We started calling this lathe "FrankenLathe" because there were times when it seemed to have been cobbled together from parts of several different machines. The color scheme certainly looks that way. Ed Richley is helping me bring this beast back to life (hmm, we really do need a Jacob's ladder for the shop...)
It's a 9" swing, 4' bed length South Bend model B9W, s/n 127990. Shipped to the Electric Vacuum Cleaner Company in Cleveland, 8/27/43. It conforms to the requirements of the War Production Board: says so right on the bed. (I have no idea what this means). Plain headstock bearings. Underneath drive; in the full title photo you can see the bench that I got it with, along with the disassembled motor drive: that's a huge 1/2 HP DC motor that's controlled by a Variac. The bench is, uh, less than stable. You can also see the home made tool post that a former owner (Dino Santi) made for it.
I got a copy of the "serial card" from South Bend (just call the parts department and ask), and it turns out that the machine is aptly named. At least the headstock, saddle and tailstock have been replaced with parts marked with different "unit codes" than are on the serial card. No one can come up with drawings for the headstock (unit code UH111), but Rose in Parts found a photo that seems similar, supposedly used on lathes built from 1936-1941(!) The serial card is stamped "REVISED", though it's not clear what was revised on it. It's confusing in itself, since the lathe is clearly a model B (standard change gears) but there is an entry for a gear box unit code...
Here's another badly lit view of the lathe and its accessories. (Note that I had removed the spindle and feed screws before taking some of these photos - they did come with the lathe!)
It seems to be in good shape, some wear in the bed, and lots of dirt. The white parts have two different coats of white. The green parts have, for the most part, two greens, red under that, and then the original South Bend gray. The compound and cross feed had red, with gray under only part of it. Needless to say, it's coming completely apart to be cleaned and painted. I'm trying to avoid rescraping - I'd like to use this lathe sometime soon.
There are a lot of things that we haven't been able to figure out: the earliest drawings I've been able to unearth are from 1957, and the headstock is quite different from the one on my lathe. For example, the cover goes down through the ways. There are two holes where I've removed drive screws - these must have held a label plate once upon a time, but I don't know what it would have said - the threads and feeds plate and lubrication plate are intact on the end cover. To make things more interesting, the two drive screw holes are not aligned horizontally! At least we can be sure that the two 1/4-20 holes below the cover were once used to hold a drum-style motor controller; they line up perfectly with the used Dayton controller I found surplus.
From the side, there are more mysteries. There's a hole in the top of the casting next to the quill - it's been drilled, spot-faced and threaded 3/8-16. If it was a back drive lathe with separate guards, I'd think that the quill guard bolted here, but it's not. I've seen one photo (in How to Run a Lathe) of a 9" lathe with a motor controller mounted here - but then what are the two holes in the bed under the headstock? (One reader has identified this as the mounting point for a handlever collet closer; this seems to be the right answer. It goes along with the hand-lever tailstock to indicate that this was used as a second operation machine in a manufacturing setup for at least part of its life.)
Lubrication of the spindle is also curious. There is a capillary oiler coming up from each oil cup. But there is also some sort of circular felt oiler around the edge of the bearing; it is anchored in place by a small spring clip that fits into a hole that goes down to the oil cup. I've never seen anything like this in any drawings. There was a fair bit of crud in the slots in the bearings - we've been wondering if perhaps there should be felt wicking in the slots to draw oil from this circular felt along the spindle.
A reader from the net has provided me with copies of pages from a 1941 South Bend publication, Parts List for 9-inch "Workshop" South Bend Lathes Models A, B, and C. This shows that the headstock and underneath motor drive parts that I have were specific to the Model B and C machines of that era, which may be why I can't find anything about them - most folks concentrate on the Model A with its quick change gearbox.
The original motor is another mystery. The serial card indicates that it was a 1/2 HP, 1725 RPM, 115v 1 phase motor, GE part 5KC63AB494. I called a local motor company who didn't have cross reference information that went back far enough to read the number. Ed and I figured out from the bolt holes in the motor mount casting that it must be a NEMA 66 Frame; Grainger doesn't list any 1/2 HP motors on that frame. The woman I talked to said that if I buy a new motor, I'd probably want a 3/4 HP at least anyway - the definition of horsepower hasn't changed, but the reliability standards have, and a new 1/2 HP motor won't be as sturdy as the original.
I called GE Motors in Ft. Wayne, IN and asked about the number. I got shuffled around a bit, finally to a fellow named Jim who knew it was built in the 40s, and put me on hold while he got out "the old 3x5 card with the particulars". He came back on, saying "Oh, that was a dandy. It was a special we made just for South Bend, instantly reversing, capacitor mounted specially, with a relay for something that I can't tell ... yes, a real dandy. We're not making anything like that any more." He faxed me an outline drawing that shows the case dimensions, and a recommendation for a motor in current production with a connection diagram. I'd really like to understand how "Instant Reversing" single phase motors work! Despite much searching, all we've found is vague references to "special relays". The connection diagram indicates that there is some solid state component that controls the start winding and capacitor, and no centrifugal switch at all...
Several people have emailed about these motors.
Dick McBirney wrote
"I bought a South Bend 9" Model C brand new from SB in 1956 with an instant reversing motor, and I still have it. Since it was made in 1956, I seriously doubt that it has any solid state components. From the day I got it brand new, the instant reverse was a little hesitant to work. From full speed ahead, I would throw into reverse with one motion. The first time I did it, the motor growled (without reversing), the shop light dimmed, and I hastily shut it off. By carefully experimenting, I found that I could coax the instant reverse into operation as follows:
1 - hit forward and let it come up to speed
2 - shut it off and let it coast until I heard the centrifugal switch (it SOUNDS like it has one) cut in
3 - hit reverse and let it come up to speed
4 - shut it off, let it coast until I heard the centrifugal switch cut in
5 - repeat steps 1 thru 4 several times, then the instant reverse would work.
However, I used it so rarely that it was more of a curiosity than a necessity, so I have not used the instant reverse in years."
Peter Haas wrote:
Here is some data from another SB 9 ...
General Electric A-C Motor
Temp Rise: 40 deg. C
Time Rating: Cont. PBK
Internal switch: General Electric "Reverswitch", U.S. Patent 2,246,183
A combination of the double contact centrifugal switch, which actually "remembers" which direction the motor was started and the GE "Reverswitch" are used to obtain the instant-reverse characteristic.
Today, a solid-state switch would likely replace the "Reverswitch".
For motor operation details, Rosenberg's book is probably the best.
The second edition is young enough to cover the later developments, but
still old enough to cover the earlier developments.
Third edition, new, Rosenberg, et. al.
Second edition, used, Rosenberg ...
Specifically, an IR motor has a single-pole double-throw centrifugal switch, one which "remembers" the direction of rotation, and "prepares" the motor for a possible instant reverse. The motor itself starts as a normal capacitor start motor, and it runs as a normal induction run motor. Most IR motors are small enough that these don't need to be capacitor start/capacitor can.
Anyway, I've never seen an IR-ing capacitor start/capacitor run motor. Perhaps IR-ing such a motor is injurious to the run capacitors. When an IR motor is running the special centrifugal switch has already been "prepared" for an IR, even if not needed.
Should such a motor be reversed, instead of it behaving as a normal induction run motor, which would be to continue to operate in the started direction, forever, until all power was removed, the IR motor engages the start winding while disengaging the run winding. The start winding is now fully energized and is working against the rotating armature, slowing it down very quickly.
When the armature has come to a stop, the centrifugal switch engages the
run winding, and now the start and run windings are working together in
the new direction, and the motor comes up to speed in the new direction
When up to nearly synchronous speed, the centrifugal switch drops out,
and the motor is now operating as an induction run motor in the new
This sequence has also resulted in the special centrifugal switch being
reset for the original direction, so it is capable if being instantly
And the entire cycle is complete.
The referenced G.E. "Reverswitch" seems to be a G.E. development to speed up the instant reversibility by introducing a current sensitive relay to drop out the run winding at just the right point in the above-described cycle so the reversing is exceptionally fast. Without a "Reverswitch" the reversing is slower, but it still reverses as if "plug reversed", it just takes a lot longer. The combination of the special double-throw centrifugal switch and the "Reverswitch" makes the motor reverse in a couple of revolutions, or less.
I ultimately decided that I didn't need instant reversing - it's of questionable usefulness except in some very specific threading operations that I'm unlikely to do. I found myself a nice old 66 frame motor and mounted it up on the original mount with a suitable drum controller.
If you want a modern IR motor, you can get a dandy 3/4 HP from GE. It will be 56 frame, and pricey. Grainger sells them as part 6K880.
Then there's this accessory - the picture is awful, but it's the best I could do. It's cast, seems to have South Bend gray at the bottom. There's a shank coming out of the machined face, ending with two thin nuts. There's a fairly deep hole on the right side of the machined face; there's also a not-so-deep hole in the body at a right angle to the shank. It seems as if this could have held some sort of dial or stop that was meant to be rotated - one of the holes could hold a spring and ball for a detent... doesn't appear in any catalog that we've found.
I'm still searching for answers - if you happen to know, or have an idea of where I might find out, or better yet have parts drawings or photos of one of these lathes in its full glory, or know where I might find the proper motor, please send me mail.
The first step was turning the assembled lathe, covered in four different exterior colors (grey, green, white and red) and more colors underneath, not to mention a lot of dirt, into an exploded parts collection. I've been gathering parts drawings from a bunch of sources - South Bend was the least helpful of all these, it turns out. They've apparently redrawn the 9" parts book on a CAD system and sell a poor copy of a bad printout. Too bad; the old line drawings were much nicer.
Luckily, a few folks from the net were kind enough to send me copies of their parts books, from 1965 and 1957. That covered most everything, except the headstock. Grant Erwin came through on this - he sent me a copy of South Bend bulletin 19-R, covering the 9" Workbench Lathes. It has a photo of the parts of my headstock, complete with (outdated, unfortunately) part numbers. Even South Bend doesn't seem to have a copy of this! Big Ralph and Wanda in the parts department are having a lot of fun helping me track down some history on this machine.
Dave Sobel was kind enough to send me three different styles of saddle clamp to try, after the first one he sent didn't fit. He's pretty confused by this lathe - he says that he's only seen that headstock on wide bed lathes, and he's only seen about 5 or 6 underdrive lathes of this vintage, ever. No wonder South Bend doesn't have parts drawings any more!
The most effective method I've found for removing paint is a trisodium phosphate solution, hot. (TSP powder is available at most hardware stores near the paint. Be sure to get the real thing, not the "politically correct" substitute which is usually marked TPS.) TSP will attack aluminum and magnesium, but doesn't seem to bother brass.
I bought a galvanized bucket which I fill with solution and parts, and set it on a hot plate. I let the solution boil for about 20 minutes, then rinse with water to cool. I then do a final rinse in the parts washer.
This gets pretty much all the paint off (it definitely removes all the oil). The original South Bend gray paint is very tough - some of it comes off completely, some needs scrubbing under the mineral spirits bath in the parts washer.
Bigger parts (too big for the galvanized bucket) get soaked in a plastic pan. I boil water on the stove and make up the TSP solution hot, and pour it over the parts. Then I let them sit for a day or two. This works - the top layers of paint come off in sheets - but it is not as effective on the SB gray.
The final chemical approach is to use Jasco water soluble paint stripper. This has methylene chloride and other nasty stuff in it - it is very effective but very unfriendly to your skin. Keep running water handy. We used this on the bed: it did very well causing the top layers of green and red paint to wrinkle and flake off, but all it did to the bottom layers of gray was make them soft. To get that off, as well as all the various grease and grime, we ended up sandblasting (mask the ways first!). Almost 50 pounds of sand later (a lot!) the bed looked pristine.
Parts that are going to be repainted get sprayed down with WD-40 or Liquid Wrench. Parts that are finished steel and just need to be assembled are wiped down with Microil (a fine instrument oil from Kano Labs).
The first thing I attacked was the 5" 3jaw chuck, which had a fair amount of surface rust. I took 00 steel wool and Liquid Wrench and went to work. This was very effective - it got all the big chunks off. Patience pays off here. The final pass was done with a fine wire brush mounted in a Dremel tool.
On bigger cast iron pieces that are going to be painted, I take a different approach. First I use steel wool and oil to knock off the big chunks. Then I use naval jelly (which works a whole lot better than I remember). This gets off the fine rust in the pits and leaves a phosphate coating that prevents immediate rerusting. Finally, a large wire wheel in a buffer motor gets off the remaining rust and paint, leaving a nice shiny surface. On castings with a lot of surface area (usually too heavy to hold up to the buffer) I use a stripper wheel that's some variant of Scotchbrite with an embedded abrasive, mounted in a die grinder. For tight corners, a wire brush in the die grinder is great. Wipe down with WD-40 to prevent re-rusting while waiting for paint.
Truing and making parts
The first thing we started was cleaning up some of the machine parts, namely the bushings for the cross-feed and compound feed lead screws. They had both been pretty badly scarred from years of people using the wrong tool to try to remove them (pliers, apparently). We made up a fixture to hold them in a lathe and took some fine cuts to remove the biggest scars, leaving a nice satin finish with only a few blemishes.
We also trued the surface that faces against the graduated collar, since this is a critical dimension. We were lucky in that the surface that faces the casting indicated true - all the high spots and divots and wear marks were on the collar side. Make another fixture (inside threads are so much fun!), face it square, thread in the bushing and face it square. They look so much nicer.
The next project was the countershaft and casting. Originally, the shaft ran directly in the cast iron, and both parts were pretty worn. We decided that we didn't really want to bore the original casting to install bushings. So I got some 3/4 x 7/8 x 1" long sintered bronze bushings which are a nice light press fit. We turned the existing shaft down to 3/4, except for the end that holds the motor belt pulley - it is secured by a tapered pin, and we couldn't figure a good way to make a sleeve and ensure that the pulley, the sleeve and a new shaft would all line up properly on the taper. A sleeve for the cone pulley was easy to make, although the cone pulley is running just a bit out of true; we'll have to revisit this. I'm also not quite sure how I'm going to oil this arrangement. There are oilers in the casting, and they fed into a felt wick that ran across the shaft. I could do that now, but I'm not sure how well the oil will soak through the bronze bushings; I'd like to make sure that the felt touches the shaft. It'll probably be fine...
It was very nice to see some pieces go back together, even though I know they have to come apart again to be painted.
We managed to find some 66 frame motors - even a GE one. No instant reversing, though. Now we have to figure out just how all the pieces of the puzzle fit together - none of the drawings in How to Run a Lathe or Keep Your Lathe in Trim show this particular arrangment. The "obvious" way requires a lot of depth, and it looks like it also requires a spacer on the shaft that isn't shown in the parts book...
It would be nice to figure out how IR motors work, and see if we can configure this one to do it. There must be some extra circuitry inside, both to detect direction and to engage the start winding in the opposite phase (as a brake?) even when the centrifugal switch is out. Or something else. We have yet to find a book that describes the guts of an IR motor. We found one book that says "special relays can be used", we found a product note from 1957 describing a shaft-mounted switch, used in conjunction with two start windings, and we have the modern connection diagram which seems to have a solid state controller and no centrifugal switch!
The first parts I "refinished" were the various name and data plates. Removing the drive screws that hold the plates on is an interesting exercise in itself. They have a very steep multi-start CW thread, and are driven in with pressure. If you can apply pressure from behind, they will twist slightly as they come out of the hole.
Most of them weren't blind, and could be driven out from the back with a "suitable drift", as all the British auto workshop manuals call them. If you don't have some, get a set of drive pin punches - mine are General, from a hardware store, though I've replaced a couple with Starrett as I broke them. I didn't know better; now I'd just buy Starrett. They're not appreciably more expensive.
A few of the screws were in blind holes. For those, take a very thin screwdriver that you don't care much about (I have a set from Radio Shack that fits this description perfectly) and use it as a wedge under the corner of the plate. Tap, wiggle, lift enough that you can grab the head with a pair of pliers; twist CCW and pull and you've got it. You may have to reflatten the plate. If you get really desperate, use a fine kerf saw or cutoff wheel to make a slot in the head of the screw and use a screwdriver.
The plates on the bed were mostly not blind, but there wasn't enough clearance to use a punch. I used the 90 degree end on a scriber and a small hammer to get them started out, then pliers to grab, pull and twist the head.
The plates are all etched and painted - brass, steel or aluminum, etched to leave raised areas, with the areas that have been etched away filled with paint. A few had been coated with some sort of clear lacquer that has yellowed with age. I started with "Nevr-Dull" wadding - it's a cloth wadding that's been impregnated with some sort of mild solvent. It smells and feels like lighter fluid, but the can claims it's non-toxic. It's great stuff - softens the dirt without affecting the paint. Most of the dirt just wiped away; I used a dental pick to clean out the letters.
I made the mistake of using lacquer thinner to try to get the yellowed coating off the back of one of the plates - it worked fine, but some of it seeped over the edge onto the front and softened the paint. It wasn't possible to save it all, so there's a thin spot there now.
The Nevr-Dull leaves an OK finish, but removes some of the shine from the paint (the metal looks great). I restored the shine by going over the surface with Simmichrome polish. I then sealed it all with Nyalic spray (a clear finish that claims not to yellow with age). The plates look great!
Masking parts to be painted is very tedious work. I'm pretty certain that South Bend painted and baked the castings before machining - I don't really have that option! So all the machined surfaces need to be protected before spraying. It took an afternoon to mask off the saddle, apron, cross-slide and compound. I use masking tape of various sized for most places; apply the tape, burnish it down, and use a sharp single edge to trim the tape against edges. Holes get filled with rolled up masking tape or modeling clay ( not a good idea, it turns out). For very large areas, tape the edges and trim, then use a large piece of paper held with tape at the edges to cover the interior. When masking, try to keep in mind that you will have to hold these pieces to paint them - hanging is best in the long run, so that you can manipulate the pieces for coverage as well as allowing them to dry undisturbed.
One person wrote to me about a product called Liquid Frisket. This is used by air brush artists (a "frisket" is their name for a shape used to mask an area). It can be painted on and trimmed, and removed easily after painting. I never tracked any down to try it (check art supply stores).
Welding Cast Iron
The headstock cover had several cracks in it; some had been vee'd out and filled with some sort of brass braze. Upon trial assembly, it was pretty obvious why it was cracked; the cracks were in the portion that covers the large back gear, and when the headstock cover is opened, this portion contacts the headstock casting. It's not very thick, and I can imagine that repeated vigorous flipping open and shut, as might happen in a production shop, would quickly lead to damage.
Ed took the headstock cover away and welded up all the cracks:
"This cover had evidently been dropped. There was one area of several square inches in the back where a pivot screw mounts which had been completely replaced by a steel part made to fit roughly where the original piece had been lost. Furthermore, another piece of cast iron, presumably from the original, was brazed in place next to the steel. Brass brazing material was holding this mess together, resulting in what looked like someone's first practice project. The vee-d out areas of the steel-iron joint were missing braze metal. There were cracks. There were a few gaping voids. Chris decided that we had little to lose by letting me work on it.
"In my vast experience with oxy-acetylene welding I've managed to log a grand total of maybe 20 or even 30 minutes behind the torch. As the headstock cover is made of cast iron, I felt that it would present no real difficulties for such an old pro as myself. Thus, armed with my highly refined skills and a large supply of All-State No. 3 Cast iron rods, I proceeded to improve upon the efforts of the previous master craftsman whose work was less than impressive.
"Using some carbide burrs on a die grinder, I removed some brass. The general idea was to replace all the brass with cast iron filler rod, and to make the whole thing look better. This was to be done one region at a time, so that a re-alignment of the pieces would not be necessary. Okay, now I had a few regions near the edge that could be back-filled. I popped the casting into the oven at 450F, and set up a wheelbarrow outside with some sand in it to use as a heat blanket/place to drop hot things.
"When the roast was done, I took it outside and started torching. This filler metal is supposed to melt at a lower temperature than the base metal. It does, but not by much. Thus, I managed to melt through, and to recede the edge, which was pretty thin. Thin metal is not my specialty. Neither is gas welding. So, this is how you learn. On somebody else's precious antique.
"After a bunch of new metal was in place, I let it cool a little, and ground it away to make the surfaces uniform. I went too far. Furthermore, the fusion wasn't too good. Being impatient, I didn't re-pre-heat, and just slowly brought in the torch. This was an OK thing to do (much to my surprise) as long as I went slowly. I couldn't get the knack of laying a bead. I just could get a well-established "puddle" like with arc welding. So, I discovered that putting down a blob, and then heating a region until it fused was the best I could do. It doesn't always flow easily and often beads up. Lots of flux helps (I used "Anti-Borax" cast iron flux. It's red.), and also stops the little explosions (which leave voids and blow holes in the work. So, it takes some effort to make a blob flow over a region. Then, moving on to the next region, I repeated the process. The brass managed to melt and most of it fell through or could be flicked away. That was good. I managed to put a peanut-sized hole in the cast iron at one point, and so had to fill it. Using the torch with an in-and-out motion after that enabled me to control the temperature so that it wouldn't melt the base metal, but would flow the filler. This is hard to describe, but it's easy to get a feel for when to get the heck out of there for a second. Then you come back in and try again.
"After about a rod and a half, I had added a lot of metal. After a pass, I'd put the hot end down, and toss some sand over it to slow the cooling. After it was cool enough to handle with welding gloves, I'd do some grinding. More grinding on a bench grinder and with a carbide burr left the edge a bit rough. So, I added lots more here. Some of it actually stayed. A few iterations later, I have a reasonably good replica of the original casting, with a lot more iron in it than when I had started.
"The next day I took a look at it and found...a pinhole! Musta gotten real thin there, or not fused real well. OK, now that I knew what I was doing, it would be easy to fix. Once more in the oven, and then to the torch. I added metal, flowed it, and added some more. A little grinding, and it was just fine. A few little voids remain, but I won't risk ruining it for that. "
He's too modest: it's beautiful! There were a few voids left in some places, and some rough edges - cosmetic stuff, really. I filled them in with JB Weld and will file them to shape before painting.
All the parts were masked, but we hadn't quite worked out the painting process. What we finally decided was to make hanging hooks out of coat hanger wire to hang the individual pieces from the garage rafters. A few pieces, like the headstock casting, were too heavy to support this way - the formed hooks would just straighten out. Those were threaded onto a rod and hung by string! (Ed thought it looked like Christmas.)
Finally, we were ready to paint. We waited for a weekend where the weather report called for two sunny days. The first thing to do was build a cheesy spray booth out of 2x4s and a drop cloth. The idea was that a handful of pieces would be moved from the garage to the booth, I'd paint one side, flip the plastic for access to the other side, and we'd rotate to the next batch. It didn't really work that way, because the plastic was a pain to manipulate without getting it stuck to a freshly painted piece. Most parts got painted by me holding onto the hanger with one hand and rotating. The big pieces got one side painted, were flipped on the cross beam, and got the other side painted.
The bed was the hardest, since it's the biggest. We supported it on a couple of crates inside the booth, and I crawled under the plastic to paint the second side. We worked out a carefully choreographed maneuver to turn it over (the bottom was painted first) that avoided touching any non-masked surface. It took about four practice tries to get it right, but it worked perfectly when it needed to.
The last preparation step before painting was to wipe the pieces down with lacquer thinner. They were then treated with MetalPrep, which leaves a thin phosphate coating to help prevent corrosion (and supposedly improve paint adhesion).
I feel compelled to add here that when dealing with these paints, you should wear adequate protection. Read the label directions; most modern paints call for a forced-air system. Catalyzed paints have cyanide byproducts in them. Wear gloves and eye protection and a good respirator. If you can smell the paint while you're spraying it, your respirator isn't doing its job.
It has been reported to me that the gray paint that South Bend sells is either Pittsburgh or Fuller O'Brien Light Grey Machine Enamel, depending on who wins the contract that year. By all reports, it's really good paint and goes on well with a brush. It also isn't nearly as toxic as the paints I describe here.
Other machine refinishers report excellent results with Benjamin Moore Industrial Coatings - either the alkyd base or latex based system.
I primed everything on Saturday. Everything got a single coat of PPG DP40 primer, mixed 1:1 according to the directions. I shot it with a "big" spray gun, with a quart cup, at about 45 psi. This took about two hours with Ed and Pat shuffling parts to me to paint, and consumed about a pint and a half of the primer.
DP40 is great stuff - very durable in its own right, very resistant to sags (important for an amateur painter like me!) It leaves a rough surface that is just great for holding a top coat. The can claims a pot life of 72 hours, so there's no rushing, and you have up to a week to put on the top coat. (I've found that you can delay this even longer, but you'll want to rough up the surface with a Scotchbrite pad and use some sort of chemical prep, like PrepSol, to soften the surface.) We found a few inclusions (insects or rag lint) and sanded those out with 400 grit sandpaper.
Sunday we painted. This was going to be trickier, because the DuPont Chroma One paint ("high solids, single stage acrylic urethane with activator") claimed a pot life of 45 minutes. It also wants two coats with a 15 minute flash-off period between them. That sounded like a lot of running around with parts and not much room for error.
Since it seemed like I was more likely to run out of time than paint, I used a detail gun with a 6 oz. cup. We activated about 8 oz. of the paint (since it's a handy quantity for the 3:1 mix ratio) and got to work. First, I sprayed the bottom of the bed; I figured that it didn't really need two coats up in the cast webs, and all the rest would be recoated from the top. We turned the bed, I started on the top, and Pat started bringing parts. (Ed couldn't help on Sunday.)
It wasn't as bad as expected. We got two coats on the bed, everything that clamps on or attaches to the bed, the underdrive castings, and the handlever tailstock pieces, before the gun started gumming up. I had poured the remaining 2 oz. of paint into the cup in order to not expose the siphon bottom, but there was some left. It was quite liquid - the real problem was that the paint orifice on the gun (and probably some of the internal passages) had started clogging.
The instructions for the Chroma One are designed for HVLP spraying; since I was using a conventional gun, I had to experiment a bit. I ended up at around 40psi with the color valve mostly open; as the paint thickened, I opened the valve even more to keep from getting a "dusty" surface. This paint is even better than the primer - it was easy to put on full wet coats, I got lots of visual cues before the paint started to sag (I still ended up with two or three, but they're hardly noticeable - the paint fills really well) and the finish is amazing. The parts that are fairly smooth look like they're made of plastic. The parts that are rough cast have been filled a fair amount and have a nice gloss. I'm convinced that with enough time to spray multiple coats, and a good holding fixture, you could make any rough cast piece look like it was made from thermoplastic.
DuPont recommends force drying the parts with a 2000W infrared unit. I don't happen to have one of those, but I rigged up the next best thing: two 250W infrared lamps on a timer. I left the parts in the beam for about four hours; I suspect it sped up the curing process a bit, but air drying is probably also fine.
I found four parts (out of about 40) that needed to be retouched; given my skill level and the working conditions, I think that's pretty good. In all but one case, the problem was not enough paint; the remaining piece had some big dirt chunks. I found two very slight sags that I decided to leave. The masking tape came off cleanly the next evening. The modelling clay was not really a good choice. It was a nuisance to get out of holes (especially blind ones, where I couldn't just push it out) and had a tendency not to make a clean parting line under the paint. In once case, the paint had flowed so smoothly across the clay that when I pulled it up, a small section of paint next to the clay came up as well. (Happily it was on the saddle gib, where the touchup is out of sight.)
The next warm day, we scuffed the areas that needed to be retouched with 320 grit paper, wiped down with lacquer thinner, and spotted in the areas with the gun running about 30psi. Then I set the gun back up to 40psi and painted the headstock and covers, saddle, apron, compound, cross slide, and various handles and clamps. This took about 8 oz. more paint for two full coats. I did it in two 4 oz. pours; the remaining paint was put into the freezer in a covered container in a (successful!) attempt to stretch the pot life. The spray quality was about the same, i.e., awesome; I took the extra precaution of wiping down the jet nose of the gun with a rag soaked in lacquer thinner from time to time, to keep it from clogging.
I baked all the handles after they cured. With other enamels, this has the effect of smoothing out imperfections in the surface, and sometimes darkening the color a bit. I didn't notice any such effects; I went up in temperature to as high as 400 degF, for as long as an hour. My hope is that it at least improved adhesion!
All the bare steel handles were polished on a buffing wheel. First with emery compound on a sisal wheel, finished with white rouge on a loose section. The emery got rid of some of the pitting, though it wasn't able to remove the worst sections - 50 years of abuse have made their mark permanent. The handwheels look amazing!
Assembly begins, surprise, surprise, with more cleaning. If the masking tape has been on too long (some of mine had been on for four months) it leaves a gummy residue. In some places, the MetalPrep had leached past the masking and left a powdery white residue. Removing the masking tape sometimes left a raised edge of paint. All this needs to be cleaned up before assembling.
I used a single-edged razor blade to clean up the worst of the masking tape residue and cut off the paint edges. Kroil or Nevr-Dull removed the various residues. I got the machined surfaces to look very nice by cleaning with Nevr-Dull and polishing with a small wire wheel in the Dremel tool - looks like a satin chrome finish. I used Microil, an instrument oil from Kano Labs, as an assembly lube.
I started with the tailstock - it seemed somehow like the easiest place to begin. It went together quite easily - the only puzzlement (still unsolved) was how to avoid scraping the paint under the bolts while the handlever moves. My guess is that the shoulder bolts really don't have a long enough shoulder; they *should* tighten down completely and still allow free movement (ideally with a washer). Maybe the new paint is too thick? Anyway, it sure looks a lot better than the old part did with white paint and rusty handles.
Another evening, another assembly. This time, I started with the compound and worked down to the saddle. Lots of surfaces to clean, lots of parts to find in boxes and put back together, new way wipers to make. Turns out that one of the dials was missing its set screw, and both were missing the brass shoes. There was some play in the feed screws and some dings on the flat surface of the dials (where it mates with the shoulder bushings that we'd carefully machined flat).
So Ed and I spent another evening in the shop, puzzling over the design (for example, it's a bit of a surprise that the dial is part of the bearing surface when feeding, so you can't spin the dial to reset it if there's any pressure on it). Baby got new shoes, and I'll "upgrade" to knurled thumbscrews.
Cleaning up the dial was easy. The trick was removing the slop. There's no thrust bearing in the original assembly to take up wear, so we made one. 10 mil thick brass was about right, thinned down on emery cloth on a surface plate. Ed measured the cross-feed clearance carefully, put it all together, and found it was way tight - after a couple of iterations (hmm, 10 mil slop, 5 mil spacer, it's still tight?) he realized that the slop was in the feed nut, not the screw assembly. Nothing to do about that! The compound needed about a 4mil spacer, so it was worth the effort.
Next up was the rack. This was easy - it just needed a very thorough cleaning, since it was covered in grease! (This was not the first place I found grease on the machine - there was a lot in the apron.) I also took this opportunity to reattach all the name plates to the bed. Boy, it's starting to look really nice. Since I had some time left, I put the change gear bracket and reverse gear assembly together, too.
Assembling the apron was the biggest challenge. I don't have any drawing that matches it completely, and getting it apart was a matter of guesswork and luck. I was so worried about getting the worm and clutch back together that I failed to realize that all the gears had to be on first.
I had figured out the right way to get oil to the worm (not the way it was, nor the way it's shown in any of the drawings I have), got the clutch assembly and halfnuts on (those studs are a really tight fit, and why are they Whitworth threads?) and the handle in place (after spending a long time trying to figure out how the interlock and retaining pin work) and then found that I couldn't put any of the gears on. There really is no extra space back there.
So I took it all apart and started over.
Now, at least, I pretty much understand how it works. There's some play in the casting around the handwheel shaft, and a little in the gears, but it feels really nice. The engagement of the crossfeed is not very smooth or solid; maybe it will be better while the worm is turning. The long interlock pin is worn (or maybe it's the shifter casting) so that you can engage the crossfeed (but not the longitudinal) while the halfnuts are closed. The easy fix, I think, is to make a new pin that's a scrid longer (it looks to be drill rod with the ends tapered).
It looked so much like a lathe when I attached the saddle that I couldn't resist: even though it was well after midnight, I put on the change gear guard and the headstock cover. Boy does it look great.
The headstock cover definitely needs some sort of padding at the ends of its travel. The more modern ones have little rubber feet; I think I'll just glue some leather or felt to the bottom of the front edge. I'm not sure what to do about the back yet, probably something similar.
Next was some fine tuning - there was a sticking point in the carriage travel, which we diagnosed as a slightly bent tooth on the otherwise pristine rack. Some quick work with a file seems to have put it right. Then we gave the lead screw a final cleaning and installed it.
Now there's a complete gear train. We hooked up a set of change gears, oiled everything, and started spinning pieces. Sure enough - the lead screw turns, the thread dial spins, the cross- and longitudinal feeds work! There's a fair bit of slop and backlash, but it's a 50 year old machine. I suspect that some can be adjusted out, some will go away when I get around to making new feed nuts (or half nuts, which are really in bad shape), and some will just be there forever.
I'm not really happy with the spindle bearings. We had to hammer pretty hard to get the spindle out; I also had to hammer pretty hard to get it back in. Tightening the rear bearing bolt freezes the spindle, and it's impossible to move the spindle side to side to check for the thrust bearing setting... this will require more work.
Took the whole thing apart again. Sure enough, one of the little side wicks had dragged in to the bearing area. It's the one by the thrust bearing, which gets dragged in by the splined/threaded portion of the spindle.
I tried three or four times to get it in without interference. A couple more times and I probably would have succeeded - but by then I had torn the wick badly enough that I gave up. I didn't have any more wicking material and I decided I didn't care. If oil leaks out there, it just leaks into the headstock area where it will eventually run into the chip tray.
Dave Ficken of Meridian Machinery suggested using the tailstock to push the spindle into the headstock, since it presses it in dead straight. What a grand idea! This worked really well, up to a point - getting the bull gear fully onto the Woodruff key took more force than I could apply that way (either the headstock or the tailstock would start slipping on the bed). But by then things were well lined up and spinning freely, so I pulled out the trusty Bakelite hammer.
If you look at the side view of the headstock above, you can sort of see the channel for side wicks that I'm talking about - there's a big hole for the capillary oiler, and then directly in line with that a smaller hole that goes back into the reservoir. There's a fine brass spring clip that goes into that hole - the parts book calls them "oil return clips". At that point on the bearing, there's a V channel that goes all the way around the bearing. I put a 1/8" diameter wick in there, on all four, well, three, bearing sides. The roughly circular black things sitting by the box of TSP are what came out when we originally disassembled the headstock - it seemed to be string, oil soaked and filthy! Maybe pipe cleaner would be a better choice next time around; at least it would stay in place. Apparently these wicks aren't used in later SBL headstocks.
I got the bearings all adjusted to spec - just about .001 play everywhere. There's more drag than I would have expected, but since the numbers are what South Bend calls for, I figure it's OK. Good thing, since this headstock doesn't have the laminated shim packs found on later lathes - there's just a thick steel and thin brass shim at each end, and removing thickness is both permanent and inexact!
The bench was next. The former owner (Dino) had cobbled together a drive assembly that had things in a pretty weird configuration. I really wanted to try to hook up the drive the way South Bend originally intended, or at least as close as we can figure out without any real documentation. (The combination of underneath drive and model B seems to have been quite rare - I haven't been able to find any parts drawings that show this particular drive in detail.) Even though we're going to build the One, True Bench some day, following the ideas in an HSM article by Ralph Walker (A Lathe Cabinet/Stand, July/August 1989), I wanted to test sizes and heights and actually use the lathe before we get to that.
Much meatball carpentry ensued. The lathe had to move forward on the bench to make room for the underneath drive hanger to mount in front of the rear leg. But that meant cutting through and remounting some of the supporting members at the front of the bench. The bottom of the cabinet had to come off to make clearance for the motor. One of the side braces had to come off to allow the motor base adjusting screw to stick out - it was in exactly the wrong place. In fact, if there were several ways to do something, in all cases, Dino chose the one that made mounting the lathe properly the most difficult! But now it's all mounted and things even seem to line up.
Yahoo! Seven months and a day later (to be known as a "Richley month of evenings"), FrankenLathe spins again!
First task last night was to wire the motor. I think Ed took the motor apart three times to figure out how it was wired, fish out things he dropped into it, and make repairs. By then, he had sussed out how to wire up the weird drum controller we found (it was a good deal, but the inside is most assuredly *not* a standard 3PDT switch - we probably won't be able to use 220V in the future, but that's a different story). We did the smart thing and wired and tested it on the bench ("Sure, this 18 gauge lamp cord will be just fine for supplying power to the switch"). Even had the rotation direction right the first time!
Next, we laced up the belt. Old lathe, old methods - we cut off the wire hooks and got out the (synthetic, alas) cat gut. After puzzling out the drawings in "How to Run a Lathe" and finding the carpet needles, it was pretty easy.
It was late, but there was no way around it - we had to remount the motor and spin this sucker, even though the wiring will have to be disassembled again. Quick, four bolts and much grunting, and it's on there. Fiddle the pulleys, switch it on ... yahoo! Oil everywhere! It's a lathe! Everything turns, it seems to take loads OK. The bearings get a little warm but not hot. After adjusting the gear lash, all the horrible noises went away. Pushing the spindle protecting ring up against the headstock casting seems to have stopped the spindle from slinging the oil (a drop or two a second!) - I'm starting to wonder if having those wicks is really a good idea, since they seem to bring oil out to the edge of the casting where it can be thrown... who knows?
Much alignment and checkout remains before we dare take a cut, but this is real progress.
I spent some time on the wiring; now there's a real power cord going to a nice metal box with a contactor that is pushbutton controlled, just like a real machine tool. There's even a switched utility outlet for the lamp.
A number of people have written and asked me about wiring their lathe; everything I know about this topic, I learned from this article, by Robert Lamparter. There are lots of good how-to articles on the Metal Web News.
The bench has some sides, and I'm considering building a couple of drawers.
Dave Sobel sold me a nice 6" four-jaw, so tooling is pretty much complete. We spent an evening trying to figure out alignments; the spindle nose is perfect, the tailstock is very good. The saddle/cross-slide alignment is within SBL's spec, but barely; it's hard to tighten the gib in such a way that the drag is even for the full range of travel. We don't have a test bar, so we can't really test the headstock alignment to the ways, but we came up with a cute idea of how to make an adjustable tool do do this.
We finally took that first cut. The belt needs to be re-laced; it makes two distinct "whumps" every time around as the laces hit a pulley. Deeper grooves are needed. The wear in the bed shows up in these cuts, unfortunately - the diameter varied by about .0006 over the fairly short length we tried, going up and down. There is hope of fixing some of this, but I believe it's mostly something I'm going to have to live with.
We also figured out why the belt would sometimes start slipping at high speed - the spindle bearings were too tight! Apparently they'd heat up enough to cause the spindle to drag heavily, stall the belt, cool down again and start spinning. I added a 0.001" brass shim to both the front and rear spindle bearing shim packs, and this seems to have gone away. Some readers of rec.crafts.metalworking have indicated that it's not uncommon to have the bearings adjusted such that you have to loosen the clamp bolts to run at high speed; this seems a bit strange to me, but if I find that I'm getting chatter at low speed, I'll try taking the shims out and see if that's the problem.
I re-laced the belt, since it was really too long. The turnbuckle was close to the end of its travel and the belt was still "loose". It seems that the turnbuckle needs to be actively stretching the belt, not just letting the weight of the motor and countershaft hang on it. If it's not tight enough, the three pieces of the turnbuckle are just resting against one another, and there is a set of clunks from the pieces the countershaft bounces every time the laced joint passes over one of the pulleys!
How to Run a Lathe says to adjust so the cone pulleys are as close together as possible, measure around them with a steel tape, and subtract 1/2" - for this belt, at least, subtracting a whole inch would have been closer to the mark.
I figured out a key trick to getting a quiet laced joint - unfortunately, I figured it out about halfway through. Next time for sure. Each of the grooves on the back side of the belt gets two passes of gut (see HtRaL). It makes a big difference to try to arrange the two passes so they lie side by side in the groove, rather than atop one another (or worse, crossing). I guess SBL figures no one laces belts any more, so lots of details like this are missing from the instructions!
It's beautiful, but is it accurate?
I've been trying to do little bits of work on FrankenLathe, mostly with a lot of frustration at how much of a taper I would get. Of course, I hadn't really spent much time trying to get things square, so I got what I deserved.
This past week, I did a whole lot more work. For starters, I took a 16" piece of 1/2" drill rod (the largest I can chuck in a collet) and faced and center drilled each end. I wanted a long piece to get beyond the obviously worn areas on the bed.
I suspended this between centers, mounted my last word on the compound, and zeroed it at the headstock end. "Zeroed" is used loosely here - the center hole was about .001" off. I ran the saddle back and forth, and watched with amazement at how far off things were. After several iterations of adjusting the tailstock setover, I was able to get zero at both ends, with quite a bit of change (as much as .010") along the span. I chalked it up to a very worn bed and bent rod...
But it still cut quite poorly. I decided to take the next step. I cut up some 2x2x1/8" angle, drilled and welded nuts on the four pieces, in order to fasten one to each leg and take a levelling pad. This was important for several reasons - it gets the legs (each is a 2x4) up out of the small puddle that's forming in the back corner of the garage, it raises the table to take some strain off my back, and it allows me to level the bed with the lovely master precision level I bought from J&L a few months back.
Which is what I did. It took a bit of doing, and another pair of eyes to get right. It's still only sort of right - the wooden bench is flexible enough that moving the saddle from end to end causes the reading on the level to shift! But things are much better. I had to reset the tailstock setover again, quite a bit. The maximum error along the bed is now closer to .0025" - still a lot, but much closer to something that can be dealt with. (I suspect that Ed will get his way and we'll rescrape the bed - he's going to want to do something with the straightedges he's making molds for.)
Part of it is the learning process - I've learned an awful lot about how machine tools go together, what makes them good, what makes them bad, how they go wrong. I've met a lot of great people. I've made the machine my own - there's a part of me in there, rather than it just being an appliance that I own.
But there's another part to it, too, which Errol Groff recently articulated in a way that made it crystal clear for me: "One of my reasons for collecting the darn things is to imagine the toolmakers that might have used them in the past. The image of these men building the tools and machines that saw our country through two world wars brings a shiver to me."