LumberJocks

PVA shrinkage over years

  • Advertise with us

« back to Joinery forum

Forum topic by tvrgeek posted 11-20-2013 03:38 PM 2245 views 0 times favorited 15 replies Add to Favorites Watch
View tvrgeek's profile

tvrgeek

23 posts in 1110 days


11-20-2013 03:38 PM

Topic tags/keywords: question

My main hobby is designing loudspeakers. Most of my smaller units are made from MDF. Yea, MDF for it’s acoustic properties. Anyway, the conventional way to do things is with PVA ( pick your brand, I use Titebond). Plenty strong with no fasteners. BUT here is the issue:

You can sand, seal, fill, prime and lay on as many coats of Centari as you wish. They come out perfect, but a year later, the seams will all show as a tiny trough. Question: What do commercial shops use when the have painted MDF and flush seams? I have a dozen test blocks with every glue I could find, but it would be nice to not have to wait a year to see which ones work. PVA shrinks in 3 or 4 months. That is all I know now. Some sort of thermo-set stuff?


15 replies so far

View shipwright's profile

shipwright

7164 posts in 2258 days


#1 posted 11-20-2013 04:08 PM

Hide glues won’t shrink. That’s one of the main reasons I quit using PVA.

-- Paul M ..............If God wanted us to have fiberglass boats he would have given us fibreglass trees. http://thecanadianschooloffrenchmarquetry.com/

View joeyinsouthaustin's profile

joeyinsouthaustin

1294 posts in 1533 days


#2 posted 11-20-2013 04:15 PM

Bondo joints as part of the finishing cycle. Don’t rely on the glue as the filler, just for the holding. You can design your joints for this. We do special boxes for a scientific firm that test microphones. We do a 1/16” R.O. on the joints, then join TB II and clamp etc. Bondo, sand, prime, bondo, then on to finial prime finish etc. Our final is a conversion varnish, FWIW.

Or as shipwright recommends, try hide glue, or even gorilla glue, if you can get around the mess.

-- Who is John Galt?

View b2rtch's profile

b2rtch

4822 posts in 2509 days


#3 posted 11-20-2013 04:59 PM

I had glue joints failing with TBII,
I am not sure why. I use enough glue.

-- Bert

View JAAune's profile

JAAune

1634 posts in 1777 days


#4 posted 11-20-2013 06:23 PM

I use epoxy to seal MDF joints if I really want to minimize cracking. A thin epoxy soaked into the joints and MDF fibers will do more than just hold the joint together. It will also harden up the MDF a lot.

-- See my work at http://remmertstudios.com and http://altaredesign.com

View tvrgeek's profile

tvrgeek

23 posts in 1110 days


#5 posted 11-20-2013 08:40 PM

No problem with failure or cracks with PVA, just the shrinkage. If you bondo over them, the bondo will stick and a tell-tail trough will eventually show through the paint. Some suggest cutting a deep V and then filling that, but the shape and location of my joints make that not practical. I need a glue that does not shrink after a week or so.

Making the MDF stiffer is not actually a good thing for my application. It is the very low modulus that keeps the panel resonances at a very low Q. Even using varnish as a sealer is an issue as it soaks in about 3/8 inch and changes the acoustic properties of the material as it does make it quite a bit stiffer. Good for furniture, bad for loudspeakers.

Is hide strong enough for an un-reinforced joint? Before I go out and buy a glue pot and all the fixin’s. It would be good to know. I know how many chair legs I have fixed over the decades, so my impression if hide is it is very brittle and not very strong. I have “reclaimed” a lot of hardwood assembled with hide and it is always the glue that failed ( a good thing in that case) Is that what is used in production with RF to melt it? What is this unibond 800 stuff?

View Rick M's profile

Rick M

7907 posts in 1840 days


#6 posted 11-20-2013 09:01 PM

I would just use polyurethane glue/construction adhesive. Supposedly it’s not as strong as PVA glue but I built a cabinet butt jointed with construction adhesive and screws then ended up not needing it. Removed the screws and figured I could just knock the butt joints apart but the plywood splintered and broke before the glue, eventually I cut the sides apart. Made a believer out of me.

-- http://thewoodknack.blogspot.com/

View tvrgeek's profile

tvrgeek

23 posts in 1110 days


#7 posted 11-20-2013 09:13 PM

I had not thought of that. I’ll do a test block to see. Some good old PL plus.

View joeyinsouthaustin's profile

joeyinsouthaustin

1294 posts in 1533 days


#8 posted 11-20-2013 09:26 PM

Our round over method is the same as the “deep v” Those boxes are irregular octagons. I have seen urethane glue push out, as it retains flexibility. Like caulking a butt joint. But your sample method will tell overtime. So make sure and post the results a year from now to fill the canon on this one. I will be watching.

-- Who is John Galt?

View tvrgeek's profile

tvrgeek

23 posts in 1110 days


#9 posted 11-20-2013 09:35 PM

Roger that.
I wonder how one posts a picture on this forum. I don’t use those public photo sharing thing-ies. Not only an I an old guy, but I know too much about computer security.

View joeyinsouthaustin's profile

joeyinsouthaustin

1294 posts in 1533 days


#10 posted 11-20-2013 09:44 PM

I set up a separate folder for images. Just click on img icon and upload from computer. To date I have not had a security compromise through my firewall from this site. I am no IT manager but people on these threads are better than some.

-- Who is John Galt?

View JAAune's profile

JAAune

1634 posts in 1777 days


#11 posted 11-20-2013 10:51 PM

Well if epoxy is a bad choice for speaker boxes then that eliminates the most obvious solution. One of the issues with MDF is that it’s pretty stable across the length and width but the thickness is more prone to changing dimension than solid wood as humidity changes. The epoxy would minimize the effect of moisture on the edge of MDF which is why it would have been ideal if not for the acoustic issue.

So this leads me to wonder how much of the problem is caused by the shrinking PVA and how much is caused by the MDF itself. An MDF butt joint will result in one board changing dimension in a different direction more than the other side.

Before I go further, let me say that I have no idea how your joinery is designed so I’m just working off assumptions in the following paragraphs.

I would probably create several samples and try a few different things. First, I’d do rabbet joints instead of butt joints to minimize the influence of “cross grain” interaction between the two panels. I’d also try using epoxy but keep the glue only at the joint. Another option is to try different grades of MDF. My lumber supplier offers waterproof MDF (Extira), ultra light MDF and the standard stuff. There’s also Gorilla glue which expands rather than shrinks and it cures in the presence of moisture.

To speed up the testing I’d create something similar to a drying kiln with plywood, insulation and a light fixture. Since it’s also possible that the shrink/swell cycle is causing problems and not just the shrinking I’d probably use a humidifier of some sort to introduce moisture back into the samples after drying them. Just torture test the samples and see which ones are doing the best job.

-- See my work at http://remmertstudios.com and http://altaredesign.com

View tvrgeek's profile

tvrgeek

23 posts in 1110 days


#12 posted 11-20-2013 11:42 PM

Most of my joints are 90 degree or something close butt joints, edge to face, or just a 50% rabbit to help alignment. I then rout at lease 3/4 roudover of all the edges. So, things like the standard lock-miter joints don’t really work as I would just be routing most of it away. Another joint is to laminate a dozen or so band saw cutouts to build up odd shapes I sand down to a rounded form. That leaves the surface 100% edge! The insides of my boxes have a full matrix, so the joints have a lot of help. Rarely will I have any panel over about 4 inches square unsupported.

I made one set of test blocks with end to end joints ( 2×4 inch) to test glue strength. Super glue gel and Gorilla glue were the only ones where the glue failed totally before the MDF. The super glue gel filled the joint and squeezed out, but when I broke the joint, it had retreated and had very little actual surface. Unsuitable for porous surfaces. Gorilla is probably strong enough, but it seems porous. West 407 is strong, but it does not get really hard. Poly resin seems to work and if baked at 170 for an hour, it does get pretty hard. One finish technique is to use poly resin as the sealer coat and of course, bondo is poly based so it fuses rather than epoxy bonding. You get easier feather edges with poly.

I am aware there are different qualities of MDF and I am sure what I get at the big orange store is the lowest quality junk one can buy, but without a truck or a lot of bucks, that is what I have to deal with.

As I seal both the inside and outside, I doubt I am seeing much difference in movement. Never seen a crack or tear, just the thin line sinking in over the joint. This is much like the pinholes you see in an automotive finish after a year or so if the shop did not let it sink in long enough and fill them. Unlike the TV shows, a friend of mine who did real classic restorations would let cars sit in primer a year before finishing. When you are buying a 30K paint job, pinholes are not acceptable!

I was really surprised that PVA shrinkage was so well known over here in the woodworkers forums, but over in the speaker building forums it is such a mystery. I have been building speakers for about 40 years, and until I started doing very rounded edges I could not veneer ( edge diffraction) and went to automotive buffed out painted finish did I ever discover the problem. Most of my earlier boxes ( still box shaped) were fine wood lined with MDF and just an oil finish.

View 404 - Not Found's profile

404 - Not Found

2544 posts in 2430 days


#13 posted 11-20-2013 11:58 PM

Is this happening with creep resistant PVA?

View JAAune's profile

JAAune

1634 posts in 1777 days


#14 posted 11-21-2013 01:45 AM

You could try Unibond 1. It’s a much harder PVA glue than the Titebond glues. It would probably qualify as “creep resistant”. I think Titebond is stronger but I’m sure the Unibond is good enough considering the low strength of MDF.

An alternative is the costly and toxic Unibond 800 which is an excellent glue. It requires mixing and has a shorter shelf life (less than a year) but is alcohol based so it isn’t as prone to shrinking as PVA.

If you really want to cure that PVA glue quickly there’s always radio frequency. I’m a bit reluctant to use that myself because it seems similar to having a microwave in your hand. It would probably help a lot in pre-shrinking the PVA though. I imagine it’s also pricy to purchase an RF unit but I’ve not done much research on that yet.

Gorilla glue is pretty weak unless the joints are perfect. If everything is super tight, then it’s strong. Gaps just get filled with a spongy foam which has no strength. Perhaps the porous MDF prevents the Gorilla glue from getting a good, thin glue line.

-- See my work at http://remmertstudios.com and http://altaredesign.com

View GOOD LUCK TO ALL's profile

GOOD LUCK TO ALL

418 posts in 1188 days


#15 posted 11-21-2013 02:16 AM

I built a dining table out of MDF. Had three layers of 1” thick material. Full Bullnosed edge
(1 full pc for the top and 2- 4” wide pcs for the build up.)
1st table failed just as you said showing the seams after a short period of time. I made the second table using the polyurethane gorilla glue and never saw the table again.
I also sealed the table with automotive primer putting extra on the edges.
I assume since I never got it back that it worked.

Have your say...

You must be signed in to reply.

DISCLAIMER: Any posts on LJ are posted by individuals acting in their own right and do not necessarily reflect the views of LJ. LJ will not be held liable for the actions of any user.

Latest Projects | Latest Blog Entries | Latest Forum Topics

HomeRefurbers.com