The 120 tooth saga n the Isaacs clock

The biggest wheel on the Isaacs clock is a 120 tooth monster. As I said before, when I made the pinions I inadvertendly went from module 0.6 to 0.9 and I am unwilling to redo the pinions. I calculated that the biggest wheel on the clock  it would fit on the lathe i.e. I could "swing" it. So it did, but this whole 120 tooth "great wheel" is a monster fully up to Godzilla's standards. I calculated that a 120 tooth wheel would fit the lathe. I had no idea how close that calculation was.

I cut these things out on the bandsaw. Then I turn them down to the exact diameter on the lathe. It is something like (N + 1.76)*M (M is the module and N is the number of teeth, and if anybody is screaming about this remmeber these are cycloidal bears, not involutes). As you can see it clears the little Taig by about 2mm and I had to file it a little to get it to fit!

Next we need a pair of backing disks for  the wheel. I like the kind of board called masonite or its like. It does not matter if it is a lot off-center, its function is to support the wheel in its odyssey on the mill.

Now my dividing head has a height of 50 mm. Unfortunately at Module 0.9, the radius is about  110mm so I am a bit short. What to do? Why, riser blocks.

Above 100 tooth wheel for comparison.

A 100 tooth just barely fits without blocks. So... First I tried putting it on the left side of the mill. Note the riser blocks. Note the clamping. Also note I have not supported the wheel. Two big mistakes in one go. Unsupported Lexan wheels flex and lead at best to inaccurate cuts. At worst to "crunchies" where you wreck the wheel beyond repair. Worse yet, I am "climb milling." the cutter is fed in the same direction as the cutter moves. Sometimes this does not matter. On all the previous wheels it didn't. But this is a huge wheel relative to my equipment. So remove the vise, go to the right side of the mill.

Much better. Now I am "conventional milling" and the cutter moves opposite to the feed. I am still having problems. Stay tuned.

Cutter project II. Getting there

I have been remiss in posting, but other things have interfered. As usual.

So we continue with making four-tooth cutters. We are getting there, but not quite. Still, progress. I am using 3/4" (20mm) supermanium steel from Lowe's. First step is to turn it round. The exact diameter is not critical. Then we slice off a 5mm slice. Hardest part of the whole thing, because The Taig does not like to part 20mm of steel. Then we put it on the drilling jig, and drill four holes in it. Equally spaced, if our drilling jig is up to snuff. I made on the mill, it better be! Then we mount the slice on the eccentric arbor which I have described previously. Here it is. As you see, one piece of the round sticks out. Exactly what we want.

Now we proceed to turn a certain amount of material off. I have to watch the dial on the cross-feed very carefully. I mark the dials with a sharpie. What we want is a sort of square, but with rounded sides. We have to rotate the blank 90 deg after the full cut has been taken. After a while we start getting the shape of a square. Again, sharpie keeps me on track. After a series of cuts, we have to unbolt the piece, rotate it 90 degrees, and make some more cuts. The circle of holes go successively into the pin in the eccentric mandrel.

A closer view. Getting to square.  When I get a sharp corner I am done. Now comes the hard part. We have to form the radius at the bottom of the cutter. For this we use a form tool shaped like the radius of the cutter. It is less than 3mm. The form tool is plunged in. But setting it up on center is very difficult. I have since learned about something called the button method, which I intend to try. The Taig has no way to measure transverse feed. Unless you fit the compound. The Taig compound is very flimsy and I don't think it would work. A dial indicator would be great, but nowhere can I fit one in; the lathe is mostly aluminum. So my radii left something to be desired.

Next we take it over to the mill. I made a fixture -- a square wih a hole dead center -- so I can turn the thing over. Using an abrasive Dremel cutoff disk we "gash" the teeth. I did not get any pix, but it is in the previous episode. In the next pic, I have gashed radially at the corners.

Then we rotate the piece 45 deg and gash again. One advantage of using the cutoff tool is that your cutter faces are automatically sharpened. You can see one of the teeth pointing at you, one in profile to the right, and even see the relief on the teeth. The relief is the sole reason we went to the eccentric arbor and the holes. Without relief, the cutter won't cut.

Still have to work on the radius. But this is much better than my previous attempt. Onward.

The cutter project (really part of the clock)

As I may have said, I got tired of cutting gears. This requires so much concentration that I wanted a vacation. So I have undertaken two new projects. One is to install my (Taiwanese) Emmert vise on the woodworking workbench. The other, which goes back a while, is to make mutipoint clock gear cutters. The first project is a future  post. This post deals with making clock wheel (gear) cutters.

Now gears come in two major flavors. One is the involute form found every mechanical contraption that uses gears. The other is the cycloidal form, found almost exclusively in clocks. The involutes are better at transmitting power. The cycloidals have less friction, and so are favored in clocks.

Gears of any flavor are cut by (guess what) cutters. You can buy these things. They are expensive. Not only that, and mainly, none of them will fit my Proxxon micromill. So I am making my own. The main requirement is that I have to fit them to Cecil B. de Mille, my mill.

In making these cutters I am basically following Dean's writeup.  All cutters need relief. Just like a kitchen knife. It is difficult to slice anything unless your knife is curved. That's relief. The edge does not drag once the nain part goes through. If you really want to make cutters, you must read this writeup.

The way we do this is to build an eccentric arbor. This provides the relief.
But first we have to turn a wheel blank. I started out with the idea of turning it to fit the Proxxon 3mm collet. 

Nice idea. But for various reasons it did not work, so plan B. We will make disks, and turn them on an eccentric arbor, but my eccentric arbor is much smaller an Dean's device. An eccentric arbor is a cylinder, but with the center (a 6-32 screw) offset from the true center by 4mm. In the real center is a broken drill bit about 2mm in diameter. Very small. It is an anti-rotation and indexing pin.

Next step is to drill four holes in a cutter blank. The holes form a square 4mm to a side. Dean & co. suggest making a drilling jig. I did the first one on the mill. Afterward I did a proper drilling jig, because drilling holes on the mill by plunging is like dentistry. Painful. I used my handy setup plate in the mill vise. You can barely see the blank.

So with four holes in the blank, pick one, put the center hole into the screw, put one hole in the pin, and you have the setup below. It looks off-center, does it not? It is. It is supposed to be. When we put the whole megilla into the lathe, we will turn sort of a square with round sides.

When you are doing this you have to be very careful with the depth of cut. I experimented with setting up dial indicators to do this.

Impossible. They don't make them small enough, and I have no room to fit everything in. So I made a stop. It took a day, well worth it.

The stop is a dovetail that fits the ways of the Taig, an has a screw for fine adjustment, unfortunately a 6-32 screw because it is difficult to find  metric screws this small (about 5mm) this small in Alaska.

Next job is to make a radius forming tool. This tool will form the radius (which has to be exact). I had not reckoned with Lowe's 3/4" "mild steel." It is made out of supermanium. I think Lowe's supplier slipped up that day, and threw in a round bar made out of Titanium SuperSteel, because I made this tool out of Dremel shanks, which I know can be hardened.

Now we mount this in the middle of a square .25" bar and mill off exactly half of the tool, leaving a very sharp edge. We did this on the mill. Note the gear sitting to the left. It's the one I haven't completed. We use this tool to form the radii on the cutter. Then we use this tool to cut the radii. The first few radius tools I made were eaten up by the Lowe's Supermanium. I finally resorted to a broken Proxxon endmill. Teutonic technology proved superior to supermanium. I got a radius, Crude, but good enough for practice purposes. Now let's cut off the waste. Dean et al. use a slitting saw, but I used a Dremel abrasive cutoff disk mounted on the mill.

 
This has the great advantage. You don't have to grind the cutter; the cutoff does it for you. My finished cutter has two good teeth on it, the others were ruined by stupid mistakes.

I have learned a lot from this. I think I can make my own cutters now. More to come. Pic of cutter bit blurry, new camera. Will improve.

Onwards with wheel cutting

In the saga of the Isaacs clock , we now go on to the big wheels. There are two 96-tooth wheels, and (I think) a 100 or maybe a 120 tooth wheel, the latter in extreme range of what I can turn on the lathe. I am now making the wheels out of Lexxan instead of acrylic. Acrylic shatters too easily. With our homemade fly cutter it was possible to cut the first 96-tooth gear. Note the masonite backup disk on the wheel, this helps damp out the cutting shock.




So on to the second one. First step is to turn the blank. I try to do a spare, but it is  not a good idea to turn them together because the turning process tends to melt the Lexxan and then you have two welded wheels, which is not a happy situation.



So this gear was (still is) mounted on the dividing head, and I have cut about 8 teeth, but gear cutting is an absolutely frustrating situation and you should only do it when you are wide awake and capable of extreme concentration. So I decided to take a break, and the disadvantage of this is that I will have to "pick up" the cut. So I have only half of the mill available.

Then I decided to take a break. I would make a multi-point cutter. Comercial cutters are made with very large holes. Maybe 7mm. My mill takes a 3mm arbor. So the cutters are a saga all by itself, which will be the subject of the next post.

Cutter-making requires an eccentric arbor. I will explain this more fully in the next post, but it involves turning off-center. But here's a shot of the making of the arbor, in case I forget to include it in the next post.

It has taken a week to get this thing up, thanks to Google for their user-friendly (hostile) interface with blogger. Sorry.

 and one 100-tooth wheel. The 100 toother is going to be a real deal.

Divide and conquer. Maybe.

Well, it has been a while and I have not posted. Life gets in the way. In the last post I had tracked my wheel (gear) cutting problems down to to bent shaft in the dividing head. Since the dividing head is a Topsy project (she just growed) this is a retrofit and rather difficult.Below, I am checking the runout on one of the wheels. Just as large as ever it was, a whole mm. Nothing new. Nothing for it but to remake the shaft. Turning it is easy, but..

Thing about the shaft is the dimensions. One end has to fit the gear. That is 5mm right on the money; the printer I took apart to get these gears is that dimension. The other end has to be a duplicate of the Proxxon mill spindle. This is M8x0.75, also metric. So I remade the shaft. Here it is, parted off. Note I am using both my homemade steady rest and a Dremel tool to part off the 5mm end. This gives a nice clean cut and wil not distort the shaft. I claim a new parting-off method.

Now, the big job is to put a front support on the shaft. For this I used a piece of my lovely Aluminum block that I ordered from the Internet.

 The problem is that I have to bore the 8mm hole exactly in line with the original 5mm hole in the rear. Lacking a jig borer (they cost a fortune) I used a transfer punch, and the above lashup shows how I got it done.

A runout check shows that I am down to a runout of .30 mm or so. Not really good enough but better than the full mm I had to begin with. What I will have to do is to watch my depth of cut. The wheel is perfectly round as the lathe can get it; the dividing head has a wobble! By varying the depth of cut, maybe I can compensate for this. So the setup looks like this:

 The dividing head is bolted to the mill table and "trammed" i.e squared to the table. In the mill spindle is my homemade fly cutter.  Let's try a 50 tooth wheel, I have two of them to do. Result:

Eureka! I have a wheel. It was like pulling teeth. Literally. One mistake and the whole wheel is trash. And I made lots of mistakes. Acrylic is totally unforgiving of mistakes, say forgetting to lock the dividing head. It explodes.  So I have switched to Lexan, much more resilient.

In the next episode we will cut the big wheels.

Dividing head woes

Things have been very awry. My clock wheels are in trouble. Fortunately they are acrylic, or I would be out a fortune. Problem is that while some of the teeth look fine, others have flat tops. After pondering this one for a while I finally came to the conclusion that the dividing head is running out. This means it isn't centered. It is wobbling. So let's check.

Observe my new elegant mini-dial indicator holder. It will hold both my Imperial supersensitive indicator (shown above) and a conventional DTI. It is being used to record the runout (wobble) on the wheel, which is is on a mandrel (shaft) held in a collet. These are my wonderful ER collets. They have essentially zero wobble.
 

 'Nother shot same thing, same results. Runout about .002" or about 4 "cents" (.04mm). Uh-oh. The wheel is quite acceptable for clockwork. Time to check the runout on the shaft of the dividing head. This is quite a production. The dividing head is mounted on the mill. This, except for the base, is made of non-ferrous metal, and the indicator base will not adhere. So I had left the vise on the mill. In it I clamped a piece of angle iron, aand the dial indicator will adhere to that!  So by now I had acquired a metric dial indicator, and checked the runout again. Horrors. A whole millimeter!
The runout is all in the shaft of the dividing head.

Then  I took the dividing head apart and checked its shaft. It needed no dail indicator to show it was bent. So I made a new one. I used my steady rest to hold things still. Here I am, parting offf the result. I claim a new method of parting off. I use my Dremel tool holder, one of the very fragile  cutoff wheels, and spin the lathe one way and the work i t'other. Got a nice clean part, and a very narrow kerf.

So the next thing was to do something about this. It is a very small dividing head, so my next idea was to add a new outboard support.  It is an aluminum block. Here it is under construction.

The results on divde head 2.0 are not encouraging. I measure the runout  on the new spindle:

It still comes to 30 cents. While better than a whole dollar it is not very good. Impasse. While I am figuring out what do do about this, I added a new feature to Cecil B. de Mille. Behold my Z-axis Digital Readout.

Just a super-cheap digital plastic caliper, a bit of angle aluminum, and some drill and tap, and now I know where my Z-axis is. Of course I can always use the dials. But the Count himself (on Sesame Street) would get confused  by the number of turns you made. And one turn off is a whole millimeter off. Worthwhile addition.

Big wheels on a small lathe

No matter how large a lathe you have, sooner or later you will come up with something too bigto turn. When I went to module 0.9 on the clock, I calculated the size of the biggest wheel and found I could swing it (that is, get it on the lathe without colliding with the bed). So I said, let us go ahead and do this thing with module 0.9. The following picture is the 50 tooth gears made so far. None of them are stellar. So we foresee problems ahead.

First thing was to cut the blanks out for the midsize wheels. I did this on the bandsaw. The first thing I did was to cut out masonite-type board slightly smaller than than the gear itself. When I put it on the lathe, it was very obvious that the regular toolpost was not going to reach to the rim and cut it. Impasse. Next morning I started designing a fixture that would move the toolpost -- and realized as I did I already had one, the compound slide. The Taig compound is very flimsy, so I usually don't use it unless I have to cut tapers. So up with the compound and success:

plus I have an extra 25 mm (1") by moving the tool to the outside groove.

So the last problem was to make a new dividing head plate. I did this my usual way, with my PostScript program. I used my optical center punch. Then off to the drill press. Just to make sure everything was centered, I used a fixture.

Then I put the dividing head back together and tried some more 50 tooth blanks.  None of them were stellar either. There is something wrong with the dividing head. But that will wait until the next episode.

Life at module 0.9

In the last episode, we nade some lantern pinions. Measurement revealed that they were actually module 0.9. So this is really a  blessing. Since I am making my own cutters, the larger module will be easier to work with. So I am now embarking on the process of cutting the wheels. The first problem is to make the cutter.  I use Dremel 3mm tool shanks from expended Dremel tool cutters. Cheap, and they are good steel. they can be hardened and tempered. I am making a fly cutter, a one-point cutting tool.

Now a fly cutter does not actually cut teeth. What it does is cut the space between teeth. A clock tooth is supposed to have a cycloidal profile. This is the profile generated by a circle rolling on another circle. Yuk. However, this is approximated by a straight cut with a "rounded  over" circular radius at the tip. The radius is something like 1.7 mm at module 0.9.

So as a first task I made a button gauge.

I turned down a piece of steel to the proper radius, say 1.7 mm. I am too tired to go consult my notes in the shop. I drilled two holes the proper distance apart. This was done on the mill, you could never hit it by eye. The button gauge will be used see if I am on target with the radius. There is the problem of depth of cut, but if I overdo this I can always grind it off. Off to the mill.

Here we have an expended Dremel shaft put into a homemade fixture, a piece of square stock with a setscrew to hold it in place. The fixture is clamped in the mill vise.

I have available 3mm, 2mm, and 1mm. end mills. These are diameters. Hmm. If I were to cut 1.7 radius I would need a 3.4 mm cutter. Unicorn. Uncomfortable. But the 3mm guy will go 1.5 mm aand for now that will do. It is quite difficult to center up the cutter. But above you see it taking shape. So I did this. Now we heat it up red hot and quench. This will harden the steel.

I use my handy furnace and water-quench, and then temper, a difficult job on a piece smaller than your little fingernail.

Having done this, we take a test cut on a leftover blank we happen to have. The diameter is completly off, we just want to see if the cutter works at all.

So I mount this random blank on the dividing head and cut a few teeth. The diameter is wacky. But it does work -- i.e. it cuts teeth. Spacing all wrong of course.

Next step is to turn up a proper blank on the lathe. I cut them out on the bandsaw. The scrollsaw would be better, but it melts the plastic so the bandsaw wins.

Now we can cut teeth properly. I did a whole bunch of them. There are so many errors you can make. You can forget to tighten the dividing head, for instance. This will chew the blank. You can forget to loosen the dividing head, which will mean slippage in the gear train. Maybe I should loctite the worm. But I don't want to do this yet.

Anyway, at the end of several days work,  I came up with some wheels.

The leftmost wheel is complete chowder, as Tom Lipton would say. As we go left to right, we see gradual improvement, as I correct my mistakes, so the rightmost wheel is almost usable. But there are two problems. The tooth profile is off. Also the spacing is irregular. The tooth widths vary. This is a problem with my homemade dividing plate. In the next episode we del with these problems.

The saga of the lantern pinions

In a clock, the gears that convert the movement of the pendulum to the movement of the hands are of two types. If the gear  has 12 teeth or less it is called a pinion. If it has more than that, it is called a wheel. The Isaacs clock has 8 toothed gears for the pinions and other numbers for the wheels. Pinions are small fiddly things, about 6mm diameter. That's about 1/4" for the metrically challenged. Now there are several ways of doing pinions. First is to buy a commercial pinion cutter. Messrs. Thornton in England will sell you one, at what I consider an exorbitant price, 40 quid or about $80. Second, make tour own pinion cutter. I am really challenged here, because my mill is a real micro. The largest collet it will take is 3.2 mm (1/8") so the  7mm diameter of the hole in Messrs. Thornton's cutters is far too big for my tiny Proxxon mill. Second, make your own cutter. I looked a lot into this and they are quite a complex problem -- again because I have such a tiny mill. I will deal with this some other day. I can do it, I think, but I will have to rescale a lot of things.

The third way is to make lantern pinions and this is what I did. Essentially a lantern pinion is a very small hamster cage. It is two circles for the side of the cage, and 8 bars to the cage. Eight bars work out conveniently to 45 degrees at a side.

So I made up a wheel divided into 45 degree increments. A production,  but possible. I then used my aformentioned Dremel tool holder to drill the 8 holes. Simple, eh? Not really. First I had to make a mandrel, a shaft that fits into my "crocodile," the ER-16 collet on my Taig. I threaded it US 4-40 because that is the smallest tap and die set I own, about 2.4 mm. Then I had to make a special nut to fit the 4-40 thread and not interfere with the boring of the holes. A standard 4-40 nut is too big. The diameter of the hamster cage is 6.1 mm at module 0.6.

So now we turn up a bunch of hamster cage circles to the proper diameter, which is about 12mm. This can be done en masse, four sides at once. Then I laboriously cut up some music wire into cage bars. Regardless of its name, music wire has nothing to do with music, and worse, it is often called piano wire, although it has little or nothing to do with pianos.

The first result is shown above. It is a valid lantern pinion. It is sitting on top of a ski wax container. I use the ski wax on bandsaw blades and it really helps.

Now I made up an index stop out of an old saw blade and a broken Dremel mini-drill. I have lots of those, they are are very easy to break. The ones I am using are about 0.7 mm but unfortunately the wire is 0,77 mm.
The index stop is saw blade attached to a magnet., super-glued to the saw blade. I works.

And fortunately, looking through my supplies, I found a wire (from Michael's) same gauge as the music wire, slightly less stiff, and far less expensive. And much more obtainable. I have bought out Lowe's supply.

So here is the final mise en scene (forgive the lack of a grave accent). These are the tools I used to make 9 lantern pinions. I should only need 7, but better safe than sorry. There are pliers, of course. Then is my Archimedes drill. This has a piece of music wire in it, which is used as a drill/reamer to bring the holes in the cages to final size. It was quite a feat to grind that thing properly so that it would actually drill.

 There is an 8mm wrench that belongs to the mill. I use it to cinch up the pin vise, the invaluable object on the right, which holds the wire while you get it through the holes. Sitting in the pin vise is the last of the hamster cages.

When it was all over I measured the diameter of the pins in the cage. It was supposed to be 6.11 mm and came out to 7.7mm. Ouch! This is a major blunder. A real Bozo, as Tom Lipton would say. However I think it is a blessing disguised as a blunder. I worked out what the module actually is, and is 0.9 instead of 0.6. I think this module will be much easier to work with. Of course I will be into a redesign of the clock because the spacings will be different from the plans. But since I can calculate all of this, the redesign will not be too bad a deal. I can still swing the biggest wheel on the Taig. Stay tuned.

Creeping up on a new clock

First of all you should be aware that the wooden clock is alive and well. It is being finished at John's cabable hands. There are a whole bunch of fiddly parts that have to be finished.And there is still the pendulum, and the weight, and the pulleys. Lots of stuff to do. The backbreaking work is over, though.

But until they are finished, I can't work on anything else. So, as I have said before, I am jump-starting my next clock. This is the Isaacs clock.

The real problem with making clocks is making the wheels, i.e. gears. The real problem with making wheels is the wheel cutters. So since my last post I have been obsessed with making cutters, I am following a page which I supplied before, so I will not bore you with the details. I will say that it has been quite difficult. Now you could order these cutters form Mssrs. P.P Thornton in the UK for a modest 40 UKP apiece. Gulp. No. I must make my own cutters.

The cutters in question are described in Dean's pages, which I linked before. THe trouble is actually cutting them. Of course, I am not using "1/8 tool steel" as Dean suggests, because this is Unobtanium in Alaska. So I am using an old lawnmower blade. Very tough stuff.  It resists even my carbide tools. Plus it is an interrupted cut. The lathe tool cuts only sometimes. Since it it made on an eccentric arbor, that is, the center of rotation does not coincide with the center of the tool, it is very difficult even with carbide tools, which is what I am using. I persevered. I learned. Next time I will use my newly acquired angle grinder to rough the thing out. These things are at the extreme limit of the Taig lathe.

In the meantime, I read, on a clockmaking forum, of the equipment of some fellow clockmackers. This included a gentleman with an 18" South Bend lathe. That's about 450 mm swing and would have breezed through the cuts that strained my 55 mm lathe. Wish I had an 18" South Bend. "I like to take big cuts," said the owner.

But eventually, about a week or so later, I wound up with a cutter blank.

This image is a little confusing. There are four holes filled in with JB-weld. They don't count. The ones that do count are really half-holes. Notice the shape of this thing. It is a sort of a square with rounded sides. That is the effect of the eccentric arbor. The rounded sides provide relief.  See Dean's page, previously cited. It took forever to get to this point.

However, we got there. The next stage is to make a form tool to shape the cutter teeth.

This whole thing is unwinding the Industrial Revolution. You need A. But to make A you need to make B. But to make B, you need to make C... and so on.  Either that, or pay Mssrs. P. P Thornton Ltd. forty quid for one cutter. Ouch. I will make my own, and and a very interesting experience, too. A fine fate for a recycled lawnmower blade.