Once the magic smoke comes out, things don't work any more.

John Kasunich
jmkasunich@fastmail.fm
GPG Key

Postings:

• most recent
• Shoptask CNC
• Spindles
• Photography
• Programming

Index (titles only):

• Shoptask CNC
• Spindles
• Photography
• Programming
• All of 2008
• All of 2007
• All of 2006

If you are into RSS, you can Subscribe to a syndicated feed.

• EMC
• Van Norman

• Alex
• Chris
• Jeff

Mon, 08 Sep 2008

Spindle Project - Part 2 - Deciding on a Taper

The "Spindle Project" is my attempt to make something useful (and perhaps somewhat profitable) out of some surplus machine tool spindles that I bought in late 2007. This posting is about choosing a spindle taper and toolholder for the spindles. (See this posting for background on the spindles themselves.)

My existing milling machine is a Shoptask 3-in-1 that I have converted to CNC using EMC2 as the control software. It has a #3 Morse Taper on the mill spindle. I use " Tormach Tooling System" (aka TTS) toolholders on the machine, with a 3/4" MT3 collet in the spindle. The TTS system is a nice arrangement for small to medium mills - it is quite rigid for its size, and very repeatable in the Z direction. The toolholders have a flange which engages the spindle nose as the collet draws them in. Below is a drawing of a typical TTS holder, and a photo of my selection of TTS tooling. (Click drawings for a scalable PDF file, click photos for higher resolution.)

Since I already have some TTS tooling and plan to get more, it is naturally a leading candidate for these spindles.

Before I focus entirely on TTS tooling, I wanted to examine any other possiblities. The first thing to come to mind was of course R8. You can get R8 tooling anywhere, at very competitive prices. The main problem with R8 on these spindles is the length. As the drawing below shows, the back of the collet goes way up into the spindle.

The back of the collet would extend so far into the spindle that the step where the drawbar springs sit would be removed. That means I couldn't use the springs - I'd have to use a regular drawbar.

I would also have to figure out how to enlarge a 0.236" (6mm) hole that is over 2-1/2" inside the spindle, and make a precision 0.950" bore 4" into the spindle. The spindles are hardened (at least case-hardened), so the machining wouldn't be fun. I have a carbide boring bar with cermet inserts that can cut hardened steel, but it is far too large to get into the 6mm hole. I have no idea if an ordinary drill would touch it, and I don't want to think about what a carbide drill might cost.

And finally, the 1.25" diameter at the gage line leaves very little of the "step" on the front of the spindle. I would have to remove the step and machine the front of the spindle flat. That would move the taper another 0.160" deeper into the spindle, making all the other problems even worse. So, R8 is out.

NMTP-30 taper tooling is roughly the same diameter as R8, but considerably shorter. Tooling is less common and more expensive, but not outrageously so. So I dug out Machinery's Handbook and drew up a 30 taper toolholder. The result is below.

Although I would be able to use the drawbar springs, and wouldn't have to do a precision bore 4" into the spindle, the 30 taper has all of the other problems that make R8 impossible. And it costs more. So NMTB 30 is out.

The Tormach Tooling System toolholders only extend 1-3/8" into the spindle - that seems perfect since I really don't want to try machining deep inside the spindle. The standard way of holding TTS tooling is by either an R8 collet or a #3 Morse Taper collet. Tormach sells both collets and they aren't very expensive. I already know that R8 is too long, so I measured up the MT3 collet from my Shoptask, and added it to the drawing.

The MT3 is better than R8 in several ways. It is short enough that I could use the drawbar springs. It doesn't remove so much metal from the nose, so I can keep the 0.160 step on the front of the spindle. I would need to make a 0.160" or slightly thicker "washer" that would mount to the spindle nose, and be turned or ground square for the toolholders to seat against. That won't be too hard - there are already four M3 threaded holes in the spindle nose that would work perfectly. The washer would also cover up some other holes in the spindle face to make it a bit smoother and neater.

Unfortunately, the MT3 collet is still too long - I would have to clear out the 6mm area of the spindle. It wouldn't be a critical dimension, since the collet only bears against the taper for the outer 1-1/2" or so, but it would still be difficult to do.

Also, if you look closely at the drawing, you can see that the Morse Taper surface intersects the original taper near the end of the spindle. The existing taper is mirror smooth and very hard. If the spindle is case hardened, my new taper will transition from very hard to less hard material. That will make it hard to get a good fit and finish.

The conclusion is that I could probably use the stock MT3 collet, but it will require some very tricky machining.

Since the main problem with the MT3 collet is its length, I started considering a shorter collet. TTS shanks are short enough that it should be possible to remove 3/4" or so from the back of the collet and still have enough flex for it to close properly. However, simply cutting down the existing collet won't work - the threads don't go deep enough. I started desgning a collet that I could make from a hardened and ground MT3 to MT2 reducing sleeve.

However, tonight I realized that if I'm making a custom collet, I might as well go all the way. A Morse taper is about 0.6 inches per foot. The existing HSK taper in the front of the spindle is 1.2 inches per foot. If I make a collet to fit that taper, it will probably be easier to release from the spindle. In addition, I won't need to do as much critical machining. I still need a nicely fitted straight area in the back of the collet, but I won't need to deal with the taper, and I won't have to worry about the transition between hard case and softer core. The drawings below show the collet, and how it fits into the spindle.

I have a small electric oven that gets hot enough to harden drill rod, and is big enough to hold this collet. I'm leaning towards this approach, in part because making a precision collet would be an interesting challenge.

I haven't decided for sure whether to do a collet as shown above, a shorter MT3 collet, or the full length MT3 collet. I just ordered a MT3/MT2 reducing sleeve, as well as a chunk of 1" drill rod. We'll see what works out best.

I haven't just been drawing and thinking. Over the last weeks I've built some tooling needed to machine the spindles, and have done some experimental turning and boring. Those results will be the subject of my next posting.

(posted: 08 Sep 2008 01:09) (permalink)

Mon, 01 Sep 2008

Spindle Project - Part 1 - The Spindles

The "Spindle Project" is my attempt to make something usefull (and perhaps somewhat profitable) out of some surplus machine tool spindles that I bought in late 2007. In the process I hope explore some areas of machining that I haven't done before, including hard turning, grinding, and extreme precision work.

I have sixteen spindles, all more-or-less identical (details later). I initially bought two, just for the bearings. When I figured out what they were I went back and bought the rest. A picture is worth a thousand words, so here are side and end views of an assembled spindle (click to enlarge):

After examining the business end for a while, and doing a lot of googling, I figured out that the spindle taper is HSK32. HSK tapers are relatively new and at least for the home-shop world, very exotic. Toolholders start at $200 and go up - way up. The premise behind HSK tooling is that it doesn't just make contact on the taper. It is designed to contact on both the taper and the flange at the same time, for increased repeatability and rigidity. The HSK design is also suitable for very high speeds.

The somewhat exotic and high-speed nature of HSK was my main clue that these were probaby some very high-class spindles, with correspondingly high-class bearings. Although most of the spindles had tags on them indicating that they needed repair, none were tagged for bearing problems. Most were things like damaged retention mechanisms, drawbar sticking, etc. So I decided to take a chance, and bought the whole batch.

The first step after I got them home was to remove the retention mechanism. That consists of a tapered piece that is screwed onto the drawbar, six fingers that expand inside the toolholder, and a crown-shaped piece that sits in the back of the taper and hold the backs of the fingers. Several spindles were missing one or more of these pieces, and some had broken or bent fingers, etc. The tapered center part unscrews, and the fingers and "crown" come out with a pair of needlenose pliers, leaving this:

Some more carefull fiddling around and I figured out how to get the rest of the spindle apart. Since HSK spindles are pretty much always used with automatic toolchangers, the drawbar is spring loaded. Some have a long stack of Bellville spring washers, others have helical springs that are sort of like a stack of Bellvilles where each layer is twisted and connects to the next. The helical versionis a lot nicer to handle - two spring sections instead of 132 individual washers stacked in a precise series-parallel pattern. I had to make a simple tool to allow me to compress the drawbar and retain it in that position while I loosened a setscrew and removed a cross pin - then I could release the spring pressure and the drawbar came out the back. The two photos below show the results of the second stage of disassembly:

In the first photo, the retention "stuff" is assembled as it goes into the spindle - in the second, the fingers, crown, and tapered part are separated. The drawbar in the photo has the helical springs (of my 16 spindles, 10 have helical springs, and 6 have Bellville washers).

I'm about 99% sure I'm not going to be using the retention stuff or the keys, so those pieces are bagged and tucked away. The drawbars, drawbar springs and various other bits and pieces will probably be reused, so they've been carefully cleaned, sorted, and stored.

The only thing left is to remove the bearings themselves from the spindle. So far I've only done that to two spindles. I figure if I don't mess with them, I won't mess them up. I might be able to do whatever work I need without removing them. If not, I'll remove them when it becomes neccessary, not before. The picture below shows one of the two that I did carefully take apart:

From the left, the pieces are: front ring (part of a non-contact labyrinth seal), front bearings, outer race support ring, inner race spacer (long tube), back bearings, and inner race nut. The bearings are two pairs of matched 25 degree angular contact ball bearings, each pair arrainged in "back-to-back" configuration. The clamping nut loads up the entire stack, with the clamp force going through the first back inner race, both back outer races, second back inner, the long spacer tube, first front inner, both front outers, second front inner, and into a shoulder on the spindle shaft.

The two spindles I've taken apart so far have bearings from two different makers, but they are basically the same thing. The front pair are trade size 7008, 40mm ID, 68mm OD, and 15mm thick. One vendor's datasheet for the front bearings is here. The back bearings are size 71908, 40mm ID, 62mm OD, and 12mm thick. Vendor data is here. A general page for that vendor's spindle bearings is here, and from there I was able to download their 11-megabyte spindle bearing catalog with lots of good engineering info about the care and feeding of high-end bearings. The other vendor doesn't have the same detailed data, but a magic decoder ring for their part numbers is here - the number on the front bearings is VEX40/NS 7CE3 DD/3. Both sets of bearings have ceramic balls in steel races, and are ABEC-7 grade, or the European equivalent. While googling for data on these bearings, I found a 2006 price sheet that listed the back bearings at 276 Euros each. At today's conversion rates, that is about $400 per bearing!

After much measuring and investigating, I made a to-scale EasyCad drawing that shows all the important spindle parts, along with several partial sections to show all the various holes and details near the front of the spindle. Click on the thumbnail image below to get a scalable PDF version.

The drawing does not show the retention pieces, or the key. Section B-B does show the slot that the key fits into, and the long hole for the spring loaded push-rod. The holes in section A-A are for coolant - apparently HSK tooling can accept high pressure coolant either through the center of the drawbar, or through the toolholder flange.

In my next posting, I'll describe my plans for these spindles.

(posted: 01 Sep 2008 00:45) (permalink)