Bringing the Art of Scraping into the 21st Century?

[Sparky_NY]

Originally Posted by handlewanker Hi John, sigh, once again I will state, not out of wanting to be awkward but just to produce the end product as is required, that if you use ANY ABRASIVE method you will by the nature of the abrasive material make a ploughed field, (multiple vee grooves) of the surface and this will leave you with millions of sharp pointed crumbly peaks that under the sliding of one piece of metal on another just shear off and so you end up with a rapidly lower, although miniscule, area of non support. Abrasives will produce grooves and sharp jagged peaks, whereas scraping will result in a series of undulating hills, more condusive to oil film retention and load bearing. Ian.

Read this whole thread with great interest. I have a well worn bridgeport that I would like someday to tackle. I have no experience at all and have read everything I can find on the subject of scraping and other possible methods. The table has the usual sag that no doubt will require a lot of material removal to correct.

Now for the probably dumb question. Wouldn't the sharp peaks, grooves etc. described above show up when the part is blued up against the surface plate? The pictures of the "dremel" method surfaces appear to be lacking any such sharp edges or grooves, just a nice oval spot pattern.

Also, I read that the scraping method also leaves sharp burrs at the end of a stroke thus the need to lightly stone the finished hand scraped surface.

What am I overlooking?

[Oldmanandhistoy]

Originally Posted by Sparky_NY Read this whole thread with great interest. I have a well worn bridgeport that I would like someday to tackle.

Hi,

A bit late replying to this as I didn't get an email reminder.

I would say if you have an old Bridge port the iron will be of good quality so I would stick to conventional methods.

If you want to give my method a go for removing iron faster you could always finish off with a scraper.

Good look,
John

[whitis]

Originally Posted by ironDigit but seriously i guess i was just looking for a way to do this without the appropriate(read expensive to me)master surfaces and squares.

I am no expert but I have been reading the tome, Conelly's Machine Tool Repair and Applications of Hand Scrapping as well as some other metrology stuff.

You can have your master surfaces for the price of three dimensionally stable (preferably) pieces of metal plus labor (each). You can start with nothing precise and make your surface plates, straight edges, squares, v-blocks, dovetail gauges, and parallels with no external reference. The will have random uncalibrated dimensions but with enough pieces and time they will be flat, square, and parallel.

The trick is, you need to scrape the three pieces relative to each other. With two pieces, you could wind up with each mirroring the convexities and concavities of the other. But three surfaces cannot spot true to each other unless they are all truly flat (within tolerance). And you want three, anyway, because otherwise, how do you know that your precision master surface is STILL a precision master surface. You use one as your working master and periodically verify it against the other two.

You use each one in turn as the temporary master and spot the others to it, scrape the other two, then for the next pass you choose a different one for the temporary master.

1: Spot B and C against A
2: Spot A and C against B
3: Spot A and B against C
4: Spot B and C against A
5: spot A and C against B
6: spot A and B against C
...
You also swap the pieces end for end.

If your hunks of metal aren't very stable, you have to repeat this process frequently.

Likewise, I believe if you have a flat surface plate, three blocks can be tested on the flat surface plate for being right angles. I.E. Spot B and C against A on the reference plate for both flatness and angle (which will show up as not flat). Of course, you could get 90 degrees in one axis but not the other but there are ways to compensate for this, too. Like using more than two adjacent surfaces on each block and rotating them. If you didn't first prepare flat surfaces to use at this stage, you could probably use 5 blocks (90+90+90+90=360), 4 at a time. As more surfaces are involved, the various rotations involved could get complicated. Using the same N+1 technique, you could probably also make 45 and 60 degree references.

The trick is getting the dimensionally stable pieces of metal so you don't need to keep adjusting them. They need to be of a shape such that they don't bend under their own weight and they need to be aged or stress relieved. Even rotating them between the oven and the freezer for a few cycles can reportedly help age them. Cast iron is more stable than hot rolled steel which is more stable than cold rolled. Aluminum is a bit soft and expands a lot with temperture. Engineered shapes such as angles, I beams, square/rectangular tubes, trusses, and castings have better strength to weight ratios when it comes to holding their shape against gravity.

Using the sanding technique or careful use of a power scraper, you can even start with pieces that are way out of alignment to start. Though if you remove drastic amounts of metal, this changes the internal stresses.

[Mcgyver]

you know, I've done a lot of scraping and am of the opinion that this is generating a flat surface by abrasives in a dremel is not a good idea. If you take a finished hand scraped piece of work and indicated from high to low you might barely move the needle on a 10th indicator. the difference between high and low is very very small. As this machine surfaces wear a tiny bit, there's a another bearing point a fraction of tenth below ready to step up a do its duty. What variation between high and low are you getting with abrasives?

Next, I push back on the use of a medium stone. Use a file with the points of the teeth stoned off. it will skate across the surface but any burrs protruding with find themselves between the teeth and get cut off...OR use something like an Aransas stone that so hard/fine you won't affect the overall surface. A medium stone just sounds like it is capable aggressive removal rates and that is not what we want. Of course there is latitude; roughing a 40' planer bed or finishing a bench top instrument calls for a different technique. Understanding what you are doing and apply common sense makes scraping easy

whitis, if you want to do original generation of flats, and then plan on redoing it occasionally because of movement, you are special please take pics. Seriously though, while generation might be of interest it makes little practical sense when you can buy a top quality Starrett plate for a couple of hundred bucks....the hunks of cast iron'll cost you that

To save you a ton of work on scraping for squareness: make a cube, get two sides perfectly parallel P1 & P2 and another side flat, A. put the cube on the plate on A. Use a tenths indicator at the top of P1 & P2 with the ball of the surface gage arm at the bottom of the cube. the difference indicated between P1 and P2 is how out of square you are - scrape A until there is no difference and you are square

Also, I read that the scraping method also leaves sharp burrs at the end of a stroke thus the need to lightly stone the finished hand scraped surface.

The burrs are very small and mostly can't even be felt - you want to remove them so they don't give a false reading - we're generating a surface flat the to 10ths so little things matter. a quick rub with an arkansas stone and they're gone. Abrasives cut exactly the same principal as any other cutting tool, lathe tool, mill cutter, scraper etc, albiet often on a smaller scale. Point being you may just as well be left with burrs from the abrasive technique that, given the small scale we're working with, would affect things. Scraping is something you just have to do, thinking about it to much will hurt...it's a very simple process that makes a lot more sense once you done it

here's a pic of a tool makers cube i have square and parallel to a tenth via hand scraping. The cube is an Eclipse (quality English and probably 20 years old) so it is very stable. I have a few extras of these tool makers cubes if someone is interested.

Finished cube (well, ok not technicall a cube, but thats what they're called )





How to check for parallelism






How to check for squareness.



With enough bearing points and careful work, this becomes a master precision square....or you can be rough and ready with it and clamp stuff to it for grinding which will come out almost equally as square

[DMF_TomB]

For specialized applications requiring slots parallel to 0.00001" or 10 millions of an inch. We used an capacitance gage traveling the slot to measure and graph it out full size.

We used phenolic blocks sanded flat on a piece of sandpaper on a granite surface plate. To these block we used double sided tape and applied 400 grit and 600 grit sandpaper to. We change sandpaper often. We could easily use micron paper 10 to 100 times finer and get a surface that you can see your reflection in if we wanted too. In fact the reflection of zebra striped lines and checking the reflection quality is also a way of checking surface quality. Also we work in a temperature controlled clean room as dust can seriously effect readings.

Now the surfaces we got were just short of a mirror finish. When assembling the equipment even fingerprints and even drops of water will affect the slot width. I am guessing this is on the order of 10 to 100 times finer than scraping.

take you scraped block and measure it with a gage that can show 10 millions. you will see how flat it is compared to sanding. also we were working with titanium as iron and steel would loose surface quality due to oxidation within 24 hours.

now the true test of gage block quality is when clean, you can wring them together and they will stick to one another. I believe gage blocks are lapped similiar to making glass lenses by an abrasion process.

do your scrapped objects stick to one another ?

[Mcgyver]

The use of abrasives as you refer to hasn't much to do with the subject of the thread; that is using a dremel to effect localized material removal in lieu of scraping. Besides, it’s still apples to oranges, lapping doesn’t do what scraping does and techniques used for gauge block manufacturing aren’t of much use where scraping would be used. I guess one could argue if a fine enough abrasive was used the dremel technique wouldn't have a depth of cut anymore than that of scraping....this would be very difficult though as it is the geometry of the scrapers cutting edge that controls depth cut. With such a fine abrasive that its taking a tenth or so, it'd think it would be quite time consuming. Scraping btw is probably quicker than the uninitiated assumes it to be; the work on that cube for example would be measured in hours not days or weeks to its to highest standard of accuracy one likley able to do in a machine or home shop

We used phenolic blocks sanded flat on a piece of sandpaper on a granite surface plate. To these block we used double sided tape and applied 400 grit and 600 grit sandpaper to. We change sandpaper often. We could easily use micron paper 10 to 100 times finer and get a surface that you can see your reflection in if we wanted too. In fact the reflection of zebra striped lines and checking the reflection quality is also a way of checking surface quality. Also we work in a temperature controlled clean room as dust can seriously effect readings. Now the surfaces we got were just short of a mirror finish. When assembling the equipment even fingerprints and even drops of water will affect the slot width. I am guessing this is on the order of 10 to 100 times finer than scraping.

Finer in what sense? Its not difficult to put a mirror finish on something with abrasives, but try to keep it flat and parallel over a substantial area and it can be challenging. There are ways to do so but they are not practical for the making of shop tooling. Do you make your 8" angle plates or set up your machine tool bearing ways to 10 μin’s? Really though, they are not competing techniques and ideally the engineer knows a variety of techniques to apply as the situation demands

also we were working with titanium as iron and steel would loose surface quality due to oxidation within 24 hours

gauge blocks are made of tool steel are to millionths of an inch and seem to hold up ok.

Originally Posted by DMF_TomB now the true test of gage block quality is when clean, you can wring them together and they will stick to one another. I believe gage blocks are lapped similar to making glass lenses by an abrasion process. Do your scrapped objects stick to one another ?

there's no claim to make anything to gauge block accuracy, nor is the thread about gauge blocks…..but how accurate is scraping? Depends on the accuracy of the surface plate of course, but a surface well done won’t create much movement on a 10th’s indicator’s needle (so it’s not far off 10 millionths). At that point, being with a 10th of thou over a large area, you are approach the accuracy of the surface plate, at least quality tool room ones likely to be found in our shops. Light pieces will wring with a bit of oil and as you can see from the photo’s its reflective. Still, all that is irrelevant as scraping is different than lapping as is different from the dremel idea of this thread

interesting work you are doing re the parallel slots. When you gauge it to 10 μin, what do you do - map the high spots and concetrate the abrasive action toward them?

[RICHARD ZASTROW]

Just to add to Mcgyvers post, #102: Generally, we don't want the surfaces to "wring" together. The objective of most scraped surfaces is to provide a surface, as accurate as is REQUIRED FOR THE APPLICATION, that allows movement without "stick-slip".

Normally that will include lubrication which the minuscule hollows of scraping provide.
ex. oil fills these hollows.

If the surface is "gauge block quality" the mating surfaces would "wring" together and stick.

I realize there are "plastic" surfaces that slide well and flat air bearings whose surfaces would function well at super smooth, flat etc conditions.

That's overkill and impractical for a general machine shop environment.

Maybe practical for a Cat Herder LOL

Dick Z

[JerryFlyGuy]

Kinda late to the part [re: the overall thread] and it's been an interesting read. I'm curious as to how large an area can be scraped accuratly? Does it require an extra large surface to do it accuratly?

Curious more than anything

J

[handlewanker]

Hi Jerry, scaping is done to produce a surface that is........flat ,round, oval, whatever.

The reference surface is what will reveal the scaped surface imperfections, and if you have a large area to be scraped or rescraped, you need a correspondingly large area, be it a table or straight edge to show up the imperfections of the surface your're trying to get to a similar surface condition.

A lathe bed is something that will try your skill due to the various planes that won't allow flat table application, but must be true to one another as the design dictated.
Ian.

[JerryFlyGuy]

Ian, so are you saying that if I've somethng that is 4ft long I need a 4ft long reference? Can not a smaller reference be used over a larger area w/ the same result?

[Oldmanandhistoy]

Originally Posted by Mcgyver What variation between high and low are you getting with abrasives?

I can achieve less than +/- 0.002mm (best I can measure with the equipment I have).

I say don’t knock it until you have tried it, using wet and dry paper belt material and a light touch minuscule amounts can be removed very accurately.

Btw I have learned a great deal about scrapping from you on the zone and would just like to say thank you for your time posting.

John

EDIT: In fact it is a lot less than 0.002mm as I can detect almost no movement on a 0.002mm DTI

[Mcgyver]

thanks John,

Originally Posted by Oldmanandhistoy I say don’t knock it until you have tried it, using wet and dry paper belt material and a light touch minuscule amounts can be removed very accurately.

fair enough, I haven't tried it....its the uncertainty of the depth of cut that concerns me. You say there is no movement on a .002mm indicator; well that's pretty good then if there is less than .002 mm between highs and lows.

scraping has kind of self regulating depth cut built in by the included angle of the cutting edge. I'm thinking there isn't this same regulation with a say a die grinder in that depth of cut will vary, I'd think, quite a bit with dwell. then again if you're getting less than .002mm high to low across the work maybe it doesn't although i dont't understand why....