Calibration results from my WSKO-BGA: actual apex angle vs. tool setting

[Cyrano]

When I use my WorkSharp Ken Onion with Blade Grinding Attachment to re-profile a blade, I don't obsess about the absolute value of the edge apex angle. As long as the final edge is sharp and in the right ballpark for angle, I'm happy; I don't care much whether the apex is 30 degrees inclusive or 35 degrees inclusive.

When I use the tool to touch up an existing edge, it's a different story. I know from experience that if my sharpening angle doesn't match the existing edge apex angle, I'm going to have trouble:

• If the tool is set to an angle significantly more acute than the existing apex angle, I'm going to remove a lot of steel, and broaden the edge quite a bit, before my sharpening hits the apex.
  
• If the tool is set to an angle significantly more obtuse than the existing apex angle, I'm going to form a double-bevel edge. The resulting edge might work OK or not, depending on the specific circumstances, but it's not what I want to achieve.

To avoid such problems in the future, I recently performed an experiment intended to calibrate my WSKO-BGA in terms of the final apex angle it produces as a function of the tool angle setting. This calibration would include whatever systematic effects are intrinsic to the tool, combined with the random and systematic variations I introduce when I use the tool.

The experiment consisted of sharpening ten 20-gauge steel strap ties:

• Five of the samples were sharpened using a coarse WorkSharp standard belt, at angle settings of 10, 15, 20, 25, and 30 degrees.
  
• The other five strap ties were sharpened using a coarse WorkSharp extra-stiff belt, at the same five angles.
  
• All ten samples were de-burred using a WorkSharp linen stropping belt with Dialux green polishing compound.
  
• All ten samples were then measured for apex angle.

Here are the results:

I was surprised and delighted to get such clean results.

• The results for the standard belt show a much better linear correlation than I would have expected, given the number of uncontrolled variables I create when using the tool.
• The results for the extra-stiff belt form nearly a perfect linear relationship. Had I not done this work myself, I'd say these results are too good to be true, and were fabricated by the investigator.

(In the cross-sectional images, irregularities and distortions seen at the very apex are caused by residual burr, resulting from the sub-optimal sharpening procedure I used in this experiment to save time. With proper use, the tool routinely produces clean, well-formed apex geometries.)

I know from experience that even slight changes in blade orientation with respect to the belt (in pitch, yaw, or roll) can cause big variations in the geometry and variability of sharpening results. Thus, I don't believe the absolute values of my results are directly applicable to any other user of this tool.

What might be true about the tool itself, and thus applicable to other users, are these observations:

• There is a good linear relationship between the tool setting angle and the resulting edge apex angle. This means the tool is well-behaved, and can deliver consistent results.
• The linear relationship does not intersect the origin (x=0, y=0). This means one should expect an offset between the tool angle setting and the actual edge apex angle.
• There is more variability (i.e., weaker linear correlation) for the standard belt than for the extra-stiff belt. This might be expected, as a more flexible belt might translate more variation in blade-to-belt pressure than would a stiffer belt.

 

[cbwx34]

Interesting test... although it seems the difference between the setting and actual angle is a bit high... 8-10 deg. per side in most cases, if I'm reading it right?

I've bounced back and forth between the BGA and other guided devices... and it seems more consistent than what you're showing. Pressure and distance from the wheel (I tend to sharpen closer to the lower wheel), may be the difference? Any idea on pressure and where you sharpened on the belt? Also which wheel distance "setting" did you use? (I use the closer distance between the wheels).

 

[Cyrano]

... it seems the difference between the setting and actual angle is a bit high... Pressure and distance from the wheel ...may be the difference? ...

Bravo! I believe you are correct.

In the following diagram, the green line represents a perfectly flat belt, while the pink line represents a very flexed belt. Assuming a consistent blade angle of zero, and a consistent intended grind angle,

• On the flat belt, placement of the blade along the belt has no effect on the resulting grind angle.
• On the flexed belt, placement of the blade along the belt has a big effect on the resulting grind angle. Depending on where the blade is placed along the belt, the resulting grind angle will vary from (less than the intended grind angle) to (more than the intended grind angle.)

Any idea on pressure and where you sharpened on the belt? Also which wheel distance "setting" did you use?

I aim for light pressure, which would induce relatively less flexing of the belt -- but I have no idea what that means in physical units of force. My "light" may equal your "heavy."

My belt geometry is worst-case for flexing. I aim for the midpoint between the two pulleys, and I have the pulleys set to the larger of the two possible spacings.

 

[cbwx34]

My belt geometry is worst-case for flexing. I aim for the midpoint between the two pulleys, and I have the pulleys set to the larger of the two possible spacings.

Ah... probably why there's such a difference... especially because of the pulley settings. The shorter spacing makes a big difference, as far as I can tell.

If you feel like repeating the experiment... be curious what the difference would be, if you did nothing more than reset the pulley space. Looks like you're able to document it pretty accurately.

 

[Cyrano]

... If you feel like repeating the experiment...

Count on it! The engineer in me won't rest until I figure this out.

 

[adamlau]

So little money, yet so many interesting things that I want to try, the WSKO being one of them

 

[Cyrano]

I repeated this experiment using the same materials and procedures as before, except that the pulley spacing (which had been set to the wider position) was set to the narrower position. As before, the resulting apex angles were measured using cross-sectional microscopy.

Here are the combined data:

Observation: Within each experimental cell, the data fit a linear regression essentially perfectly. The R^2 values are extremely high, which is something one never expects to see when dealing with real-world data. As the combined system of (Work Sharp tool + how I use it) is remarkably predictable, the tool is clearly precise and consistent.​

This chart is too busy to allow easy analysis, so I broke it down to show the effects of changing each variable independently:

Observation: At either setting of pulley spacing, results from the standard belt show a greater slope than results from the stiffer belt. The practical consequence of this behavior is that using the standard belt will allow a greater range of achievable apex angles than would be possible using the stiffer belt.​

Observation: When using the standard belt, changing the pulley spacing has very little effect on the apex angle. I have no idea why this is the case.​

Observation: When using the stiffer belt, changing the pulley spacing has a small effect on the apex angle. When compared to the results immediately above, these results seem counter-intuitive. Versus the standard belt, shouldn't the stiffer belt allow less variation due to pulley spacing, not more?
​

As the data fit nearly perfectly to linear regressions, one can look at the regressions only (i.e., hide the actual data points) to make it easier to assess the practical implications of these results:

When sharpening blades, I am usually less concerned about the exact profile of the edge than I am about the absolute value of the apex angle. In most cases, I prefer a convex edge to a flat V-grind, but that is typically of less importance to me than whether the apex angle is 30 degrees inclusive or 40 degrees inclusive.

Using this rationale, the best choice for most of my routine sharpening will be whatever gives me the greatest range of achievable apex angles. From these results, that criterion is best met using a standard (not extra-stiff) belt with narrower pulley spacing. For how I use the tool, this combination allows a more acute apex angle, and a more obtuse apex angle, than any other combination.

 

[cbwx34]

Thanks for the update... I just don't understand the results.

Here's what I did to make a comparison. I took a knife and sharpened it at 20 deg. setting on the WSKO-BGA, with a stiff 220g belt... making sure I had an even bevel with all Sharpie removed. Then I remarked the bevel and stuck the knife in the W.E. and set the angle at 20 deg., and made a couple of light straight up and down passes with an 800g stone... same as if I was checking to set the angle. The Sharpie was removed almost perfectly... with a bit of Sharpie left near the edge and shoulder of the bevel...

... in the picture you can see the WSKO-BGA angle at the 20 deg. setting, the W.E. at 20 deg., and the Sharpie removed.

So, why you're seeing such a large difference in your testing is puzzling to me.

Edit: p.s. At 21 deg. on the W.E.... I was hitting the edge... even checking under a pocketscope.

 

[Cyrano]

Thanks for the update... I just don't understand the results ...

Do you know the feeling of having eaten corn on the cob and having some stuck between your teeth? That feeling drives me crazy; I can't think about anything else until I've flossed.

I'm having that same feeling right about now.

Would you be willing to send me one or more samples of your sharpening, so that I can do cross-sectional imaging on them?

 

[cbwx34]

Sure. The next time I go to the hardware store, I’ll get a few of the “strap ties” you linked in your first post.... I’ll duplicate what I did today... and unless something changes, I’ll send them to you... if that’ll work?

PM me your address if so (no hurry... it’ll be a bit before I get them).

 

[David45]

So, what degree do we set the BGA to and where do we hold the knife edge on the belt (between the lower wheel and the middle wheel or between the two topmost wheels) if we want a 20dps result, or a 15dps?

 

[brasileiro]

So, what degree do we set the BGA to and where do we hold the knife edge on the belt (between the lower wheel and the middle wheel or between the two topmost wheels) if we want a 20dps result, or a 15dps?

I don’t have worksharp but I saw this video 10 minutes ago:

 

[David45]

Thanks.
Long center for more belt flex for high curve blades.
Short center for less belt flex for straight blades.
I take it the degree setting is “inclusive” and we are to sharpen between the top rollers.
The diagram in post #3 has the position of the apex a little below the midpoint between the rollers.
I hope I got that right. I only used the BGA for really big blades, axes and lawn mower blades. Now I have a new challenge. If I got it wrong, please don’t hold back.

 

[Cyrano]

.... I’ll duplicate what I did today... and unless something changes, I’ll send them to you... if that’ll work? ...

Yes, thanks.

To help explain the difference I'm seeing, I was hoping that the position of the blade along the belt (relative to the two pulleys) would have a strong effect. Alas, no:

 

[Sonnydaze]

I just received this WSKO, and have not turned it on yet. Good grief, perhaps I should have stayed with my Apex Pro sharpener. It takes a lot longer, but at least I know what I'm doing...
Actually, I really intend to use the WSKO for the kitchen knives, as it's obvious that there is some inherent variability just due to the design of the unit. I wouldn't use this (yet) for my large Sebbie, for instance.
You folks went to an awful lot of work, and I appreciate all of the data.

 

[David45]

Cyrano,
Maybe you should consider changing your screen name to Myth-buster!
Your results are very illuminating. It appears the settings on the KOWS BGA are WAY OFF.
So, can you tell us what setting to use if we want a 40 degree inclusive edge? Somewhere around 15?
For 30 degrees- 11?
I couldn’t find any guidance in the BGA setup information paperwork.
Thank you so much for your work.

 

[Cyrano]

... it's obvious that there is some inherent variability just due to the design of the unit.

I'm not ready to draw this conclusion.

• My results, while weird in having a large offset between setting angle and resulting apex angle, are remarkably self-consistent. If the product had a design flaw which introduced a large amount of variability, I don't see how I could have achieved such results.

• Another forum member (cbwx34) has produced results which do not show the offset I"m seeing.

With help from others, I will figure this out.

 

[cbwx34]

Just out of curiosity...

How much of the blade am I actually seeing in your pictures? I'm wondering if maybe you're measuring such a small area that you're catching the "worst possible" angle the belt is making (i.e., the maximum angle the convex belt produces?)

Have you tried what I did... and just compare the results of the WSKO-BGA, against another sharpener? (W.E., EdgePro, Sharpmaker, etc.)

Also, not that it really matters... how are you creating the pictures of the edges? (Is it an impression of the edge into something)?

 

[Cyrano]

... It appears the settings on the KOWS BGA are WAY OFF.

I've seen no evidence that anyone else is seeing results similar to mine. Until such results have been reproduced by at least one other person, we should allow for the possibility that the cause of the offset I'm seeing is unique to my system.

"My system" includes the tool combined with all the many variables I introduce when I use the tool, including

• roll angle of the blade with respect to the belt
• pitch angle of the blade with respect to the belt
• yaw angle of the blade with respect to the belt
• pressure of the blade against the belt
• position of the blade along the belt relative to the two pulleys
• belt tension
• belt speed
• belt usage hours

So, can you tell us what setting to use if we want a 40 degree inclusive edge?

All I can tell you at this point is what setting I need to use to achieve a 40 degree inclusive apex angle:

For the reasons described above, I don't know if my results will have good predictive power for your system.

 

[Cyrano]

... How much of the blade am I actually seeing in your pictures? I'm wondering if maybe you're measuring such a small area that you're catching the "worst possible" angle the belt is making (i.e., the maximum angle the convex belt produces?)

The photos have a scale in the lower-left corner. The units are microns, where 1 micron = 0.001 mm. For these images, my measurements are based on approx. 0.5 mm of edge near the apex.

Have you tried what I did... and just compare the results of the WSKO-BGA, against another sharpener? (W.E., EdgePro, Sharpmaker, etc.)

No, I have not. That's why I plan to use my measurement technique on samples you provide.