PDA

View Full Version : Great Scraping Videos



lazlo
05-19-2008, 01:19 PM
Here are 3 short hand-scraping videos from a gentleman named Nick Mueller on German YouTube:

"Don't use a Chinese Tombstone." :D

Handscraping: Preparing the Surface Plate (http://de.youtube.com/watch?v=8kWbIxB4z3s)

Handscraping: Sharpening (http://de.youtube.com/watch?v=uIvxXMDeCIc)

Scraping in a lathe's bed (http://de.youtube.com/watch?v=uHF7TtHVSWE)

Your Old Dog
05-19-2008, 04:22 PM
Thanks lazlo, first time I've seen the process.

GrahamC
05-19-2008, 04:32 PM
Lazlo, thanks for posting. Learned somethine new today.

cheers, Graham in Ottawa Canada

laddy
05-19-2008, 04:58 PM
Man, That does not look like fun! Wouldn't it be great to use a random orbital sander and have the job done in minutes!!!!!

rantbot
05-19-2008, 05:04 PM
Scraping in a lathe's bed (http://de.youtube.com/watch?v=uHF7TtHVSWE)
Note that, despite what it says at the end, the goal is not to "scrap" the part.

Charles Ping
05-19-2008, 05:28 PM
Nick's written english is usually very good - odd to make a mistake like "scrap"

2ManyHobbies
05-19-2008, 05:34 PM
Man, That does not look like fun! Wouldn't it be great to use a random orbital sander and have the job done in minutes!!!!!

Heh, new this year, lathes that produce "organic" looking work. :D
Little bulge here, little taper there, and between 16" and 22" you can even turn Morse using only the lead screw. *TWITCH*

J Tiers
05-19-2008, 10:37 PM
Interesting, but I am not so sure about all of that...... at least for the first part.

I am certainly no expert, but I understand the geometry, and the process..... and have spent many hours scraping....

The marking was quite heavy, maybe mostly for the video...... a lighter marking shows up the spots better, visually, and does not make them seem big.

The general bearing was reasonably good, spots per sq inch (!) although somewhat uneven. That surface really may not have needed scraped for bearing, which appeared to be the first part.... But then, that particular scraping would have merely moved-around the spots, since he didn't 'split" any, and scraped pretty much all over.

Then also, no de-burring shown after scraping and cleaning the part..... He should have rubbed over a fine stone or burr-file to knock down the fins and burrs, or it won't mark-up right.

The adjustment of tilt..... He could have just started off by "shoveling" off the high side, without marking, because it appeared there was a considerable 'step", unless that was in very small increments on the dial. It takes a LOT of effort to take down a surface even 1 thou (.04mm).

Then also, he absolutely left one end alone, instead of scraping heavy then lighter and finally leaving the edge. So there should be a "knuckle" where he k=left off scraping. Maybe he wanted to correct the tilt and then scrape it flat, but I am not so sure that works quite.

But, certainly shows the process.....

Bruce Griffing
05-19-2008, 11:30 PM
J Tiers - I agree with your comments. He also did not demonstrate the fact that each pass of scraping should be done at 90 degrees to the last pass. The method that Forrest demonstrated for sharpening the scraper is easier and does not depend so much on a steady hand. We just used a wooden block cut to the required angle and held the scraper flat against the block and against the wheel. Pretty easy to set up and do.

lazlo
05-20-2008, 10:26 AM
The marking was quite heavy, maybe mostly for the video...... a lighter marking shows up the spots better, visually, and does not make them seem big.

Nick's mark-up is about as heavy as Mike Morgan does in his video, and less than Rich King does in his video.


The adjustment of tilt....

Then also, he absolutely left one end alone, instead of scraping heavy then lighter and finally leaving the edge. So there should be a "knuckle" where he k=left off scraping. Maybe he wanted to correct the tilt and then scrape it flat, but I am not so sure that works quite.

That's actually the textbook method of correcting for tilt: you scrape the high side flat in several overlapping passes, and then re-scrape the whole surface, and re-check for flatness.

The Los Alamos scraping book that someone posted on PracticalMachinist goes even further: he makes "zones": several successive plateaus that he scrapes flat, and then when he thinks he has the geometry correct, he scrapes the whole area flat.

lazlo
05-20-2008, 01:36 PM
Several folks have asked by private message for the Lawrence Livermore scraping book. I'm not sure if this is proper Netiquette, but here is Jarno Seven's post from PracticalMachnist which contains the links with the Lawrence Livermore Laboratory scraping guide. I've edited out the .PDF's which are hosted on his private web page. If you want to download the document, click over to PracticalMachinist.

The "Step Cut" process of correcting geometry like Nick Mueller is doing, by scraping one or more plateaus on the high side, is described on page 23 (the second .PDF download):

http://www.practicalmachinist.com/vb/showpost.php?p=681916&postcount=38

08-03-2007, 12:53 AM
Jarno Seven

Regarding books on scraping, earlier in this thread both "Machine Tool Reconditioning & Applications of Hand Scraping" and "Foundations of Mechanical Accuracy" were mentioned and, of course are both outstanding books.

The booklet that is linked to below in PDF format is titled simply "The Art of Hand Scraping", by Robert Wade. It was prepared in 1981 for training purposes at Lawrence Livermore Laboratory, and I do not believe that it is a copyrighted work.

It is concise and I felt very descriptive of the methods you would have seen in use if you were to see high accuracy machine tools being built during that era.

I had to scan the booklet in three sections to get it to feed through my scanner. Total is 46 pages.

There is no substitute for having an accomplished scraper show you their technique, however if this is not an option, the guidance provided may help get you started on the correct path.

aboard_epsilon
05-20-2008, 03:20 PM
how about you guys find all links to scraping hints

put them here and i will combine them all the webpages into readable pdf...
which someone can host

here's number 1

http://www.bugattirevue.com/revue23/scrape.htm

all the best......markj

alsinaj
05-20-2008, 08:52 PM
Does anyone know what grade carbide is needed for scraping? Carbide blades from Dapra work great, but they are expensive. I tried several kinds of generic rectangular carbide inserts, but none would hold an edge. Under magnification I could see that the edge was chipped, even after just a few strokes.

J Tiers
05-20-2008, 10:26 PM
Laslo,

Maybe so about the tilt, but I do it the other way, and find that it works well.

As far as the marking, whatever..... for the video was one choice, and may be the reason..... For very precise work the marking is with a medium that is lighter than condensed breath........ I tried it, but don't need that sort of precision.

The heavy marking I have tried was just not helpful....... Then again, that may be a different medium, hi-spot that heavy smears out marks for me.

Forrest Addy
05-20-2008, 11:17 PM
alsinaj

The grade I've always used is C2 which is a straight unalloyed tungsten carbide. The generic C2 carbide is available in a wide rangse fo sizes from MSC's "unground carbide blanks". I use 3/32 x 1/4 x 1 blanks for my home made scrapers. I think I've made over a hundred of them. I silver braze the egde of the carbide on the end of the scraper shank and grind and lap from there. It pays to flatten and stone the carbide faces to a near mirror finish before brazing.

MSC #: 04120754

or look at MSC catalog page 723 or this PDF:

http://www1.mscdirect.com/CGI/NNPDFF?PMPAGE=723&PARTPG=GSDRVSM&PMT4NO=43233058&PMITEM=04157061&PMCTLG=00&PMT4TP=*LTIP

Correcting an alignment by selecticve scraping involves a few minor tricks. One is to determine the low corner(s). The other is to step scrape a slope to make the correction. If you know the true angle of the correction you can establish a slope by calculation. I prefer to scrape out small hollows at the very corners to provide stopping points. I calculate the depth for the correction at each corner and use a depth mike to deternine their depths. I scrape across the slope in a series of parallel cuts leaving a "shingle." Then I scrape the surface down until the hollows clean up. Follow the scraping with a surface plate or other accurate reference so you don't get undulations in the surface.

I've said it before: correcting the alignments of a machine tool's elements is a problem of the simultaneous solution of several unknowns and knowing where to start. Many start with the milling machine's table or the compound slide of a lathe. Wrong. Wrong. Wrong. You start with the immovable reference - the spindle axis - and bring the next fixed way system into alignment with it: the column ways on the mill or the bed ways on the lathe respectively. Then scrape and stack axis element on axis element until you run out and every element is in alignment, the way bearings are linear, mutually square, properly fitted, and the gibs fitted and assembled.

The thickness of the transfer medium (marking medium) is also a variable. Before you start you cross scrape the worn surface to 80% clean up leaving a checker board of witness marks. Only then do you apply the scraped surface to the reference. Put it on thick enough to cover but thin enough to see the teexture of the plate through it. Use a ink roller (burin).

Well, I'm gonna rewrite Connelly's RMC if I'm not careful. Scraping takes hours and days. Target fixation is an inevitable hazard. If the scraping work not gone about right a neophyte scraper hand can go a long ways down the wrong path before he discovers he's lost. What I'm trying to say is: plan your work then work your plan. And do it in writing.

TGTool
05-21-2008, 12:34 AM
alsinaj

The thickness of the transfer medium (marking medium) is also a variable. Before you start you cross scrape the worn surface to 80% clean up leaving a checker board of witness marks. Only then do you apply the scraped surface to the reference. Put it on thick enough to cover but thin enough to see the teexture of the plate through it. Use a ink roller (burin).



Ahem. I don't want to be the nomenclature police and I have no qualifications to correct Forrest, but I believe you intended to say brayer. Burins are the engraver's pointy tools.

Forrest Addy
05-21-2008, 03:06 AM
Ahem indeed. Breyer!! Aw crap. And I have a box of them in the shop. Old age guys.

Peter N
05-21-2008, 03:38 AM
I'd like to have a carbide scraper, but I don't really do enough of it to make me go out and get one.
I still use an old file, annealed, hammered flat, reshaped, re-hardened and ground. Crude I know, but it worked well for me in limited applications.

Picture of a work in progress when restroing the Myford follows, and I'll apologise in advance to the scraping experts for the crude appearance.

http://www.btinternet.com/~p.neill/Scraped_cross_slide.jpeg

Peter

Forrest Addy
05-21-2008, 07:28 AM
Looks good to me Peter.

Only results count and yours seem to be progressing right along. I notice you have a nice full width bearing that's coming in hopefully in time to take out those long scratches. So many would-be scraper hands roll off the outside corners by over-zealous scraping but your blue pattern goes right up to the edge of the way bearing. You're also scraping at angles to the way axis and getting a nice checkerboard (checquerboard?) appearance.

If you're successful with your modified file then that's plenty good enough. I expect you have to touch up the edge frequently. Carbon steel, even high speed steel doesn't last more than a few minutes of steady scraping. I don't know if you can get carbide blanks in Blighty but if you can, think about getting some to braze on your scrapers. Carbide outlasts carbon steel about 50 to 1. Only you can decide what's affordable or worth the effort in your shop but to my mind the increase in productivity in a time consuming process places a low threshold on acquiring carbide scrapers and the means of sharpening them.

You didn't scrape under those teensy Myford dovetails with the scraper in the picures, did you? Don't you have a scraper shaped with a bevel almost like a chisel to get under the angle of the dovetail right into the inside corner? Have you had to relieve the inside angle of the dovetail yet?

I made a little plane looking thing that has an angle more acute then the dovetail and a pointy tool that extends just beyond the dovetail. The object is to relieve the inside corner a bit below flush so the mating slide doesn't ride up on the unscrapable part in the very corner and give you screwy readings.

J Tiers
05-21-2008, 08:18 AM
Correcting an alignment by selecticve scraping involves a few minor tricks. One is to determine the low corner(s). The other is to step scrape a slope to make the correction. If you know the true angle of the correction you can establish a slope by calculation. I prefer to scrape out small hollows at the very corners to provide stopping points. I calculate the depth for the correction at each corner and use a depth mike to deternine their depths.


Yes yes yes.........

pcarpenter
05-21-2008, 12:58 PM
I have had the best luck using the stepped method to change the plane of a way surface. To maintain a plane that is already correct perpendicular to the direction you want to change, you can leave an untouched portion (really just a line) at the edge as a reference that establishes the already "good" plane. In effect, you are pivoting the plane and your straightedge around this known good as you "step" your way down in the direction you want to go. Boy....this sure is hard stuff to put to words. I could show what I was just trying to describe in about 2 seconds with a straightedge:D

Maybe this is really the same technique Forrest is talking about with regard to hollowing out near the edge?

The video does not show the author stoning between passes....which produces hugely erroneous results. In effect, without stoning, you are marking up all the burrs you kicked up with the scraper and not the plane you are hopefully establishing underneath.

With regard to the scraping strokes that seem to exceed the marks, I was tought a much longer stroke for all work up to the point of "pinpointing". You learn to manipulate the down force to take the highest points off quickest by hitting them harder, but you are actually covering more area with the stroke than you are really making a big dent in. When just trying to get bearing everywhere, what you don't cut this pass you will end up cutting in one of the subsequent passes anyway. As I have mentioned before, I use a marking method that rubs a dry marking compound *off* rather than on. I sometimes use a contrasting agent like a very faint haze of blue on the straightedge. This produces easily 4 tiers (often more) of depth the the mark, making it easy to figure out what needs to be hit hard on a pass and what is "next lowest"...and next etc. You just vary your left hand force as you move across the various levels of marking.

As for the angle thing, I found, and was tought (and later read in Connelly's book) that the alternating angle value is not so critical unless you are on finish passes and want that nice checkerboard appearance. Up until that point, the only critical thing is that the angle be different for alternating passes, to prevent the scraper from following previous scraper strokes and creating "hollows"

From looking at the Bugatti page above, it appears that he is mostly after a scraped appearance as he is laying out rows.....sort of a variation on say "engine turning".

Note to self: Never buy a lathe from "Laddy the RO sander guy":D

Paul

lazlo
05-21-2008, 01:07 PM
I have had the best luck using the stepped method to change the plane of a way surface. To maintain a plane that is already correct perpendicular to the direction you want to change, you can leave an untouched portion (really just a line) at the edge as a reference that establishes the already "good" plane. In effect, you are pivoting the plane and your straightedge around this known good as you "step" your way down in the direction you want to go.

Good description Paul -- that's basically how the Lawrence Livermore Labs book explains "step cutting."


From looking at the Bugatti page above, it appears that he is mostly after a scraped appearance as he is laying out rows.....sort of a variation on say "engine turning".

Yeah, the Bugatti scraping is really frosting. Very pretty though! :)

Forrest Addy
05-21-2008, 01:17 PM
"Yes yes yes........."

Oh! I guess I should go further.

Then by not scraping the corners that just begin to bear and by following the scraped surface down with the surface plate or other reference, you can tilt the surface down into alignment while keeping it flat and linear. This works well for vertical columns, knees, and saddles into parallelism and perpendicularity. You do not do this by surface plate and scrape reference alone. You check in process with levels, dial indicators, home brew checking fixtures, and home made versions of the King Way Alignment System.

As I said, establishing the alignments of way systems is part of that problem you solve simultaneously for a number of unknowns.

I can teach anyone to scrape flat, parallel, sqare and to dimension. However when that's accomplished I sense my students stepping back and congratulating thems selves of learning the whole techniquue. I try to disallusion them but it's like pulling teeth to get a neophyte to understand the necessity to carefully plan, formally on paper, the step by PITA step sequence of operations, checks and events whereby the worn way systems of a machine tool are restored to their original fits, linearities, alignments, and bearing with respect to the spindle axis.

Then there are the subsequent steps of correcting the now displaced lead screw axes, refurbishing quills and the bores they are fitted to, analysing the spindle for deficiencies and overhauling/replacing the bearings, clean work, the electrical work, motors, shafting, bearings, transmission, ect.

Painting is seldom a problem. Most guys attack the paint first even when it's better to defer the final coat until the rest of the work is done.

The hardest part of all is to keep the neophyte from charging into a machine tool rebuild without a formal survey of machine condition where every deficient condition is assessed, measured, and quantified on paper. If you don't know where you're starting from it's damn difficult to map out how to get to where you want to go.

Experienced machine tool rebuilders can wing it but they are experienced. Whether they are conscious of it or not they still have a detailed plan but it's in their head.

lazlo
05-21-2008, 01:30 PM
The hardest part of all is to keep the neophyte from charging into a machine tool rebuild without a formal survey of machine condition where every deficient condition is assessed, measured, and quantified on paper.

I think that's where Connelly's Machine Tool Reconditioning shines: he lists, step-by-step, how to measure the existing (mis) alignment of the machine, and what order to scrape everything back into alignment.

For a lot of home-shop type scraping though, like Peter's saddle, you're really just "scraping-down": flattening a surface that's already in alignment from the manufacturer.

Forrest Addy
05-21-2008, 03:06 PM
"...flattening a surface that's already in alignment from the manufacturer." OK but know the risks in doing so. The cross-slide may not finish up square with the spindle and afterwards parts may be faced slightly convex or concave.

Rule number one for machine tool rebuilders. Take nothing for granted.

Rule number two: See rule number one.

pcarpenter
05-21-2008, 05:24 PM
Rule number one for machine tool rebuilders. Take nothing for granted.


That's for sure! In my own experience with my ongoing rescraping work on my Bridgeport knee, I found this firsthand. I was pretty green at the point I was getting my initial advice about establishing references in the case of my mill. I did not yet own Connelly's book and was going only on advice. Here's where Connelly's book would have offered an important clue. There's a section in which he points out that worn machine tools tend to scrape back in close (if they are not very worn) very quickly. Provided their initial planar relationships were OK, they tend to just wear low in the places of greatest travel. I had done some quick rubbing and relatively minor scraping work on the knee while it was still off the column and found that I quickly got bearing at each end of both the dovetail and flat ways. It took very little scraping (relatively) to make these both look very good and put them in near perfect relationship to one another with bearing across the full length of both ways. Measuring across pins in the dovetails put me within a few tenths from one end of the knee to the other. Then we hung the knee on the column.

Let me start by saying I am on my third round on the saddle to knee ways. We first verified that the column ways were indeed in good shape. This was a reasonably safe assumption based on negligable viewable wear (there's that word again) that proved true. We then installed the knee, ran it to the top of the dovetails where it could be scraped on from one end to the other, established a "normal" clamping force for the column to knee gib and then worked out a means of measuring how far nose up or nose down the knee was. Lots of experimentation made it completely clear that the ram would have to be locked as we measured first one end of travel and then the other. I made up a nice rigid indicator holder and used a tenths-reading dial indicator with a 1-2-3 block as a "puck" to remove randomness in the readings caused by the scraped surface.

This worked out to be a very practical way of making the measurements that would be used to verify the correct planar relationship of the flat ways. I asked my mentor if I should pause in the process and finish re-buiding the head so that its mass would figure into the equation. He indicated that it should not be necessary.

That was the wrong answer as everything is made of rubber....even a cast iron machine tool. Ram flex was enough after installing the maybe 350# head, that all of the "step work" I had done to make the knee .0005" high at the nose was overkill by a several thousandths. Sigh...now to step it all back the other way. My only hope is that I won't have to re-center the yoke that holds the cross-feed nuts.

So....my mill knee was scraped in with relative ease, then messed up by attempting to change the planar relationship to a "correct" one...and then all that was undone again. Likely it was "just right" after only that first, minimal work on the knee. Forrest...as you stated, a plan is important, but it needs to be a plan based on correct assumptions:rolleyes: . In all of this, I most certainly remain less experienced than my mentor....but I won't hesitate to stand on my own two feet and verify my hunches next time.

Paul

lazlo
05-21-2008, 05:51 PM
There's a section in which he points out that worn machine tools tend to scrape back in close (if they are not very worn) very quickly. Provided their initial planar relationships were OK, they tend to just wear low in the places of greatest travel.

It took very little scraping (relatively) to make these both look very good and put them in near perfect relationship to one another with bearing across the full length of both ways.

So....my mill knee was scraped in with relative ease, then messed up by attempting to change the planar relationship to a "correct" one.

Doh! Unless you're a professional machine tool rebuilder, or a masochist (same thing?), always scrape straight down! :p

The manufacturer probably had the alignment right at the factory, and the ways tend to wear in the centers, so the edges of the ways usually keep the right geometry, unless (like you say) the machine is very worn.

If the latter part is true, it's probably more cost effective to just buy another machine :rolleyes:

J Tiers
05-21-2008, 09:35 PM
If you scrape without assessing the whole machine, yes you may get into trouble...... (er......make that WILL get into trouble).

BUT, if you have a genuine limited problem, such as a warped compound base, I see NO REASON to expand that to a complete re-scrape for "academic" reasons....

Scrape the *&^%$ thing so it sits flat, and get on with work..... If you can make any errors to the "material safe" side with regard to possible future work, so much the better. That may take some more measuring....

But otherwise, you can't use the machine, and may get into a "freeze" where you feel you have no choices but either leaving it be, or going on a two year project..... That helps nothing.....