Is the early wood the weak link to cause set?

[avcase]

Tuomo,
You show a good example of one of the complications of testing with natural materials. With these test samples, I am applying a load for about 5 seconds and then letting it go all the way to zero. I noted the immediate set or offset after letting off the load, and then gave it about 30 seconds to a minute to settle and stabilize and then I recorded this second offset. Then I increase the load and do the measurement cycle over again.  I would get a little different stress-strain curve if I let down either quicker or slower.

I find it confusing to say energy storage exceeds 100%. An ideal maximum force-draw curve should look like a rectangle instead of a triangle.   112% compared to a triangle is really 56% compared to a rectangle.

Alan

[avcase]

How does this tie in with Steve's thoughts of avoiding set and hysteresis?

This is a good question. I haven’t been stressing these wood samples far enough to see a major permanent change.  A majority of the “set” I have measured so far is mostly temporary and recovers over time after the load is removed. What I should do is load the sample higher to where a more distinct yield point is exceeded.

Alan

[avcase]

The strain gauges I am ordering are good for measuring up to 2% strain.

That is lots for working on a conventional bow limb. Maybe not quite enough for the wheelie bows, but that isn't our concern here.

The gauges cost about $1-$2 apiece.  The gauges require a little amplifier that directly interfaces with an Arduino microcontroller.  For a few more dollars, I can even integrate a Bluetooth module to wirelessly beam the measurements directly to my phone!

Technology is awesome. If I told you how much gear I packed into the middle of a large portal crane to strain gauge that machine you would be horrified. Now all that is on one small amplifier and a phone app! Despite complaints to the contrary, we live in amazing times.


Mark

Mark,
Some of the components have come into the “lab” which I am tinkering with.  I’m still waiting on the actual strain gauges. I should be able to get a resolution of around .002%, which is good enough for this kind of work.

1.85% strain is about the maximum strain that unidirectional glass can handle on the compression side of the limb, so these strain gauges will work pretty well for a glass bow too.

Alan

[mmattockx]

Some of the components have come into the “lab” which I am tinkering with.  I’m still waiting on the actual strain gauges. I should be able to get a resolution of around .002%, which is good enough for this kind of work.

Please post some pics and maybe links for the bits you are using when you have it all in hand.


Mark

[avcase]

Please post some pics and maybe links for the bits you are using when you have it all in hand.

Mark

Yes, I will do this as soon as I have something working.

[avcase]

While waiting for the strain gauges, I decided to take these test samples all the way to breaking. I wanted to see if there was a clear point where pushing the wood further was not worth the gain.  This first flat grain sample results was pretty inconclusive. It held together pretty well and was pretty close to linear almost to the breaking point.  I am very impressed with a bending strain of nearly 1%.  The lines on the graph show the first test that I did nearly two weeks ago compared to the one I did tonight (in blue).

It broke in an interesting way. It almost looked like it was neatly cut on the tension side, but that was where it transitioned from a late to early growth ring.  Time to break some more!

[willie]

Did it seem to loose some stiffness at .8%? I seem to see a slope change in the graph. No permanent set even if you pinch the break back together?

I am kinda disappointed it broke so soon, but not surprised as chasing rings on osage seems necessary with bows. My previous testing with some birch stock of square cross section needed a decent radius on the back corners to make it take very much strain past the proportional limit,  and with your thin/flat aspect ratio, any cupping would create a higher local stress.

[avcase]

Did it seem to loose some stiffness at .8%? I seem to see a slope change in the graph. No permanent set even if you pinch the break back together?

I am kinda disappointed it broke so soon, but not surprised as chasing rings on osage seems necessary with bows. My previous testing with some birch stock of square cross section needed a decent radius on the back corners to make it take very much strain past the proportional limit,  and with your thin/flat aspect ratio, any cupping would create a higher local stress.

There wasn’t a significant change in stiffness at all, right up to when it broke. I was quite surprised how linear this wood behaved. 

I am not having an issue with any obvious cupping with these test samples as they are loaded. The corners are not rounded off, but I figured they would start taking a set before this became an issue. I have many more that I will be breaking before I am done.

[avcase]

This graph is interesting. At higher stresses, I am starting to see some trends comparing flat grain to edge grain test samples. Edge grain has higher elastic modulus and better consistency than flat grain, but the edge grain samples do start taking a greater amount of short term and long term set at lower stresses than the Flat grain. Any ideas on why this might happen?

[Digital Caveman]

By the way, you are measuring the strain at the surface of the sample, right?  Are you measuring on the 'back' or 'belly', or is this the same?

[willie]

Edge grain has higher elastic modulus and better consistency than flat grain, but the edge grain samples do start taking a greater amount of short term and long term set at lower stresses than the Flat grain. Any ideas on why this might happen?

looked like it was neatly cut on the tension side, but that was where it transitioned from a late to early growth ring.

the early ring ran out pretty close to where the load was applied ie @ maximum bend?

I do notice that once the samples are pushed well into the elastic range, values seem to average out as indicated by the convergence of the two curves.

If you consider the possibility of one datapoint being out of range, the graphs could be seen as having the proportional limit somewhat close to each other @ .63%

[Selfbowman]

This graph is interesting. At higher stresses, I am starting to see some trends comparing flat grain to edge grain test samples. Edge grain has higher elastic modulus and better consistency than flat grain, but the edge grain samples do start taking a greater amount of short term and long term set at lower stresses than the Flat grain. Any ideas on why this might happen?

The early wood comes all the way to the surface of the back with edge grain. Instead of one continuous ring on the back. While the edge grain is stiffer like a I beam. I bet this as a complex composite would be a awesome bow.

[mmattockx]

Edge grain has higher elastic modulus and better consistency than flat grain, but the edge grain samples do start taking a greater amount of short term and long term set at lower stresses than the Flat grain. Any ideas on why this might happen?

My thoughts:

The stronger fibres (early wood/late wood?) in the edge grain samples are not as well distributed as with the flat sawn samples. When you chase a ring you get a continuous layer of strong fibres to carry the tension loads on the back. This is similar to your flat sawn samples. When you look at edge grain (ie - quarter sawn) the strong fibres and weak fibres are approximately equally distributed through the thickness. This means both of those fibre types share the loads relatively equally. If you look at the portion of the material that carries the bulk of the load through the thickness, the outer 30% of the thickness carries 50% of the load. In the flat sawn samples you might get a large proportion of the stronger fibres in that outer 30%layer and that reduces the set the sample takes. The edge grain versions don't get that same advantage and they show set sooner because of it.

By the way, you are measuring the strain at the surface of the sample, right?  Are you measuring on the 'back' or 'belly', or is this the same?

Alan is not measuring anything, he is calculating the strain based on the deflection and sample thickness. Assuming the neutral axis is mid-thickness on the limb (that assumes the stiffness is the same in both compression and tension), the strain will be the same on both the back and belly. The maximum strain will always be located on the surface of the sample, regardless of where the neutral axis is located.


Mark

[Digital Caveman]

Ok, I see.  I feel like I should know that, but I'm still working through basic statics.

[mmattockx]

Ok, I see.  I feel like I should know that, but I'm still working through basic statics.

No worries. He is getting set up to measure strain directly but is not there yet. I look forward to seeing how that all goes.


Mark