Home Made Bearpaw Spine Tester

dvd8n

Supporter
Supporter
AIUK Saviour
#6 - It’s Over-Engineered in the Wrong Areas and Under-Engineered in Others

Mk II would have simpler, lighter columns. The columns in the current version have legacy heft that’s just not needed. On the other hand the beam between them could do with being stiffer. Every thou of flex in the beam translates directly to one point error in the spine. I can live with that but it’s something to bear in mind.
 

dvd8n

Supporter
Supporter
AIUK Saviour
#7 - The User Interface

I was always surprised that the Bearpaw tester managed with only one button. Then I saw a video which showed that the Bearpaw was using fist bumps on the end supports to drive a menu system. Now, that’s an interesting choice on Bearpaw’s part to reduce component count and hence decrease costs and improve reliability, but my life has enough idiosyncratic user interface challenges already. Having to set my wife’s dashboard clock twice a year is enough.
 
Last edited:

dvd8n

Supporter
Supporter
AIUK Saviour
#8 - The Hypothetical Mk II

If I were to build a Mk II, it would have the following changes:
  • The end columns would be less sturdy and less elaborate.
  • The beam would be stiffer – extruded aluminium would provide lightness and stiffness.
  • I’d use two strain gauges, either 1 kg or 2 kg rather than 5kg – 1 kg would probably be enough if the end supports were less bulky (as all weight in the supports takes away from the capacity of the strain gauges). 2Kg would give a better margin for stiffer arrows but it’d have half the precision.
  • I’d thread the central pillar so that it could be adjusted in height by twisting it, which would be a much better solution for adjustment.
All in all it’d look a lot more like the Bearpaw.

Funny that.
 

dvd8n

Supporter
Supporter
AIUK Saviour
Ok, the difficulty of setting the height of the central pillar finally got too much for me. I took the central support apart and reworked the pillar:

IMG_20210420_135230.jpg

The idea was that I'd thread the bottom collar and the bottom of the plinth to allow a micrometer-like adjustment of the pillar; a new collar with locking grubscrews would be fitted at the top.

That was the theory; of course reality was a little tougher. The collar was easy to thread; the pillar was another story. I'd used stainless for hardness and corrosion resistance but stainless is hard to cut with hand tools - I tried for an hour but eventually had to give up and swap in some mild steel instead. I'll get a stainless bolt sometime, cut the head off and swap that in.

The result's not quite as neat as it was but it's much easier to adjust so I'm happy.

IMG_20210420_163518.jpg

To adjust you loosen the grubscrews in the unthreaded top collar then turn the pillar to screw it up and down. With the M10x1.5 thread that I used that works out at about 25 degrees of rotation giving one thou of adjustment up or down. A finer pitched thread would have been better but I used what I had.

If you look hard at the photo you can see that the rain came on just as I was finishing up. So that'll be the mild steel rusting already, then :rolleyes:
 

dvd8n

Supporter
Supporter
AIUK Saviour
The adjustment's still not something that I'd want to do on an arrow by arrow basis but I can see me doing it occasionally for different types of arrows.
 

Stretch

Well-known member
This kind of nuttiness is definitely what makes the Internet worthwhile. Keep it up!

Stretch
 

dvd8n

Supporter
Supporter
AIUK Saviour
Well, I was playing with the tester trying to see if I could spot the axis with the low spine to identify where the carbon fibre discontinuity was, to see if I could index my arrows. Long story short; no I couldn't :rolleyes:

Investigation revealed that the reason was that the tester just didn't give repeatable results. It just wouldn't give results that were precise enough and repeatable enough to spot a low spot in the spine. The results were broadly correct but just not consistent.

Testing showed that the issue was that I had only put a load cell at one end. I was measuring the forces at one end then extrapolating the forces at the other end (basically by multiplying by 2). But if you don't press the arrow dead on in the middle then that strategy fails. Worse than that; because I was multiplying the forces by two I was doubling the errors. I'd already spotted that this was an issue and I'd tried to mitigate the problem by reducing the area of the central pillar but seemingly that wasn't enough.

So I bit the bullet and re-engineered the tester to use two 2kg load cells rather than one 5kg cell at one end.

And the result is, after a major re-engineering of the mechanics and a re-write of the software (drum roll 🥁)...

IMG_20210427_152427.jpg

... something that looks basically no different.

I struggled for a long time to get the new version to work; results were really erratic at first. It seemed like the load cell at one end was over-reading then topping out at 1kg and refusing to read higher. It looked for all the world like the mechanics at one end were bottoming out; that red herring took literally days to get past.

Eventually I disassembled the whole thing and found this:

IMG_20210427_152848.jpg

See that bottom edge covered with what looks like hot glue? With the frayed wire sticking out? Yep, that's the problem.

That's where the resistance matrices live - near the area of the cell designed to bend and flex. It's also the area that all the instructions and all the data sheets tell you to be really careful with and not to damage :( Data sheets which I had read, digested, and thought, "aaaaaah, it'll be fine." Of course it could have been damaged in the post but to be honest there are other more likely causes :rolleyes: (kicks hammers and sharp screwdrivers out of sight).

I swapped in a new load cell and life was suddenly a lot better. I had sensible, and more importantly, repeatable results.

Another week goes by and I inch closer to the simple utility of two nails in the wall and a rock on a string for a weight.
 

dvd8n

Supporter
Supporter
AIUK Saviour
I also put some effort into tidying up the wiring. I cannibalised an old PS2 mouse for its 6 core shielded cable and pre-wired moulded on mini-din plug. Of course I didn't notice the huge arrow and moulded in word "UP" on the plug until I was finished. Which, needless to say, was pointing down due to the orientation of the socket that I had randomly chosen 🤦‍♂️ .

If I ever have to disassemble the whole thing then I'll flip the socket over, but 'till then I'll have to put up with it mocking me every time that I measure an arrow.
 

dvd8n

Supporter
Supporter
AIUK Saviour
I mean, it was a 50:50 chance! Come on! Surely I deserve a win once in a while.
 

dvd8n

Supporter
Supporter
AIUK Saviour
According to the great Sir Pterry, million to one chances happen nine times out of ten, but I find that the corollary to this is that a 50:50 chance never pays off.
 

dvd8n

Supporter
Supporter
AIUK Saviour
So, an advantage of an electronic spine tester over a mechanical one ought to be that you can do some mathematics on the readings.

Spine readings are not consistent around the circumference of the arrows. For example, given a number of readings around the arrow it ought to be possible to calculate the average spine of the arrow and the likely direction of the maximum spine.

Three readings around the arrow at a spacing of 120 degrees would give a triangle of forces from which it ought to be possible to extrapolate much useful data.

Ok, so it would seem that a likely place to start would be with the centre of the triangle. Seems obvious, right? But what's the centre of a triangle? Well, it's the centre, right? Seems obvious, doesn't it? "Well," I thought, "I'll just look up on the internet how to calculate that."

A quick search and I found the "Online Encyclopedia of Triangle Centres"


According to that invaluable resource, there are currently just short of 40,000 definitions for the centre of a triangle.

40,000!

All curated by one person. And people say that I have too much time on my hands!

So, I set to and evaluated all 40,000 definitions for their usefulness and applicability to my problem.
 
Last edited:

dvd8n

Supporter
Supporter
AIUK Saviour
No, not really. I started reading, got as far as #3, thought "that'll do" and got on with it.

#3 is the 'Circumcentre of a triangle', the point at which all three vertices are equidistant. So, given a triangle of forces, this would be the notional centre of the arrow from which the spines are equal. That's my thinking, anyway:

IMG_20210502_194613.jpg

So, given three spine readings around the true centre of the arrow at 0,0 of 500, 600 and 300 you can calculate the point which is equidistant from the triangle vertices and at which the spines in all directions are all the same.

You do this by halving the sides of the triangle, constructing a line at 90 degrees to each side. The points at which they cross is the 'circumcentre', the centre of the circumscribing circle. In the example above it's at 182, 50 and the radius of the circumscribing circle, and hence the 'average' spine is 485. The distance between the two centres gives an idea of the maximum variation of the spine and the direction between the centres gives an idea of the direction of the maximum spine.

That's my thinking, anyway.
 

dvd8n

Supporter
Supporter
AIUK Saviour
You'd think that in 2021 someone would have worked out general purpose equations for this, but no, you have to work through the geometric steps in code :rolleyes:

Anyway, a week's work gave this: IMG_20210502_120505.jpg

IMG_20210502_120526.jpg

The display is saying that the test arrow, a 400 arrow with readings varying above and below 400, had an average of 398, a 1.47% variation with the maximum of 218 degrees round from the cock feather.

Time will tell how accurate (or useful) this information is...
 
Last edited:

PetrolHead

Member
AIUK Saviour
dvd8n, thank you for this thread. The maths is mind boggling to me but fascinating! Spine is something I have always thought is critical to be consistent across a set of arrows but with the actual number being largely irrelevant as long as you can achieve the same number across the set of arrows (and assuming the arrows work for your set-up).

When I have tested this before, I have looked for the high and low points of deflection on each arrow and found a point within this range across a set of arrows to then nock and fletch with the cock fletching set along the same number across the whole set. Does this fit with your thinking?
 

Stretch

Well-known member
I love this thread too but more as an academic exercise than something that would make you shoot better.

Forgive me if my retelling is a bit off as 1993 is a long time ago. A friend of mine (GB junior, Europa Cup Gold medalist etc) had 3 dozen ACE spined by Beiter. He got one set of “as close to perfect spine match as you can measure and weigh”. One set of “pretty damn close” and one set of “inside Easton tolerances but not as good as the other sets”. I may have known what that was in billionths of an inch and grains in 1993 but it has long been filed in forget. All had index marks to show best spine orientation for the sets. I think the best sets had an all around spine uniformity too.

In use he found no difference not even from the best to the worst. He was only (only!) shooting 320 to 335 ballpark at 70m (so I’d class as bloody good but mortal).

So in summary, fun to do but shouldn’t make a significant difference to your scores if the manufacturer did their job right. So I think this kind of thing is great for finding duff arrows (which can usually be done shooting them as bare shafts) but unnecessary in 99.9% of the time.

So I do think it is worth shooting all your arrows as bare shafts when new and rotate nocks on any that don’t group until they do* but after that... I typically do 30m as it is far enough to show up rogues. 70m if you’re really keen.

* I have been blessed to be shooting x10 since 1997 and there has not been much rotation and zero discards but appreciate that other arrows may not be as well behaved.

Stretch
 

dvd8n

Supporter
Supporter
AIUK Saviour
dvd8n, thank you for this thread. The maths is mind boggling to me but fascinating! Spine is something I have always thought is critical to be consistent across a set of arrows but with the actual number being largely irrelevant as long as you can achieve the same number across the set of arrows (and assuming the arrows work for your set-up).

When I have tested this before, I have looked for the high and low points of deflection on each arrow and found a point within this range across a set of arrows to then nock and fletch with the cock fletching set along the same number across the whole set. Does this fit with your thinking?
I really just started doing this build for the challenge to see if I could, and things got out of hand. And getting it to read in ASTM units was part of the challenge. As part of the challenge I also made it read out the AMO 2lb at 26" standard, wooden arrow poundages, and grams force on the towers (which is no standard at all but just as meaningful when just comparing arrows which as you say is all that you want most of the time). It could just as happily read in Dave units and it'd be just as useful 90% of the time.

The huge majority of the time I shoot Easton FMJ carbon allys and don't bother testing them at all; I just assemble the parts and go shoot. To be frank, at my level, that's enough, the kit is already way above my level without this level of tweaking. That's probably true of most archers. There are, however, loads of articles and videos on the internet talking about how you should index your arrows when building them, aligning the point of highest spine with the button, or at 12 o'clock, or something. I'm not convinced that it would benefit me as someone who can't reliably hit the side of a barn when shooting inside one, but on the other hand, it's essentially free if you have a spine tester, so why not? That was my motivation for putting in the average spine code; the theory is that it'll tell you the direction of maximum spine from only 3 readings rather than having to do a whole bunch around the arrow (I still need to test this by the way - the fact that I'll need to defletch some good arrows is putting me off - those fletches cost a quid an arrow!)

There is an elephant in the room though. Have you looked at Easton's arrow specs? They make a really big thing about how awesomely straight they are. X-10s are ± .0015”. And the weight is ± 0.5 grains per set. But the spine tolerances? Have you noticed how they don't make a song and dance about that? I wonder why. With a woody there's theoretically scope to tweak the spine with a bit of sandpaper but carbons and allys are what they are. I remember watching Jake Kaminski talking about building arrows and saying that he set aside arrows outside of his desired spine tolerances for blank bailing and fun shooting. That's nice Jake - I have arrows set aside for fun shooting too - they're the ones that I've whanged off trees and rocks. I can't afford to sort arrows like that. So is it worth indexing an arrow with a tiny circumfrential spine variance that is starting with an overall spine value that exceeds it in tolerance? Almost certainly not - but I'll likely do it anyway.
 

dvd8n

Supporter
Supporter
AIUK Saviour
I did have the idea that the software could work out the direction of the optimal spine - i.e. the direction at which the spine is (say) exactly 500 for a 500 arrow. The task is already 90% done finding the max spine value and direction. But the truth is that the accuracy needed would probably be outside the scope of hobbyist strain gauges and lab quality gauges would make the device too expensive. You'd be much better off spending the money on X10s than trying to optimise lower quality arrows.
 
Last edited:
Top