All About Spines

Finding spines: empirical data

In this section, we look at how the different methods of spine-finding fared on different kinds of shafts.

When I was preparing the spine-finding instructions for the NeuFinder 4, I went through all three measurement processes -- feel finding, FLO, and differential deflection -- for a variety of shafts. I tried to pick one of each of the following types of shaft:

Conventional name-brand sheet-wrapped graphite shaft -- TrueTemper EI-70.
Known large-spine shaft -- AJ Tech Z-43.
House-brand shaft from a budget component company -- Hireko Select Gear.
Steel shaft -- Mercury Savage steel.
Filament-wound graphite shaft -- Mercury Savage graphite.

These were shafts I had on hand. I made no effort to select from among the shafts on hand to prove a point, except for the deliberate, explicit choice of the large-spine shaft; I knew the AJ fit the description, and selected it for that reason. The two Mercury shafts were on hand because I was testing shafts for Mercury at the time. While not that widely known, Mercury produced among the highest-quality shafts I have tested (quality being defined as adherence to specifications with tight tolerances).

Note that the "feel-finding" was both by actual feel and then repeated using the NeuFinder 4 as a naive deflection gauge -- that is, without pre-loading. The NeuFinder 4 was also used to perform differential deflection. Here are representative examples for each type of shaft. For each shaft (at least the shafts with enough spine to matter), the results are displayed in two forms:

A table of "events" you encounter as you go around the circumference of the shaft.
A color-coded diagram showing the results of all three types of measurements.

Fairly typical graphite shaft - EI-70:

The TrueTemper EI-70 is a conventional sheet-wrapped graphite shaft. The specimen measured had a modest spine of about 4cpm magnitude.

Angle	Feel Finding	Differential Deflection	Result of FLO test
55º		NBP	FLO
70º	NBP		Slight wobble
145º			FLO
155º		Spine	Very slight wobble
200º	Secondary NBP		Wobble
235º		180º from NBP	FLO
310º	Spine (only one)		Slight wobble
325º			FLO
335º		180º from Spine	Very slight wobble

Observations:

FLO gave results consistent with the precepts of physics. The spine plane and NBP plane were 90º apart.
Differential deflection agreed with FLO for the NBP, and was 10º off FLO for the spine. Because of the fairly small size of the spine, this reflects the precision problems associated with differential deflection.
Feel-finding gave results that were only slightly related to FLO, and not at all related to engineering mechanics. It provided only one spine, not two opposed spines. The two NBPs were not opposed, and were 15º and 35º from the true NBP. Interestingly, the strange feel-finding results could be explained by a residual bend of about .05" in a direction of 110-120º.

Known high-spine "supershaft" - AJ Tech Z-43:

Over the years, AJ Tech has become notorious for high spines. This was one of the more extreme I have measured: 17cpm. Some people refer to shafts with extreme spines as "supershafts", and claim that their performance when properly aligned is superior to a shaft without spine. I have not seen data to either support or refute this claim, so I remain a skeptic.

Angle	Feel Finding	Differential Deflection	Result of FLO test
30º	NBP		Some wobble
40º		NBP	FLO
130º	Spine	Spine	FLO
90º	Not near anything		Serious wobble
210º	NBP		Some wobble
220º		180º from NBP	FLO
310º	Spine	180º from Spine	FLO

Observations

FLO gave a spine plane and NBP plane at 90º to each other. These planes were easy to find, since the wobble away from the FLO planes was quick and massive. Hey, it's a 17cpm spine.
Feel finding gave pretty good locations for the spine and NBP: identical to FLO for the spine and only 10º off for the NBP. Explanation: the size of the spine is so large that spine effects swamp out any pollution due to residual bend.
Differential deflection gave results identical to FLO. Explanation: the size of the spine was large enough to swamp out any problems due to resolution of the instrument.

Cheap house-brand graphite shaft - Hireko "Select Gear":

Hireko is a component importer; its best-known house brands are Acer and Synchron. They sell solid, but unremarkable and inexpensive, components. This shaft, made for the Hireko brand by Aldila, came from the 1990s when nobody was being very careful about spine -- and it's a budget shaft at that. Not surprisingly, the spine was 7cpm -- pretty large by today's standards, but certainly not a supershaft-class spine.

Angle	Feel Finding	Differential Deflection	Result of FLO test
80º		Spine	FLO
90º	Spine (only one)		Very slight wobble
170º			FLO
175º		NBP	Very slight wobble
200º	Secondary NBP		Wobble
260º		180º from Spine	FLO
330º	NBP		Wobble
350º			FLO
355º		180º from NBP	Very slight wobble

Observations

Once again, FLO identified planes 90º apart, as physics says they should be.
Once again, differential deflection gave results similar to FLO: identical for the spine plane and only 5º off for the NBP.
Feel finding did OK with one spine; but it only found one spine, which is a serious mistake. It found two NBPs, but they were not opposed. Worse, they were 20º and 30º respectively from the true NBP.

Steel shaft - Mercury Savage steel:

The shaft showed FLO in all directions, and differential deflection showed the same stiffness to within 30g. (That's about 1cpm, as verified on the frequency meter during the FLO test.)

Feel-finding, on the other hand, pointed to a really significant spine. When feel-finding was repeated as naive deflection measurement, it indicated a load difference of 150 grams. This is seriously misleading. The apparent spine is really just residual bend.

This shaft is typical of most steel shafts, in that there is very little actual spine but feel finders identify it as a "strong type 1" shaft because of residual bend. But there are exceptions. I have never seen a spine as much as 3cpm on the steel shafts I have used from TrueTemper, Mercury, nor the original Apollo. (Some company has bought Apollo's name; I can't vouch for the shafts they are building.) But I have seen reports of 3cpm and 4cpm spines in steel shafts from other (premium) brands. All I can say is, "Premium? Maybe in price. But not in quality."

Filament-wound shaft - Mercury Savage graphite:

All the filament-wound shafts and all the steel shafts tested behaved the same. (See "steel shaft" above.) FLO and differential deflection showed them to be without spine for all practical purposes. Feel-finding showed a Type 1 shaft, and was often sharp enough to imply a strong spine.

Feel finders revisited

More on residual bend

Now we have seen data showing how feel finders react to real shafts -- how you need a strong-spined, straight Type 2 shaft (like the AJ Tech) for them to give anything close to the proper spine directions. We already know that a little residual bend can turn a very-low-spine shaft into an apparent Type 1 shaft as seen by a feel finder. But how do we explain Type 3 results like the EI-70 and the Hireko shafts above? The answer is still residual bend. Let's look at an example.

Here's a shaft with a medium-size spine. The N and S points are those read in a feel finder. All the usual things are in all the right places. So this is a Type 2 shaft. What this means is that the spine completely overwhelms any residual bend. Since the spine is of medium size, the shaft must be pretty straight.

Let's see what would happen if the shaft has a little residual bend. The bend is represented by the black arrow in the diagram. It is mostly left, but a little above straight left. That leaves it just above being aligned with the left S.

This changes how the feel finder perceives the N and S directions and magnitudes. As we saw earlier, the inside of the residual bend looks to the feel finder like an NBP, and the outside of the bend like a spine. So this added bit of residual bend:

Makes the left spine S less prominent; it is somewhat cancelled by the apparent NBP caused by the bend.
Makes the right S more prominent, by the same reasoning.
"Attracts" both Ns in to direction of the bend.
Since the bend is closer to the upper N, the upper N is a little more prominent and the lower N slightly less so.

This is a distorted Type 2, and has already begun to lose its value as an alignment instrument. Both the Ns have departed from the actual NBP of the shaft.

Let's increase the residual bend in the same direction. At some point, the effect of residual bend is roughly the same as the effect of spine. At that point:

The left S has been canceled out by the residual bend.
The right S has been further reinforced, and changed in direction to better align with the bend.
Both Ns are significantly attracted to the bend.
The upper N, being closer in direction to the bend, becomes more prominent, and the lower N less so.

This is now a Type 3 shaft.

As the residual bend continues to increase, it eventually reaches a point where:

The lower N disappears.
The upper N nearly aligns with the residual bend.
The right S also moves to align with the residual bend.
The magnitude (prominence) of both the remaining N and S increase as the residual bend increases.

This is now a classic Type 1 shaft.

Something to note: A Type 1 shaft is actually more pathological than a Type 3 shaft, not less. So Type 3 behavior is not something that should preclude the use of a shaft (as some feel who have written about spine alignment). It is caused by residual bend, usually in pretty small amounts. How small?

To answer that, we need to estimate the amount of residual bend that would roughly equal the effect of the real spine, when viewed in a spine finder. That amount of bend would produce a Type 3 pattern in the spine finder. And it isn't all that hard to compute a ballpark figure.

Before I start on that, let me clarify what I mean when I talk about the amount of residual bend. There are two ways to measure it:

In the middle of the shaft, as shown in the upper diagram above. This was the measurement cited earlier in the section on feel finders. If you detect residual bend by rolling the shaft on a flat table, this is the measure you detect.
At the tip, as shown in the lower diagram. For my feel finder measurements, I used an NF-4, but "naively" rather than with differential deflection. The NF-4 deflects the tip by a known amount, so that is the measure I will use in the calculations below.

But you can convert easily enough between the two. For the same shaft, each millimeter of mid-shaft bend corresponds to four millimeters of shaft-tip bend.

Now let's estimate the amount of residual bend that caused the Type 3 behavior in the EI-70 shaft. Type 3 behavior occurs when two percentages are roughly equal:

The percentage of spine, expressed in terms of stiffness or spring constant. Usually, we know the size of the spine as a frequency. Stiffness varies as the square of frequency, so we have to take that into account.

In the case of the EI-70, we have a 4cpm spine, on an overall stiffness of 230cpm.
That is a percentage of 4/230 = .017 = 1.7%.
Because stiffness is proportional to the square of frequency, the stiffness ratio is 3.5% (roughly double the frequency percentage).

The residual bend, expressed as a percentage of the deflection the spine finder places on the shaft.

The NF-4 in naive mode deflects the tip of the shaft 2".
The percentage of residual bend should be the same as the spine stiffness percentage that we found above. So the residual bend should be 3.5% of the deflection.
Thus the residual bend is 2" * .035 = 0.07".

That's a shade more than 1/16". Remember that this is the measurement at the tip; the corresponding residual bend measured at mid-shaft is only 1/64".

This is a rather small amount of out-of-straight to detect. You aren't going to detect it by sighting down the shaft, nor by rolling the shaft on a flat table. You can probably measure it if you have a stable lathe or spin indexer for the shaft butt, and a mirrored ruler at the tip. I have made such an instrument, and it is capable of measuring this level of residual bend -- but not much smaller. So shafts that appear very straight may well have enough residual bend to completely throw off the value of any feel finder measurements.

How can feel finders be more useful?

This afterthought is pure speculation, based on data from only two shafts. It is intuitively satisfying, but we now know how misleading intuition can be in telling us how spines behave. So maybe I shouldn't say anything. But here goes...

Where feel finders reported Type 2 shafts, they gave useful results. If the angles are right (180º, 180º, and 90º), then their output can be trusted.
Where feel finders reported Type 1 shafts, they gave totally useless results, and can safely be ignored.
How about Type 3 shafts, specifically shafts that report one spine and two NBPs? There are two shafts in the sample that fit that description. Can the feel finder output give us a useful hint where the spines and NBPs really are?

Based on the two shafts in question (the EI-70 and the Hireko), let me postulate rules for interpreting feel-finder output where there is one spine and two NBPs:

Take N1 (the stronger NBP, per Bill Day's nomenclature) and the reverse of N2 (the weaker NBP), and find their average. By "reverse", I mean 180º away from N2. N1 and the reverse of N2 should be fairly close, at least in the same 60º sector. Call their average direction N_avg.
If N1 is tangibly stronger than N2, then shade N_avg by 5º in the direction of N1.
The resulting N_avg is one NBP. The other NBP is 180º away from N_avg, and the spines are 90º away from N_avg.

I am not going to measure a lot of shafts to see if this holds true for more than the two shafts I measured. Maybe Colin Dick or Jerry Ballard will be motivated to follow up.

Last modified -- 11/30/2008