Hand Scraping - How can it be more accurate than machining?

[ww_kayak]

I'm in the "that cost way too much" crowd. I didn't even know it existed until I got a quote of 10k to restore a 900 dollar Bridgeport and said to myself, "wow, there must be another way". That's when I started searching the forums and found the reference to scraping. It also just happen to be at a time when Forrest was offering a class in the NE. Actually I would have missed the class if Forrest hadn't broken his leg

2 years later, I'm finally going out in the shop TODAY to start scraping a Bridgeport Column

[Your Old Dog]

I haven't read the entire thread so if ten people have already said this forget that I did!

In my mind, scraping would be better because you don't have to rely on the wear on a grinder or mill you are going to finish something on. My mill may have a belly sag in the bed that I would never realize. However, a surface plate may easily be flatter then any of the beds on my equipment on which the grinder wheel and cutters depend. After all, it's only job in life is to be pee-in-a-pan flat

[J Tiers]

The scraping process is incredibly crude, and totally incapable of producing anything like a smooth surface or precision flatness. It is simply impossible for that process alone to produce the result desired.

The scraper gouges a swath of material away. The scraper is not flat on the end, so the gouge it produces is rounded. The resulting surface is therefore a series of gouges, in a random pattern and of random depth.

You cannot simply scrape a surface flat. End of story.

Because of that, scraping is inevitably associated with various other items besides the scraper. These are necessary to correct the horrible distortions of the surface which the scraping process produces if uncorrected.

That is the secret of how the scraping process ends up with a flat surface.... The extra equipment tells you where to apply the incredibly crude process, and where NOT to. By correctly following the directions given by the comparison straightedge etc, as shown by the blue (or red, etc) spotting material, you can remove the high spots, and bring the surface down to where the "surface" is a series of flat spots with the "residual gouges" between them.

The net result is a surface defined by those flat spots, which is as flat and planar as your comparison device is. If the straightedge is curved, you can scrape a very accurate curve to match, but never a flat surface.

Luckily, geometry allows the creation of a flat surface of any desired flatness by the principle that if A=B, and B=C, and C=A, then A=B=C. A may have a curve "up" , and B may have a matching curve down, but C cannot then match both. Only when all are flat can C match A and A match B AND B match C.

A side note:
The existence of scraping marks on a machine may either mean that scraping was used to bring the machine into alignment, OR that someone made marks all over the surfaces in order to make you assume that a proper re-alignment was done.

This is usually obvious when a surface with visible linear wear marks also has "Nike swoop" marks on it that cut across the wear marks...... Most of the tricksters don't put a full surface of scraping marks on... that would be too much like work. They are satisfied with crude "flaking" to fool the unwary.

[Evan]

Luckily, geometry allows the creation of a flat surface of any desired flatness by the principle that if A=B, and B=C, and C=A, then A=B=C. A may have a curve "up" , and B may have a matching curve down, but C cannot then match both. Only when all are flat can C match A and A match B AND B match C.

That is true if, and only if the reference surfaces are used in a variety of random orientations. There are many possible surfaces that can fit two others if the orientations used for reference are symmetrical revolutions or simple translations of each other.

For instance, any surface of this shape fits any number of others of this shape if the orientations are limited to selected 90 degree revolutions.

[J Tiers]

(NO I will NOT get into this..............NO EVANMENTS)

Yes, there are some pathological cases.

However, the slightest error in re-positioning will start to show uneven marking. A simple end for end turn, which is part of the system, eliminates many errors, and shifting sideways, lengthwise, or minor angle-offs eliminate most everything. I have no idea how to precisely re-position the pieces so as to exactly and precisely NOT detect that type shape, unless it is under the tolerance.

For residuals which are not detected by repositioning, the total error presumably must be low enough to be within tolerance no matter what shape it is.

[Mcgyver]

the act of transferring the markings means one is moved relative to the other an small and random amount, so i can see how this could ever be a practical consideration

[Evan]

It most certainly is a consideration on a surface where placement is constrained by the shape of the surface such as lathe ways. There may not be much room to slide the straight edge in either direction and of course, longitudinal placement is the only option. When lapping three surfaces together it is essential that the contact areas be randomised or a non flat surface will be produced. When using only two surfaces the natural tendency as many telescope mirror grinders have discovered is to produce something that approaches an oblate surface.

[lazlo]

A side note:
The existence of scraping marks on a machine may either mean that scraping was used to bring the machine into alignment, OR that someone made marks all over the surfaces in order to make you assume that a proper re-alignment was done.

That's why a Biax power flaker (that just makes half-moon flaking marks) goes for a lot more on Ebay than the Biax power scraper

I bought a pair: a Biax flaker and scraper, recently on Practical Machinist. I got several messages asking to borrow the flaker. None to borrow the scraper

[lazlo]

the act of transferring the markings means one is moved relative to the other an small and random amount, so i can see how this could ever be a practical consideration

He's talking about the classic potato chip warp, which are easier to get when you're scraping long, thin items (like a camelback/scraping master).

That's why all the classic scraping manuals tell you to use square plates, and mark the corners when you're doing the three plate method. You're supposed to rotate each plate 90° each time you switch plates, which avoids matching potato chip warps.

[doctor demo]

That's why a Biax power flaker (that just makes half-moon flaking marks) goes for a lot more on Ebay than the Biax power scraper

I bought a pair: a Biax flaker and scraper, recently on Practical Machinist. I got several messages asking to borrow the flaker. None to borrow the scraper

Lazlo, can I borrow Your scraper? .

Steve

[Teenage_Machinist]

Clamping, stress relief, and other difficulties can cause inaccuracies in machining with HSM equipment. An HSM can put together everything needed for scraping with 1 Enco order costing under 50 bucks.

[Evan]

Here is what I found in the kit I bought from an old scraper hand.