Replies

1. 
   rob_ss | Jun 25, 2020 05:57am | #1

   I am quite new to steam bending, but would suggest a low probability of success from that.
   By the time you have applied the glue, your wood will be cold and unlikely to bend well. It is hard enough to get one piece into a form, never mind three.
   It does I suppose depend on the kind of gear you have - https://www.youtube.com/watch?v=mHAbShWQ7JI could handle this job. If you could make something like that you'd have a good chance of success.

   Bent wood also has a tendency to spring back substantially so your form needs to be a bit 'tighter' than the final piece. This would be mitigated some if you could actually glue them together, but it would be almost impossible without some serious equipment.

   You could potentially bend one piece, let it cool, apply glue then apply the next, but when the first piece springs back, the force required to bend the next AND re-bend the first cold piece will be very high.

   Hide glue would not mind the moisture or the heat, but a polyurethane would be a good and very strong alternative.

   Given that you need a very precise profile shape, it might pay to reconsider the size and use 3/16 inch or so strips and glue-laminate them. If the curve is mild you might get away with 1/4 inch.

2. 
   remodlrj | Jun 25, 2020 07:21am | #2

   Thanks for your reply Rob. The form is actually the handrail that is attached to the tops of the uprights. Its more of a skeleton frame that creates the shape and gives me something to build onto. The problem is that the existing stairway never had a handrail installed and it comes to close to the wall in a few spots so I cant shape it adequately if I just build it up with laminations. I am hoping to steam, apply one piece and clamp. apply glue and then the next piece, then work to the top in stages. I know its going to be a challenge. But I plan on doing test pieces in oak as I have a lot more of that on hand. But once applied, it will be there for good.

3. 
   _mj_ | Jun 25, 2020 08:36am | #3

   It sounds like your "form" will not be up to the task. 5/8 cherry will require a long steam and a lot of clamping power to bend. If the form fits the staircase now, springback will make your new rail too wide to fit. Testing with oak won't tell you what will happen with the cherry in terms of steam time or bending. Test with cherry.

   If you really want to try to use the existing rail as your form bent lamination with thin plys and non-creeping glue like Unibond has a far better chance of working out. I would not use hide glue on something with internal stress designed into it.

4. 
   DRBWWing | Jun 25, 2020 11:04am | #4

   I think you need to clamp all the laminations at once. You may need to rethink the form and make a stronger one with a little bit of "over curve" to it, as there will be spring back.

   Also, there is a limit to how tight a curve you can do before the wood splinters.

   If you go in stages like you're talking about, the risk is the amount of curve is going to diminish due to spring back.

   Once its bent when you glue it up, reclamp the laminations into the form. As mentioned a hard drying glue like a resin glue will work well.

   Epoxy is my glue of choice for most laminations. I have no experience using in laminations, but glue will probably work, just be sure its no higher than 192.

5. 
   remodlrj | Jun 25, 2020 01:48pm | #5

   I appreciate your feedback(s), I was thinking that the piece on the outside and inside would create opposing forces so the form should be plenty strong enough. That was made out of 1/4 " laminations with spacers. My issue was would the hide glue be strong enough. I have only done veneer work and chair repair with hide glue to date, so I'm unsure of how it will react. I've done a fair amount of bends with thin laminations but with the space limitations I have it would be difficult to get the outside shaped the way I want. I guess the only way to approach it is with experimentation. I'll try test pieces with overbends after they dry, and with hide glue while they are 'hot' in the interest of science. If the latter doesn't work I'll find the right 'form' formula. But all tests will be done at least initially with oak as my shop is overflowing with it and cherry is not as plentiful. I'm also trying to avoid a trip to the wood warehouse if possible. I do thank you all for your input, it's given me a lot to consider.