>Somebody on WC or some other forum posted a link to the 19th century Joshua Rose text on machineshop practices. I finally found the section on handfiles and wanted to share it. I've regressed to fooling around with handfiles and plane sole flattening ... again. I thought I'd share the link for the 2 or 3 people on the site (or on the planet) who might have an interest.
Go to the Chapter 26 pdf link. File talk starts on pdf page 25 of 40.
>From another list, talking about plane sole flattening (should one want to) as per my site:
http://www.geocities.com/plybench/flatten.html
BugBear
QUOTE....
I like the method because you can vary the metal removing process.
Recently I've been having good results using an "abrasive chisel(tm)" which is a lath of beech, around 8"x1"x1/4", with the end fully rounded in the 1/4" dimension, and cambered in the 1" dimension.
Wrapping a cut 1" strip of abrasive around this (length ways) gives me a spot cutting tool. When the abrasive clogs or wears, I simply rotate it 1/4", bringing fresh abrasive into play.
This tool removes metal effectively, and is comfortabnle to use. I normally use 80 grit AlZi for most of the work, only changing to 180 grit when the entire sole (even the original low spots) has been touched by the 80 grit.
People have formed the impression (possibly due to my own emphasis) that my process is about accuracy. It *is* accurate, limited only by your patience.
However, it's also fast. In the early stages you can confidently shovel off metal from the high spots. This is because of the CERTAINTY the surface plate brings to the party.
Researching files, that Joshua Rose piece makes me think there are only two files suited to (actually designed for) flattening, both Swiss pattern. One is the "hand" and one is the "pillar". Both are said to taper in thickness from mid-file to tang and to end. I see from the Grobet site that's the way they are still made apparently. I'm really intrigued by what's said about the Valtitan's. The Simmonds here in the USA may also offer the same. I'm not so sure about the origin of Nicholson Swiss pattern files. I did find in my file collection one 14 inch second cut that has tapered thickness from middle to end but flat to the tang and worked on a Bailey #5 with it for awhile. Last night I fooled around with a little 4" #2 barrette file, cleaning up a Bailey around the mouth. Eventually it stopped cutting. I picked up a piece of steel and the file cut well. Made me wonder what the cause of that was.
It's very curious to me that there is so very little detail around about file types, methods and so forth. I realize I know next to nothing about it. Much like when I tried to use my first hand plane.
Yep, it was me. No ones cares about "print and cut" for flattening except me!
Researching files, that Joshua Rose piece makes me think there are only two files suited to (actually designed for) flattening, both Swiss pattern. One is the "hand" and one is the "pillar".
Another way I've seen recommended is to wire two files together attheir ends with a washer or coin "of suitable thickness" between them at the middle.
This creates 2 convex filing faces and (I'm told) won't snap the files.
Never tried it :-)
I would love to hear how you get on with your plane. Keep us posted.
"Yep, it was me. No ones cares about "print and cut" for flattening except me!"
Not true! I have found that it makes a big difference, especially for a smooth plane, to have a very flat accurate sole. I've even taken to flattening the face of my frogs this way. The frog flattening doesn't seem to matter as much with the old thin flexy Stanley blades, but if you use a new thick stiff blade, it can make a big difference.
I was reading "around", and found chapter 27 of Volume II.
In it he discusses the usual "3 plate" method of originating a surface plate, mainly working by filing, then scraping.
However, he then discusses taking the process further, as follows (page 134):
The filing process was continued with fine Groubet files, and testing the plates, rubbing them together sufficiently to mark them without the use of oil. Very short file strokes must be employed, and great care taken to apply the file to the exact neccessray spots and places.
Then, instead of using the scraper, No 0 French emery paper was used, wrapped around the end of a flat file. The plates being interchanged and trued with No. 0, No. 00 was used, and the testing and interchaging repeated. These grades of emery were then wrapped or folded around the curved end of a piece of wood, the plates interchanged and rubbed together as before, and the emery used as described for the scraper. Subsequently, Nos. 000 and 0000 French paper were similarly applied until the plates were finished.
.
.
.
In skillful hands this process very far surpasses, both in the superiority of its results and in rapidity of excecution, the scraping process, leaving a brilliant polished surface, so smooth that it feels as soft as satin, and the contact becomes so complete that no bearing marks can be distinguished.
So abrasive paper on a stick ain't new.
BugBear (who did invent this idea independently, honest!)
>Ya, it's turning into a fiasco. The plane isn't a #5, it's an early style round-side 605 I found in the basement the other day. I bought it cheap and planned to rehab it. Here we are a couple years later.
After putting up the milling machine thread last week, I realized that approach is very flawed and has no use to most on this forum. It occurred to me to put together another piece on rehabing this 605 with description of all my mistakes using hand methods to flatten the sole.
I gave up (temporarily) when I found most of the pics I was taking were blurring. I originally draw filed the sole, checking only with a Starrett straightedge (longitudinally, cross, and diagonally). When I thought it was perfect, I blued it on my granite surface plant. Geez! Why'd I do that? At this point, I've easily 3 or 4 hours trying to discover the best reference technique and the best filing technique. I'm still not there. I realize there are some fundamentals I'm ignorant of. That led me to the Rose book.
At the same time I'm filling a type 10ish #4 Bailey. It's flat enough but it does not work well on wood. Very puzzling. If I slap the toe the plane "twangs". If I pull the frog or loosen it completely, no twang. I've blued the frog/body embossments, done a little filing, but still not getting the twang out. The plane isn't going to work well until I get the twang out. You guys been down this road?
The blueing pattern shows the outcome of very careful draw filing. Terrible! Granted the area in front of the mouth is flat. That's not good enough. Sure the plane works okay, but I think it could work better.
>Well, I just used the Frei link and ordered two. $50! I swore to myself last year I wouldn't buy another tool until at least after the 2008 elections. This is completely on your recommendation. But don't give the pressure another thought.
Give us your opinion on this: The skin on these 80 and 100 year old planes is hard. Most of the real effort of this approach seems to be in getting below the skin. Basically, no Nicholson I have will get through it, at least in some spots. How deep do you think the skin is? In 0.001"? Once through the skin, I notice spots that cut well for awhile and then don't. Other spots cut fine. It doesn't seem to just be the file. Would you say this is local carbide density or some local variablity in hardness or some other characteristic? In fact, are files even the best things to be using on old cast iron soles that aren't more than 0.005" or so out of flattness?
Uneveness in thin-walled gray iron castings *LINK*
David Barnett - SW Florida
>"Once through the skin, I notice spots that cut well for awhile and then don't. Other spots cut fine. It doesn't seem to just be the file. Would you say this is local carbide density or some local variablity in hardness or some other characteristic?"
There is undoubtedly wide variability in thin walled castings such as bench plane bodies, just as you've described. What I know about such things, however, is sparse at best. I'd conjecture (scant knowledge never stopped me from taking the plunge) that how castings cool and how they are cooled is key, as this might determine how the graphite flakes and silica crystals are distributed. Lowered surface temperatures and mechanical operations (mostly forging) can 'steel' it into increased hardness through local carbon depletion, but I don't think that applies here. The 'skin' on bench plane castings is usually ground away before much else happens to the body. How much gets ground away and where may be a factor. How much would case, work or age-hardening and oxide formation add? Don't know. One thing, though, is oil penetration and iron's ability to hold onto it can make it surprisingly more challenging for a not-so-bitey file to get purchase. But again, I don't know. Not much help, huh?
As for whether files are the best flatteners for plane soles, I'd say "no". BUT... they are completely adequate for the task, reasonably affordable (compared to mass-production or high-tech solutions), fast enough, and entirely pleasing to use. It's been a long time since I've flattened a bench plane, and I was always a scraper kind-o-guy. But starting with a file makes good sense and gets you close in short time. So if you don't have a Wilton Square Wheel handy, a file's just the thing, and a hard sharp Swiss (or German or French) file is best of all. As far as handtools go, that is. I mean non-motorized hand tools. When my wrists were more supple and my health more robust, scraping was my idea of a fun during those New England winters. These days, I'd likely hit the high spots with a flexy angle grinding disc, my foredom, or my favorite: amodified 1x42" belt sander (with my caster tire contact wheel). Then I'd have a go with the scrapers. BTW, for anyone who's not all that comfortable with a straight machinist's scraper, a triangular-ended scrapers with a slight bend is very inuitive to control and can work nearly as fast. Doesn't leave quite so pretty a surface, though.
>Bob; you're probably "ok" for now, since your plane is (clearly!!) in the early stages of becoming flat.
But I think your blue is very thick, and is marking too large an area.
A little (at this stage) less blue will give a more informative marking.
As the plane becomes flatter, you'll want thinner and thinner blue on the plate.
Fortunately the act of "taking a print" removes some blue, so if you spread the blue around (to make it uniform) prior to each print the process is often self-controlling.
If you haven't already, I recommend reading Stephen Thomas and Forrest Addy's posts from Badger pond, held on my site
> Most of the real effort of this approach seems to be in getting below the skin. Basically, no Nicholson I have will get through it, at least in some spots.
Heh. In that case use something much tougher than a steel file.
Coarse (e.g. 80 grit) AlZi, either on a small block (for big areas) or wrapped a round a wooden pseudo-chisel (see other posts) will cut anything up to and including Rc63 tool steel!
>"But I think your blue is very thick, and is marking too large an area."
Paul, let's discuss this a little more for those who might be reading along.
The $20 Chinese granite surface plate I have has an alleged flatness of 0.0005" per 12". I have no way to check this and obviously wouldn't spend the dough to have it checked. I just accept it's somewhere in that vicinity. The "resolution", or size of an individual file "scoop", of metal removal with me using a flat file is probably somewhere around 1/2" square or more...probably more like 1" square. Thomas and Addy are talking about at least an order of magnitude greater resolution. They are talking about scraping with a resolution of 20-40 spots per square inch and scoops well under 0.0001" and using a reference surface much more precise. At my most accurate, I'd be doing well to have repeatable spots at 1 or 2 per square inch, but still only based on accuracy of 0.0005" at best.
I agree with you that it is those irridescence, silvery, "lighter" blued areas that are true contact points. The large dark blue areas are "false" and represent excess bluing squished into areas that wouldn't otherwise mark.
The reason that a thick excess of bluing helps me is that the true and false markings give me an idea of the big picture. What is marked are the mountains (true) and the foot hills (false). The plains and valleys remain unmarked. I plan on removing the mountains and the foothills and I want to do it fast, in big broad sweeps, not one mountain or foothill at a time. I want to get at the level of the broad plains view fast. That's why I'm fiddling with the files in the first place.
Of course, when it comes to techniques and approaches everyone has his own way. The fun of these kinds of exchanges is that we all have a chance to make our case as articulately as we can and then leave a decision about the best case to the jury.
>Bob; clearly you have a good understanding of the process, and the principles of the process.
I still think your blue is too thick - but not by much. I normally aim to mark around 30-40 percent of the sole. I guess you're up at 45-55 (by eye). This might per personal preferance as much as an absolute.
Controlling and exploiting blue thickness is one of the judgements involved in the process; if the blue is too thin, you're only marking mountains peaks (not mountains) and the work will go slowly.
Of course, if the blue is massively thick, you mark almost every part of the surface, and (thus) gain little information about how to proceed with the filing.
At the same time this thread was active, I posted a question to a large group of machinist types over on the PM forum. The core question was: How can I make a file more convex...for flattening plane soles.
It's an arduous thread to go through, largely because quite a few people seemed to misunderstand me and thought I was either nuts or was trying to get files straight.
In the end, I performed an experiement and illustrated the outcome with a picture. The posts from others that follow the picture are really brilliant and have all sorts of other applications/relevance.
>Ever see the movie, Twins, with Arnold Schwarzenegger and Danny DeVito?
The long post in that USENET thread from Michael Knights-Whittome was really superb.
Two things especially caught my attention:
"In passing - you can see light through a
quarter of a thou with a straight edge against the light."
I've often wondered what the rule of thumb limit was, but have never read anyone giving a guess. Unfortunately, he doesn't say what the contact width is (thickness of the straightedge). I notice that with my Starrett, about 3/16 thick, I can block out light but if I twist it a little to rest on the edge, I see variablity in light transmission. I'm tempted at times to buy a bevelled straight edge from starrett and compare.
The second thing of note was his saying curved files are hogwash and you need a perfectly flat one. Rose, as you know, emphasizes the geometric impossibility of having a file perfectly flat and that if it's anything other than convex you can't expect flattness in your workpiece. Picturing the geometry, I favor the Rose advice.
My Valtitans came Wednesday mid-day, after ordering on line Monday from Frei. Seems reasonable for an efficient process from California to Oregon.
These were the "hand" files. They visibly taper in thickness from middle to tip, but not from tang to middle. I'm temped to buy a cheaper and bigger "pillar", which I believe has the greatest chance of having both tapers. But, I'm wary of falling into my old habits of buying too much stuff.
I tried one on of the Valtitans on a Bailey #4 and on an old Spiers that's still "in need". I wasn't impressed. I was depressed.
So I'm seeing, for the moment, if I can flatten 2" diameter ends of 2" long piece of steel round. I can, if I'm a little careful with where the convexity of the file is (by feel) when starting and finishing each cut. It's very easy to see with a straightedge if I'm knocking off (chamfering) the edges and I'm not. Ten minutes of filing, then a minute or two lapping on a diamond plate, then a minute or two on a very flat soft Arknasas stone, then a one minute polish on a 8000 Norris waterston (dry). A completely flat mirror. I'm anxious to get a few of these done to see if I get "suction" with these "gage blocks" by wringing them together.
While waiting for the files to arrive, I regressed to scraping the sole I showed in the picture. Hadn't scraped from a year or more. I had unusually good and fast luck getting 6 point per inch or so after 10 or so scrape/print cycles. Maybe this is because I was already down around 0.001" flatness from filing as Knights-Whittome suggests.
The only scraper I get much of anywhere with is a piece of carbide I have attached to a Harbor Freight type scaper from India (which in itself is worthless except as a long handle).
Does your transition to sandpaper on a curved implement mean you've decided files may not be the best process from start to finish?
>Does your transition to sandpaper on a curved implement mean you've decided files may not be the best process from start to finish?
Yes; I have not yet tried the "distorted" file trick, which sounds to have potential. Bending a common file sounds easier and cheaper than sourcing a rare file.
But I was having trouble with the files leaving deep "scratches" that were hard to control and difficult to remove. My plane soles were flat but not cosmetically nice.
However, files (and scrapers) are very good at removing local "highs" (blue areas to you and me ;-) without risking surrounding area.
I tried abrasive on small blocks (1" square with soles of varying convexity). These worked quite well (AlZi, 80 grit), but I had no feeling in my fingers for around 4 days after using them for 2 hours - the holding pressure was too high, since they didn't have handles.
The pseudo-chisel (wooden lath with cambered end) I described is more confortable to hold and manipulate, but only works VERY small areas.
I may try to come up with a bigger version for the early (rougher) stages of flattening.
In all cases, the abrasive techniques are not quite to easy to localise as files and scrapers. Removing slowly (when needed) is easy though - just use finer grit.
I still haven't found a "perfect" solution, and I have no more planes to flatten. I can only work on #4 and #5, since I only have a 9x12 surface plate.
I may have to borrow a friends plane, just to experiment on.
>I know what you're are saying about the surface after scraping or filing. Last night I did use the paper-on-a-stick method to take the more gawdy relections and irridesences off a scrape job. And, I have the same problem with not enough unflat smoothers. But I have a new idea: make them slightly convex length wise, try them out as "improved" smoothers and if they don't work as well, reflatten them. Did you ever read the very erudite Forest Addy rebuttal against flat sole?
Earlier today I read your exchange with Ian on UK and noticed his disappointment, or impatience, with scraping steel soles. Do you have any insight about his use of the term "snow balling": "On another note BB....have you ever flattened a mild steel soled plane by scraping or just cast iron soled. i had a go but stopped rather quickly when the snowball effect started...just wondering if it was me or something that is notoriously difficult."
The only thing I can think is that it is a metaphor for picking up a chuck of previously removed steel debris (either with a file or scrapper), impailing it on a tooth or edge, and plowing an ugly furrow with it ... i.e, "pinning" in file talk. Still, "snowballing" doesn't seem such an apt for that. Do you now what he means by "snow balling"? I understand so little British jargon.
