefer not to have to re-flatten, so
what glue should i consider that is readily available to us mere mortals, and indeed maybe even from a shed

you do though wonder whether the wood movement now is due to the internal conditions of the room, the wood and its initial storing and drying, and finally the way in which it has been machined, and then assembled and finished, and whether as i said before these new variables need more careful consideration than historic data passed down as gospel.

paul

[paulchapman]

The top that you are making is quite small, so I doubt that much will happen to it over the years unless you make a real c*ck-up of planing and assembling the boards. I'd go with PVA and join the boards with biscuits or loose tougues. It works for me every time

Cheers

Paul

[engineerone]

c*ck, up,
planing,
me ;D

as if

paul

[blacksheep]

Apr 19, 2008 12:39:20 GMT Sgian Dubh said:

There is a further wood movement factor to consider in edge joinery. Water based glues such as PVA, hide glue, urea formaldehyde, etc, all introduce water to the glue line. This water is very localised, but it does cause wood to swell creating a slight ridge at the line itself, perhaps no more than 6- 10 mm (1/4”- 3/8”) wide. If you glue up a panel with PVA glue at 9 am, take it out of the clamps at, say, 12 pm and immediately get to work preparing the panel for polishing, and actually follow up with polish you’re likely to encounter a problem. A week or two after polishing you’ll notice a slight furrow tracking every glue line.

If you’ve used a series of glued in biscuits to reinforce the joints you’ll see a slight furrow, and evenly spaced roughly elliptical shaped depressions that match the position of each biscuit. The chances are that you’ll get away with this error on vertical surfaces because there is often little or no light glancing off the surface. But you won’t with table tops, and the glossier the polish the more the furrows and depressions stand out. Because biscuits are manufactured of compressed beech wood, they work in part by swelling with the water introduced during the gluing up procedure. This makes them fit tightly in the slot and can exacerbate any swelling caused by water alone.

Sufficient time is needed between introducing moisture to the wood causing it to swell, and preparation of the surface for polishing. The wood around the joint line needs two or three days to lose the added moisture and shrink down to match the level of the rest of the surface. I like to leave edge joined panels at least three days to settle, and a week is better still.

I have only discussed biscuits in this section, but similar considerations are required when using other edge joint reinforcement or alignment aids, eg, dowels or loose tongues. Non water based glues such as epoxy resin or polyurethane types don’t cause the same problems, but these types aren’t always appropriate for the job.

Interesting stuff...

I have long used PU glue for edge-joining boards, simply because I find it easier and quicker to set than PVA.

It would appear from your excellent explaination that I have unwittingly made the right decision.

For once! ;D

Phil

PS I've never found any need for biscuits or splines. For alignment I use a big rubber mallet!

[paulchapman]

Alan Peters, in his book "Cabinetmaking - The Professional Approach", wrote an interesting piece on this subject. I don't have the book, but the gist of what he said was that he uses PVA because of its flexibility and because wood never stops moving, and some sort of mechanical device (such as a loose tongue) for strength. It made perfect sense to me and I've also always done it that way.

Cheers

Paul

[Sgian Dubh]

Apr 19, 2008 19:43:06 GMT engineerone said:

... many varnishes since they are basically plastic products, they do not breath in the same way as shellac does.
one also wonders whether they actually cause moisture to be retained within the wood, or through the glue line.

No finish is impervious to moisture at its microscopic level. Wood still absorbs and desorbs moisture at that level. A film finish such as varnish, shellac, lacquer etc, merely delays moisture exchange.

... i am wondering how much i should consider the striking nature of the grain, and how much the actual colour, since i actually do not want to stain the cherry too dark. in fact i would prefer it to be clear varnished, and then weather itself.

i would guess from what i have seen, it will take a couple of years, and then i may have to strip the varnish and refinish the top to take account of the change of colours in the different pieces. i would obviously prefer not to have to re-flatten, so
what glue should i consider that is readily available to us mere mortals, and indeed maybe even from a shed.

PVA is fine. Just be aware of its characteristics and work with them, not against them.

Arrange the strips in a pleasing way. Go ahead and do your finishing and leave it to age gracefully. Over time the colour of cherry evens out, although cherry has a tendency to blotch.

you do though wonder whether the wood movement now is due to the internal conditions of the room, the wood and its initial storing and drying, and finally the way in which it has been machined, and then assembled and finished, and whether as i said before these new variables need more careful consideration than historic data passed down as gospel.
paul

Nothing has really changed with the nature of wood and how it responds to its surrounding conditions over the last 250- 300 million years, although the first real tree, Archeopteris, evolved about 350 million years ago. The conditions you describe above are all valid considerations when you're designing and making a piece of furniture, or any other wooden item really.

How wet is the wood when you start? What sort of RH does your workshop exhibit? What are the typical seasonal RH conditions at the final location? What MC will the wood exhibit in the workshop as you make the piece and how will that change when the piece is in service in the final conditions? How is the wood converted-- radially, rift cut, figured or tangentially, or a bit of all those? How will the type of polish you apply affect the hygroscopic nature of wood? And you need to be aware of the hysteresis factor, hysteresis meaning delay. Wood doesn't react immediately to changes in RH, it takes time for the change to show up as change in width, thickness, warping, etc.

I'm not expecting you to supply me with that information so that I can give you answers. They're questions you really need to ask yourself. Then you need to know how wood works to intelligently assess how to proceed. That bit of text I put up above might help a bit, but it's only a small extract from a larger manuscript I'm working on. In the meantime, until I finish my bit of verbal diarrhoea, your best bet for information on timber and its nature is probably R Bruce Hoadley's Understanding Wood. Slainte.

[blacksheep]

Apr 20, 2008 8:50:11 GMT Sgian Dubh said:

How wet is the wood when you start? What sort of RH does your workshop exhibit? What are the typical seasonal RH conditions at the final location? What MC will the wood exhibit in the workshop as you make the piece and how will that change when the piece is in service in the final conditions? How is the wood converted-- radially, rift cut, figured or tangentially, or a bit of all those? How will the type of polish you apply affect the hygroscopic nature of wood? And you need to be aware of the hysteresis factor, hysteresis meaning delay. Wood doesn't react immediately to changes in RH, it takes time for the change to show up as change in width, thickness, warping, etc.

Blimey.

Makes me wonder if I'm not better off just sticking with veneered MDF!