Ok so having a weird problem with some curved panels I'm making up. These are 3/8" panels composed of 3 pieces of 1/8" cherry roughly 10"w x 20"l. The 1/8" skins were freshly resawn and fairly flat. Glued with TB2, taped up in sets of 3 and popped in the vac bag for about 2 hours. When they come out the glue is set fine, but as soon as the tape is taken off they warp!

I've double checked and the form is good, no warp at all. And as I said the individual skins were fairly flat going in, just not coming out. Also they don't all warp, and the ones that do warp do so to different degrees.

Anyone have this problem before? I'm going to have to make up another batch of these as many are too far out to use for door panels, and would love to figure this out before wasting more cherry and time.

Thanks,
JeffD

JeffD:

Unsusual, to be sure. It sounds to me as though the skins are not making full contact in the form; is there squeeze out between the plies under pressure?

If so, it could be that just the edges are getting pressure and the internal areas are not. If the form is male only, have you laid a straightedge across the laminations when under pressure?

How tight is the curve? Could springback and incomplete lamination be conspiring to contort these panels?

And does it need to be solid cherry? In the past, I have taken 3 plies of 1/8" Italian bending poplar and veneered the outside; very rigid and no springback. You could press the plies first and then veneer, or do it all at once for panels that small.

Just trying to help.
TonyF

Jeff,

What do you mean by "warp"? Cupping, twisting or springback?

That is a fairly wide lamination that may be influenced by moisture in the glue, especially if the material is flatsawn or milled from boards with initial twist. You might try using epoxy to rule out water as a factor.

Most of my curved lamination problems seem to be related to water based glue, although I have certainly used it on many projects with decent results and obviously much less material expense.

LikeTonyF. I have often used bending ply as a core for pieces like that with good results.

I will be interested to hear how you solve this.

Thanks Tony, this is the first time I've had this problem and it's a pain not knowing why. These are very small panels and there's plenty of glue applied with just a touch of squeeze out....just as I normally get for this type of press. As far as I could tell they were sucked down tight as a drum and no hollow spots I could find.

This is a pretty subtle curve at roughly a 54" radius, so not a lot of force required and using yellow glue I don't get much, if any, springback normally.

I thought about using Italian poplar as well but thought I'd go upscale, (and save myself a trip), just using all cherry. Stuff is pretty cheap right now so it seemed like a no brainer. I have to do another half dozen tomorrow so may try something different. Maybe do a 5 ply panel and use something more flexible like poplar for the core?

thanks,
JeffD

Kevin, I think you responded while I was typing. The warp is....well warp, as I take the tape securing the panels together off, the panels will kind of pop apart from each other. These are panels for a wood door and are convex or barrel shaped. if I turn one upside down on a flat it rocks across diagonal corners. The worst offenders are probably a good 3/8" out, meaning if I push one end down the other pops up that much. Way too much for a door and they're too stiff to bend into place.

The stock was fairly flat to begin with, then milled and resawn and sanded to thickness.

All the S&R stock I did came out normally, though it was 6 plies and much thicker. Also was bent long grain vs across the grain. Maybe using more plies would help?

thanks,
Jeff

Too much glue not enough time. Leave them in for a full day or overnight.

I suspect the veneers are saturated with moisture.

I think all of the barrel shaped panels I have done have had ply cores, and though I have had some slight movement with some that I used pva glue on I never had any unuseable ones.

I spoke to a friend who did a project very similar to yours recently. a mahogany bowfront cabinet with a frame and panel door in which the panel was 1/2" thick, three plies of resawn veneer and glued with epoxy. He had no problem with it.. He also recalled some much larger panels done in the past with powdered plastic resin glue that worked well, so maybe water is not such an issue. I am not sure what the water content of mixed ppr is relative to Titebond.

Two hours is not very long for Titebond to set under stress, so the the bond is not fully cured and the moisture is still probably concentrated at the glueline. It definitely doesn't have much chance to leave the wood in a vacuum press (unless it is a very leaky one). Adam's suggestion to give it more time is good.

I think the real culprit is your choice of glue. I've done many glue-ups in a vacuum. First, when you're doing a curved glue-up and you need to be certain of no spring-back, use a UF type glue. They form a very rigid glue line with very little spring back. As another person mentioned, with curved work it's best to leave it in the press for 24 hours when using UF glue. Titebond on curves will always produce a higher degree of springback and will not have the same strength to eliminate the natural tendency of the wood to move how it wants to.

I also think the glue - water - is causing/contributing to the problem. The short time in the bag is not helping you any.

TBII is a lot like TBIII in that it stays a bit soft and will allow creep. Even TBI will allow some creep. TBIII encourages it.

I'd use UF glue or even urethane for the less water, and more rigid glue lines.

And a core of bending ply will make this a breeze.

The only thing I can think of that hasn't been mentioned yet is the stacking order of your parts. Are you laminating the layers back together in the order they were resawn from the original board?

If so try doing some flipping and reversing of the 1/8" strips. This should help balance out any skewed grain direction taking place in the original board.

Before remaking (if not already done) look carefully at the grain direction......both faced and edge.......on the boards to be resawn. Obviously the straighter the better.

Also I'm no fan of UF adhesives. We try to avoid it if at all possible to keep formaldehyde dust out of the air. A good alternative is epoxy for a rigid glue line.

BH Davis

Thanks to everyone for the responses! I thought about this a lot last night and this morning and feel that it's probably a combination of things. Possibly the moisture from the glue fully saturating the inner panel while only one side of the outside panels causing an unbalance.

The shorter time in the bag could be a problem also. I haven't had issues in the past, but the panels possibly being out of balance b/c of the glue maybe they need more time? Plus the immediate warp as soon as the tape was released makes me think the internal stresses may have still been there even after additional hours. And lastly, thinking maybe they needed more time I popped them back on the form with clamps overnight. I figured the last batch to come out may still cure correctly if I get them back to position....no such luck.

I'm not overly worried about using the TB glue as I've been using it for years. In my experience it allows very minimal if any springback. It cures relatively fast with a reasonable enough open time. It's quick and easy to apply and doesn't kill tooling. Other glues may work better in some situations but have negatives that make me less likely to use them. I cringe whenever I have to use epoxy.....messy, stinky, takes forever to cure and then take that and have to put it into a bag....yikes! If I have to I'll try it, but only as a last resort.

The thing to keep in mind is the stile and rail stock has already been done this exact way, except for more layers as it's thicker stock. So I'm not re-inventing the wheel here, just have to figure out what's the major difference between the S&R stock and these panels?

I milled fresh stock today and am going with a 5 piece glue-up this time. I didn't try it the first time as I thought I'd be OK with 3, and I honestly wasn't sure how thin my old beast of a sander could reliably sand. Turns out it can handle the thin skins just fine. I hope with the 5 piece I'll be OK. There should be less "pull" from any single skin and should be more balanced with more skins exposed to the moisture of the glue.

I've left the skins on the form overnight to try and give them a little pre-bend. I'll glue them up tomorrow morning and leave them in the bag for most of the day. Also I think I'll take BH's advice and go back through the skins and check orientation. I did not do this the first batch figuring the glue would overcome it, but looks like I was dead wrong there!

Oh and I did consider bending ply, but it would have eaten up half of my day making the trip to get a sheet of the stuff. My new shop unfortunately is not near any of my suppliers. Of course if this next batch fails I'll regret that decision too!

thanks again, I'll post back when I have results....good or bad!

JeffD

JeffD,

Its your choice of glue combined with the clamping method.

You are using a waterbased glue and sticking it in a sealed plastic bag for 2 hours. Its not a catalyzed glue. Alot of water needs to evaporate in order for it to cure. Its not going to happen that quickly under those conditions.

I never noticed real difference in springback between glues. We have always used Titebond 1 or epoxy.

Logically once the glue is cured there is no movement in the glue line. So why would the glue choice contribute to springback?

If the glue does not cure or remains soft TB3, I would think it could increase the springback.

Thanks Adam, I agree with you, as long as the glue is rigid spring back should be minimal, and that's what I've experienced with TB. I've heard others claim some glues are better than others but it's always worked fine for me.

As far as "curing" or drying, or hardening, or whatever one wants to call it, I'm still unsure about how long it really takes to get to a 100% fully dried glue joint. Realistically the glue is stuck inside 2 pieces of wood, so drying by evaporation regardless of in or out of a bag is going to take quite a long time...which is probably why the 24 hour timeframe came up. However in our business most of us go with "long enough", for example typical cabinet door goes in the clamps for 30 minutes then stacked and left alone. Since there's no stress the joint stays together fine. These panels on the other hand have stress added to them, so I need to leave them longer. The question is how much longer? With the curved stile and rail stock 2 hours seemed to be plenty, maybe b/c of the extra laminations, or maybe some other factors.

So for this newest batch I left them in the bag for 7 hours....Friday was a short day:) They look fine so far but I left the tape on them overnight and will unwrap tomorrow morning. I only did 2 sets this time so if there's failure I'll do the next set with West System.....almost used it today, but had to at least do one more try with TB first. I also noticed other factors before gluing up which I'll address later after I know what has, or hasn't worked.

Anyway have a great weekend, will post back later!
JeffD

I too believe that evaporation is not a factor in glue dry time. As JeffD said the moisture is locked between two wood surface with very little edge exposure compared to the surfaces involved.

So where does the moisture go? I believe it absorbs into the wood. This will swell the wood fibers and and have the typical moisture affects on the wood.........i.e. warped or twisted boards.

This is substantiated by a strange affect that occurs if you strip laminate a curved moulding. An example would be a 3/4" x 3 1/2" moulding for a half round window.

If you glue up 1/4" x 1" strips for the half round with regular Titebond for example the moulding springs IN as vs. out. The only explanation for this is moisture penetration into the wood causing it to swell and force the spring in. In our experience I would expect about 1/4" of spring in on a 36" diameter half round moulding.

BH Davis

My best experience when veneering with water based adhesive is letting the panels set for ten days to two weeks. Glue is as rigid as it gets. And dried totally while held in form, flat or curved. Not much moisture is lost when in clamped cauls or worse vacuum bag. So long time to dry.
You can even only veneer one side of a drawer front if you let set long enough. Stays flat.

Hi Guys, figured I'd give a quick update. My idea of using the additional thinner plies and leaving them in the bag longer....FAIL! They came out worse than the first set! So I still don't have an answer, but will probably try another set with epoxy when I get a chance.

The good news is a couple of the 1st batch of panels I glued up that I though weren't going to cut it "relaxed" enough over several days they were usable. Basically if I put them on a flat surface and pushed 3 corners down the 4th was up by about 1/8". Once the doors were glued up they held flat.

So I was able to get the project out the door, but not able to resolve the problem. I still have about 4 extra sets of skins though so will make a couple more attempts.

thanks again,
JeffD

View higher quality, full size image (4032 X 3024)

View higher quality, full size image (4032 X 3024)

Like others here I tend to use 1/8" bending poplar for curved laminations. At 1/8" your cherry laminations can start to exhibit characteristics of solid wood. I cross ply the skins. An excellent tip from Darryl at Vacuum Pressing Systems is to put single plies between the layers of bending poplar at 90 degrees to the curve. I have used this successfully on quite a number of curved panels of different radii. The idea is that the lengthwise orientation of the 90 degree pieces resists compression and better locks the curve in place. My panels stay put with esentially no springback. I also use UF from Vac Pres Syst.