Lam Bow Perry Reflex Experiments

[mmattockx]

Instead of polluting the What Did You Do Today? thread with all my ramblings I figured I might as well put all this in its own thread. Maybe someone down the line will find it of some use.

First off, the form I am using:





The spine is a 3.5" x 2.5" x 0.25" steel angle that was left behind at our acreage when we bought the place. The posts are 2x4 construction lumber that I milled up flat and cut a radius on so that I can clamp the bow to the post at any angle without the clamp twisting the bow. The adjustable posts let me experiment with many glue up shapes without having to continuously make new forms. If you know what profile you want to make a one piece form is a better way to go overall.

Note that all of what follows is mostly applicable to wood bows. The mechanics of it also apply to glass lam bows, but the glass lams are so strong that it is not particularly necessary or important for them.

For those that don't know what Perry Reflex is, read The Traditional Bowyer's Bible Volume 3, pages 91-95. What Dan Perry developed was a method of making lam bows where he would take the belly lam and pull it forward into reflex when gluing on the back lam. This results in a reflexed unbraced side profile. The effect of this was to preload the belly surface into tension on the unbraced bow. This preload has a couple of effects. First, it forces the core of the bow to store more draw energy than it normally would in a conventional flat bow. Second, it reduces the maximum compressive stresses on the belly surface at full draw. This reduces the amount of set a bow will take. Both of these improve performance.

The downside to doing this is it results in higher back stresses due to the reflexed side profile of the bow. There is also a limit to how much reflex you can have in a bow before it gets tricky to tiller and becomes difficult to brace without a lot of effort and equipment.


Mark

[Roy from Pa]

 

[mmattockx]

Now to my experimenting.

Doing structural calculations on the effect of Perry reflex shows that the benefits continue increasing well past the point where a bow made with Dan's original method (a straight belly pulled forward into reflex) would become pretty unruly due to all the unbraced reflex.

Also in TBB V3 was a section by Tim Baker on the DuoFlex bows. In it he suggested adding Perry reflex to the DuoFlex design by deflexing the belly lams during the initial splicing together of billets. This got me thinking about ways to increase Perry reflex without ending up with too much unbraced reflex.

Inspired by Tim Baker's thoughts, I ended up with the idea of making a multi-lam bow where the belly lams and core were initially glued into deflex and then pulled forward into reflex when gluing on the back lam. This way, the belly could have the preload effects of more reflexing, but the finished bow would only have moderate reflex that doesn't over stress the back and leaves it easier to tiller and brace.

My first design glues up the belly and core lams with the nocks having ~6" of deflex, then glues on the back with ~3.75" of reflex. Assuming some spring back off the form, I hope to end up with close to 7" of reflexing effect at the nocks, but only 2-3" of overall reflex.


Mark

[mmattockx]

For my first attempt at this I am using red oak boards for the wood. I like red oak for it's beautiful appearance, it is an OK bow wood and it is cheap enough that I can scrap a few without crying over the cost.

EDIT - The first bow design I am trying is a pyramid back profile, 67" NTN, target weight of 40-42# @ 28".

This is a dry fit practice run for the first glue up. The core lam, risers and belly lams are clamped in place. The deflex and reflex shapes are in circular arcs. This is because the bow design has a pyramid back profile that will bend in an arc of circle tiller. For a back profile that gives an elliptical tiller you would want to do the deflex and reflex using elliptical arcs.




The actual glue up:




The bow off the form:




This attempt was a failure due to gluing problems. I used Titebond III because it is more than strong enough, is much cheaper than epoxy and has no toxic fumes. Being water based, it caused the lams to cup heavily, resulting in large gaps on the edges of the lams.






It wasn't all a lost cause, though. I learned that the bow kept a bit over 5" of the 6" of reflex that the form has. That is better than I expected. I also got to experiment with bending the bow limbs forward into the 3.75" of reflex I planned for gluing the back lam on. They will easily bend that far with minimal effort, so I know that the second glue up will not cause problems in that way.

I also had problems with getting the belly lams to form to the ramps on the risers due to their thickness. I boiled them and clamped them onto the riser to pre-bend them to the ramp radius, but there was lots of spring back and they still needed a lot of clamping force to get them tight to the ramps and the fades off the riser. I also had issues with feathering the fades out to zero thickness, but that was due to my inexperience with making a riser block and nothing to do with the Perry reflex part of the experiment.

The lam thickness I tried is one of those places where a lack of experience with this design is hurting me. For a same wood bow (with all lams from the same wood), Dan Perry recommended having the belly lam be 2/3 of the thickness and the back lam 1/3. So that is the ratio I was using on the first attempt. That is what resulted in the 0.150" belly lam that didn't want to bend up the riser ramps easily.

Since I started, I have done more analysis on the effects of Perry reflex and it turns out that the effect is increased with thinner belly and back lams. This means Dan's recommended 2/3 belly - 1/3 back ratio isn't that important and I can go thinner to make it easier to bend them onto the form. The downside that I can see (so far) is that the thinner lams won't hold the deflex and reflex shapes off the form as well. This can be compensated for by increasing the amounts on the form, but it is not something that is easy to calculate and I am going to have to make a few bows to see how much extra is needed.


Mark

[Roy from Pa]

Maybe you should give unibond800 a try.

I get a gallon delivered for about $35.00

[mmattockx]

After my difficulties of forming the belly lams up the riser ramps I have been rethinking my riser design. Instead of the shape used in the first attempt, I am thinking of a multi-part riser similar to many FHLB designs use. In the picture below the original riser is on the right, an alternative is shown on the left.




The first riser ramps had a radius of just over 6.25". The second riser has bends of 10.375" radius.


Mark

[Mad Max]

How long is this bow going to be Nock to Nock? EDIT
67"

[mmattockx]

Maybe you should give unibond800 a try.

I get a gallon delivered for about $35.00

It's still water based, though, isn't it? I think the problem is water being combined with wide, thin lams. I am thinking I am going to try another run with the TBIII and use aluminum pressure strips to keep the lams flat. If that doesn't work out then I think I will be stuck with epoxy.

I can't easily find the Unibond 800 up here, but I can find the DAP Weldwood equivalent. Any experiences with it?

How long is this bow going to be Nock to Nock? EDIT
67"

Good point. I edited my third post to include the bow specs for the first design.


Mark

[Flem]

Biggest problem I see is those pinche' clamps. A bunch of 2" C-clamps would work better, barring a more effective setup. Those spring clamps aren't for glue-ups, which I am sure you know. It looks like you are not using a pressure strip on both sides? How are the clamping pressures distributed? I know in a different thread, you said you calculated the fade length, but it looks like they still ramp up too fast from the feathered edge. Obviously that would be eliminated with a different design
Also, there is no reason you can't thin out the lams as they go up the ramps. It would make that bend easier to achieve, if you ever decide to try this design again.
There could be other issues, those are just the one that are obvious from your pictures.

[mmattockx]

Biggest problem I see is those pinche' clamps. A bunch of 2" C-clamps would work better, barring a more effective setup. Those spring clamps aren't for glue-ups, which I am sure you know. It looks like you are not using a pressure strip on both sides?

The spring clamps were used when I ran out of the F clamps. Definitely not my first choice. I went out and bought a dozen 3" C-clamps the other day to solve that problem. I was not using pressure strips on that one but I will be on the next one.

Also, there is no reason you can't thin out the lams as they go up the ramps. It would make that bend easier to achieve, if you ever decide to try this design again.

Yes it would. I thought of that but had no good way to taper them. Next time I go this route I will use thinner lams and maybe add a taper as well.


Mark

[Roy from Pa]

Unibond800 has a low water content.

[mmattockx]

Unibond800 has a low water content.

That would be helpful. I will have to see if I can find any up here. Where do you order yours from?


Mark

[Roy from Pa]

Check your PM

[Mad Max]

I would make sure the sun is not shining on the bow while it is drying.

[Roy from Pa]

That's what the piece of plywood behind the bow is for, he was taken pichers first..

[mmattockx]

I would make sure the sun is not shining on the bow while it is drying.

There's something I never thought of. The sun has so little energy up here at this time of year it doesn't warm things up much, but I will keep that in mind for when things get hotter towards summer.

That's what the piece of plywood behind the bow is for, he was taken pichers first..

I wish I was that far ahead of things...


Mark

[Mad Max]

 

[LittleBen]

I’ve built a number of wood bows “like glass bows” as you asked in your previous post, and I’ve dealt with some of the problems you’re taking about.
Titebond will swell the wood and the back and belly laminations will cup. To counteract this I’ve had success using a lot of clamps and only putting them on the edges of the limbs. Also, titebond adds a lot of water to the bow, so I would let it sit on the form for a week, then let it dry for another 3weeks or so. Trust me, it seems dry but it’s not. For these reasons I started using EA-40 epoxy. I did not bake it. Let it cure at room temp for 2days on the form. Had only one failure with EA-40 and it was in a bow I traded so I didn’t inspect after. I suspect the power lam was too thin. Sorry woodcarver, I still feel terrible about that.

Regarding the riser fades, I ran the core and belly lam through the riser so the fade was more gentle, others like woodcarver have had success running lams up the fades, but I think he thinks the lams at the fades to make the bend.

Also, much like woodcarver does, I pre-taper all three lams (I use hickory backing, various cores, and typically Ipe belly). They come off the form so close to final tiller that I round the corners and can typically go straight to low brace. Tillering is very quick, basically just a scraper and some sanding. As a result, the lams are well proportioned all the way to the tip. I taper each lam to .002”/inch. That seems to be about the sweet spot.

I don’t pre-deflex them. I just glue the lams up at once. One point I would note is that I think that perry reflexed Tri-lam bows take more set than a self bow, so you need to start with more reflex than you might otherwise. Maybe I’m just doing it wrong though and others may have different experiences.

[Mad Max]

Those 4 on the table are all wood? Nice

[mmattockx]

I’ve built a number of wood bows “like glass bows” as you asked in your previous post, and I’ve dealt with some of the problems you’re taking about.

Thanks for the reply and the info on your experiences. Those are some beautiful bows, I hope to get mine looking that good at some point.

I think I am going to switch to epoxy or one of the urea glues before I am done. I still haven't figured out how to get everything glued up and clamped fast enough before the titebond starts to set and the lams are all cupped. I may try one more with titebond and pressure strips to see if I can keep it all straight, but I am not at all sure it is going to work out. If that doesn't go then it will be epoxy or urea glue next up.

I know what you mean about how wet the wood gets. My first glue up felt damp for days afterwards and our humidity is super low inside in the winter, so it is the best drying conditions possible.


Mark