[Newbie] A metal scraping epiphany (and confession)

Rex Walters

H-M Supporter - Diamond Member
H-M Lifetime Diamond Member
Joined
Apr 21, 2015
Messages
566
I have a confession to make.

This is hard to admit. (Heavy sigh)

I ... I'm ... I'm a reformed woodworker. There! I've said it.

I'm in the middle of an absurdly long thread about making a set of straightedges. It's basically a public diary of my monomaniacal pursuit of metal scraping. I'm hopeful that readers of this diary think me slightly less insane than my family and friends, but if not, so be it.

Anyway, ....

The more I worked at scraping in these straightedges, the more vaguely familiar the whole process seemed. Eventually it hit me (and I've never seen mention of this anywhere):

Metal scraping is almost the exact equivalent of using hand planes and spokeshaves in woodworking. To more of a degree than might at first be apparent.

Bear with me here. This was by far the biggest epiphany I've had since I first tried my hand at metal scraping.

Not many people other than crazy handtool enthusiasts like myself attempt to dimension lumber by hand any more, but those that do should be familiar with something called a scrub plane. It's the very first plane you use to to make a roughsawn surface flat and smooth. Unlike the larger, wider, longer, flatter planes you use later in the process, the blade is fairly narrow. The end of the blade also has a very pronounced radius (about the curvature of a quarter or maybe a half dollar), it has a fairly short base, and it's used with a wide open throat.

A scrub plane removes a lot of wood quite quickly. It takes deep bites and produces very thick shavings. You use it to knock down major high spots to remove any twist, cupping, or bow. The resulting surface may be relatively flat (all the high points in the same plane) but it will have a very pronounced ripple.

Furthermore (toot doodle doo!) you push a scrub plane along a line at 45 degrees to the board you're flattening. First in one direction, then 90 degrees in the opposite direction. EXACTLY like metal scraping.

Check out this video of a scrub plane in use, starting at about 2'56" -- the process and resulting surface should be very familiar!

Once you've "flattened" a board with a scrub plane, you start removing the cross-hatched ridges it leaves behind using planes that have progressively wider blades and progressively less curved cutting edges.

Sound familiar?

This epiphany suddenly clarified something that I'd completely misunderstood from Richard King's scraping class.

I've no doubt that this is spelled out clearly in the handout from the class that I've somehow managed to misplace (I'm no longer capable of hanging onto physical printouts, apparently). But the process of metal scraping should proceed from fairly narrow blades with a very tight radius for roughing, to a quite wide blade with a very large, gentle radius (and light touch) for finishing.

I'm embarrassed to admit that I had this all mixed up. I thought you roughed with wide blades, then used narrow blades to pinpoint and finish.

I "finished" one of my straightedges to a very high degree of flatness using a fairly narrow blade and far too pronounced a radius.

My epiphany occurred when I temporarily quit working on reference gages, and started scraping in the top (non-bearing) side of my lathe cross-slide for a break. This wasn't a critical surface, so I wasn't pinpointing or trying for more than 20 PPI or so.

Here's a photo partway through the process:

IMG_0415.JPG

(If you look closely, you can see a pretty substantial hole just above the circular slot for the compound).

I used a much, much wider blade with a slightly wider radius for this surface than I last used for the straightedge. I was amazed (and, frankly, disheartened) to discover by touch that this surface was considerably smoother than that of the straightedge, even though I was absolutely certain (by spotting on the surface plate) that the straightedge was flatter with more PPI.

(I was disheartened to realize that I still need to put several more scraping cycles into my straightedge).

Put simply, and it seems obvious now, you should start roughing by creating fairly narrow and deep gouges with a narrow, tightly curved blade, then finish with quite gentle and shallow shavings using a wide blade with a very gentle radius.

I think I was also making a pretty major mistake by bump scraping with far too much pressure, creating much deeper scrapes than are called for in final finishing (if narrow and short as they should be).

One last point regarding the plane analogy: like a wood plane, the scraper does not take off uniformly thick shavings. You can actually feel this as you "snick" through a high point. That's the whole point: the edge of the blade should follow a relatively continuous "scoop" regardless of intervening bumps and troughs. You start cutting a shaving before a blued up high point, then snick through it. Your hands and ears will actually feel and hear a slight "tick" as you hit this increased resistance.

I hope this is useful to someone else. This honestly feels like a huge breakthrough in understanding for me, after literally a couple months of scraping.

Regards,
--
Rex
 
Last edited:
Interesting. That is also not how I had interpreted the use of different radii and width cutters for scraping. Your discussion makes good sense, but I need to think it out and try it out to see... I have also been under the notion that wide, large radius tools are for scraping large surfaces, and small, tight radius blades are for tight areas, like dovetails. That still seems to make sense to me intuitively. Your ideas just overturned my understanding of the process, and I think I am confused. :(
 
You still need to use narrow tools for narrow areas, for sure, but I now think the radius should decrease as you get closer to finishing.

Arguably, it's a nit. It's the high spots that are the bearing surface either way. The low spots just hold oil in a bearing surface. Who cares if they are a tenth or two lower than necessary? A tighter radius tool definitely digs a deeper groove than a wider radius, though.

Still, the finger feel between the top of the cross-slide and the straightedge I pinpointed with a narrow and tightly radiused blade was startling. The former was visibly smoother as well as in touch and feel.

It sure seems like a precision gage should have the smoothest, flattest surface you can create.

Regards,
--
Rex
 
Also stationary fit-ups that are scraped into plane and position. For stationary fits, say a vertical machine column, PPI is not as important as total solid support, and oil is not an issue. A grid of grooves would definitely not equal a flat surface.

Look at the scraping in this video:
and note the nearly nonexistent radius on the tools. Yes, there are some very large areas there, but they are certainly not working toward oil pockets...
Here, too:
on smaller work.
 
sorry to be the bearer of bad news.
you start roughing scraping with a large radius (90 or 120) and wide blade (30 or 25 respectively),
and finish scrape with a small radius (60) and small nose blade (15/20).
 
sorry to be the bearer of bad news.
you start roughing scraping with a large radius (90 or 120) and wide blade (30 or 25 respectively),
and finish scrape with a small radius (60) and small nose blade (15/20).

Why would that be bad news? I'm just trying to learn.

Why is that the case? Seriously, no doubt you're correct -- I'm not trying to pick a fight. Why is a tighter radius better for finishing when demonstrably in my hands a tighter radius produces deeper scrapes?
 
Hi Rex,
i didn't get the impression you were trying to pick a fight, we're good. :grin:
sometimes the written word is subject to interpretation beyond what is stated.

in pertinence to the radius,
simply put the tighter the radius , the smaller of a point you can scrape
you are directing more force into a smaller point , in effect.

i'll attach the Dapra Biax manual so you can draw your own conclusions.
the same truths are universal, whether hand or biax scraping

i look forward to seeing more scraping and more discussion :)
 

Attachments

  • biax- english.pdf
    3.5 MB · Views: 27
Last edited:
... note the nearly nonexistent radius on the tools. Yes, there are some very large areas there, but they are certainly not working toward oil pockets...

Love those videos, thanks for reminding me of them. (My wife is Japanese, by the way, and two of my grown kids currently reside in Japan. I have a deep connection to Japan and speak the language fairly well -- at least as well as I scrape!)

My understanding (newbie, frequently wrong, all the usual caveats) is that scraping is superior to lapping for many applications, but especially for dynamic bearing points like the ways in machine tools. The advantages include:

  1. You can scrape much larger surfaces than would otherwise be practical to lap.
  2. Scrapes have depth. Even with 0.0001" or less scrapes it provides a place to retain oil. (Note that this differs from flaking which intentionally creates very deep grooves on the order of 0.001" to hold reserves of oil.)
  3. A lapped surface presents 100% bearing and is less accurate as it wears. This is because the softer areas wear more quickly and you cannot control which areas are hard and soft. It's also because 100% of the surface is constantly wearing due to friction. A scraped machine way presents a dense grid of about 50% bearing points all mostly on the same plane, with grooves between that can hold oil. As a scraped surface wears, more points are exposed and the ways are more likely to stay in alignment.
The precision shown in those videos is simply staggering. They aim for flat within 5 um (0.0002") over a full meter and as demonstrated in the video often achieve far better than that (I saw about 1.8 um deviation around the 4'18" mark).

I sent that same video to Richard before the class, by the way. He replied, "The Research Institute I taught at in (http://scraping.pmc.org.tw) contacted that company in Japan and my company and they picked me to teach in Taiwan (2nd time too, as I taught there in the 1980's)."

As you point out, they are using very, very wide radius scrapers and making quite shallow but consistent depth scrapes. Note that the scraping marks are about 5X to 10X wider than they are long! That's very different to what I've been doing, roughing or finishing. (To be fair, the program quotes them on how it takes years to acquire the skill, and not everyone does. My goals are far more modest.)

Anyway, what matters most is the end result. I know what I'm after with my lathe as well as with the straightedges. I think I want higher PPI and far more bearing percentage with the latter as they are merely reference surfaces and should see very little wear. I believe a wider radius will help me reach that goal, but others should draw their own conclusions.

Still learning.
--
Rex
 
i'll attach the Dapra Biax manual so you can draw your own conclusions.

Thanks for that. I bought used and hadn't downloaded the manual.

Dapra appears to agree with me that larger radii create less depth.

Bottom of page 6 in the section on precision scraping: "However, flatter recesses are obtained with a large blade radius."

Cheers,
--
Rex
 
you are welcome!
by varying the stroke and speed , you can create different results with the same blade too :)
 
Back
Top