ESI Q and A Forums Home
 Search       Members   Calendar   Help   Home 
Search by username
Not logged in - Login | Register 

Saddle trees and changing back shape
 Moderated by: DrDeb  
 New Topic   Reply   Print 
AuthorPost
David Genadek
Member


Joined: Sun May 13th, 2007
Location: Spring Valley, Minnesota USA
Posts: 425
Status:  Offline
 Posted: Sat Feb 13th, 2010 04:42 pm
 Quote  Reply 
Adam,

 

I'm wondering what your rationale is for the curves one and two and also why you're curving the bar up in the middle and reducing bearing surface in the main weight-bearing area at three?

David Genadek

Attachment: TopBarsCG.jpg (Downloaded 401 times)

Last edited on Sat Feb 13th, 2010 04:48 pm by David Genadek

David Genadek
Member


Joined: Sun May 13th, 2007
Location: Spring Valley, Minnesota USA
Posts: 425
Status:  Offline
 Posted: Sat Feb 13th, 2010 04:51 pm
 Quote  Reply 
Next step would be to plot out some crown (?) profiles for the lower bar profiles.


What does putting crown on the bars do other than reduce bearing surface and make it easier to Rawhide?

David Genadek

AdamTill
Member
 

Joined: Tue Mar 27th, 2007
Location: Calgary, Alberta Canada
Posts: 289
Status:  Offline
 Posted: Mon Feb 15th, 2010 03:45 am
 Quote  Reply 
Hi David,

1 and 2 are based off the same thing - the measured flexibility of the equine spine. I took the numbers and graph from here:

http://nicholnl.wcp.muohio.edu/dingosBreakfastClub/BioMech/BioMechbend2.html

..and plotted the center yellow line, which represents the curve in the area of concern. I put the center of curvature at the point midway between the riggings, since those are the anchor points which determine the pivot points for the saddle when the horse bends.

If you take the center yellow line and move it to the edge of the bar, you get lines 1/2. Seems to work nicely, since even with Tindur bent over to touch his own flank (flexible boy here), his neck *just* kisses the inner edge of the bar.

Curve 3 is arbitrary at this point. I'm still noodling with the seat shape, and intersection of that shape and Tindur's barrel will determine if it's required.

On the subject of crown, my though there was just to add enough curve so that when Tindur bends, the shape will glide without possibly catching an edge. I'm thinking only something in the order of 0.125" over 5" or so.

David Genadek
Member


Joined: Sun May 13th, 2007
Location: Spring Valley, Minnesota USA
Posts: 425
Status:  Offline
 Posted: Mon Feb 15th, 2010 03:46 pm
 Quote  Reply 
Adam,

In regard to number one and two, I agree that that curvature must be in their and you may be able to get away with doing it that way on single tree. By moving the one curve further away from the spine you necessitate a really wide gullet which can also have the effect of widening the seat and making it difficult to get a narrow twist for the rider.  If your doing a single tree 2 won't be problem as you can figure your Cantle  position and make the surfaces match up at one particular location. You can't do that when you're designing for multiple seat sizes so you have to fill the space but still keep the shape.

The three curve is a curve that is there to make it easier when laying up the tree. It will not make the seat in the narrower it will only add dead air space and you can see this very clearly in your drawings. Since you're doing this in the computer and you'll be able to use indexing there is no need to put that curve in and as a percentage think about how much bearing surface that adds.

My bars are concave to the degree that the muscles of the back are convex and we just put up radius on the edge of the bar to allow for the movement.

David Genadek

AdamTill
Member
 

Joined: Tue Mar 27th, 2007
Location: Calgary, Alberta Canada
Posts: 289
Status:  Offline
 Posted: Tue Feb 16th, 2010 03:08 am
 Quote  Reply 
>In regard to number one and two, I agree that that curvature must be in their and you may be able to get away with doing it that way on single tree. By moving the one curve further away from the spine you necessitate a really wide gullet which can also have the effect of widening the seat and making it difficult to get a narrow twist for the rider. 

To keep weight down I'm going to go with a half-seat saddle, so that should make it easier to keep the twist within reason since I don't have to flow the seat pattern all the way forward.

This is my initial thought for a cross section at "4":



One idea I'm batting around is having the stirrup leather exit the bar inside the lower skin. I'd build a pocket within the bar, and wrap the edges with kevlar thread to prevent bursting of the pocket. The advantages would be a smooth underside to the bar without a notch, and elimination of the stress concentration that a notch brings.

As you may know, a plank with a notch cut into it is actually weaker than a plank that's uniformly as thin as the thinnest section of the notched version. That's because the notch itself introduces a stress concentration, since the plank can't bend uniformly. Having the leather exit the "middle" of the bar keeps the upper and lower skins continuous, which is handy. Should be easy enough to layup as well.

Rough model:
 

>If your doing a single tree 2 won't be problem as you can figure your Cantle  position and make the surfaces match up at one particular location. You can't do that when you're designing for multiple seat sizes so you have to fill the space but still keep the shape.

I can see how that would make things difficult. That said, I'm going nutty enough thinking of the shapes involved in a single tree, so I'll leave making the more flexible commercial versions for folks like you.

>The three curve is a curve that is there to make it easier when laying up the tree. It will not make the seat in the narrower it will only add dead air space and you can see this very clearly in your drawings. Since you're doing this in the computer and you'll be able to use indexing there is no need to put that curve in and as a percentage think about how much bearing surface that adds.

Good thoughts again, thanks. I took my earlier model back out to the barn this morning, and looked at expanding that area. I came up with the following:



Pink is the old line. If I widen that any more, then the curve as the shoulder meets the barrel starts to make the bar surface get a little funky, but I think I've gained some very useful area regardless - thanks!

>My bars are concave to the degree that the muscles of the back are convex and we just put up radius on the edge of the bar to allow for the movement.

Good rule of thumb, thanks, and it's very similar to the number I quoted before (which was just going by eye).

These are the final mold forms, for now, with the gullet shape in the area under the half-seat. I've come close enough that I need something 3D to sight some things in by eye, so I think I'll start on the male mold in the next few days. Plan is to make a framework, plank it with wood, 'glass the outside surface, then fill the underside with concrete to prevent it from warping and/or crushing under vacuum. Heavy but substantial, and cheaper than the epoxy/sand mix I've used on other molds.



Thanks much!
Adam

AdamTill
Member
 

Joined: Tue Mar 27th, 2007
Location: Calgary, Alberta Canada
Posts: 289
Status:  Offline
 Posted: Mon Mar 1st, 2010 03:04 pm
 Quote  Reply 
Well, this weekend was a balmy 9 degrees C so I spent a bunch of time out at the barn, but did get a good chunk of time to spend on the saddle tree mold.

I planked one side, and took a couple of photos in case anyone was curious what I was babbling about before.

Front end is facing left:



Front end facing right (planks left long for now):



Should hopefully be able to finish the planking by next weekend, then finalizing the shape with sanding templates prior to fibreglassing.

AdamTill
Member
 

Joined: Tue Mar 27th, 2007
Location: Calgary, Alberta Canada
Posts: 289
Status:  Offline
 Posted: Mon Mar 22nd, 2010 02:59 am
 Quote  Reply 
Okay, so my last time estimate was a bit off - spring is here after all, so I've been out in the fresh air as much as possible. Going to press hard on this project now, however, since the saddle I'm riding right now slips so much that working at faster paces becomes a bit nerve racking.

In case anyone is curious or has suggestions, I thought I'd keep this thread up to date.

Planking done top:


Bottom:



Sand fill started:



I decided to go with sand in the end, since I couldn't find the right concrete mix (HydroCal). The epoxy/sand mix is expensive at $1/oz of epoxy (~60 oz), and the sand is heavy, but it ensures that the mold will never twist or distort. I used marine-grade laminating resin with an 8-hour cure/1 hour working time hardener.

Start by painting the planks with epoxy to bond them together, and saturate the wood with glue. Try to catch any drips that leak through to the top surface before the epoxy cures, since you're have to sand them back otherwise.

To minimize the amount of sand and keep the weight of the mold reasonable (ended up at ~30 lbs), I added blocks of pink polystrene insulation into the gaps and packed around them with the sand/epoxy mix.



Make sure to skim any protruding bits of sand off, or the mold won't sit level when you go to check the shape later on (I almost had a heart attack when I missed one spot, and thought my mold had twisted).

One layer of fibreglass prevents getting your hands sliced up on the sandpaper-like cured epoxy/sand mix:



Done. This is before sanding the planks smooth, and you can see where the bits of epoxy have leaked through and been wiped off with a paper towel before they cured.



AdamTill
Member
 

Joined: Tue Mar 27th, 2007
Location: Calgary, Alberta Canada
Posts: 289
Status:  Offline
 Posted: Mon Mar 22nd, 2010 03:12 am
 Quote  Reply 
Now the rough planks are sanded as smooth as possible. This is done after the sand is packed in, so that the pressure of the sandpaper bowing the unsupported planks doesn't give dips in the balsa.

I used 6 formers to check the shape as I went, and it worked well. A drywall sanding sponge with 80 grid sandpaper wrapped around it worked well for the rough sanding, and the drywall sponge itself (well used - about 300 grit) for finishing work. The sponge adjusts better to the contours then loose paper used by hand does, and is less likely to gouge or leave dips.



Since the epoxy between the planks sands differently then the balsa itself, and even different densities of balsa between adjacent planks sand differently, it's almost impossible to get a perfect surface without a little bit of filler.

Here I've used a nice drywall putty that goes on pink, and dries white. It sands easier then the balsa, so I use it for final contouring. Sands easily, so can go on liberally here.



Lastly, the sanding sponge is used to get the surface perfect. I tend to close my eyes and do the last sanding checks by feel alone, since I can pick up tiny ripples, twists and dents better that way.

Side view:



Top:



Last step for now is to apply a good coat of fibreglass to the finished mold surface. This will ensure that the surface is dent-proof for the most part, and will allow it to be polished later on.

I used about a 3 oz cloth I believe, and added graphite to the epoxy to see where it was soaking up a bit more easily. After this cures, there will be one more round of filling (to fill the weave) and polishing, possibly a last very light glass layer, then the mold should be ready to use.

I'm riding with Josh this week, but I can see the end of this part of the project in sight soon.


AdamTill
Member
 

Joined: Tue Mar 27th, 2007
Location: Calgary, Alberta Canada
Posts: 289
Status:  Offline
 Posted: Tue May 4th, 2010 06:21 pm
 Quote  Reply 
Hi folks,

Big update - seems to have been a worthwhile project in the end!

First, couple of notes on how the mold is finshed off. After another 'glass layer on the top, the edges are finished one at a time (just used on heavy layer of cloth on the edges):




I also went back and filled the end pieces with sand. I thought it would be tough enough without, but turning the mold on edge was making ominous cracking noises.




Finally, after the last round of fibreglass, I did a fine-levelling check. The mold wasn't warped, but the bit of sanding on the base left things a little off kilter. A couple of shims and a few patches of auto putty levelled things out permanantly. This way when I use the mold to make the rest of the tree later on, I have reference points to measure from a level table.




Lastly prior to using the mold, I used a little putty in the corners here, then sanded it to shape. That just blends the transition between the bars and bridge piece better, ensuring no sharp corners. Corners are generally the first place to get stress cracks, if they're going to form.




After that, I gave the mold a few coats of high quality mold release wax, buffed them out, and set the mold aside. The release wax I use is termed "Dolphin Wax", but that's not its official trade name.

AdamTill
Member
 

Joined: Tue Mar 27th, 2007
Location: Calgary, Alberta Canada
Posts: 289
Status:  Offline
 Posted: Tue May 4th, 2010 06:24 pm
 Quote  Reply 
Next step was to start cutting carbon cloth. I used a combination of 80/20 carbon/s-glass cloth (4.7 oz), and 5.6 oz carbon cloth...mainly because I had it in the shop. If I did it from scratch I'd use a combination of 5.6 oz uniweb carbon and the 5.6 satin weave carbon cloth.

Layup schedule is an outer layer of fibreglass for scuff resistance, then 12 layers of cloth on the bottom. After building out the core of the bars, I'll add another 14 layers on top, for a total of 26 for each bar.






Here's a shot of the vacuum pump system. Its an adapted refridgerator compressor, and has an adjustable vacuum cutoff. I left it set at 19" Hg since that was what I was using before, but it could be set higher if I wanted the pump to run more often. As it is, that much vacuum means that there's the equivilant of about 4500 lbs of force spread evenly over the surface...roughly equivilant to the weight of my truck.






The cloth gets wetted out with epoxy one layer at a time, then as much excess as possible is removed with a credit card squeegee. THe cloth has to be completely wet, but any excess resin/glue actually weakens the final structure, since the cloth "floats" in the resin. The resin itself has almost no strength - that comes from the cloth. Excess is trimmed before going into the bag.



Here's the mold ready to go. The tube you see in the background is a vacuum resevoir made from a piece of sewer pipe - it helps the pump to not have to cycle on and off as often. The resevoir is connected to a small brass valve, and then to the vacuum bag itself...made of thin plastic. The edges are sealed on 3 sides, and then the blue/white clips are used to seal the forth side when the mold is slid inside.

The white fabric is called breather material, and allows the air to be drawn out of the bag. The white film is a porous release material, which helps the bag to avoid sticking to the mold. The roller is used to work excess epoxy out to the sides of the bar surfaces, where it will be ground off after things cure.

The wrinkles are almost unavoidable, and will fill with epoxy. That's just cosmetic however - it gets ground off as well.

Mold stays under vacuum at least 24 hours.



Finshed product:


AdamTill
Member
 

Joined: Tue Mar 27th, 2007
Location: Calgary, Alberta Canada
Posts: 289
Status:  Offline
 Posted: Tue May 4th, 2010 06:28 pm
 Quote  Reply 
Since the cloth wraps around the edges, it takes a lot of pursasion to come off the mold. The edges get carefully ground down with a combination of a dremel tool and bandsaw, making sure to not grind into the mold itself.

After freeing a couple of the edges, then you can slowly start to pry the mold free. If the wax is good, then it should pop free pretty easily. It also pays to design the mold so that there aren't any inside corners which can lock the part to the mold.

This is the end result:



After trimming some of the flashing off the part:




Bottom:



...and the pattern showing what the bar shape will look like after finish trimming:



The finished thickness of the material is about 3mm. Being pessimistic and assuming that the fabric to resin ratio is only 50%, that's still equivalant to a bar 3" thick out of good wood (assuming the layup mentioned for the top surface, and no strength contribution from the core).

Since I had done the design work and back tracings about 6 months ago now, I was a little leery of whether the fit would be good. Seems to still be fine thankfully!

Good fit in the center:



Clearance at the shoulder:



Clearance at the end of the bar:



I'll update again when the tree is all done.

David Genadek
Member


Joined: Sun May 13th, 2007
Location: Spring Valley, Minnesota USA
Posts: 425
Status:  Offline
 Posted: Wed May 5th, 2010 03:31 pm
 Quote  Reply 
Adam,
    This is really fun for me to see. I've started a new line of trees I am fiberglassing. I just invested in a vacuum pump I'm trying to figure out how to use all this stuff. I asked the people I purchased my supplies from to give me everything I needed however my vacuum pump runs continually. They did not recommend or think that I needed a resin trap. I'm questioning this because the pump produces a lot of exhaust to the extent that when I came in in the morning the whole shop was filled with a smoke.
    I am also having trouble with the wrinkles and had a very large problem getting the air to evacuate from the bag when I had the entire 3-D shape of the tree and it.
It seems the front of the cantle will suck down and then block the flow of the air to certain parts of the back. I did not use any filler material but was planning to on my next attempt.  My other thought was to make U shaped hose with more outlets so that I have air sucking from multiple locations with in the bag.
    This vacuum bagging really sucks.  I'm looking for any and all advice I can get on it.

David Genadek

AdamTill
Member
 

Joined: Tue Mar 27th, 2007
Location: Calgary, Alberta Canada
Posts: 289
Status:  Offline
 Posted: Wed May 5th, 2010 04:40 pm
 Quote  Reply 
Ask and ye shall recieve:

http://www.badger.rchomepage.com//vacbag.html

This gent is local, and he knows his stuff (we met in the same flying club). That's where I know my composite work from - airplane wings. The only change I made from his setup is to run a vertical hose off my oil inlet/outlet...then the oil runs out, up the tube, and back into the pump. Been 10 yrs or so, and I haven't had to top up the oil.

If you got your continuous running pump from a place with "aerospace" in the name, then I know the one. It's a cheapo little pump...designed for intro systems. It avoids the need for a vacuum switch by running continuously. Not sure what you mean by exhaust - is yours gasoline powered or something? :)

My pump has the adjustable switch, and runs for maybe 15 seconds every 5 mins or so if I have a good seal...more often if the bag seal is a bit more iffy. Having the switch isn't a big deal if your core is solid (other then noise and wear on the pump), but I do work on lighter materials like foam which can crush under too much pressure. The fridge pumps are tough and quiet, since they're designed for use in the home.

On the wrinkle side, I sympathize...tough to eliminate. That's why I made a mold, so my finished surface is set by the mold shape. You can eliminate the effect of the ripples by encasing the part in heavy mylar, but you're limited to surfaces that have curves that are primarily 2 dimensional, otherwise the mylar will bridge the curves.

That's why the wrinkles come in the bag as well...you have to pucker the bag a bit to get it to stretch into certain areas. It's also why I have the brass valve and roller. I'll draw light vacuum, seal off the valve, stretch and roll the bag over the part, then draw a little more vacuum off the valve. It's MUCH easier that way then to draw full vacuum all at once. The brass valve is just from Home Depot too...about $4.

If you need to get vacuum to parts of the mold where it isn't making it right now, that's where the breather cloth comes in. I use a dacron fabric, but paper towel will even do in a pinch. Just use a release film to prevent the cloth from sticking, and make about a 1" wide strip of breather to snake into the spot where the vaccum isn't sealing properly. In my photo you'll see that's how I get proper draw from the quick release bag fitting in the top left corner down to the part.


Another trick I didn't bother with due to laziness is to use a porous material kite quality kite dacron over the finished part. If you use that cloth on the part, then a backing of paper towel, the excess resin will be forced through the breather and soak into the towel. That's a good way to ensure a good resin/cloth ratio.


Finally, unless you're just going for an awful pun :) then just know that bagging is a LOT of trial and error. I've been through yards of fabric and gallons of epoxy learning how to do it. Just be glad you're not trying to transfer paint from mylar to a bagged part...that's a whole new nightmare!

Last edited on Wed May 5th, 2010 04:47 pm by AdamTill

David Genadek
Member


Joined: Sun May 13th, 2007
Location: Spring Valley, Minnesota USA
Posts: 425
Status:  Offline
 Posted: Wed May 5th, 2010 05:06 pm
 Quote  Reply 
Thanks Adam!

I was going for an awful pun but am glad to hear I'm just experiancing a normal learning curve.

I did get a cheapy beginner level pump it is a 1/3 horse with an air displacement of 4CFM. They talk about the exhaust in the directions but I had no idea it would produce the amount it did.  I might not be using the gas Ballast correctly I'll just keep pluging away through the learning curve.

David Gendek

David Genadek
Member


Joined: Sun May 13th, 2007
Location: Spring Valley, Minnesota USA
Posts: 425
Status:  Offline
 Posted: Wed May 5th, 2010 05:56 pm
 Quote  Reply 
Adam,
I have noticed some of the other tree companies are using the chopper to spray on the glass. It seems to give a good finish but I am wondering how the sprayed glass compares to the fabric in regards to strength? 
    I sat down with the president of the company that I buy my resins from and showed them the tree project and he really didn't even think we needed to use glass that the resins would strengthen the wood enough by itself. They suggested if I wanted to overkill it all I would need to do would be to glass the seams.  I am currently using a fabric and wrapping it around the entire bar so have a single thickness on the bottom but two layers of cloth on the top of the bar.  I would love to know your thoughts on this.

David Genadek


 Current time is 09:50 am
Page:  First Page Previous Page  1  2  3  4  5  6  7  ...  Next Page Last Page  




Powered by WowBB 1.7 - Copyright © 2003-2006 Aycan Gulez